Loading...
The URL can be used to link to this page
Your browser does not support the video tag.
Home
My WebLink
About
Energy Storage Systems Program Report for FY99 2000
SANDIA REPORT SAND2000-1317 Unlimited Release Printed June 2000 Energy Storage Systems Program Report for FY99 John D. Boyes Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185 and Livermore, California 94550 Sandia is 2 multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under Contract DE-AC04-94AL85000. Approved for public release; further dissemination unlimited. (Fh) Sandia National Laboratories Issued by Sandia National Laboratories, operated for the United States Department of Energy by Sandia Corporation. NOTICE: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government, nor any agency thereof, nor any of their employees, nor any of their contractors, subcontractors, or their employees, make any warranty, express or implied, or assume any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represent that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government, any agency thereof, or any of their contractors or subcontractors. The views and opinions expressed herein do not necessarily state or reflect those of the United States Government, any agency thereof, or any of their contractors. Printed in the United States of America. This report has been reproduced directly from the best available copy. Available to DOE and DOE contractors from U.S. Department of Energy Office of Scientific and Technical Information P.O. Box 62 Oak Ridge, TN 37831 Telephone: (865)576-8401 Facsimile: (865)576-5728 E-Mail: reports@adonis.osti.gov Online ordering: http: //www.doe.gov/ bridge Available to the public from U.S. Department of Commerce National Technical Information Service 5285 Port Royal Rd Springfield, VA 22161 Telephone: (800)553-6847 Facsimile: (703)605-6900 E-Mail: orders@ntis.fedworld.gov Online order: http://www.ntis.gov/ordering.htm SAND2000-1317 Unlimited Release Printed June 2000 Energy Storage Systems Program Report for FY99 John D. Boyes, Program Manager Energy Storage Systems Department Sandia National Laboratories P.O. Box 5800 Albuquerque, New Mexico 87185-0613 Abstract Sandia National Laboratories, New Mexico, conducts the Energy Storage Systems Program, which is sponsored by the U.S. Department of Energy’s Office of Power Technologies. The goal of this program is to develop cost- effective electric energy storage systems for many high-value stationary applications in collaboration with academia and industry. Sandia National Laboratories is responsible for the engineering analyses, contracted development, and testing of energy storage components and systems. This report details the technical achievements realized during fiscal year 1999. Acknowledgment Sandia National Laboratories would like to like acknowledge and thank Dr. Imre Gyuk of the U.S. Department of Energy’s Office of Power Technologies for the support and funding of this work. Our appreciation is also extended to all of the contributing organizations that have been instrumental in the success of many ESS Program activities in FY99. Contributors Sandia National Laboratories: Akhil, A., Analysis, Substation Power Quality Atcitty, S., Subsystems Engineering, Power Electronics Butler, P., Former ESS Program Manager Clark, N., Subsystems Engineering, Zinc/Bromine Corey, G., Integration and Implementation Crow, T., T. Brown, G. Rodriquez, Technology Evaluation Jungst, R., Subsystems Engineering, Lead-Acid, and System Field Evaluation McIntyre, M., Analysis Rovang, D., Analysis, Substation Power Quality Contractors/Subcontractors: Achenbach, D., F. Russell, Metlakatla Power & Light Anderson, M., M. Crow, University of Missouri — Rolla Bitterly, J., U.S. Flywheel Systems Cole, J., P. Moseley, C. Parker, International Lead Zinc Research Organization Collins, J., Southeastern Electric Exchange, Inc. Day, A., Boeing Dolan, M., B. Haschke, J. Ripple, D. Rivard, Tech Reps Eidler, P., P. Lex, R. Parry, ZBB Technologies, Inc. Francis, I., Just Do IT Frisby, P., D. Lockard, State of Alaska Division of Energy Hammond, R., University of Arizona Huang, A., Virginia Polytechnic Institute Hunt, G., J. Boehm, GNB Technologies Ianucci, J., Distributed Utility Associates Johnson, P., B. Hill, Arizona Public Service Kern, G., Ascension Kovis, J., C. Backus, Brockway Standard Corporation Lovas, T., Chugach Electric Neal, J., S. Lindenburg, National Rural Electric Cooperative Association Norris, B., Gridwise Engineering Company Oweis, S., Saft Roberts, B., E. Camm, S&C Electric Schoenung, S., Longitude 122 West, Inc. Sen, R., C. Namovicz, N. Miller, K. Klunder, Sentech, Inc. Sloan, R.B., Energy United Symons, P., Electrochemical Engineering Consultants, Inc. Taylor, P., L. Johnson, L. Bush, K. Reichart, M. Farber, J. Dunleavey, E. Skolnik, J. Philip, P. DiPietro, Energetics Torres, W., Puerto Rico Electric Power Authority Whisenant, S., Duke Power Co. Whittaker, C., J. Newmiller, Endecon Engineering Winter, R., Powercell Contents DP Executive! SUMMALY osecec<scccsssssccsssesesesessessesees pesasenestsnsnssetesecsed ete eceteneessueeensessstuescuensusenseenssesesrecetresesesesres=s Introduction. Highlights. Integration... System Development. System Evaluation Components Development . Evaluation... Analysis. System Studies Technology Assessments .. SeHOeUYANHYK Se pte bet et T —_ 7 2. Integration... Introduction. System Development .. Advanced Battery Energy Storage System Alternate Renewable Generation and Storage System Designs to Improve Battery Performance..... Small Village Storage System Integration... Renewable Generation and Storage and Related Projects. System Evaluation.................2:::sce00 Mobile Power-Quality System . VRLA Battery Test at Vernon............ Field Test Data Management Project . Field Test Data Management for Grid-connected Systems.. Field Test Data Management for Renewable Generation and Storage for Off-grid Systems...... 2-46 PREPA System Lessons Learned... Peru System Field Test Monitoring. PV/Battery Hybrid Controller Field Test Data Analysis from PV/Hybrid Controller Testing Integration and Testing of Energy Storage with Flexible AC Transmission System Devices RAPS Testing Methods Development. 3. Components.. 3-1 Introduction 3-1 Development.. 3-1 Advanced Energy Storage Development... Battery Simulator Development and Validation.. . 3-14 Valve-Regulated Lead-Acid Reliability Improvement Project 3-15 Emitter Turn-Off Switch Development for Power Conversion System... .. 3-16 Development of Intelligent Controls and Control Strategies for Renewable Generation and Storage . Evaluation .. Valve-Regulated Lead-Acid Battery Evaluation. .. 3-19 ABSOLYTE [IP Testing ............:sseceeeeeeeeeee .. 3-19 ABSOLYTE II Testing .... 3-22 Yuasa Laboratory Testing for Hybrid Environment Intermediate State-of-Charge Testing ..............0000 es System Evaluation of a 200-kWh Zinc/Bromine Battery at “The Barns” ...........:.ccscsesesesesesesesnseeeeenes 3-28 CONTENTS 7 4. Analysis Contents (continued) Introduction System Studies . Opportunities Analysis—Phase II Value of Storage for Restructured Utility Industry Utility Operating Cost Analysis ............. International Energy Agency Annex IX. Transmission Power Quality Study . Technology Assessments ...........:.0:0+00000+ ss Performance and Economic Analysis of Superconducting Magnetic Energy Storage, Flywheels, and Compressed Air Energy Storage Systems Project. Long- Versus Short-Term Storage Study .............:::cscs0000+ Power Conversion System Magnetics and Functionality Analysis EESAT Conference WB ROG CL ONCES coccsccecccocacecscosesecseccececovesecesceccscscaccesscesseestecelorceccerseteenessccspecesroceTeucTecaeuuuseeCceceseceteceesaseeTatesDlTeacsecveceseeetee 5-1 Appendix A: Data From Vernon BESS Operations . Appendix B: Savings Due to BESS Only Appendix C: Presentations and Publications ...............ssssssessseseensescesnenesesssnenenesencsneassnenssnsacensncnssnneneneaceesessesaces li CONTENTS 2-10. = 2-12. 2-13. 2-14. 2-15. 2-16. 2-17. 2-18. 2-19. 2-20. 201 DOD: DiI: 2-24. D5: 2-26. 2-27. 2-28. 2-29. 2-30. Di. 2-32. 2-33. 2-34. 2-35. 2-36. 2-37. 2-38. 2-39. 2-40. 2-41. 2-42. 2-43. 2-44. 2-45. 2-46. 2-47. 2-48. 2-49. 2-50. Figures Schematic of 50-kWh Battery Module............ Battery Module DC Interconnection Scheme.. Connection between PCS and System Controller... Load Profile for Distribution Circuit 2594 for Beach Substation — June 26, 1998 to July 1, 1998.. The Effect of Temperature on kVA for the Beach Substation.... Load Profile for Brewer Circuit during June 1999....... Cost Breakdown for 400-kWh Zinc/Bromine ABESS. Regional Map Showing the Three Selected Villages Relative to Other Major Cities and Villages in ALAS kAlstecneenereseteatsasreereneressrsresrarentuersessrermerttes Chistochina Power Smoothing Fall and Winter. Chistochina Power Smoothing Fall and Winter, 45-kW Diesel. System Performance for One Week in Selawik — Fall and Winter. Fuel Price vs. Power Cost. System Design (approximately to scale). . System Electrical Schematic. Mobile PQ2000 in Transit....... The Mobile PQ2000 Installed and Operating at S&C Electric Company Average and Peak Plant Demand at Vernon for the Months of July, August, and September 1999... MP&L Power Plant Output Including BESS between 9 and 10 a.m. on July 1, 1998. BESS Operation between 9 and 10 a.m. on July 1, 1998. .0.....eeseeeeseeseseseeeeeseeeeseeneeees Brockway Standard Lithography Plant Layout of PQ2000 Installation (not to scale).. PQ2000 Located behind Brockway Plant. PQ2000 Host Computer. ..... PQ2000 Industrial Processor. . PQ2000 Discharge Events by Season in 1998 PQ2000 Discharge Events by Time of Day.............cecesesessssseseseseseseeeeeeeeeees Brockway PQ2000 Discharge Ampere-hours Comparison in June 1998 Brockway PQ2000 DC Current Versus Time in 1998. .............. Brockway PQ2000 Charge and Discharge Comparison in 1998... Brockway PQ2000 Normal Versus Duty Cycle Charges in 1998. OGPS at Wawona Point in Yosemite National Park, Consisting of a 1080 PV Array, 16-kWh/24-V VRLA BESS, a 5-kW Propane Generator, and a 4-kW Inverter. Wawona Point Daily Battery Current...........0..0.2.. Wawona Point Daily Discharge Amp-hour Error... Wawona Point Sample of Rapid Charge and Discharge Cycling Current.. Wawona Point Average Range of Daily SOC. Location of RGS at Grasmere. ...............-000+ Grasmere Battery SOC Based on Raw Data. Grasmere Battery SOC: Minimum Six Hours without Reversal. . Grasmere Battery Degree of Overcharge and Undercharge. .. Grasmere Battery Operating Temperatures over 30°C.... Grasmere Battery Operating Temperatures over 35°C. Battery Storage Building in Puerto Rico. Project Timeline and Key Milestones...... BESS Rapid Reserve Response on May 2, 1 Disconnecting Cells...... Removing Failed Cell Installing Good Cell. .... Battery String Voltage vs. Time, May 2 through 8, 1999 (Julian 122-128). . Battery String Amp-hours and % DOD vs. Time, May 2 through 8, 1999 (Julian 122-128). Block Diagram of STAR Solar Hybrid Facility. ..............cccccccsscscsessesecesesescsescseseseseeeseeeeeeeee CONTENTS iii 2-51. 2-52. 2-53. 2-54. 2-55. 2-56. 2-57. 2-58. 2-59. 2-60. 2-61. 2-62. 2-63. 2-64. 2-65. 2-66. 3-1. 3-2. 3-3. 3-4. 3-5: 3-7. 3-8. 3-9. 3-10. 3-11. 3-12. 3-13. 3-14. 3-15. 3-16. 3-17. 3-18. 3-19. 3-20. 3-21. 3-22. 3-23. 3-24. 3-25. 3-26. 4-1. 4-2. 43. 4-4, Figures (continued) Power Flows for Components in STAR Hybrid Tests. . UMR Statcom/BESS System Layout. ............. Battery Rack (left) and DAS (right) at UMR. . StatCom Front Panel at UMR.. UMR StatCom Sensor Rack... UMR StatCom Converter Rack. . Typical Sine-triangle PWM. . Modified SPWM. ............4++++ Relationship between Modulation index k and StatCom/BESS Output Voltage... Measured Results for Active Power Control...... Measured Results for Reactive Power Control One-Machine-Infinite-Bus Power System with StatCom/BESS. Generator Rotor Angle. ............:::::00000++ StatCom/BESS Terminal Bus Voltage. Active Power in Line 2 REACH Ve) PO Weds IN] LGRING 2 ofeceenceesesceenectrseasnencsnssearsnerstansnsastmtetesttrertetensvacestrssstesrerneecsmensecaee Capacity Test (100% DOD) Results for Small Saft Cells Cycling at 10% DOD at 40°C. Typical USFS Flywheel—Cutaway View. USFS Concept for Utility FES Building with 1.15-MWh Energy Storage Capacity. . USFS Utility Flywheel Multi-MW Footprint Parameter Study Energy Output per Gallon of Diesel Fuel vs. Fractional Utilization of Generator Capacity. = Boeing Conceptual Design of Flywheel System for Target Application...............:::cesecesseseeseesssesesesteeeeeees Rim Weight of Thin-Rim Flywheel vs. Specific Strength (yield strength divided by density) for Steel and Composite Flywheel Rims Delivering 6 kWh of Useable Energy. Useable Energy Accounts for 75% DOD, 20% Power Electronic Losses, and Strength Derating of 2.0 for Fatigue and actor, Of SAtet yi ccttecescctesscactesecsveressssres tussccueeswesteesossseeesoaceeesssoaeanssovensstcvensess rar ccesveestesavessontsensoseneteese 3-13 Power Consumption in Magnetic Bearings. Boeing HTS Material was Tested in Bearings at Argonne National Laboratory; Comparison is to Active Magnetic Bearing Data.. Picture Of ETO1045HS) Model os ccttececeseeessssensess Boost Converter Based on a Standard Phase Leg. Photograph of Completed Phase Leg. .. An Overall View of the Converter..... Waveforms of ETO System Running at 100 kW. ABSOLYTE IIP Capacity vs. Cycle Number for FY99.. ABSOLYTE IIP Capacity vs. Date, Showing Periods of Open Circuit. EOD Cell Voltages for the ABSOLYTE IIP—Cell 3 is Bypassed......... ABSOLYTE IIP Boost Charge and Capacity Cycles, After EOL Reached at Cycle 699. . Voltages Measured and Calculated Using a Cadmium Reference Electrode in Cell 7 of the ABSOLYTE IIP. ABSOLYTE II Capacity vs. Cycle Number for FY99. ABSOLYTE II Capacity vs. Date, Showing Periods of Open Circuit EOD Cell Voltages for the ABSOLYTE II. Cell 4 was Bypassed after Cycle 446. Nasal Exide DG Xs Capacity FiStonysteccscenesononsteeceeserr snes setersecterercerereccecemcctesterereeets The Barns Implementation Plan. ... PowerBlock Performance Data. PowerBlock’s Response to Grid Failure.. PowerBlock State-of-the-Art Power Electronics A Map of Annual Average Wind Power, along with Wind Power Classifications . Average Annual Wind Power in Missouri Average Wind Speeds at Anemometer and Hub Height for a Representative Day in January... 4-12 Average Power Generated by the Z-50 Wind Turbine Generator (WTG) on a Representative Day in January. pA veraging, Method iw asi Used: cooctercencsereneastessconsese sore snsecocasepeecetnecesasetcececasescocetersessstersersa eters) 4-13 iv CONTENTS 4-5. 4-6. 4-7. 4-8. 4-9. 4-10. 4-11. 4-12. 4-13. 4-14. 4-15. 4-16. 4-17. 4-18. 4-19. Figures (continued) Operating Cost Savings with a BESS Used for Spinning Reserve Only by KCPL (1995) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). . Operating Cost Savings with a BESS Used for Spinning Reserve Only by KCPL (1996) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). . Operating Cost Savings with a BESS Used for Spinning Reserve Only by KCPL (1997) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). . Operating Cost Savings with a BESS Used for Load Leveling Only by KCPL (1995) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). . Operating Cost Savings with a BESS Used for Load Leveling Only by KCPL (1996) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). . Operating Cost Savings with a BESS Used for Load Leveling Only by KCPL (1997) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). . Operating Cost Savings with a sed for Load Leveling with Spinning Reserve by KCPL (1995) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations)... Operating Cost Savings with a BESS Used for Loa veling with Spinning Reserve by KCPL (1996) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations)...............cceesesseseseeeesesesecsesesseseseceesescsusscstsaceeecucecscateneeeeaeeneneaeees 4-17 Operating Cost Savings with a BESS Used for Load Leveling with Spinning Reserve by KCPL (1997) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations)................csececesesesesesesesesesesesessesessscscscseseseseersescsescaearanenenseeeeesess Plot Showing the Modeled Benefit of Mitigation at Different Levels... The Modeled Cost of an Energy Storage System. Cumulative Benefit/Cost. ..........:cesceeeeseeeeeeeeees The 13-State Region Under Study Defined as the Southeastern U.S. ... Power and Energy Characteristics of Energy Storage Products. . Response Characteristics Of ESSS. ..........:cccsssesesseseseeeteeeeseeeeneees w 4-14 we 4-15 4-15 -- 4-16 «4-16 CONTENTS v 2-1. 2-2. 2-3. 2-4. 2-6. 2-8. 2-9. 2-10. 2-11. 2-12. 2-13. 2-14. 2-15. 2-16. 2-17. 2-18. 2-19. 2-20. 2-21. 2-22. 2-23. 2-24. 2-25. 2-26. 2-27. 2-28. 2-29. 2-30. 2-31. 2-32. Tables 400 kWh Zinc/Bromine Battery Summary Specifications PCS Summary Specifications..............000+ Beach Substation Distribution Circuit 2594 ( Data for Summer Load Circuits Performance of 50-kWh Zinc/Bromine Battery With and Without the Control System and Pumps Powered by the Battery Stacks Effect of Module Modifications on Performance... X-ray Spectroscopy Results for Separator Materials. Alaskan Village Diesel System General Overview- Evaluation Results Summary of BESSs That Were Modeled a Summary of Current Quoted Price for Similar Equipment, and Projected Future Costs for Systems at Chistochina. Summary of Results from Three Alaska Village Analyses . Prototype AC Lighting System Specifications ... Prototype System Controller Specifications Prototype System Specifications ... Prototype Design Specifications... Proposed Demonstration Sites...........::0:se:esesee 2-28 July 1999 Data from Vernon BESS Operations . August 1999 Data from Vernon BESS Operations September 1999 Data from Vernon BESS Operations . PVAIADASS EAT ANINCUCTS seers terteresttretteserstrsrereeee File Descriptions of Electronic Deliverable. Overall Cost and Benefits from BESS in Metlakatla System Cumulative Values for PQ2000 Operation at Brockway: December 1996 through July 1999 Brockway PQ2000 Downtime Causes and Duration Summary of Glitches Avoided by Brockway PQ2000: December 1996 through July 1999 Wawona Point OGPS CSI 21X Datalogger Input Channels .. Wawona Point OGPS Database Field Description................. Grasemere Battery Parameters ...............:00+ PREPA BESS Costs by Components ($M). Comparison of Yuasa Battery Manual Cell Readings: April to June 1999 STAR Hybrid Test Facility 30-kW Trace Inverter Set Point Log . STAR Yuasa Battery Bank Capacity Test Comparison ...........- Energy Density of Li-ion Batteries Compared to Other Batteries . Off-grid Storage Applications, Their Requirements, and Potential Markets to 2010 According to BOGAN G trsssecesceseconssnseeeetentcntstrscstecstigrenssecaccscecacacccesess cescseccesacoscccrscrecssssastesaceaccececvecterereeterzesetiss<s sx TUTTTATST 3-10 Boeing Preliminary Specifications for Storage Component of a 12-kW Wind-Generation System with Backup and Peaking Capabilities ... ABSOLYTE IIP and ABSOLYTE II Test Regime.. Yuasa Exide Dynacell DGX Life History Table...... Module and String Data for Batteries on Test-Type A Module and String Data for Batteries on Test-Type B Module and String Data for Batteries on Test-Type C ... Definitions and Categories of Electric Power Applications of ESSs. Definitions for Applications Requirements ................. Summary of Significant Applications Requirements .. Capability of Technologies to Serve Electric Power Applications.. 4-6 Operating Cost Savings Range ($K/yr) Estimates for KCPL (DYNA. B Monte Carlo, 12 Iterations) .............:cscescesessesssessesscscescssescesessecseescessessesseaeesseseeaeeaeesceasnsesssseseseececssasaencae 4-8 Operating Cost Savings Range ($K/yr) Estimates for KCPL (DYNASTORE Simulation Method: Monte! Carlo; 2; terati Ons) x ssozcxscaseecaconceesossssusussessatecatcvesseqevsrszostcoscrestasesensasenstsessoensinensosqeovess soso easainat 4-13 vi CONTENTS 4-7. 4-8. 4-9 4-10. 4-11. 4-12. 4-13. 4-14. 4-15. Tables (continued) Power Quality Installation Application Utility Substation Application Specifications... Initial Cost Breakdown for a 2-MW, 2-sec Power Quality Application Initial Cost Breakdown for a 2-~MW, 2-MWh Substation Application Potential Participating Utilities Energy Storage Technologies Costs and Efficiencies . Comparison of Commercial Maturity and Cost Certainty for Energy Storage Technologies . Applications, Characteristics, and Appropriate Technology . EESAT 2000 Conference Schedule - 4-21 4-21 4-23 4-23 4-28 4-31 4-32 4-34 = 4-38 CONTENTS vii Acronyms and Abbreviations ABESS advanced battery energy storage system AES Advanced Energy Systems (a company) or advanced energy storage AGM adsorbed glass mat AOP Annual Operating Plan APC Applied Power Corporation APS Arizona Public Service ASD adjustable-speed driver ASU Arizona State University AVEC Alaska Village Electric Cooperative BES battery energy storage BESS battery energy storage system BEWAG Berliner Kraft und Licht BOP balance of plant CAES compressed air energy storage CAS compressed air storage (ae complete capacity cycle CMC Complete Meeting Concepts Ce. In reference to PV model CPES Center for Power Electronics Systems CRADA cooperative research and development agreement CSI Campbell Scientific Inc. CSV comma-separated variable CE current transformer CVCI constant voltage, constant current D/A digital/analogue DAS data acquisition system di/dt delta current/delta time dv/dt delta velocity/delta time DC direct current DLC double-layer capacitor DOD depth of discharge DOE U.S. Department of Energy DPQ Distribution Power Quality DUA Distributed Utility Associates EECI Electrochemical Engineering Consultants, Inc. EESAT Electrical Energy Storage Systems Applications and Technologies EOD end of discharge EOL end of life EMC electric membership cooperative EPRI Electric Power Research Institute ESA Energy Storage Association ESD electronic selector device ESS energy storage system ETO emitter turn-off thyristor EU Energy United FACTS flexible AC transmission systems FES flywheel energy storage RY, fiscal year GE General Electric GNB GNB Technologies, Inc. GTO gate turn-off thyristor-oriented HEV hybrid electric vehicle vill ACRONYMS AND ABBREVIATIONS Acronyms and Abbreviations (continued) HTS high-temperature superconductivity vo input/output IEA International Energy Agency TEEE Institute of Electrical and Electronics Engineers IGBT insulated gate bipolar transistor IGCT integrated gate commutated thyristor ILZRO International Lead Zinc Research Organization IPC Idaho Power Company ISO independent system operators ISOC intermediate state of charge KCPL Kansas City Power and Light Co. KPU Ketchikan Public Utility kVAR kilovolt ampre reactive kW kilowatt LANL Los Alamos National Laboratory LBX an operating mode LED light emitting diode Li-ion Lithium-ion LL load leveling LL+SR load leveling with spinning reserve LVD low-voltage disconnect LVDR low voltage disconnect and reconnect MB magnetic bearing MG motor/generator MOU memorandum of understanding MP&L Metlakatla Power and Light MW megawatt NASA National Aeronautics and Space Administration NCDC National Climatic Data Center NEMA National Equipment Manufacturer’s Association NERC National Energy Reliability Council NIST National Institute of Standards and Technology NOAA National Oceanographic and Atmospheric Administration NRECA National Rural Electric Cooperative Association NREL National Renewable Energy Laboratory O&M operating and maintenance o/c open circuit OGPS off-grid power system OPT Office of Power Technologies ORNL Oak Ridge National Laboratory PAR project authorization request PC pilot cell PCE power cost equalization PCL premature capacity loss PCS power conversion system PEBB Power Electronic Building Blocks PG&E Pacific Gas & Electric Company PI proportional integral PLC programmable logic controller PNM Public Service Company of New Mexico POA plane of array PQ power quality PQDIF power quality data interchange format ACRONYMS AND ABBREVIATIONS Acronyms and Abbreviations (co ntinued) PREPA Puerto Rico Electric Power Authority PS peak shaving PSEL PV System Evaluation Laboratory PT potential transformer PV photovoltaic PV4U Photovoltaics for Utilities PWM pulse width modulation R&D research and development RAPS remote area power supply RFI request for information RFP request for proposal RFQ request for quotation RGS renewable generation and storage rms root mean square rpm revolutions per minute SAMSOM Solar and Meteorological Surface Observational Network SCADA supervisory control and data acquisition SCC Standards Coordinating Committee SCE Southern California Edison SEE Southeastern Electric Exchange SEM scanning electron microscopy SES a company (manufacturer) SMES superconducting magnetic energy storage SNL Sandia National Laboratories SOA safe operating area Soc state of charge SOW statement of work SPI Superconducting Partnership Initiative SPQS Substation Power Quality System SPWM sine-triangle pulse width modulation SR spinning reserve STAR Solar Test and Research StatCom Static Synchronous Compensator TBD to be determined T&D transmission and distribution Tc critical temperature TGA thermogravimetric analysis TPQ transmission power quality UE&C United Engineers & Constructors UL Underwriters Laboratory UMR University of Missouri-Rolla UPFC unified power flow controller UPS uninterruptible power supply USFS U.S. Flywheel Systems VAR volt amp reactive Voc volt open circuit Vr disconnect voltage VRLA valve-regulated lead-acid Vir reconnect voltage WTG wind turbine generator XPS x-ray photoelectron spectroscopy XRD x-ray diffraction ZBB ZBB Technologies, Inc. x ACRONYMS AND ABBREVIATIONS Energy Storage Systems Program Annual Report for Fiscal Year 1999 1. Executive Summary Introduction The U.S. electric utility industry is undergoing revolutionary change as a result of deregulation and competition, seeing limitations on installing new con- ventional generation and transmission and distribution (T&D) equipment and greatly reduced resources for research and development (R&D). The United States Department of Energy (DOE)/Office of Power Tech- nologies (OPT), through its Energy Storage Systems (ESS) R&D Program at Sandia National Laboratories (SNL), continues to work cooperatively with the elec- tricity supply industry and the manufacturing sector to develop energy storage systems that will play a vital role during and after this transition period. In doing so, the ESS Program is furthering the goals of the DOE by developing technology that can be used by industry to (1) strengthen the nation’s energy security in terms of electricity supply, (2) reduce the environmental impact of electricity generation and T&D through the increased use of renewables, and (3) increase the global economic competitiveness of U.S. industry with more reliable, higher quality, and low cost electricity. The ESS Program is conducting focused R&D, leveraged by U.S. industry, to make possible the wide- spread use of energy storage systems for electricity system applications and for renewable generation appli- cations. Its goal is the development of new energy storage systems with superior performance and higher energy densities at competitive prices. The program includes research on a portfolio of storage technologies such as batteries, flywheels, and superconducting mag- netic energy storage (SMES). The ESS Program balances the R&D of promising new technologies with focused analytical and educa- tional tasks. The ESS Program is organized into three interrelated elements: e Integration — System Development — System Field Evaluation e Components — Development — Evaluation e Analysis — System Studies - Technology Assessments Highlights The FY99 Annual Operating Plan (AOP) for the ESS Program was used as a guide for program imple- mentation. The Mobile PQ2000 highlighted the hard- ware testing projects by being successfully moved 600 miles and set up for operation at an industrial polymer processing plant in five calendar days. Three signifi- cant new projects were initiated. The ESS Program is teaming with the State of Alaska Energy Authority, Chugach Electric Cooperative, other Alaska utility or- ganizations, and the National Rural Electric Coopera- tive Association (NRECA) to design, build and demon- strate a 20- to 30-kW energy storage system for small Alaskan villages. The development of a high power semiconductor switch and emitter turn-off (ETO) thy- ristor, which began late in FY98 with Virginia Tech, shows significant promise of increasing the capability of the power conversion system (PCS) for energy stor- age systems. A transmission power quality (TPQ) 1. EXECUTIVE SUMMARY 1-1] study was initiated with the Southeastern Electric Ex- change (SEE) to study the power quality of the existing transmission system in the Southeastern United States to provide data necessary for design of future energy storage systems. The program continued its interna- tional relationships with the completion of the Interna- tional Energy Agency (IEA) Annex IX work, participa- tion in the Sixth International Conference on Batteries for Utility Energy Storage, and continued cooperation with the International Lead Zinc Research Organization (ILZRO) on valve-regulated lead-acid (VRLA) battery reliability and remote area power supply (RAPS) initia- tives. The following sections describe the key highlights of the program for FY99. Integration System Development Advanced Battery Energy Storage System (ABESS): The zinc/bromine battery is an emerging technology that has many attributes that make it attrac- tive for energy storage applications. The main advan- tage of the zinc/bromine battery system is good gra- vimetric energy density, which results in a modular, transportable battery system with sufficient capacity to be placed anywhere on the utility grid. The battery is made almost entirely from plastic (high-density poly- ethylene), which makes it cost-competitive with lead- acid batteries at high-volume production without the hazardous manufacturing and recycling concerns. Also, the battery operates efficiently over a wide temperature range, functions under intermittent charge/discharge conditions, and can experience complete discharge hundreds of times without damage. The objectives of the ABESS project are to design, fabricate, evaluate, and optimize a zinc/bromine battery system suitable for electric utilities. The soundness of the battery technology was demonstrated, and new larger cell stacks, designed for an electric utility battery, were developed during previous contracts between ZBB and SNL. The end product of the present contract is to demonstrate a 400-kWh system at a utility installation. Based on the results of this testing and utility interest, larger systems may be tested in the future. In FY99, a 50-kWh module was successfully tested to demonstrate the operation of a single module con- nected to the controller with and without utility power. This strategy will simplify the design of the control for the battery system by enabling it to function as an un- interruptible power supply or as a RAPS. Alternate Renewable Generation and Storage (RGS) System Designs to Improve Battery Perform- ance: An investigation of alternative configurations to optimally use lead-acid batteries in renewable hybrid systems began in the third quarter of FY98. The pur- pose of this project is to devise alternative configura- tions for renewable hybrid systems that will allow more optimal charging of the lead-acid batteries in these sys- tems. The effort during FY99 involved building increas- ingly sophisticated versions of the breadboard hardware and then testing this hardware with more fully devel- oped versions of the software required for its operation. When the breadboard was first developed, testing could be performed with cycles at a shallow depth of dis- charge (DOD). As the development progressed, how- ever, it was necessary to test it with cycles of 50% DOD, so that the development would more closely mimic lead-acid batteries in the field. In the latter part of the second quarter and the re- mainder of FY99, the alternative configuration system was operating more or less continuously. Operation during this time consisted of automated cycling with discharges and charges each lasting approximately seven hours in duration. In the last few weeks of FY99, the cycling was continuous, under automated control, with almost no interruptions for weeklong periods. Small Village System Integration: Chugach Electric Cooperative, the State of Alaska Energy Authority, other Alaska utility organizations, and the NRECA in collaboration with the ESS Program are designing, building, and demonstrating a 20- to 30-kW energy storage system for small Alaskan villages. The purpose of this integration, test, and field evaluation project is to quantify the benefits of battery storage for small villages that have diesel-only generation. This study is collaborating with a similar study by ESS and the U.S. Agency for International Development. A system feasibility study was initiated in FY99. If ap- propriate, a prototype system will be assembled and undergo initial testing and debugging at a test bed in Anchorage under controlled conditions. The feasibility study was initiated in June 1999. A contract was placed with Sentech, Inc., to conduct the study and prepare a report describing the results. The SNL and Sentech study team visited Alaska and had discussions with the director and staff of the Alaska Energy Authority, with Chugach Electric Cooperative staff and management, and with Alaska Village Electric 1-2 1. EXECUTIVE SUMMARY Cooperative (AVEC) personnel. The perspectives of each organization were obtained, and there was support for the project. The objectives of the Alaskan organi- zations for this project are to find ways to use advanced technology to save diesel fuel and associated costs, re- duce pollution and environmental threats from fuel transportation and storage, and improve the economic conditions of the more than 200 remote communities in Alaska. Renewable Generation and Storage and Related Projects: In FY98, the ESS Program issued a request for proposal (RFP) for the first phase in this project. Contracts were awarded to three companies to conduct three-month studies that focused on integrating energy storage systems with renewable energy generation. These studies were completed in FY98 and the results are summarized in SNL’s ESS Program Report for FY98, SAND98-0883. In FY99, the final reports for each of the three contracts were published: e AeroVironment, Inc.—SAND99-0936, Solar- powered Systems for Environmental Remediation e Solarex—SAND99-1477, Investigation of Synergy Between Electrochemical Capacitors, Flywheels, and Batteries in Hybrid Energy Storage for PV Systems e Ascension Technology, Inc.—SAND99-0935, System and Battery Charge Control for PV- powered AC Lighting Systems The ESS Program determined that the results for Phase I were satisfactory enough to justify moving to Phase II of the project. Consequently, a second RFP was issued soliciting bids for follow-on work. Three proposals were received and evaluated by ESS Program staff and staff from the Photovoltaics Program at SNL. One contract for Phase II work was awarded to Ascen- sion Technology, Inc., now a division of Applied Power Corporation (APC). Ascension Technology’s final re- port for the Phase II work was submitted at the end of FY99 and will be published in FY2000. The results of Ascension’s Phase II work, as documented in the final report, are summarized in Chapter 2 of this report. System Evaluation Mobile PQ2000: On May 13, 1999, the Mobile PQ2000 (2 MW for 15 seconds, trailer-mounted sys- tem) was removed from the Virginia Power Iron Bridge facility in Richmond, Virginia, and began its journey to the S&C Electric Company plant in Chicago, Illinois, where it will undergo field testing. The unit, which was installed by S&C to protect its polymer product fabri- cation plant, arrived on Friday, May 14, 1999, and was on line and fully operational by Tuesday, May 18. VRLA Battery Test at Vernon: In the beginning of FY96, a contract was awarded to GNB to perform testing over a four-year period of a final deliverable from a development contract. Costs for the test are being shared by GNB at 50%. The 250-kW/500-kWh deliverable battery has been incorporated into a 3.5- MW/3.5-MWh battery system that was installed by GNB at its lead-recycling center in Vernon, California. To match the rest of the battery system planned for Vernon, and because of its established production capa- bility, the battery design chosen as the deliverable was the ABSOLYTE IIP. The primary application for the battery system at Vernon is to provide emergency backup power to critical loads at the facility dealing with environmental (air emission) controls. It is also being used in a peak-shaving mode for demand reduc- tion that will lower electricity demand charges for the facility and take advantage of lower off-peak energy costs. The battery system has been in operation since the first quarter of FY96. Tests of the peak-shaving mode began in July 1997 and continued in FY98 and FY99. A primary objective of these tests is to find the opti- mum trigger power level for the battery system to sup- ply the plant demand during peak shaving while the battery energy storage system (BESS) continues to per- form its primary function of supplying backup power for emissions equipment at the lead smelter. At the beginning of FY99, practical engineering problems with ground-fault detection circuits that were an issue at Vernon in the past were resolved so that ground-fault alarms were no longer continuous. The peak shaving trigger level was progressively decreased during the winter months until it reached 1275 kW in April. The optimum winter peak shaving setting is estimated to be near 2925 kW. In May, the trigger was raised to 3100 kW to accommodate the utility’s longer summer season on-peak demand. Choice of BESS peak shaving settings was handed over to the local engineering staff at Vernon in June 1999. Therefore the 3100-kW setting was maintained for several months to allow the local engineering staff to gain more understanding of the facility's on-peak demand characteristics. Peak shaving will begin in the first quarter of FY2000 at this same aggressive level, with anticipation of going lower as the winter season approaches. Finally, in August the Vernon BESS interface computer was replaced to ensure Y2K 1, EXECUTIVE SUMMARY compliance. A new Windows NT-based system was installed with upgraded software and hardware components. The new interface computer passed all Y2K date tests successfully. In September 1999, the peak shaving trigger setting was lowered to 3025 kW. During summer on-peak hours, this setting was expected to be near optimal level of peak shaving at which maximum cost savings can be attained while still maintaining the BESS within the proper sate-of-charge (SOC) window to continue peak shaving on any given weekday of the month. Field Test Data Management for Grid- Connected Systems and for RGS Off-Grid Systems: In FY99, SNL initiated a project to develop a database for fielded battery energy storage systems for off-grid, stand-alone and grid-tied systems that are currently in operation. The purpose of the database is to provide information on the operation and management of these existing systems in order to learn how to design future systems and avoid problems identified from this evaluation. The database is intended to provide practi- cal field performance data for potential customers. Four contractors were selected for the initial phase of the project, two for off-grid analysis and two for on- grid analysis. Their first task was to identify a site that would provide historical data for an energy storage system in operation that met the criteria for analytical tasks called out in the contract. Secondary tasks were also defined to determine the adequacy of the data ac- quisition system (DAS) used for the fielded systems, determine the adequacy of the battery management strategy applied to the fielded system, and to make rec- ommendations for ways to increase the reliability of the DAS for fielded systems. Four sites were selected and the data obtained and analysis begun by year’s end. PREPA Systems Lessons Learned: A 20-MW, 14-MWh BESS was installed in June 1994 at the Sabana Llana substation near San Juan, Puerto Rico, primarily to mitigate under-frequency load shedding. The owner-operator, Puerto Rico Electric Power Authority (PREPA), and the architect/engineer and supplier firms learned many valuable lessons, from planning through the first four years of operations, about the largest “commercial” battery system in op- eration in an electric utility application. PREPA col- laborated with the ESS Program to compile a “lessons learned” report for the facility. The report, Lessons Learned from the Puerto Rico Battery Energy Storage System (SAND99-2232), published in September 1999, is available in both English and Spanish. Peru Small Field Test Monitoring: In July 1997, the ILZRO, the Energy Storage Association (ESA), the Solar Energy Industries Association, and the country of Peru signed a memorandum of understanding (MOU) for a collaborative project. The objective of the first phase of the project was to perform a feasibility study for photovoltaics in combination with storage, power electronics, a genset, and controls for use in remote villages in the Amazon River valley. The feasibility study included a discussion of sys- tem and component sizing, performance and economic benefits. The cost of the study was shared by the MOU signatories and the ESS Program. Results of the ap- proximately six-month study are documented in an ILZRO report. The MOU signatories have now started the hardware demonstration phase and are pursuing funding opportunities. Photovoltaic/Battery Hybrid Controller Field Test: The evaluation of the first photovoltaic (PV) battery hybrid controller at the SNL PV System Evaluation Laboratory was completed in FY95. A project plan was developed that called for the installa- tion of the second prototype control unit at a site that could provide at least a 15-kW PV array, a 200-kWh battery, 30-kW genset, and a variable load that would simulate the electric power consumption of a small vil- lage. A multiyear operational test plan was success- fully negotiated with Arizona Public Service (APS) to perform a complete system field test of the controller at the APS Solar Test and Research (STAR) Center in Tempe, Arizona. The overall objective of the DOE/ SNL project at STAR is to develop an operational strat- egy for solar hybrid systems, which include energy storage, that will maximize performance and minimize life cycle cost of these systems. An MOU and a loan agreement were finalized in FY97. The MOU provided for the loan of a state-of-the-art inverter/controller and a DAS for up to three years. APS obtained a tubular gel valve-regulated lead-acid battery under a special cost sharing agreement with Yuasa, Inc., to expand the test results to determine whether the gel technology battery was best suited for the hybrid operational environment. A special hybrid test facility was constructed at STAR to house the battery, inverter/controller, and natural gas genset. The STAR Center, with nominal loads of 2 to 35 kW, provides the load for the test program. Two multiyear contracts were placed in FY98 to provide operation and maintenance for the DAS and in-depth analytical support for data generated by the system. By the end of FY99, after more than two years of service at the STAR Center, the consensus is the Yuasa DGX85-11 VRLA-Gell batteries have performed well. 1, EXECUTIVE SUMMARY Data Analysis from PV/Hybrid Controller Testing: Following a competitive bidding process, which began late in the first quarter of FY99, a contract for analytical support was awarded in early February to Electrochemical Engineering Consultants, Inc. (EECI), of Morgan Hill, California. EECI was given the task of providing analyses of data collected at the APS STAR Center for the Hybrid Controller Field Test Project. The objectives for the data analysis that is under way at EECI include: 1. Assist SNL and APS personnel in fine-tuning the Trace controller so that the electricity loads are supplied reliably and efficiently. 2. Ensure that the Yuasa cells in the energy storage battery at STAR are being maintained, repaired, and operated in a way that will maximize life and performance. 3. Make a preliminary determination of the usefulness of gel VRLAs in the solar hybrid application. The work performed by EECI, beginning on Feb- ruary 4, 1999, has met all three of the objectives for this project. For the first two objectives, this progress is indicated by the results from the testing of the solar hybrid system at STAR. Integration and Testing of Energy Storage with Flexible AC Transmission System Devices: Flexible AC transmission system (FACTS) devices offer in- creased flexibility in decentralized control of transmis- sion systems. As the vertically integrated utility struc- ture is phased out, centralized control of the bulk power system will no longer be possible. Transmission pro- viders will be forced to seek a means to get local con- trol to address a number of potential problems such as: e Uneven power flow through the system (loop flows), e Transient and dynamic stability, e Subsynchronous oscillations, and e Dynamic overvoltages and undervoltages. Adding energy storage to FACTS devices will give them active (real) power capability. Several significant advances were made in the last quarter of FY99. The most important progress has been the incorporation of the BESS into the static synchronous compensator (StatCom) setup. The StatCom is a well-known shunt- connected, commercially available FACTS device. The University of Missouri-Rolla (UMR) project team has designed and built a StatCom and has also designed the topologies necessary to integrate a BESS into a Stat- Com. The incorporation was delayed because of prob- lems with premature failure of several of the battery modules. This was corrected and the BESS is now fully functional. The StatCom/BESS system has been successfully operated in synchronism with an external AC source. RAPS Testing Methods Development: The ESS Program continues to coordinate with ILZRO on sev- eral collaborative projects. Each project is being co- funded by ESS and by ILZRO at varying levels of cost sharing ranging from 50-50 to 80-20, with ILZRO tak- ing the lead in some projects and ESS in others. A project to define international standard test-cycle regimes for RAPS has scheduled goals for drafting test procedures and reviewing the procedures. During FY99, the RAPS testing methods develop- ment project evolved significantly and began to cement relationships with standards organizations. The goal is to prepare documents that will lead to RAPS design and certification guidelines. With major involvement of the project’s co-sponsor, ILZRO, and Energetics, contrac- tor to both SNL and ILZRO on this project, substantial planning was done, documents were prepared, and standards organization interfaces were established. A major decision was made to integrate this project with the existing infrastructure of the Institute of Electrical and Electronics Engineers (IEEE) Standards Coordi- nating Committee (SCC) 21, Fuel Cells, Photovoltaics, Dispersed Generation, and Energy Storage. This com- mittee has several standards documents already in preparation or approved that will contribute to RAPS design, testing, and certification guidelines. The docu- ments in process all relate to PV energy systems with battery storage used in a stand-alone mode (no utility connection). SNL, ILZRO, and Energetics will con- tribute to the PV/battery documents, and will prepare new guidelines for RAPS systems—those that use a renewable generation resource, energy storage, and a fossil-fueled generator in a stand-alone mode. Components Development Advanced Energy Storage Development: In FY97, the ESS Program started to analyze advanced storage technologies (SMES, flywheel, double-layer capacitors [DLCs], etc.) and determine the current state J, EXECUTIVE SUMMARY 1-5 of the art of each technology for both capability and cost. The overall approach to R&D for all advanced technologies will be to (1) identify the industry drivers for each technology, and (2) estimate the compatibility and value of these storage technologies for each of the 13 storage applications (identified in the Phase I Op- portunities Analysis performed in FY94) as well as for the combined applications. The goal of the advanced energy storage (AES) component development project is to support the im- provement of AES components. Commercial battery technologies are not included in the scope of this proj- ect, but advanced batteries in development are. In FY99, the ESS Program issued an RFP and funded three companies (Saft, U.S. Flywheel Systems [USFS], and Boeing) to conduct the first phase of a possible multi-phase research project to develop and test components that are large enough to be used in field demonstrations. In Phase I, contractors identified target applications and characterized the applications in terms of benefits and requirements. The contractors then developed storage component specifications and conceptual designs for the most beneficial applications based on state-of-the-art technology. Both prototype and production-scale costs for the technology were then estimated. Finally, contractors created a component development plan and schedule to develop the technol- ogy to the prototype field testing stage. Work on Phase I was completed at the end of FY99. Saft presented results on an advanced battery technology and USFS and Boeing each presented re- sults on flywheel technology. The final reports docu- menting this work will be published at a later date. A brief summary of each contractor’s Phase I work is pre- sented in Chapter 3 of this report. Battery Simulator Development and Validation: The ESS Program initiated a collaborative project with NRECA in FY97 to develop, validate, and demonstrate simulators of power-quality and peak-shaving systems. The project is providing technical and economic data about peak-shaving and power-quality systems at elec- tric membership cooperatives. More important, the project will introduce a technology assessment that is more exact and no more expensive than a traditional “paper feasibility study.” The ESS portion of the project supports the devel- opment and validation of energy storage simulators that will mimic the operation of two BESSs: one that Brockway Standard Corporation operates for power quality (PQ2000) and one that Energy United (EU) Cooperative operates for peak shaving. The NRECA portion of the project supports field demonstrations of the energy storage simulators and the development, validation, and field demonstration of a diesel-generator simulator. Development and validation of the battery peak shaving simulator at EU’s 500-kWh energy storage system at its Statesville, North Carolina, facility was completed. A draft report has been prepared and is under review. During the validation period of August 1998 to April 1999, the utility dispatched the battery system 25 times. Initially, the simulator correctly dis- patched the virtual battery 13 times. The software was then modified to improve the dispatch of the virtual battery. VRLA Reliability Improvement Project: Be- cause SNL believes that VRLA battery technology of- fers real advantages in utility and renewable energy applications, a VRLA reliability improvement project was formulated. The primary objective of the project is to determine VRLA cycle and calendar life under typi- cal utility battery operating conditions and use modes. The ESS Program and ILZRO have established a col- laboration that addresses VRLA reliability issues. A three-phase project was designed to identify and resolve these battery life issues. Phase 1 involves a survey of the industry, in cooperation with VRLA manufacturers and users, to determine objectively and accurately the status of the technology. Phase I activi- ties were the primary focus in FY99. Efforts continued in FY99 to encourage return of the surveys, especially by the user group. Manufactur- ers were also reminded to submit data when they were encountered at professional meetings and during vendor visits. Five manufacturers of VRLA batteries com- pleted surveys that discussed the materials, manufac- turing processes, and service-life expectancies of 13 different technologies that serve in stationary applica- tions. The technologies included both gel and adsorbed glass mat (AGM) designs, and the applications included both float and cycling service. The manufacturers also identified customers who have field experience with their largest volume product, their best performing product, and their least favorable product. The data from the manufacturers’ surveys are recorded in an Access database. By the end of FY99, more than 50 end users had been contacted about the survey, and 10 had agreed to complete it. This should generate a large enough data set to carry out the trend analysis. This analysis will commence as soon as the surveys are received. 1-6 1. EXECUTIVE SUMMARY ETO Switch Development for PCS: The devel- opment of a high-power semiconductor switch, an ETO thyristor, began in the third quarter of FY98. This proj- ect entailed the development and testing of a prototype ETO. ETOs could greatly enhance the PCS, a vital part of the energy storage system used to interface between the storage component and the generation and T&D equipment. At the heart of the PCS are the topological connections of high-power semiconductor switches. To meet the demand of these devices, efforts were made to improve the gate turn-off (GTO) device in the past few years for applications rated in the megawatts. The ETO is a type of GTO device developed at Virginia Tech that could substantially advance high-power ESS applica- tions. Based on the mature technology of current semi- conductor devices, the ETO could provide a low-cost, superior solution for megawatt applications. In FY99, this project was extended to the next phase of ETO development. The objective of the next phase is to build and demonstrate an ETO in a high- power converter with thermal, electric, control, and reliability all being demonstrated and tested in one unit. The preliminary version started at 100 kVA and is an- ticipated to gradually move up to high power (1 MVA and higher). Development of Intelligent Controls and Con- trol Strategies for RGS: Both the ESS and PV Pro- grams are interested in the development, design, and testing of an improved, more versatile, and better inte- grated PCS. Power electronics and the supporting cir- cuitry and software are critical to the eventual wide- scale use of renewables and energy storage technolo- gies for many diverse applications. As documented in the recent SNL report, Renewable Generation and Storage Project—Industry and Laboratory Recommen- dations (SAND98-0591), the DOE ESS Program is embarking on an initiative to address system integration and components in RGS systems. This project includes a competitive procurement with an industry partner to develop an intelligent system controller and evaluate advanced control strategies for power conversion sys- tems used in RGS applications and will include tasks for design, applications analysis, fabrication, and com- ponent and system testing. In FY99, the ESS Program gathered information from the energy storage and PV industries and nonin- dustry organizations such as the Florida Solar Energy Center and the New Mexico State University Technol- ogy Development Institute to clearly define needs, di- rections, and scope of an improved RGS controller project before initiating a request for information (RFI). The objective was to prepare a clear taxonomy of RGS controls, explore how these controls should be organ- ized, explore which functions are more natural to im- plement within subsystems and which functions are better suited to more advanced control principles. Evaluation VRLA Battery Evaluation: Controlled laboratory tests are the best method to determine battery capacity degradation rates and mechanisms. While batteries in field tests have the same problems, the often uncon- trolled variability of the test environment slows the collection of data and makes it difficult to distinguish cause and effect. Life cycle testing continued in FY99 at SNL on VRLA batteries, which were under test for utility appli- cations, specifically the Yuasa Tubular Gel batteries, the intermediate state of charge (ISOC) test units, and GNB’s ABSOLYTE II and ABSOLYTE IIP AGM batteries. The GNB batteries were deliverables from a development contract with GNB. Results from this testing are presented in Chapter 3 of this report. ABSOLYTE IIP and ABSOLYTE II Testing: Life cycle testing of the ABSOLYTE IIP resumed with Cycle 611 in November 1999, after tester repairs were completed. Capacity continued to decline, falling be- low 80%, and testing was halted at Cycle 699. After consultation with GNB, a decision was made to try to recover some capacity. A boost charge was performed with little effect. A reference electrode inserted into one cell provided data that indicated possible positive electrode capacity limitations and negative electrode polarization. Tests of an aggressive charge profile are planned for FY 2000. Life cycle test of the ABSOLYTE II battery also continued in FY99. Air conditioner and tester repairs limited testing time. The capacity declined steadily throughout the testing. Cell 4 end-of-discharge (EOD) voltage became significantly lower than other cells, so it was bypassed. A boost charge was performed with only slight improvement in capacity. Testing was halted at Cycle 506 after several cycles with capacity below rated end of life (EOL). A total of 163 cycles were performed during FY99. Yuasa Laboratory Testing for Hybrid Environ- ment: In early January 1998, two modules of Yuasa Exide, Dynacel DGX Tubular Gel VRLA cells were received at SNL. A test plan was developed to test these special cycling batteries in a laboratory. The test pro- gram emulated the off-grid hybrid operational environ- 1, EXECUTIVE SUMMARY 1-7 ment and was run in parallel with the actual field testing of the DGX batteries at the APS STAR Center Hybrid Test Facility in Tempe, Arizona. This test program pro- vided data to compare to the Dynacel battery at STAR, which is being tested in an actual hybrid operating envi- ronment. The battery is exposed to a cycling regime that will possibly lead to the optimization of the oper- ating strategy for off-grid RGS applications. Testing began in the summer of 1998 and continued through FY99. All testing has been done in-house at SNL. Intermediate State-of-Charge Testing: The ISOC test project was initiated in early FY98, and test- ing of batteries from several participating manufactur- ers began in the first quarter of FY99 at the SNL test laboratories. Two VRLA gel electrolyte batteries and one AGM battery from three manufacturers were tested as single units and in 60-V strings to determine whether the batteries can operate in an ISOC environment with- out damage. The ISOC environment allows for better energy management in off-grid applications by not re- quiring a full 100% recharge following discharge op- erations. During FY99, the test project required inter- nal technical support to ensure continued and reliable operation of the test equipment and data management system. No external contracts were needed to support this project. System Evaluation of a 200-kWh Zinc/Bromine Battery at “The Barns”: In FY97, after completing a competitive procurement process, the ESS Program initiated a contract with Powercell Corporation to per- form in-house tests on Powercell’s Zinc-Flow™ bat- tery. The original objective of this project was to char- acterize the performance of a 9-kWh Zinc-Flow battery. The project was expanded in FY99 to include the field characterization of 100-kWh units that Powercell calls the PowerBlock. Two 100-kWh PowerBlocks were built and will be field tested at “The Barns,” a community theatre at Franklin Park in Loudon County, Virginia. This project will entail monitoring both bat- tery performance and energy use/generation of all the major system components. The electrical design for the system was completed in the third quarter of FY99. The design allows for optimum use of the BESS. A plan was developed based on this design to measure energy use/generation of the major components. The plan enables the project team to perform energy balances for the entire system. A draft plan for implementation was completed in the third quarter. Analysis System Studies Opportunities Analysis—Phase II: The Opportu- nities Analysis was conceptualized in FY94 as a two- phase project. Phase II of the project, started in FY98, is an extension of Phase I performed in FY94 and FY95 by the ESS Program. In an earlier, preliminary assess- ment of national benefits, SNL estimated that the gen- eration and transmission applications of storage could represent $17.2B in national benefits. In Phase I of the Opportunities Analysis, the T&D benefits were found to be significantly higher than previous estimates. Phase II of the study included a refinement of the technical and economic understanding of the role of energy storage in the utility industry. Increased under- standing will help promote appropriate development and more rapid commercialization of utility energy storage systems. The current information is insufficient to estimate market size with a high degree of confi- dence, especially from a system supplier’s perspective. Activity in Phase II primarily focused on the need to characterize the near- and long-term utility application requirements for energy storage. The first meeting of Phase II of the Opportunities Analysis study was held November | and 2, 1998, in Santa Fe, New Mexico. The meeting, which was at- tended by 15 representatives from the storage and re- lated industries and several Sandians and staff from Energetics, focused on specifying application require- ments for stationary energy storage, categorizing them according to a structure meaningful to electricity sup- pliers and consumers, and reaching consensus on clear terminology. On April 22 and 23, 1999, representatives from DOE/SNL and Energetics and stakeholders from various parts of the energy storage, power electronics, and electric power industries met to conduct Meeting II. From these meetings emerged definitions for ten individual applications of energy storage that are in general demand and have high value with the electric power producers and their customers. A summary of the requirements of each of the applications and defini- tions is presented in Chapter 4 of this report. Value of Storage for a Restructured Utility In- dustry: The electric utility industry in the U.S. is being restructured and is evolving from a regulated monopoly to a partially competitive, partially regulated group of electricity providers. The public’s expectation of plen- tiful, high-quality, and low-cost electricity for all con- sumers has not changed, and if anything, will grow in 1-8 1. EXECUTIVE SUMMARY the coming years. Generation, transmission, distribu- tion, and use of electricity will be performed in a great variety of new and evolving ways through the imple- mentation of creative regulatory frameworks, financial instruments, and technologies. One technology that could have tremendous impact is energy storage. A study initiated last year with Distributed Utility Associ- ates (DUA) described several scenarios likely for the utility industry in the future. DUA completed the study, and the draft final re- port was received in June 1999. The study describes a series of scenarios in a deregulated utility industry, and it discusses the possible benefits from storage in each case. It discusses ways to expand the envelope of pos- sible storage applications and suggests creative use for storage. It also presents many possibilities for commu- nicating the value and flexibility of storage. This report will be published in FY2000. Utility Operating Cost Analysis: This task was initiated during FY94 through the placement of a con- tract with UMR to use the Electric Power Research Institute’s (EPRI’s) DYNASTORE computer program to perform calculations of utility operating costs with and without BESS. Operating cost savings are one im- portant component of the battery system cost/benefit picture, along with the system capital cost and other projected utility benefits. Work in this area was continued under a new con- tract placed with UMR during the second quarter of FY99. A major part of this study is adding the capabil- ity to include wind-powered generators — in DYNASTORE calculations. The effect of renewable generation on operating costs was then investigated for the Kansas City Power and Light Co. (KCPL) system. Analysis of the situation with renewables is more complicated and is still not complete, but an indication is emerging that the operating cost benefit of storage in this case may be much larger than that previously esti- mated for conventional generation sources. The reason for this is that renewable energy sources are not con- tinuously available (e.g., PV at night) and therefore cannot be allocated to certain applications, such as spinning reserve, that require 100% availability. How- ever, the use of energy storage will broaden the range of applications to which renewables can be applied by providing the capability to ride through periods of low renewable generator output. Therefore, storage not only directly replaces more expensive sources of gen- eration, such as combustion turbines, but it also enables renewable generation to serve many of these applica- tions. Since renewables generally have very low oper- ating costs, this significantly increases operating cost savings. Savings can be quite large, four to six times greater than those from adding battery energy storage (BES) alone to a utility system. This initial study assumed that renewable wind generation is fully allocated to the application and that the wind generation is always on at an average calcu- lated level. Much more work is needed to determine the full effect of variability in the renewable generator out- put. Potential strategies for allocating the renewable resource also need to be considered in the study. International Energy Agency Annex IX: The ESS Program at SNL was designated to be the USA Participating Agent in the activities of Annex IX. This responsibility requires attending executive committee and experts meetings in the United States and abroad, coordinating United States representation by experts for the various storage technologies, identifying projects of common interest to the participants, and supporting the implementation of the projects. Phase 2 of the Annex IX Electrical Energy Storage Technologies for Utility Network Optimization work was concluded at the end of FY99. A final report, which was an overview of this work, was submitted in November 1999. Topical reports on Subtasks 1, 2, and 3 were submitted in the third and fourth quarters of FY99. e Subtask 1: Applications Case Studies e Subtask 2: Project Definitions e Subtask 3: Applications Modeling The Electrical Energy Storage: Network Applica- tions Case Studies report focuses on storage systems as solutions for specific problems. It is aimed at users who may be relatively unfamiliar with the field. Pre- existing electrical energy storage installations are de- scribed from the perspective of the end user application, as distinct from the particular storage technology em- ployed. The Project Definitions report focuses on two energy storage applications: (1) power quality and (2) primary substation applications. The report presents a cost breakdown for the various electrical energy stor- age system solutions currently available. Generic in- formation is provided without focusing on any single supplier’s solution or any particular site. The Applications Modeling report was received in October 1999, and represents the final output of Sub- task 3 of Phase 2 of Annex IX. Two electrical energy storage models were introduced. One model was a 1, EXECUTIVE SUMMARY 1-9 fully functional, stand-alone power quality applications model. The other was a software specification model for network integration for the applications associated with substation level connected energy storage systems. This model will allow the user to match and size an energy storage system to meet exact technical and eco- nomic requirements without the risk of system redun- dancy or under capacity. Transmission Power Quality Study: The TPQ study will monitor and define power quality levels of the electric transmission system for a defined 13-state region in the Southeastern United States. A previous study conducted by EPRI characterized power quality levels of the distribution system; however, very little data now exist for customers of transmission networks. A study team composed of representatives from utilities, SNL, EPRI, and the SEE conducted a scoping study that outlined the project justification, methodol- ogy, cost information, and schedule. Monitoring sites will be chosen throughout the Southeast to provide a statistically valid sample of the transmission system. Criteria for choosing sites are voltage class and light- ning strike density. Participating utilities will procure, install, download, and maintain appropriate monitoring equipment. The ESS Program will provide funding for the project, in conjunction with cost sharing from in- dustry. Coordination will also take place between the ESS program and other SNL programs that may benefit from this study. Technology Assessments Performance and Economic Analysis of SMES, Flywheels, and Compressed Air Energy Storage: The scope of the ESS Program includes a portfolio of energy storage technologies for electric utility applica- tions. The program approach has been to apply exper- tise gained from work with BES to the development of other storage media, PCSs, peripheral devices, and ad- vanced storage systems that depend on similar compo- nents. The ESS Program initiated this analysis project with Energetics, Inc., in FY97 to identify the areas in which program expertise directly applies to this ex- panded range of technologies and where such program expertise must be developed. The final report, A Summary of the State of the Art of Superconducting Magnetic Energy Storage Systems, Flywheel Energy Storage Systems and Compressed Air Energy Storage Systems, was published in July 1999 (SAND99-1854). Long- Versus Short-Term Storage Study: A study to characterize the stationary applications and technologies of short- and long-term energy storage was completed by Longitude 122 West, Inc. Applica- tions of energy storage have a wide range of perform- ance requirements. One important requirement is stor- age time or discharge duration. In this study, applications and technologies were evaluated to deter- mine how storage time requirements match technology characteristics. Comparisons were also made on the basis of capital cost for various energy storage systems operating over a range of discharge times, categorized as short term (< 2 hrs) and long term (2 to 8 hrs). Spe- cial categories of very short term (< 1 min) and very long term (a day to weeks) were also considered. The technologies evaluated included batteries (lead-acid and advanced), flywheels (low and high speed), superca- pacitors, superconducting magnetic energy storage, compressed air energy storage (CAES), pumped hydro, and hydrogen. A draft copy of the report was received and is un- der review. A brief summary of the draft report is in- cluded in Chapter 4 of this report. PCS Magnetics and Functionality Analysis: A previous study conducted by the ESS Program focussed on identifying the state of the art in PCS technology and identified several promising research directions whereby cost and footprint reductions could be achieved. Additionally, the study cited several areas that could potentially contribute to cost reduction, but available information did not permit a thorough as- sessment. Listed below are two promising ideas that could not be adequately evaluated. 1. Simplify PCS specifications by defining the minimum functionality required for a PCS used as part of an ESS. 2. Improve the magnetics used in the PCS. The purpose of the work being undertaken at this time is to further evaluate the above areas. As such, this report is an extension of and relies heavily on the previ- ous study. The objective of Task 1 is to determine if there is a standardized approach to specifying PCS functionality that can reduce cost in energy storage systems. EESAT2000 Conference: The ESS Program, un- der the sponsorship of the DOE, and in cooperation with the ESA, has taken over the responsibility for or- ganizing the Electrical Energy Storage Systems Appli- cations and Technology (EESAT) conference to be held September 18 to 20, 2000. The conference will be held 1-10 1. EXECUTIVE SUMMARY in the Royal Plaza Hotel in the Disney World Resort, Lake Buena Vista, Florida. The international confer- ence will address all aspects of energy storage system technologies including conventional and advanced bat- teries, supercapacitors, SMES, flywheels, CAES, pumped hydro, the power electronics and control sys- tems, as well as system studies and economic analysis of storage systems. Individual components, system applications, and research into advanced systems and components will also be included in the conference agenda. A call for papers was issued in September 1999, with abstracts due in January 2000. Technical and eco- nomic papers are encouraged. The component areas of interest include the energy storage system, power elec- tronics, control and data acquisition. The application areas of interest include generation, transmission, dis- tributed resources, distribution power quality, energy management, and customer applications. The research areas of interest include advanced systems and compo- nents, identification of the potential users and needs for energy storage and the future of the deregulated utility industry. 1, EXECUTIVE SUMMARY 1-11 This page intentionally left blank. 1-12 1. EXECUTIVE SUMMARY 2. Integration Introduction Under the Integration element, a strategy is being pursued to reduce the inefficient, one-of-a-kind system engineering that was historically required when an ESS is designed and built. A modular approach was rec- ommended as the preferred method of achieving system flexibility and the lowest possible cost. The major sub- system components (storage, power conversion system [PCS], and controls) are designed as separate modules. This allows integration to take place either at the fac- tory or at the user site. From a cost perspective, this modular approach permits more efficient engineering, design, and manufacturing processes to be used. Fi- nally, the large quantity of on-site labor required to assemble and start up the system at the user site is minimized. The Integration element consists of two subele- ments to fully address these system engineering needs. System development pursues development of complete, modular prototypes for end-use applications. Advanced battery systems, improvement of batteries in renewable systems, and integration of storage with village power supplies and at the substation level are major program initiatives. All are performed with industry under cost- sharing agreements. System evaluation focuses on collecting data in the field from working systems. The analysis and presentation of these data are crucial to development and verification processes necessary for new technology to be adopted. Systems, both large and small, from one end of the U.S. to the other, are being monitored and the data reported. System Development Advanced Battery Energy Storage System The zinc/bromine battery is an emerging technol- ogy that has many attributes that make it attractive for energy storage applications. The main advantage of the zinc/bromine battery system is good gravimetric energy density, which results in a modular, transportable bat- tery system with sufficient capacity to be placed any- where on the utility grid. The battery is made almost entirely from plastic (high-density polyethylene), which makes it cost-competitive with lead-acid batteries at high-volume production without the hazardous manu- facturing and recycling concerns. Also, the battery op- erates efficiently over a wide temperature range, func- tions under intermittent charge/discharge conditions, and can experience complete discharge hundreds of times without damage. In 1997, ZBB Technologies, Inc. (ZBB), the ESS Program’s industry partner on this project, participated in one of the largest turnkey advanced battery demon- strations in the United States. This major demonstra- tion, a 100-kWh zinc/bromine battery system, was par- tially funded by the DOE/ESS Program and was completed at the ZBB test facility. Following the completion of the 100-kWh battery tests in FY97, a new project was competitively placed with ZBB for the development and field testing of a 400-kWh advanced battery energy storage system (ABESS). Field testing a prototype, integrated ESS should enable ZBB to validate the technology and prove the reliability to the satisfaction of electric utili- ties and other users. Currently a demonstration is planned at a Detroit Edison site. Status The tasks completed in FY99 include: Preliminary design definition, Update to application analysis, Manufacturing cost study, Module network system testing, Qualification of a new module design, and Analysis of separators System Design: The ABESS integrates three main components that comprise the storage system: the zinc/bromine battery, the PCS, and the necessary con- trol equipment. The system has been configured to minimize transportation, installation, and maintenance costs. Once the system is located at the site, it can be connected to the utility grid by an overhead or under- ground service. The battery modules have been de- signed to allow for on-site maintenance and replace- ment of virtually any component in the battery module. . The control system monitors all system parameters and shuts down the entire unit if any safety hazards arise. System operation parameters are collected and re- viewed to resolve any unfavorable operating trends. A schematic of the battery module is shown in Figure 2-1. The energy storage portion of the system 2. INTEGRATION Figure 2-1. consists of eight battery modules, each containing three battery stacks connected electrically in parallel. Eight individual 50-kWh zinc/bromine battery modules each containing three battery stacks connected electrically in parallel supply the 400kWh. These modules are con- figured into two separate strings of four modules con- nected in series, with an open circuit voltage of 432 V (Figure 2-2). The two battery strings can be operated as two separate batteries or as a single battery, giving the entire energy storage system a great degree of charge/discharge flexibility. The battery system has a rating of 400 kWh at the two-hour discharge rate. Table 2-1 summarizes the 400-kWh zinc/bromine bat- tery specifications. The PCS converts the DC input from the batteries into three-phase, 480 Vac and is connected to a delta/wye isolation transformer. A 600-A fused dis- connect is coupled to the transformer to manually iso- late the entire storage system from the grid. The con- verter is a 200-kVA, three-phase, bi-directional, voltage-fed, current-controlled AC-DC electronic power conversion system. The PCS has a four- quadrant operation, meaning that it is capable of con- trolling its power factor while the battery is charging and discharging. A detailed specification for the PCS Schematic of 50-kWh Battery Module. has been submitted to SNL and Detroit Edison. Table 2-2 summarizes the 200-kVA PCS unit specifications. The ESS is controlled by two controllers—one con- trols the battery system and the other controls the PCS. The system programmable logic controller (PLC) is used to operate and monitor the two individual battery strings, each with its own “SYSTEM READY” signal. The PLC will coordinate the overall operation and safety of the system by comparing the measured pa- rameters to preset limits. If the measured parameters fall outside the preset norms, the PLC will adjust vari- ables, such as pump speed or heat exchanger pumps and fans (to control temperature). If the measured parame- ters cannot be modified to fall within acceptable limits, the PLC will generate either a FAULT or SHUTDOWN condition. If a fault condition should arise for one of the strings, that string would go into the “FAULT” mode, while the other string would continue undis- turbed. If a potentially hazardous condition to the sys- tem and its surroundings is sensed by the monitoring system, the controller will completely shut down the system. Conditions that would result in a complete SHUTDOWN of the entire system include an electro- lyte leak or elevated enclosure temperature. 2. INTEGRATION Tan Onl Thi wi Figure 2-2. TH tl (Th Gil String 1 432-V O/C 150-A nominal Module 108-V O/C 150-A nominal 300-A continuous 300-A continuous String 2 Total 432-V O/C 432-V O/C 150-A nominal 300-A nominal 300-A continuous 600-A continuous Battery Module DC Interconnection Scheme. Table 2-1. 400 kWh Zinc/Bromine Battery Summary Specifications Battery System Module DC interface 504 Vde maximum 126 Vdc maximum Capacity 400-kWh, 2-hr discharge 50-kWh, 2-hr discharge Configuration 8 modules, 2 parallel strings of | 3 zinc/bromine battery stacks electrically 4 modules in series connected in parallel Dimensions ~ 8' x 8’ x 20’ ~8'x3'x 3’ Weight ~ 40,000 Ib. ~ 3,000 Ib. Table 2-2. PCS Summary Specifications ” “= PCS Unit Specification a Configuration Modular, main PCS unit, filtering device cabinet and isolation transformer Type Three-phase, bi-directional, voltage fed bridge Continuous kVA rating kKVAR/KW rating Switching frequency Short circuit AC Short circuit DC Efficiency AC current distortion Power factor Operator interface inverter 200 kVA @ 480 Vac 200 kVAR, 200 kW 8 kHz 350-A continuous, 100-kA interrupting capacity 400-A continuous, 100-kA interrupting capacity 95% at full output 5% maximum at full load Unity/controllable External IBM compatible with Win95-based operator interface software 2. INTEGRATION Normally, the system controller will monitor sys- tem parameters such as the operational mode of each string (ie., charge, discharge, strip, standby, fault), string voltages and currents, ambient temperature, and enclosure temperature. In addition, the system con- troller maintains constant communication with the PCS. If either the PCS or the system controller detect a fault, the storage system (via the PCS contactors) will be dis- connected from the utility grid. The controllers com- municate with each other through relay closures repre- sented as a SYSTEM READY signal from the battery and an ENABLE signal from the PCS. As long as these two signals remain active, the storage system will oper- ate. Figure 2-3 depicts the relay communication be- tween the PCS and the battery system controller. Software, located on an optional data collection computer, will monitor the performance of the system. A program that polls specific PLC memory locations and stores this data in a file has been written. A com- puter is not necessary to run the system, but can be used to display real-time data on the computer screen via a RS232C serial line (between the PLC and the data col- lection computer). This information can be viewed off site, provided the appropriate modems are in place. The information compiled by the PLC for each of the eight modules includes parameters such as string volt- age, stack currents, electrolyte levels, stack tempera- ture, and reservoir temperatures. Update to Applications Analysis: ZBB Tech- nologies, Inc., envisions the zinc/bromine battery sys- tem as a distribution system peak-shaving device. The intended benefit is to defer conventional distribution system capacity upgrades. In addition, the system can address storage capacities over various dispatch sched- ules, power factor control, and reactive power (VAR) compensation. ZBB is conducting this applications analysis in partnership with the Detroit Edison Com- pany. The company has investigated several potential sites as locations for testing a 400-kWh zinc/bromine ABESS. They intend to demonstrate the transportabil- ity of the battery system by utilizing it for peak reduc- tion and reactive supply on a summer peaking circuit, and for load-leveling and power-quality applications in the fall/winter period of the same year. if _ Leg OCirewOstnt > BATTERY o = —, Feavorst = 7% OCinowOtot anes x = -——3 7 = FELAY CPST = SARE] & & ———— | (OC Contactor OC Contactor RELAY SPST Sere —a oo aes FESSTOR sae Fszmc FF cprosan a i Use Retav or Ontoisetainr A RELAY SPST = FELAY SPST 0k chen CONTROLER: _ a Itheinnit inthe tier mst nations x {fant Jonic low voltane of less than AV. COMPORTS 1 bas 2 RSC a7 Figure 2-3. Connection between PCS and System Controller. 2-4 2. INTEGRATION The first location under consideration for these ap- plications is in the Lake St. Clair area, in the township of Harrison, Michigan, on the Beach Substation circuit load. The Beach circuit is temperature dependent due to its high concentration of residential and commercial air conditioning load. The load profiles for this substation were evaluated over a one-week period from June 26 to July 1, 1998. The voltage profile remained fairly constant during the week, but large peaks were observed in the current pro- file. A peak in the current was observed every day, between the hours of 6 p.m. and 10 p.m. as seen in Table 2-3. The load profile during this one-week pe- riod is plotted in Figure 2-4. The kVA values in Table 2-3 are plotted versus the daily peak temperature in Figure 2-5. The temperature data were obtained from the National Climatic Data Center (NOAA) for the city of Detroit, Michigan. Figure 2-4 demonstrates that the peak in the load is highly temperature dependent, which reflects how much AC load there is on the circuit. The substation load profile in Figure 2-4 depicts the actual load that the substation observed during the seven days. This plot has some distinct peaks as well as dual peaks that could be shaved by an energy storage system. The 400-kWh system being built could not significantly impact the substation load profile; how- ever it is reasonable to assume that there are points throughout the distribution system that have a similar load profile. The 200-kVA/400-kWh energy storage system is suitable for an application near the load (as opposed to a larger substation load). Detroit Edison identified a potential site on the Beach circuit where customers have a low-voltage problem when there is a sufficient increase in ambient temperature. The Township of Harrison has experienced low-voltage problems and is located near the end of the circuit. The potential site is a small municipal complex, which has an administrative office, city sewer and wa- ter department offices, and a police station. The power for this complex is supplied by a 4.8-kV three-phase circuit, of which a single-phase 4.8-kV line feeds a residential area. A potential location for the storage system has been identified in the parking lot in front of the administrative offices. The areas that are expected to benefit the most from an energy storage system in- stallation are the administrative offices, the single- phase 4.8-kV residential line, and possibly the Sewer and Water Department offices. The design load for the 4.8-kV circuit is 3.5 to 4.0 MVA. The location of the planned battery installation is far enough away from the substation where high load periods cause low-voltage problems. The timing of placement at this site may not work well, so data are currently being collected for a number of potential sites. The intention is to determine the most suitable location for testing the battery system during the summer of 2000. Circuits that are candidates for summer load relief and/or power quality applications include: e STDF #4 (2628)-summer and winter loading at day-day rating, e New Haven (312) motor starting at the end load and voltage support, = <== aa Table 2-3. Beach Substation Distribution Circuit 2594 (Local 4.8 kV) Thursday Friday Saturday Sunday Monday Tuesday 6/25/98 6/26/98 6/27/98 6/28/98 6/29/98 6/30/98 Peak temperature 96°F 91°F 80°F 88°F 88°F 78°F Average 84°F 81°F 75°F 78°F 78°F 71°F temperature Time of peak 22:00 18:00 18:00 18:00 19:00 19:00 current Voltage at peak 128 127 125 126 126.5 125 Chart reading 5.4 4.9 3.3 4.2 4.8 2.5 Current 648 588 396 504 576 300 Peak kVA 5746 4821 3429 4400 5048 2598 2. INTEGRATION KVA Figure 2-4. Peak kVA Figure 2-5. 1000 3 4 Time (Days) Load Profile for Distribution Circuit 2594 for Beach Substation — June 26, 1998 to July 1, 1998. 6500 6000 5500 5000 4500 4000 3500 3000 2500 2000 —— 75 80 85 90 Peak Temperature (° F) 95 The Effect of Temperature on kVA for the Beach Substation. 100 2-6 2. INTEGRATION e Richmond (8420)}-Summer loading over day-day rating, e Brewer (8840)-summer loading over day-day rating 200HP motor at the end. General information for the candidate summer cir- cuits is shown in Table 2-4. Load profiles for the Brewer circuit during June 1999 are shown in Figure 2-6. These circuits show peak in the load during the day. A location from the circuit selection shown below should provide an appropriate site for testing a 400- kWh zinc/bromine system. The most promising candidate for the fall is a farm approximately 90 miles north of Detroit Edison’s Macomb Center Headquarters on the Talbot Circuit in Palms, Michigan. This location has an established grain drying facility that services the drying needs of a number of farms in the area. Detroit Edison has en- countered difficulty in reducing the amount of “flicker” from the grain drying facility at a reasonable cost to both Detroit Edison and the customer. The drying sea- son generally runs from mid-September to the end of December. The analysis of this circuit is ongoing. Cost Study: A detailed cost study was performed for the initial design of a 400-kWh zinc/bromine ABESS. Cost documentation was provided for mate- rial, labor and indirect cost to estimate the total cost of the system, including the PCS. The study also exam- ined where cost reductions are necessary to improve the cost competitiveness of the system. The study esti- mated a cost of $646/kWh for a 400-kWh zinc/bromine battery system including labor and indirect costs. The cost including the power conditioning equipment was estimated at $963/kWh. A breakdown of the costs for a zinc/bromine battery system is given in Figure 2-7. The PCS is a major portion of the total estimated cost of the 400-kWh system. The entire PCS cost is $316/kWh including the power unit, transformer, circuit breaker, and operating interface computer. The PCS for this system was custom designed, which caused the cost of the PCS to be over 30% of the entire system cost. Rapid advances in solid-state power switching devices, and anticipated benefits of mass production will reduce this cost considerably. Future prices for PCS systems are predicted to fall to in the range of $100/kWh. The cost of the system container is $28K for one unit, and ~ $23K for ten units. This corresponds to a 17% reduction in price for a quantity discount. The container is another high-cost component of the zinc/bromine battery system, but it could also be con- sidered as part of the balance of plant (BOP). The BOP for other battery technologies includes the building and support structures that would be included in the cost of the 400-kWh zinc/bromine battery system. The total cost of the battery stacks ($104/kWh) will need to be reduced for this system to be cost effective. Some cost reductions will be realized when production numbers are increased, but this still will need to be further reduced. The most expensive component in the stacks is the separator material at about $48/kWh (nearly 50% of the cost of the stacks). The separator used in the zinc/bromine battery is very similar to lead- acid type separators, and should not cost significantly more than this type of material. The major raw materi- als used in the separator are polyethylene and silica, neither of which is extremely expensive, and extrusion of the material is also relatively low in cost. Lower cost separator materials will be an important part of reduc- ing the cost of the battery system. Table 2-4. Data for Summer Load Circuits Substation STDF #4 New Haven Richmond Brewer Circuit 2628 312 8420 8840 Circuit Voltage (kV) 4.8 48 13.2 13.2 Peak Loading (kVA) 21.6 60 51 140 Rating (kVA) 1,200 3,200 8,700 7,700 Number of residential customers 239 1,124 1,648 2,347 Number of industrial customers 19 94 271 167 2. INTEGRATION MVA ° 1 1 4 1 1 1 0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00 Time Figure 2-6. _ Load Profile for Brewer Circuit during June 1999. Stack materials 10.5% PCS Stack Labor 31.9% 11.3% Electrolyte 8.3% System Labor 0.4% System Hardware 7.3% Module Hardware 23.6% Module Labor 6.6% Figure 2-7. Cost Breakdown for 400-kWh Zinc/Bromine ABESS. 2-8 2. INTEGRATION The other expensive components in the battery stacks include the injection-molded parts: endblocks, covers, and flow frames. The cost of these will be re- duced significantly when manufacturing numbers are increased. Also, a number of alternate injection mold- ers are investigating ways of reducing the cost of manu- facturing these components. The components of the electrolyte are estimated to cost $82/kWh. The cost of this will be drastically re- duced when economies of scale are realized. The zinc bromide solution costs would be reduced from $4/kg to $3/kg in large volumes, a reduction of 26%. The bro- mide complexing agent in the electrolyte, which is spe- cifically manufactured for the Zinc/Bromine battery, is very expensive at $20/kg in small quantities. When quantities of this material are increased, the complexing agent could be reduced by over 90% to about $2/kg. The remaining chemicals in the electrolyte solution would also see some degree of cost reductions with economies of scale. Scaled-up production of the Zinc/Bromine battery electrolyte would reduce the total cost of the electrolyte for a 400 kWh battery system by a total amount of nearly 80%. At this time, the drive motors for the pumps are specially manufactured ror the zinc/bromine system, at a cost of $500/motor. For high volume quantities, the 1/2 HP motors would cost $186 and the 3/4 HP motors would cost $219 a piece. This corresponds to a savings of $876 per module, and $7K/400-kWh system, or a 58% reduction in the total cost of the drive motors. Cost reductions observed for increased quantities were only included for some of the major components of the system. The cost for all components would also be reduced somewhat when scaling up production of the system. By examining the major cost reductions mentioned, the cost of the 400-kWh battery system (not including the PCS, labor or indirect costs) could be reduced from $466/kWh to $392/kWh, nearly a 16% decrease in cost. Additional reductions are predicted for the labor costs. This will be especially evident in the manufac- turing of stacks and the assembly of modules as the product moves into commercial production. A greater percentage of the manufacturing labor will become technical and workshop manpower, at labor rates sub- stantially less than the professionally qualified devel- opment engineers now engaged in the manufacturing of battery systems. The ratio of technical or non- specialized labor to professional labor will increase as manufacturing processes are stabilized into high vol- ume, repetitive processes. Module Network System Testing: A critical de- sign iteration in the control system for the zinc/bromine battery will be its ability to function without utility power. To demonstrate this feature, a 50-kWh module was tested without utility power. This strategy will simplify the design of the control for the battery system by enabling it to function in an uninterruptible power supply (UPS) or as a remote area power supply (RAPS). A commercially available isolated DC/DC con- verter was used to convert the battery voltage to 26 Vdc to run the control system. An additional DC/DC con- verter was designed and assembled by ZBB to supply the input voltages needed to run the input/output (I/O) board. A smoothing capacitor coming directly from the battery stacks runs the motor controllers. This was eas- ily done because the motor controllers have built-in isolation, which isolates the control section from the battery voltage. The motor controllers operated without interruption from the battery stacks until the battery voltage decreased to below about 30 V. Three small 8-V lead-acid batteries were connected in series to sup- ply 24 V to power the control system after the battery voltage became too low. This was needed to maintain control functions during the strip portion of the dis- charge, to keep the system active, and so that the Keyence programmable logic controllers would not need to be reset after each discharge. A diode was used to allow the control system to run from the battery sup- ply (zinc/bromine or lead/acid) with the higher voltage. This ensured that all power for the control system was taken from the zinc/bromine battery during the charge or discharge phases of the test cycles, with only mini- mal use of the backup power supply during the strip phase. For this exercise, the lead-acid batteries were not charged during each cycle. For an actual stand- alone or RAPS system, a simple trickle charging of the backup power supply would need to be considered. The battery system was also run with the control system and motor controllers running from AC power for comparison purposes. The performance data with the controls running off the battery stacks are compared to those with the controls running off of AC power in Table 2-5 indicates that the amount of energy used by the control] system and pumps is 3.8 kWh for a 50-kWh discharge. The fact that this number is less than 10% of the energy generated indicates that this system can be used without utility power. 2. INTEGRATION = = — <== Table 2-5. Performance of 50-kWh Zinc/Bromine Battery With and Without the Control System and Pumps Powered by the Battery Stacks Controls and Pumps Powered = == Controls and Pumps Powered by Battery by AC Power Discharge time (minutes) 209.9 224.8 A-hours 524.6 562.0 Energy (kWh) 49.47 53.27 Coulombic efficiency 77.7% 83.3% Voltaic efficiency 83.0% 83.2% Energy efficiency 64.5% 69.3% Qualify New Test Modules: Tests are being con- ducted to verify performance improvements and to evaluate the impact of lower cost components. A num- ber of changes in the original module design have been implemented, either to improve the performance of the module, or to examine the use of lower cost compo- nents. Several of the changes that were made in order to improve either performance or cost include: e Use of larger pumps to increase the rate of electrolyte flow through the battery stacks. e Change in plumbing configuration to improve the flow of electrolyte through the battery stacks. e Repair of the control system to improve the control of the levels in the electrolyte reservoirs. e Use of a prefabricated manifold plumbing system that is lower cost and easier to manufacture. e Testing of a new reversing valve design that is lower cost and easier to manufacture. The improved performance achieved with the vari- With the modifications made to the 50-kWh battery module, the performance increased from about 61% to 70% energy efficiency. With anticipated improvements in separator material, which will be discussed in the following section, the 50-kWh zinc/bromine battery module is expected to achieve greater than 75% energy efficiency. Separator Analysis: The separator material used in the zinc/bromine battery is one of the critical compo- nents of the battery. The important properties of the separator include the bromine diffusion rate and the resistance, both of which will have an effect on the en- ergy efficiency of the battery. The bromine diffusion gives an indication of how much self-discharge the battery will encounter. The resistance of the separator affects the voltaic efficiency of the battery. The most recent shipment of separator material from the vendor was found to have a very similar bro- mine diffusion rate to that of previously obtained mate- rials, but the resistance was approximately double that of the earlier materials. This increased resistance re- sulted in a reduction of 5 to 6% in energy efficiency when assembled into a battery. An in-depth analysis ous changes is shown in Table 2-6. was performed to determine the differences Table 2-6. Effect of Module Modifications on Performance Modification Coulombic Voltaic Energy Efficiency None 77.1% 79.6% 61.3% Larger Pumps 82.9% 78.6% 65.1% Changed Plumbing 82.5% 82.7% 68.2% Repaired Controls 83.3% 83.2% 69.3% Manifold System 84.2% 83.0% 69.8% Reversing Valve 84.0% 83.2% 69.8% 2-10 2. INTEGRATION between the two separator materials, and to examine how the material can be improved in the future. The tests performed included themogravimetric analysis (TGA), X-ray photoelectron spectroscopy (XPS), scan- ning electron microscopy (SEM) and X-ray diffraction (XRD). XPS examined the elemental composition of the surface materials, revealing a difference in the level of silica exposed at the very surface of the two membranes as shown in Table 2-7. The high-resistance separator material has a much lower percentage of silica at the surface than the standard material, which would cause the material to be more hydrophobic. Since the elec- trolyte is a water-based solution, this could be the cause of the high resistance of the separator. Table 2-7. X-ray Spectroscopy Results for Separator Materials Standard High Resis- Material tance Material Carbon (C) 39.4% 70.4% Oxygen (O) 41.4% 30.0% Silica (Si) 19.2% 9.6% The reason for this difference in relative level of organic material to silica ratio is clearly explained by the microstructural differences found with the SEM. A direct comparison of the SEM images shows that the coated silica particles are less distinguishable on the surface of the high resistance membrane because of a higher level of polymer matrix. XRD provides a measure of the crystallinity of the material. The XRD traces show two strong diffraction peaks between 20 and 25° that are indicative of poly- mers containing CH) chains (e.g., paraffin, polyethyl- ene). The intensities of the peaks show the polymers in both materials to be crystalline. The strong diffraction peaks from the polymers are underlined by a diffuse background that has a broad peak in the range 15 to 30° and indicates the presence of amorphous material in the sample. This amorphous material will comprise non- crystalline polymer material and amorphous silica. The differences in the XRD traces from the two samples show that the diffraction peaks from the well-crystalline polymer component of the standard sample are less intense than the corresponding diffraction peaks of the high-resistance sample. This indicates that the content of well-crystalline polymer material is lower in the standard sample than in the high-resistance sample. The inference from this observation is that there is rela- tively more amorphous (silica) material in the standard material than in the high-resistance sample. When considering the XRD results, it should be remembered that XRD does not compare equal masses of sample but approximately equal volumes of sample. It is therefore possible that the XRD results indicate that the standard material is more porous than the high- resistance material. All three of the analytical tests indicate that the standard separator material contains a higher concen- tration of silica at the surface than the high-resistance material. It was confirmed that the vendor made a slight variation in the separator formulation, which ap- pears to have caused the material to be high in resis- tance. The vendor has stated that the original formula- tion will be used for manufacturing any separator material in the future. Alternate Renewable Generation and Storage System Designs to Improve Battery Performance An investigation of alternative configurations to optimally use lead-acid batteries in renewable hybrid systems began in the third quarter of FY98. The pur- pose of this project is to devise systems for renewable hybrid installations that will improve the performance of the lead-acid batteries used in these systems. “Con- figuration” refers to the electrical arrangement of the major components of renewable hybrid systems such as the photovoltaic (PV) and/or wind generator, diesel generator, storage (for example, batteries), and the power conversion subsystem. Improving the perform- ance of the lead-acid batteries will extend the battery life cycle costs. The goal of this work is to devise and develop al- ternative configurations for the lead-acid batteries in renewable generation and storage (RGS) systems so that capital, operating and maintenance (O&M), and repair and replacement costs might all be reduced. Specifically, Electrochemical Engineering Consultants, Inc. (EECI), in association with SNL, is seeking to de- velop battery energy storage (BES) subsystems for PV hybrid RGS systems that offer the potential for reduced life cycle costs. Under a previous contract, two fami- lies of alternative configurations had been conceptual- ized and outlined, one directed toward equalization and the other toward the control of battery charging. The analyses from this work indicated that the alternative configurations that had been devised might ultimately lead to lower life cycle costs. In the work currently 2. INTEGRATION 2-11 being conducted, “breadboard” versions of one of the alternative configurations devised in the first phase of work will be developed and tested to provide a firmer foundation on which to perform further cost/benefit analyses. Thus, this phase of work involved the fol- lowing tasks: e Selecting the alternative configuration for further development using ongoing analyses with a model developed previously as the basis for making the selection. e Selecting and acquiring hardware for the alternative configuration to be tested. e Writing and testing the first software version that will monitor and control the hardware of the alternative configuration to be tested. e Development work on the breadboard hardware that constituted the test set-up for this phase of the work. e Building increasingly sophisticated versions of the breadboard hardware. e Testing successive versions of the breadboard hardware in an automated cycling mode, such automated cycling being controlled by the software that was developed. e Performing analyses of the testing performed to verify that the goals of the project were being met. Status The effort during FY99 involved building increas- ingly sophisticated versions of the breadboard hardware and then testing this hardware with more fully devel- oped versions of the software required for its operation. When the breadboard was first developed, testing could only be performed with cycles at a shallow depth of discharge (DOD). As the development progressed however, it was necessary to test with cycles of 50% DOD, so that the development would more closely mimic lead-acid batteries in the field. During the first quarter, it was decided that the al- ternative configuration would be the most appropriate choice to develop because it allows “smart” control of both bulk charging and of equalization. Also in the first quarter, it was decided that “off-the-shelf” components should be used wherever possible and that these com- ponents should mimic those used in presently available solar hybrid systems. Therefore, the following compo- nents were acquired for the breadboard testing: e Trojan T-105 (6-V, 225-Ah) flooded lead-acid batteries, which are used extensively by solar hybrid system suppliers; ¢ Several automotive battery chargers; e Several “modified sine wave” type DC-to-AC inverters; ¢ Twelve power resistors providing a load capability in total of more than approximately 50 A at 12 V; and ¢ Ancillary equipment required to complete the system. Additional accomplishments in the first quarter in- clude writing the first version of the software for con- trol of the alternative configuration and developing and testing the DAS. In the second quarter of FY99, assembly of the breadboard setup for testing the alternative configura- tion was initiated. In its most basic form, the alterna- tive configuration test setup consists of the following: e A battery comprised of two parallel strings each with two T-105 modules in series, e A40-A, 12-V battery charger, e A 0.5F capacitor bank to partially smooth the “full- wave rectified” current from the charger, e An approximately 0.4-ohm resistor bank for discharge, ¢ Various other control circuits, low-wattage power supplies, and cooling devices, and e A “486” class PC equipped with a digital/analog (D/A) board for data acquisition. The entire setup is controlled with the PC, using software written specifically for this purpose. The data needed to determine the performance of the alternative configuration under test is stored on the hard drive of the PC, pending later analysis. During the second quarter, many technical issues with the test setup had to be resolved before testing could begin. Heat build-up in linear power supplies, excessive ripple on the charge current, instabilities leading to oscillation in the control 2-12 2. INTEGRATION circuits had to be controlled, and refinements were made to the control software. By the middle of the sec- ond quarter, the test set-up was operating satisfactorily. The alternative configuration system became op- erational in the later part of the second quarter and the remainder of FY99, was operating more or less con- tinuously. Operation during this time consisted of automated cycling with discharges and bulk charges, each lasting approximately seven hours. In the last few weeks of FY99, the cycling was continuous, under automated control, with few interruptions for week- long periods. The only interruptions that occurred were operator-initiated. The modeling that was performed during the first phase of the work in FY98 showed the alternative con- figurations that had been conceptualized had promise to reduce the O&M costs for solar hybrid systems. An analysis of the data from the testing that has been per- formed confirms the results of the modeling. In addi- tion, the testing allowed conceptualization of several other ways in which to lower costs of operation, and perhaps lower capital costs for solar hybrid systems in the future. Small Village Storage System Integration Chugach Electric Cooperative and the State of Alaska Energy Authority have proposed a collaboration with the ESS Program to design, build, and demonstrate a 20- to 30-kW energy system for small Alaskan vil- lages. The purpose of this integration, test, and field evaluation project is to quantify the benefits of battery storage for small villages that have diesel-only genera- tion. This study will collaborate with a similar study planned by the U.S. Agency for International Develop- ment. The detailed system design will be initiated in FY99. If appropriate, a prototype system will be as- sembled and undergo initial testing and debugging at a test bed in Anchorage under controlled conditions. Following the successful completion of this phase, the prototype system will be moved to a village site se- lected by the Alaska partners. The system will be oper- ated at the site to generate data that will support the analysis necessary to quantify the benefits of storage under these operating conditions. The ESS Program expects to share the cost of about a third of the total project, with the remaining costs being shared between Chugach Electric and the Alaska Energy Authority. The feasibility and system design study was initi- ated in June 1999. A contract was placed with Sentech, Inc., to conduct the study and prepare a report describ- ing the results. A spreadsheet model will be developed and used to analyze operating data from representative sites in Alaska. The net benefit of the battery to the overall power system will be determined, along with a parametric study of component sizes and operating strategies. Status The SNL and Sentech study team visited Alaska and had discussions with the director and staff of the Alaska Energy Authority, with Chugach Electric Coop- erative staff and management, and with Alaska Village Electric Cooperative (AVEC) personnel. The perspec- tives of each organization were obtained, and there was strong support for the project. The objectives of the Alaskan organizations for this project are to save diesel fuel and associated costs, reduce pollution and envi- ronmental threats from fuel transportation and storage, and improve the economic conditions of the more than 200 remote communities in Alaska. Each organization provided the study team with invaluable data on the Temote power systems in Alaska, on diesel generation performance and operation, and on the conditions of fuel transportation and storage. A database was devel- oped by a SNL summer student from Metlakatla, Alaska, which characterizes most of the remote power systems and communities in Alaska. To analyze the potential of BES systems in remote villages, the study team first obtained actual load data from several off-grid Alaskan villages. Based on sum- mary data of installed generating capacity from the Alaska Energy Authority, the analysts classified vil- lages into three categories: small, medium, and large (corresponding to less than 1 MW, between 1 and 10 MW, and greater than 10 MW of generating capac- ity). Working with officials in the Alaska Energy Authority and local utilities (in particular, Chugach Electric Association and the AVEC), the analysts iden- tified representative villages with available load data. After site visits and discussions with operating person- nel, further analysis was performed on the data col- lected from Chistochina, Alaska, (representing a small village) and Selawik, Alaska, (representing a medium village). In order to minimize system cost, initial analyses were conducted making use of existing generating equipment. Using current configurations and common operating procedures in Alaskan villages, the study developed two operating regimes for incorporating BES into the village power supply. The first uses the storage system as a peak-shaving/load-leveling device. The diesel is run constantly, but it is run at a more efficient full loading. The BESS discharges during the peak 2. INTEGRATION 2-13 hours, meeting the load demand and recharges during the off-peak hours. The second regime uses the storage system to provide full village power during off-peak hours. The diesel only runs during peak hours and any additional time that is needed to recharge the battery. Under this system, the diesel also runs fully loaded, but incurs fewer operating hours and has more start/stop cycles. Based on preliminary modeling, the analysts de- cided that the first operating regime (peak-shaving/ load-leveling, also referred to as “power smoothing”) produced substantially higher fuel savings than the sec- ond operating regime (referred to as “cycle-charging”). Additional analysis of the power-smoothing regime indicates that the largest fuel savings occur in smaller villages with lower baseline fuel efficiencies and higher fuel costs. The study indicates that Chistochina could save approximately 12% in fuel consumption for each month the system is in operation. Larger villages, in- cluding both “medium” and “large” villages as defined for this study, take advantage of larger, more efficient diesel gensets and have a larger, more diverse load base, allowing for increased constant loading of the existing diesels (absent significant industrial loads). Hence the study indicates fuel savings in Selawik of only 6%. Economic analyses indicate that reducing BES system capital costs will increase the market op- portunities of these systems to more villages. The approach to data collection was to obtain die- sel generator and system operating performance data for common units used in remote U.S. villages. Nu- merous reports were reviewed from Alaskan Village Power Systems, DOE, and SNL. The study team con- tacted the Alaskan Energy Authority, the AVEC, and Chugach Electric Cooperative to identify appropriate sites for data collection. Based on an initial review of Alaskan generation data and telephone consultations, the analysts identified three sites from which to collect data. The site evaluation was based on several criteria including diesel fuel cost, site location, availability of operating system data, diesel capacity, and operational history and fuel consumption. The sites evaluated for the study, corresponding to small, medium, large and multi-generating units, were Chistochina, Selawik, and Kotzebue. Details on these sites are presented in Table 2-8. The map shown in Figure 2-8 illustrates the loca- tion of the selected villages relative to other major cities and villages in Alaska. Kotzebue was eliminated from the study because the low cost of electricity there would not support battery storage implementation for this ap- plication alone. Site visits were conducted from June 28, 1999, through July 2, 1999, to obtain comprehensive field data. Meetings with AVEC, Chugach Electric Coop- erative, and local site operators were held to discuss the overall scope of the project and to get support for the data collection effort. The analysts then reviewed all data obtained and compiled the data into appropriate Excel spreadsheet files. Based on discussions with the Alaskan utility operators, the analysts developed two battery/diesel-operating scenarios to model system op- eration performance. The analysts then contacted diesel generator manufacturers, including John Deere, Cater- pillar, Cummins-Onan, and Detroit Diesel, to obtain manufacturer genset operating specifications. Battery and PCS manufacturers were also contacted to obtain price and performance data on those systems. These manufacturers include GNB Technologies, C&D Tech- nologies, Exide, and Applied Power. Diesel/Battery Hybrid Model Results The battery system for each village was specified based on the gross power and energy requirements of the village and available hardware. The battery specifi- cations were then refined based on iterative runs of the model, until performance parameters were “optimized.” The battery system proposed for Chistochina consists of a single string of 40 6-V cells, each cell having a nomi- nal capacity of 300 Ah. This produces a battery system Table 2-8. Alaskan Village Diesel System General Overview-— Evaluation Results Generating System ; - Electricity Rate Avg. Fuel Price Size Capacity Village (cents/kWh)* (siGal) Small 160 kW Chistochina 36.8 83 Medium 1571 kW Selawik 46.6 1.63 Large 11,180 kW Kotzebue 21.8 85 — Rate before applying power cost equalization subsidy. 2-14 2. INTEGRATION Figure 2-8. Villages in Alaska. with a potential of 240 Vdc and 72 kWh of energy stor- age. A quick review of available marketing literature confirms the availability of lead-acid batteries meeting these general specifications (that is, capable of forming a 240-Vdc, 72-kWh battery pack, although particular cell configuration may not exactly match the modeled system). The PCS was modeled as being 95% efficient on the inverter end and 95% efficient on the rectifier end, which is consistent with commercially available, state-of-the-art PCS.* The proposed Selawik battery system consists of eight parallel strings, each containing 120 serially con- nected 2-V, 1500-Ah cells. This results in a 12,000-Ah battery, with 2,880 kWh of energy storage, also oper- ating at 240 Vdc. Once again, commercially available products can be configured to meet this specification. The PCS for Selawik was also modeled at 95% effi- ciency for both the inverter and rectifier. Table 2-9 summarizes the BESSs that were modeled. Preliminary runs using the National Renewable Energy Laboratory’s (NREL’s) Hybrid2 model indi- cated substantially greater fuel savings using a power- smoothing approach as compared to a cycle-charge approach. Therefore, given resource and time con- straints of the current study, modeling efforts were fo- cused on further study of this power-smoothing con- figuration. * Atcitty, S., S. J. Ranade and A. Gray-Fenner. Summary of State-of- the-Art Power Conversion Systems for Energy Storage Applications. SAND98-2019. SNL. September, 1998. Regional Map Showing the Three Selected Villages Relative to Other Major Cities and Table 2-9. Summary of BESSs That Were Modeled Chistochina Selawik Cell capacity (Ah) 300 1500 Cell potential (V) 6 2 Cells in series 40 120 Strings in parallel 1 8 Battery storage 72 2880 (kWh @ 1-hr rate) PCS efficiency (%) 95 95 Figure 2-9 shows the performance characteristics of one week of operation of the Chistochina system as predicted by the model. In this system, the battery is generally kept at a fairly high state of charge (SOC). However, the diesel does not have a substantial leveling of its load. This is expected, as the BESS is primarily used as a peak-shaving system. Using manufacturer’s specified fuel consumption curves, the simulated fuel efficiency of the baseline case (diesel only) is substantially better than the overall fuel efficiency reported in Chistochina in the Power Cost Equalization (PCE) report** (12 kWh/gal vs. 10 kWh/gal). Therefore, the fuel consumption curves were adjusted to more closely match the observed fuel ** Alaskan Division of Energy. Statistical Report of the Power Cost Equalization Program, Ninth Edition, 1998. 2. INTEGRATION State of Charge 10.00 LA aust ih Biwi i ist 20% 0.00 Abit A ti att Ain} jd ig ¥ ¥ W V ; -10.00 10% 20.00 O% Wednesday Thursday Friday Saturday Sunday Monday Tuesday Date (annex — "7 "Bus Ave “~~~ Average Diesel Power ~~~ Battery Power ——— State of Charge Figure 2-9. efficiencies. Using this adjusted fuel consumption, the ESS has fuel savings of 12% over a one-month fall and winter power system simulation (using data extrapo- lated from the two-week data collection effort). The fuel savings results from using a smaller engine (60 kW) coupled with the energy storage instead of the larger engine normally used in the fall and winter months. The smaller engine runs at higher loading factors than the larger engine, and thus operates in a more efficient range. The savings occur despite losses arising from the ESS of around 40% (that is, the ESS operates at 60% efficiency), and the lower baseline ef- ficiency of the smaller genset. The 1996 average fuel cost for Chistochina was $0.78/gal including the PCE. Extrapolating the one- month simulation over a nine-month period (the period during which the larger genset usually runs in the vil- lage), the village will save about $2.5K per year in fuel costs. In a similarly sized, more remote village fuel costs could easily double to $1.60/gal due to location and restricted access to the Alaskan highway system. In this case, fuel savings would total approximately $5K per year. To obtain real-world price information for this system, the analysts contacted several battery manu- factures, and were, in turn, referred to affiliated system integrators for quotes on systems closely matching the system specifications for Chistochina. Based on the best match provided by the vendors, a system similar to the one simulated for Chistochina could cost approxi- Chistochina Power Smoothing Fall and Winter. mately $60K.* This includes approximately $30K for the PCS and $30K for the battery. With enough units sold, the cost of a PCS for the system studied here could be expected to approach the uninterruptible power supply (UPS) power electronics cost based on the quotes from affiliated system integrators. An inde- pendent SNL study of PCS costs indicates that PCS units for off-grid applications range in cost from $200 to $1200/kW, with modest expectations of price reduc- tions in the near-term (2- to 5-year time horizon).** Table 2-10 summarizes the current, quoted price for similar equipment, as well as projected future costs for these systems. In a village with low fuel costs such as Chisto- china, the simple payback period for the system would be approximately 15 years. However, in a more remote village with double the fuel cost, the simple payback period for the system would approximately match the expected replacement period for the battery. In this case, the system would just break even compared with the baseline system. Note that these calculations do not assume the purchase or maintenance of a diesel genset (since this component already exists). Adequate data correlating diesel maintenance costs with engine * Powerware Plus 160 Model 130, rated at 160 kVA: $32,300 for UPS electronics. Include 240 cells of C&D XT3KC-11 batteries (40 kW for 2 hours or 105 kW for 45 minutes) for an additional $31,750. mera 3 Atcitty, op. cit. 2-16 2. INTEGRATION Table 2-10. Summary of Current Quoted Price for Similar Equipment, and Projected Future Costs for Systems at Chistochina Item Quoted Price Future Range Flooded Lead-Acid Battery $750/kW* $275-350/kwW** Power Conversion Equipment $190/kW* $175-1000/kW" * Powerware Plus 160 Model 130, rated at 160 kVA, see footnote at bottom of page 2-16. **EPRI, U.S. DOE. Renewable Energy Technology Characterizations. December 1997. Appendix A. t Atcitty, S., S. Ranade & A. Gray-Fenner. Summary of State-of-the-Art Power Conversion Systems for Energy Storage Applications. SAND98-2019. SNL. September 1998. loading factors was not obtained, and thus this aspect of the system could not be adequately modeled. If oper- ating the diesel at full load decreases maintenance costs (such as through more complete combustion and less fouling of the injectors), this would improve the eco- nomics of the modeled system. In the Chistochina simulation, the ESS serves as a “peak shaver.” That is, it operates for a few minutes out of each hour (during peak hours) to cover highly transient loads above the rated capacity of the diesel. Using the same dispatch method (power smoothing), the system could also be run as a “load leveler,” where the ESS would cover the average system peak, which occurs several hours each day. Converting the Chisto- china system from a peak-shaver to a load-leveler would require substituting a 45-kW genset for the 60-kW genset already in place. However, analysis of such a system does not indicate additional fuel savings for the village when running in this mode. Although operating in a load-leveling mode improves diesel loading, as shown in Figure 2-10, it does so at the cost of cycling more energy through the battery. The gain in efficiency from improved diesel loading is not enough to offset the additional energy losses from the ESS. Furthermore, running in this mode tends to shorten the battery life expectancy (from deeper discharges), thus making the economic case less attractive. With peak system loads five times greater than those of Chistochina, Selawik offers a substantially different test case. The average winter load in Selawik is about 340 kW (based on January sample data). This provides a good match to the existing 350-kW diesel gensets already installed in the village for the purposes of “load leveling” (as compared to the “peak shaving” achieved in Chistochina). In this case, a 350-kW genset is paired with 2.8 MWh of BES to cover the same load served by the 557 kW genset. Figure 2-11 shows the modeled system perform- ance for one sample week in Selawik. In this system, the battery undergoes very deep and somewhat irregular cycling. As is typical in Selawik (based on an analysis of the full two months of data obtained), peak loads are highest during the beginning of the week, trailing off toward a notable low during the weekend. Thus the battery is quickly drained on Monday, and is only able to partially recharge before discharging again Tuesday, although to a much lower depth. The maximum state- of-charge gradually increases during the week, and the battery is fully recharged on the weekend. This system, however, allows the diesel to operate at an optimal flat- out maximum level through the entire workweek, only going into load-following mode on the weekend when the battery is fully charged. When using manufacturer-specified fuel consump- tion curves, the system does not provide net fuel sav- ings. However, as in Chistochina, the observed* village heat rate of 10.8 kWh/gal falls substantially short of the 14.4-kWh/gal heat rate predicted by the model. When the heat rate curves are adjusted to more closely match the 10.8 kWh/gal observed village value, fuel savings increase to around 6%. At $1.00/gal, this translates into $15,000/year in fuel savings cost. In addition, the vil- lage can reduce on-site fuel storage requirements by 15,000 gallons. Note that the 10.8-kWh/gal observed village heat rate was reported for the period of July 1, 1995 through June 30, 1996. This was before the in- stallation in June 1996 and June 1997 of the 557-kW (the baseline units used in this analysis) and 315-kW generating units in Selawik. AVEC reports that since this installation, fuel consumption in Selawik has im- proved to much closer to the 14.4-kWh/gal, which was calculated in the model. Therefore, it is unlikely that the modeled system would provide fuel savings benefits in Selawik with the current generating capacity. Table 2-11 summarizes the results from the three analyses conducted for this report (Chistochina System A, using a battery energy storage system to eliminate * Alaskan Division of Energy, op cit. 2. INTEGRATION 2-17 90.00 100% 80.00 90% 70.00 80% 60.007 70% 50.00 = 60% & = 40.007; e $ 50% 5 5 30.00 2 40% 20.00 30% 10.00 Foi i. ; . ha ut 209 0.007 sesadlt A aad Aol tas atte ist 20% -10.00 10% -20.00 0% Wednesday Thursday Friday Saturday Sunday Monday Tuesday Date Bus Max— ~~ ~Bus Ave Average Diesel ——— Battery “State of Charge Figure 2-10. Chistochina Power Smoothing Fall and Winter, 45-kW Diesel. 500.00 400.00 300.00 = 200.00 N a é 2 é i I\/ © 100.004 a : ; ; Lo s YM Aw ON AA i \ yt IF ¥ | : oon 0.00 ) 7 | ¥ | + i wf 4 % uf 4 | \ } Ve. 4 20% =100.00 : we A * - 10% -200.00 0% Monday Tuesday Wednesday Thursday Friday Saturday Sunday Monday Date ——~ Bus Load ~~~ Average Diesel Power ~~~~~~ Battery Power State of Charge! Figure 2-11. System Performance for One Week in Selawik — Fall and Winter. 2-18 2. INTEGRATION Table 2-11. Summary of Results from Three Alaska Village Analyses Chistochina - A Genset: 60 kW Battery: 72 kWh System Specifications BESS Cost: $60K Fuel Savings 12% Simple payback period 7yrs Chistochina - B Genset: 45 kW Battery: 72 kWh BESS Cost: $60K 12% 7 yrs Selawik Genset: 350 kW Battery: 2.9 MWh BESS Cost: $1.2M. 6% 50 yrs transient peaks, Chistochina System B using a BESS for leveling the diesel loading, and Selawik, also using the BESS to level the diesel loading): Note that the simple payback calculation for all three systems assumes a fuel price of $1.60/gal. For both Chistochina systems, the simple payback period is the same. This results from the equal fuel savings of both systems, as discussed previously. Figure 2-12 shows actual power costs attributable to fuel ($/gal divided by kWh/gal) versus the fuel price for villages with less than 500-kW installed capacity (that is, very small villages). As expected, power costs generally increase as a function of fuel price, with the variation attributable to different generating efficien- cies. The figure also shows a solid line indicating the leveled cost of energy from a battery/diesel hybrid (rep- resenting the cost of the ESS plus cost of fuel, not in- cluding diesel capital cost, maintenance, or overhead) from the Chistochina example. Villages plotted above this line should be able to economically use BESS at current battery and PCS prices. The dotted line shows leveled energy costs assuming the 72-kWh BESS costs $45K. This cost corresponds to the low-end projections for battery and PCS costs as found in the reports titled Renewable Generation & Storage Project Industry & Laboratory Recommendations and Summary of State- of-the-Art Power Conversion Systems for Energy Stor- age Applications. Achieving these cost goals will in- crease the potential Alaskan market from six to nine villages. The broken line shows the cost for BESS nec- essary to capture 50% of the small village market (about 24 villages), about $10K for the system consid- ered (that is, a cost of 1/6" of the current $60K price). Small villages have several characteristics that en- hance the value of BESS, and maximize their ability to reduce fuel consumption. First, they use the smallest, most inefficient diesel gensets. Allowing these engines to run at their most efficient loading saves enough fuel to more than compensate for the inefficiencies intro- duced by the ESS. Additionally, smaller villages do not have a sufficient customer base to level out load spikes from individual customers. In a larger village with even as much as one hundred customers, the spikes caused by refrigerators cycling, lights turning on and off, and other uses of electricity tend to balance out and produce a relatively constant load (on a small-time scale, less than an hour). In a smaller village, diesel gensets must be sized substantially larger than even the “average peak” load, so that they can adequately cover the many transient peaks. This means that they spend more time running at lower efficiencies than units in larger villages. The analyses conducted for this study indicate that shaving these peaks and leveling the over- all load of a small village can both be effective uses of BES. Such a system can save the villages approxi- mately 12% in monthly fuel consumption for months that the system is operational (as many as nine months per year). Finally, smaller villages generally pay the largest premium for fuel because fuel cannot be pur- chased and transported at the bulk scale used by larger villages. This problem is compounded in more remote villages further from regional transportation hubs. High fuel costs coupled with high fuel savings serve to greatly improve the economics of BESS. Based on computer modeling conducted for this study, BESS can provide an economical way to reduce fuel consumption in remote villages. The economics of such systems depend on several factors, including vil- lage size, fuel costs, and operating efficiency of existing equipment. Field studies of systems operating under actual conditions will help confirm the results of this model, as well as provide valuable experience in devel- oping and operating these systems. Field studies will also allow for the evaluation of other potential system benefits, such as improved power quality and reduced maintenance costs. As systems improve in both cost and performance, they may become more economical in larger size villages. Ultimately, these systems could be an important part of reducing power costs in remote Alaskan villages. 2. INTEGRATION 2-19 0.3 8 2 bb 0.15 2 Power Cost Attributable to Fuel ($/kWh) © g 1.50 2.00 2.50 Fuel Price (S/gal) © Village Data from PCE Battery Hybrid, Current Cost = = = Battery Hybrid, Projected Cost —-—~ Batter id, 50% of Market Figure 2-12. Fuel Price vs. Power Cost. Renewable Generation and Storage and Related Projects The RGS Project is an initiative that began in FY97. This project is envisioned as encompassing the investigation of modular, integrated RGS systems ca- pable of control by utilities and other electricity suppli- ers. The RGS Project includes wind and PV generation options. An integrated RGS system will provide a new op- tion for the utilization of renewable generation. At pre- sent, renewable systems have typically been site- integrated; that is, components have been specified and purchased by a designer or an architect and engineering firm and then assembled at the final site. System inte- gration using this approach may not always result in the lowest cost or most reliable system. In the case of sev- eral recent utility battery demonstrations, site integra- tion caused significant start-up problems because of control interface mismatches, low-battery SOC result- ing from prolonged storage without a charge, and power electronics failures. There are several examples of renewable systems that have been site-integrated that have had similar experiences.” In addition, utilities have traditionally not given a high-capacity credit to renewable systems. This is be- cause of the intermittent nature of most renewable gen- eration and the inherent inability to dispatch such en- ergy by the utility. An integrated, modular renewable system with storage can address many of these issues and thereby greatly increase the environmental and economic benefits of renewable technologies. An RGS system may be a cost-effective way to increase the sta- bility of power from intermittent and fluctuating renew- able resources and provide energy upon demand when the utility needs it the most, regardless of the availabil- ity of the renewable resource at that time. Projects related to the RGS initiative include fur- thering cooperative relationships with renewable (PV and wind) programs at SNL and NREL. These pro- grams utilize batteries in many of their laboratory and field experiments and have expressed interest in having technical contact with the ESS Program. These inter- actions are leading to increased collaboration on renew- * Renewable Generation & Storage Project Industry & Laboratory Recommendations. 2-20 2. INTEGRATION able technologies and storage among the national labo- ratories and will reduce duplication of effort and result in improved RGS systems. Status In FY98 the ESS Program issued a request for pro- posal (RFP) for the first phase in this project. Contracts were awarded to three companies to conduct three- month studies that focused on integrating ESSs with renewable energy generation. These studies were com- pleted in FY98 and the results are summarized in SNL’s ESS Program report SAND98-0883 for FY98. In FY99, the final reports for each of the three contracts were published: e AeroVironment, Inc—SAND99-0936, Solar- powered Systems for Environmental Remediation e Solarex—SAND99-1477, Investigation of Synergy Between Electrochemical Capacitors, Flywheels, and Batteries in Hybrid Energy Storage for PV Systems e Ascension Technology, Inc-—SAND99-0935, System and Battery Charge Control for PV- powered AC Lighting Systems The ESS Program determined that the results for Phase I were satisfactory enough to justify moving to Phase II of the project. Consequently, a second RFP was issued soliciting bids for follow-on work. Three proposals were received and evaluated by ESS Program staff and staff from the Photovoltaics Program at SNL. One contract for Phase II work was awarded to Ascen- sion Technology, Inc., now a division of Applied Power Corporation (APC). Ascension Technology’s final re- port for the Phase II work was submitted at the end of FY99 and will be published in FY2000. The results of Ascension’s Phase II work, as documented in the final report, are summarized below. Ascension Technology began this project by ex- ploring the options currently available for stand-alone AC lighting systems. As required by the Phase I RFP, they identified a “technology gap” where existing tech- nology did not meet consumer needs. In the case of AC lighting systems, Ascension determined that no inter- mediate-sized (35 to 100 W) lighting systems were cur- rently available, despite consumer interest in the prod- uct. To meet this need for intermediate-sized systems, the Ascension team reviewed currently available AC lamp technology and developed battery and load- charging specifications for an intermediate-sized sys- tem; they then developed a prototype design for a com- plete, stand-alone AC lighting system, including the power electronics. PV modules, batteries, AC lamps, AC ballasts, fixtures, and luminaires are all mature mass-market technologies. The missing element in implementing a complete intermediate-sized PV-powered AC lighting system is an integrated controller to tie all of these components together to form a reliable, high-quality system. The work completed during Phase I resulted in a patent application being filed for such a controller. The purpose of the application is to “prevent early fail- ure of either the lighting load or the battery” in a PV- powered AC lighting system. Work during Phase II of the project focused on further developing the prototype controller to a point where it would be suitable for field and laboratory testing. The prototype controller described in the patent application has two main features that address the reli- ability of industrial-grade AC lighting systems. First, the application says, “the system reduces injection of DC current into the load and, as a result, extends the operation life of the load, particularly if the load is an AC lighting lamp or system.” Second, “the system op- erates the load in an optimal manner so that battery storage is maintained near full charge, yet the lighting load operates for a maximum number of hours per night. Smart, adaptive algorithms in the controller and inverter determine how long the light can operate each night. Initially the light operates on preset hours of operation entered by the user, but as the system learns about battery capacity and lighting load, it begins to automatically control how long the light is on so that the battery SOC is maintained high rather than low. The system learns the characteristics of the battery by measuring voltage, charge and discharge current, and temperature over a period of weeks. One of the advantages of this type of system con- trol is that the system design-day no longer has to be the worst day of the year. PV lights that are located in a recreational area that is used only in the summer can be sized for operation in the summer. As sunlight avail- ability begins to decrease through the winter months, the system will automatically operate the light a re- duced number of hours. Also, as battery capacity be- gins to decrease, so too will operating hours of the light. If an installer makes a mistake and undersizes the PV array, the controller will automatically adjust how long the lights stay on so that battery health is maintained. A 2. INTEGRATION 2-2] further benefit of this system is that it should be able to prevent premature battery failures caused by prolonged operation at low SOC because not enough PV power is available to maintain battery health. Work in Phase II included improving this system controller, designing and building a prototype AC lighting system, identifying potential demonstration sites, and developing a test plan that would adequately evaluate system parameters. Prototype AC Lighting System Components A prototype system, which uses an improved ver- sion of the charge controller described in the patent application, was delivered to SNL in late August for evaluation. The major components of this prototype system are batteries, PV modules, an enclosure for the batteries and power electronics, a “top of pole” mount for the PV array, the system controller, and the light fixture. Specifications for each of these components are provided in Table 2-12. Note that the complete system includes the lighting pole and mounting arm for the light fixture. These items were provided by SNL. The prototype controller was delivered with On/Off series charge control, but this may be changed to high frequency pulse width modulation (PWM) charge control later with firmware upgrades. The pro- totype was delivered with very basic control firmware (Table 2-13). The first step for testing the prototype will be to validate the charging and inverter power electronics. Once the hardware is deemed stable, the more sophisticated firmware will be provided for the unit. Firmware and hardware upgrades may be pro- vided periodically to SNL as product development con- tinues. An approximate scale sketch of the system design is shown in Figure 2-13. An electrical schematic dia- gram of the system is shown in Figure 2-14. Specifica- tions for the lighting system, the data collection system, and general specifications for the overall prototype de- sign are presented in Table 2-14. In addition to the lighting specifications, the sys- tem must also use Underwriter’s Laboratory (UL) listed components and be able to be installed in accordance with the National Electric Code. Those requirements are too voluminous to repeat here, but are applied to things such as DC/AC isolation, separation of circuits, grounding of the system, proper installation instruc- tions, etc. The reader is directed to UL 1741 and NEC Article 690 for further details. The prototype system was sized for operation in Albuquerque, New Mexico. The system is sized for year-round operation, all night long. During testing of the system, PV can be reduced by changing the tilt an- gle of the array. See the Test Plan below. In addition to the system specifications, Table 2-15 lists specific items that apply to the system. Candidate Demonstration Sites Each demonstration site should be a location where the cost to install PV-powered light is lower than using utility power or where high service reliability is needed, such as perimeter security lighting. Most applications will probably be in somewhat remote locations. Per- formance of these lighting systems will depend upon varying levels of sunlight from region to region, as well as varying weather and temperature. Consequently, one large demonstration system in one location would not be representative of how the product is expected to be used. Therefore it was proposed to co-locate systems with Ascension’s customers in many regions of the country. Table 2-16 shows a proposed list of customers who would be willing to provide demonstration sites. Each customer will have one system that is instrumented with a remotely operated DAS. The purpose of the DAS is to monitor battery parameters: voltage, charge/ dis- charge current, and temperature as well as plane of ar- ray (POA) irradiance. Installation plans will vary from site to site. APC will provide technical support to customers to ensure successful installations and plans to provide sufficient documentation so that an average contractor can suc- cessfully install the systems. If PV systems experts have to be sent into the field to install single lights, the systems will not be cost effective. For this application to be successful, it must be installable by the average contractor. Test Plan Should a full demonstration be funded, systems with data collection instrumentation would be installed first. As systems become operational, data collection will be done remotely from APC’s Waltham, Massa- chusetts, offices. Each instrumented system will monitor the following parameters: Battery voltage Battery current, charge and discharge Battery temperature POA irradiance coeee 2-22 2. INTEGRATION Table 2-12. Prototype AC Lighting System Specifications Battery Specifications Model Qty/system Nameplate capacity Type PV Specifications Manufacturer Model Qty/system Rating Connection Dimensions Weight Enclosure Specifications Manufacturer General SES B2600B (12 V) 4 95-Ah/battery Gel, non-spillable Siemens SR100 2 100-W/module 12V 5.6A Series connected for 24-V operation 59 x 23.4 x 2.2 (in) each 24 Ibs each SES Locking, vented, rainproof. Used in UL Listed MAPPS products. National Equipment Manu- facturer’s Association (NEMA) 3R equivalent Table 2-12. Prototype AC Lighting System Specifications (continued) Enclosure Specifications (Continued) Dimensions Mount Method Weight System Protection Wiring Method Top of Pole PV Mount Specifications Distributor Part No. Pole Mount Diameter Array Tilt Angle Adjustment Azimuth Angle Adjust 17 x 20 x 38 (in) Brackets included for round- or square-pole mount, 2 to 5 (in) diameter 380 Ibs. (approximately) with 4 batteries and controls 10-A PV breaker 30-A Battery breaker 2.5-A AC light breaker Bulkhead mounted, weath- erproof-quick electrical dis- connects for easy con- nection of PV and AC Load. No internal wiring required at time of installation other than hook-up of batteries. SES T3573 Schedule 40, 4” Pipe (4.5” actual) diameter 15 to 80° 0 to 360° 2. INTEGRATION 2-23 Table 2-13. Prototype System Controller Specifications Manufacturer Model PV Input Charger Type Charge Set Points Reverse Polarity Protection Day/Night detection AC Output Inverter efficiency AC/DC Isolation DC blocking Power On Soft Start No Load shutdown Applied Power Corporation LC-100 (prototype) 60 V max volt open circuit (Voc) 24 A max charge rate 40 V typical Voc On/off series controller 32.0 V over-voltage shut- down 28.2 V disconnect voltage (Vr) 27.4 V reconnect voltage (Vrr) 26.0 V low voltage discon- nect and reconnect (LVDR) 23.4 V low voltage discon- nect (LVD) Yes PV array voltage 125/230 V (switch select- able) 60/50 Hz (dip switch select) 150-VA design rating (Tested to 60 VA so far) 85% (will be improved as design matures) 90% design target 2-kV min transformer isola- tion Transformer Yes To be determined Table 2-13. Prototype System Controller Specifications (continued) User Inputs User Display Battery connection DC System Grounding Light Fixture Specifications Manufacturer Model Lamp Rating AC Load Power Factor Cor- rection Light Sensor Style Mount Method Dimensions Weight 16 position select switch (for operating mode) 4 pos. dip switch 50/60 Hz select Flooded/Gel TBD (to be determined) Test button for 5 minutes of light operation PV Charging indicator light emitting diode (LED) Power ON LED/misc test code display AC Light ON indicator LED 24V (nominal) 20- to 35-V operation Positive ground General Electric M-250/A2 50-W HPS 60-W typical Yes Not required, but may be used. Cobra-head style luminaire Horizontal mount on 1.25”- 2.00” pipe. 11.5 x 13.25 x 27.5 (in) 25 Ibs 2-24 2. INTEGRATION Table 2-14. Prototype System Table 2-14. Prototype System Specifications Specifications (continued) Lighting System Data Collection Specifications System Lighting load 50 W high-pressure sodium Specification Niue treanel factor Data logger type Campbell Scientific, CR 900 Light height 20 tt . Communication Cellular modem Pole height 25 tt Remote data Yes, dial up to download capability historical data as well as Pole mount OMhoedeas current operational data. Array mount Top of pole mount with ad- Measured Plane of array (POA) justable tilt angle parameters Insolation Array tilt angle Varies depending upon lo- Battery temperature cation, whether lighting will Battery voltage be provided for year-round use or seasonal use, and Charge/discharge current in taaa hor as Data logging 10 minutes, average of all * interval measurements during Hours of light Varies depending upon previous 10 minute interval. operation PV Array Sizing DC system voltage AC system voltage Maintenance interval Wind design load application. In some cases all night operation is desired; in some cases split evening and morning operation are desired, etc. The system must be capable of being sized for all night operation. Depends upon the solar resource and lighting load requirements. Typical array sizes would be 200 or 300 W at 24 V. 24V 120 V 60 Hz Yearly maintenance, system checkout, cleaning. Battery replacement, typi- cally at 4 years, but with the new controller this is TBD. 100 mph Maintenance interval Packaging Power supply Date and time stamping included. Only when data system indi- cates maintenance is required. The data logging equipment will be included in the elec- tronics compartment of the battery box along with the LC100 system controller. The data logger will operate off of the same 24-V battery bank as the lighting system. A separate battery discon- nect will be provided so that the controller can be powered off while maintain- ing power to the data system. 2. INTEGRATION 2-25 — <a Table 2-15. Prototype Design Specifications Array tilt angle 50° Design month December Design sun-hours 5.8 PV array rating 200 W Estimated battery efficiency 80% Inverter efficiency 90% (target) AC load 60 W Hours operation/night 13.92 hours/night DC load amps 2.67 A Discharge rate c/71 Days autonomy 5.1 Battery capacity _ 190 Ah @ 24V Figure 2-13. System Design (approximately to scale). Data will be monitored continuously, and average data for each period will be stored to memory every 10 minutes by the on-site DAS. During the first full year of operation, data will be collected from each system weekly. A mathematical system performance model will be developed for checking the collected data. The col- lected data will be compared against this model to de- termine if the system is performing within limits. The questions that should be answered by the collected data include: e Are the controllers operating properly to maintain batteries near high SOC most of the time? e Is there room for improvement in system perform- ance? e Could the systems operate the load longer and still maintain battery health? e Are the systems operating the load too much and drawing down the battery too much? e Are there any hardware failures? If so, of what nature are they? Because the charger is microprocessor controlled and has provision for serial I/O, providing an interface between the DAS and the system controller may be a consideration. The purpose of this type of interface would be to allow remote modification of the controller algorithm and to collect data only known to the con- troller, such as estimated SOC. After one year of testing, a report on system per- formance and lessons learned will be prepared. It is important to test for a full year; otherwise some sea- sonal operating effects may be omitted. Evaluation of the prototype system using the test plan described above continues at SNL. The next step in commercializing this product would be toward labo- ratory testing and further design iterations of the system controller. The test plan and milestones for such testing are presented in the final report for this contract, which will be published in FY2000. System Evaluation Mobile Power-Quality System The goal of the Mobile Power-Quality System Project is to further the development of prototype bat- tery systems built with commercially available and ad- vanced components and to evaluate these systems in typical utility operating environments. The project covers the design, fabrication, siting, installation, test- ing, and reporting on these systems. The systems are designed to be moved to a new location (on the same or 2-26 2. INTEGRATION WARNING !! DC SIDE OF SYSTEM USES POSITIVE (+) GROUND sorsmr BLK 2 POLE PV CONNECTOR | SBAte 12¥ Gel i-e——Hi LC-100 AC PV LIGHT CONTROLLER AC OUT.LINE 115/230 NC. ) acovrwevmat UNE oad ID No BLK Sale 12V Get + 24° No 4 RED IT NOBLE No 10 GRN Figure 2-14. System Electrical Schematic. on a different utility grid), installed, and tested. They are being developed for use by one or more utilities over several years to obtain field data at more than one site to prove reliability, functionality, and cost- effectiveness. Status On September 5, 1998, the Mobile PQ2000 system was installed at the service entry of a printing plant in Richmond, Virginia. Two hours after the system was installed, a power failure for the plant was averted as the system sensed a sag and activated to support the plant until power was restored. Three days later, a fault in the feeder lines between the Mobile PQ2000 and the plant resulted in a fire in the conduit, which led to a shutdown of the entire plant. Although the unit was not responsible for the problem and in no way contributed to the fault, plant management requested that the sys- tem not be re-installed when the repairs were com- pleted. Consequently, the Mobile PQ2000 was placed in temporary storage at the Virginia Power Iron Bridge facility where it remained at the end of the quarter while Virginia Power sought out a new customer. Following the reinstallation and start-up of the Mobile PQ2000 at the Iron Bridge facility, no further system testing or operations were performed other than maintenance of battery charge to ensure that the batter- ies remain at top of charge. Efforts continued 2. INTEGRATION 2-27 Table 2-16. Proposed Demonstration Sites Customer Location oy en — —* Black Hills Power & Light South Dakota 3/1 National Park Service Northwest 3/1 U.S. Forest Service Northwest 3/1 Bureau of Land Management West Virginia 3/1 General Services Administration Washington 3/1 Northern States Power Minnesota 3/1 Austin Energy (formerly City of Austin Electric) Texas 3/1 Northeast Utilities Connecticut 3/1 Wisconsin Public Service Wisconsin 3/1 Tucson Electric Arizona 3/1 Arizona Public Service Arizona 3/1 NREL, Outdoor Test Facility Colorado Ww SNL New Mexico 3/1 Total 37/13 throughout the second quarter to find a new owner for the system. The Mobile PQ2000 remained connected to the utility at the Virginia Power Iron Bridge Facility until a new customer location was found. In early May, a lease agreement for the system was reached between Virginia Power and S&C Electric Company. On May 13, 1999, the Mobile PQ2000 was removed from the Iron Bridge facility in Richmond, Virginia, and began its journey to the S&C Electric Company plant in Chicago, Illinois. The over-the-road PQ2000 configuration is shown in Figure 2-15. The unit arrived at S&C on Friday, May 14, at 8:30 p.m. S&C installed the unit to protect their polymer product fabrication plant as shown in Figure 2-16. On Satur- day, staff from Omnion’s production group were on site to perform the pre-energization checks. S&C installed the 480-V buss taps to feed the Mobile PQ2000 and energized the buss on Sunday. Omnion personnel completed the on-site start-up testing with a 250-kW load bank on Monday. On Tuesday, the system was fully operational and was placed on line. Several hours later, the unit responded and protected the plant from a power interruption that would have resulted in shut- down of critical plant equipment. The Mobile PQ2000 is scheduled to remain at S&C Electric through the first quarter of FY2000. Monitor- ing and analysis of system operation will be conducted by S&C and reported to SNL. Details of the perform- ance and system analysis will be available in future SNL reports. VRLA Battery Test at Vernon A cost-shared development effort was initiated in FY91 with GNB for the improvement of VRLA battery designs to meet utility application requirements in the mid- to late 1990s. Existing VRLA batteries were de- signed to meet the needs of automotive, telecommuni- cation, and industrial applications, not utility applica- tions, which have different operating conditions. The GNB effort developed and put into production im- proved VRLA batteries designed to meet these utility needs. The final deliverable from the development pro- gram consisted of a large battery (250 kW/500 kWh) whose operation was to be evaluated in a field test. In the beginning of FY96, a contract was awarded to GNB to perform this testing over a four-year period. Costs for the test are being shared by GNB at 50%. The de- liverable battery has been incorporated into a 3.5- MW/3.5-MWh battery system that has been installed by GNB at its lead-recycling center in Vernon, California. To match the rest of the battery system planned for Vernon, and because of its established production capa- bility, the battery design chosen as the deliverable was the ABSOLYTE IIP. The primary application for the battery system at Vernon is to provide emergency backup power to critical loads at the facility dealing with environmental (air emission) controls. It is also 2-28 2. INTEGRATION Figure 2-15. Mobile PQ2000 in Transit. Figure 2-16. The Mobile PQ2000 Installed and Operating at S&C Electric Company. 2. INTEGRATION 2-29 being used in a peak-shaving mode for demand reduc- tion that will lower electricity demand charges for the facility and take advantage of lower off-peak energy costs. The battery system has been in operation since the first quarter of FY96. Tests of the peak-shaving mode began in July 1997 and have continued in FY98 and FY99. A primary objective of these tests was to find the optimum trigger power level for the battery system to supply the plant demand during peak shaving. The desired nominal DOD for the battery is 30 to 40%, with the remainder of the capacity being reserved for plant support during utility power outages. Ground-fault problems resurfaced during the first part of FY98, causing the load peak-shaving function to be unavailable for part of the time. Extremely wet weather conditions during the fall and early winter were believed to be a significant contributor to this situation. Computer malfunctions in the data storage system also occurred frequently during this same time. During the spring, both of these causes for BESS outages abated enough to resume load peak shaving, although ground- fault alarms were still commonly found. Replacement of a number of suspect leaking cells seemed to temporarily reduce the occurrence of ground faults. In May 1998, aggressive peak shaving was resumed, and the trigger level remained set at 3100 kW for the last five months of the fiscal year. In September 1998, the battery SOC at the end of the peak-shaving period was approaching the desired 60 to 70% range. During FY99, the goal of this testing was to operate in the peak-shaving mode consistently enough to observe seasonal variations in the optimum threshold setting, if any. Status Peak shaving operations continued smoothly into the month of October 1998, with a reduced trigger level setting of 3050 kW. On October 15, PCP#2 (inverter) tripped for unknown reasons at approximately 3:30 p.m. This was discovered several hours later and reme- died via a basic system reset. As a result, peak shaving benefits were nullified for the month, so it was decided to disengage peak shaving for the balance of the month. The BESS, however, continued to function in its pri- mary role of UPS support. With daily peak shaving now inactive, GNB made efforts to address ground fault sensitivity concerns. On October 22, the ground fault resistors had like resistors added in parallel connections inside the ground fault cabinet, thus halving the resistance and doubling the current required to trigger a ground fault alarm. The intent of this modification was to allow higher ground fault current, which would expedite the evaporation of any electrolytic pathways generated by (a) cell leaks or (b) excess moisture in the BESS facility. Cell surface heating was computed to be insignificant, so system integrity was not threatened by this modification. By the day after the modification, ground fault alarms were no longer continuous; they tended to cycle on and off as electrolytic pathways were now evapo- rating within three hours of triggering an alarm condi- tion. This modification appeared promising and may be a positive step in achieving the virtually “hands-off” system that the BESS was designed to be. During the months of November and December 1998, engineers from General Electric (GE) performed a comprehensive preventive maintenance inspection of all non-battery components. This required the BESS to be offline several times throughout these months, so peak shaving remained disengaged until the inspection was complete. Overall, several items were discovered and corrected on the BESS as a result of the GE main- tenance inspection. Peak-shaving operations resumed during the sec- ond quarter. The peak-shaving trigger level was main- tained at 3050 kW in January. In February and March, the trigger level was further lowered to 3000 kW. On February 27, 1999, plant-take-over and per- formance tests were conducted on the BESS, which proved to be a great success and a milestone in dem- onstrating the system’s reliability. These tests were intended to achieve the following objectives: 1. Verify operation of all of the BESS components that were inspected during GE’s preventive maintenance inspection in the previous quarter. 2. Examine the smelter plant’s primary incoming breaker components (which may be examined only when the plant is taken off utility power). 3. Demonstrate that the BESS will recognize a loss of utility power and compensate appropriately as designed. 4. Demonstrate that the BESS resynchronizes and reroutes utility power smoothly and flawlessly when terminating UPS support. (Both modes of power restoration, Manual and Automatic Restore, were tested following outage conditions.) 5. Observe and document remaining concerns about long-term system performance. 2-30 2. INTEGRATION 6. Qualify as an acceptance test the overall operation, performance, and integrity of the BESS. Without question, the BESS met all requirements out- lined in the test sequence. Deficiencies identified in previous months had been resolved, and the BESS demonstrated its ability to perform in both the peak- shaving and the UPS modes of operation. BESS operation in the second quarter successfully demonstrated abilities as a primary backup system. In addition, there was further progress toward achieving optimal battery utilization during winter peak-shaving periods. Aggressive peak shaving is a way to minimize costs during on-peak hours of utility demand. It is be- lieved that during winter on-peak utility hours, an opti- mum trigger level will be achieved at approximately 2925 to 2975 kW. The first and second quarter data are presented in Appendix A. Peak shaving continued smoothly throughout the third quarter. Tables and charts showing key operating characteristics for the battery as well as plant demand are presented in Appendix A. To complete the preced- ing winter season, the BESS’ peak-shaving trigger level was lowered to 2975 kW for the month of April. Aver- age weekday DOD was only 79.9%, so optimum winter peak-shaving settings were now surmised to be near 2925 + 25 kW. For the months of May and June, the peak-shaving trigger level was raised slightly to 3100 kW to compensate for the longer hours during the utility’s summer season on-peak demand window (12:00 p.m. to 8:00 p.m. weekdays). As of June 1, 1999, the choice of BESS peak- shaving settings was handed over to the local engineering staff at the Vernon smelter facility, to which the BESS provides its support. For May and June, the average weekday DOD was only 87.1% at the 3100-kW trigger level selected, so a more aggressive setting was adopted in the months to follow as the local engineering staff gained more understanding of their facility’s on-peak demand characteristics. Peak shaving operations continued relatively smoothly throughout the fourth quarter (July through September 1999). Tables and charts showing key oper- ating characteristics for the battery and plant demand are shown in Tables 2-17 through 2-19 and Figure 2-17. For the months of July and August, the peak shaving trigger setting was maintained at 3100 kW as the on- site engineering staff became accustomed to the BESS’ overall operating characteristics while determining op- timal peak shaving parameters. BESS peak shaving operation was only moderately aggressive for these two months, with the average weekday SOC dropping to 77.5 and 78.0%, respectively, for July and August. In September, the peak shaving trigger setting was lowered to 3025 kW. During summer on-peak hours, this setting was anticipated to be the initial optimal level of peak shaving at which maximum cost savings can be attained while still maintaining the BESS within the proper SOC window to continue peak shaving on any given weekday of the month. On Sunday, September 26, inverter No. 2 experienced an instant fault due to two loose relay connections within the inverter unit itself. GE’s service group was notified the following morning (Monday), however scheduling did not permit them to repair the problem until Tuesday morning, September 28. Because of this, the BESS was unable to perform its peak shaving duties on Monday, thereby allowing the Vernon utility to realize a larger kilowatt demand from the smelter plant. It was observed that Monday’s on-peak period generated a kilowatt demand of approximately 3480 kW, so the peak shaving trigger level was adjusted to 3500 kW so as to offset any more costs (via having a demand greater than 3500 kW) for the balance of the month. The fourth quarter BESS operation was largely routine with the exception of the last three days of the quarter. Peak shaving will begin the next quarter at an aggressive level of 3025 kW, with anticipation of going lower as the winter season approaches. The BESS has been virtually hands-off, performing both (a) its pri- mary purpose of UPS support, and (b) providing ag- gressive peak shaving benefits. No alarm conditions were encountered, and attention to the BESS was minimal with the exception of periodic data acquisition. Further, in August the Vernon BESS interface computer was replaced in order to ensure Y2K compli- ance. The interface computer is what permits GNB engineers to manipulate the BESS’ peak shaving set- tings, as well as to observe its performance, alarm status, trending, event histories, overall one-line status, and data acquisition, etc. A new Windows NT-based system was installed, with upgraded software and hardware components. The new interface computer was tested for all critical Y2K dates, and passed all tests successfully. Field Test Data Management Project In FY99, SNL initiated a project to develop a com- prehensive database of current fielded battery energy storage systems for off-grid, stand-alone, and grid-tied 2. INTEGRATION 2-31 = a Table 2-17. July 1999 Data from Vernon BESS Operations oe July 1999: 3050-kW Peak-Shaving Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period Values (at left) Period 07/01/99 R 586 3065 3481 416 90 % 07/02/99 F 544 3041 3435 394 87 % 07/03/99 Sa 249 2938 3445 507 92% 07/04/99 Su 0 2821 3083 262 94% 07/05/99 M 8 2920 3162 242 93 % 07/06/99 T 391 3032 3490 458 91% 07/07/99 Ww 651 3085 3475 390 86 % 07/08/99 R 343 3030 3465 435 92 % 07/09/99 F 551 3066 3505 439 86 % 07/10/99 Sa No PS - 2891 3202 311 - 07/11/99 Su No PS - 2936 3405 469 - 07/12/99 M 639 3080 3488 408 77% 07/13/99 T 725 3087 3490 403 74% 07/14/99 Ww 722 3081 3502 421 89 % 07/15/99 R 632 3045 3487 442 88 % 07/16/99 F 807 3092 3516 424 73 % 07/17/99 Sa No PS 1148 3130 3409 279 - 07/18/99 Su No PS 616 3065 3315 250 - 07/19/99 M 805 3092 3483 391 71% 07/20/99 T 826 3099 3587 488 63 % 07/21/99 Ww 1036 3103 3470 367 67 % 07/22/99 R 885 3101 3518 417 71% 07/23/99 F 747 3099 3502 403 58 % 07/24/99 Sa No PS 1950 3220 3470 250 - 07/25/99 Su No PS 1385 3143 3627 484 - 07/26/99 M 747 3099 3503 404 79 % 07/27/99 T 777 3099 3506 407 75% 07/28/99 WwW 693 3101 3573 472 53 % 07/29/99 R 731 3094 3482 388 88 % 07/30/99 F 859 3101 3513 412 55 % 07/31/99 Sa No PS 2146 3199 3420 221 - (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 3627 Average for entire month: 765 3063 3452 78.8 % Average for weekdays only: 668 3073 3483 77.5 % 2-32 2. INTEGRATION Table 2-18. August 1999 Data from Vernon BESS Operations August 1999: 3000-kW Peak-Shaving Trigger Level No. of Average kw Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period Values (at left) Period 08/01/99 Su No PS 1333 3172 3480 308 - 08/02/99 M 568 3077 3542 465 81% 08/03/99 le 479 3074 3638 564 83 % 08/04/99 WwW 564 3071 3525 454 90 % 08/05/99 R 672 3078 3478 400 60 % 08/06/99 E 541 3074 4097 1023 50 % 08/07/99 Sa 72 2975 3493 518 90 % 08/08/99 Su 636 3076 3420 344 87 % 08/09/99 M 728 3095 3602 507 69 % 08/10/99 T 605 3099 3452 353 73 % 08/11/99 Ww 865 3097 3511 414 69 % 08/12/99 R 717 3096 3543 447 73 % 08/13/99 E 594 3066 3515 449 89 % 08/14/99 Sa No PS 1139 3141 3517 376 - 08/15/99 Su No PS 1747 3213 3529 316 - 08/16/99 M 639 3067 3476 409 85 % 08/17/99 T 513 3001 3486 485 71% 08/18/99 WwW 786 3099 3511 412 75% 08/19/99 R 839 3101 3622 521 69 % 08/20/99 F 2160 3100 3480 380 82 % 08/21/99 Sa No PS 1117 3131 3411 280 - 08/22/99 Su No PS 752 3102 3439 337 - 08/23/99 M 616 3070 4617 1547 86% 08/24/99 T 607 3094 4730 1636 78 % 08/25/99 WwW 508 3085 5079 1994 88 % 08/26/99 R 836 3096 4755 1659 74% 08/27/99 Fe 740 3079 4521 1442 91% 08/28/99 Sa No PS 838 3140 3617 477 - 08/29/99 Su No PS 1116 3106 3424 318 - 08/30/99 M 898 3096 4604 1508 89 % 08/31/99 a 854 3071 3584 513 90 % (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 5079 Average for entire month: 809 3092 3764 78.8 % Average for weekdays only: 742 3081 3880 78.0 % — —<—— 2. INTEGRATION 2-33 = Table 2-19. September 1999 Data from Vernon BESS Operations — September 1999: 3000-kW Peak-Shaving Trigger Level = No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period Values (at left) Period 09/01/99 WwW 184 2816 3361 545 93 % 09/02/99 R 28 2612 3786 1174 93 % 09/03/99 F 370 2919 3721 802 88 % 09/04/99 Sa NoPS 1012 3015 3348 333 - 09/05/99 Su NoPS 1038 3051 3318 267 - 09/06/99 M No PS 254 2998 3257 259 - 09/07/99 T 630 3020 3410 390 61% 09/08/99 Ww 854 3024 3584 560 60 % 09/09/99 R 1193 3026 3442 416 76 % 09/10/99 F 832 3021 3380 359 61% 09/11/99 Sa NoPS 2103 3165 3390 225 - 09/12/99 Su NoPS 2132 3236 3657 421 - 09/13/99 M 826 3022 3448 426 58 % 09/14/99 T 980 3027 3424 397 68 % 09/15/99 WwW 891 3025 3422 397 69% 09/16/99 R 838 3025 3487 462 52% 09/17/99 F 923 3025 3427 402 74% 09/18/99 Sa NoPS 2107 3174 3342 168 - 09/19/99 Su NoPS 1811 3092 3465 373 - 09/20/99 M 1009 3013 3210 197 90 % 09/21/99 T 1042 3025 3125 100 92 % 09/22/99 Ww 1282 3028 3353 325 89% 09/23/99 R 784 3026 3380 354 66 % 09/24/99 F 877 3026 3454 428 58 % 09/25/99 Sa 1391 3028 3336 308 91% 09/26/99 Su 2160 3267 3637 370 - 09/27/99 M 2160 3385 4010 625 90 % 09/28/99 T 443 3399 3842 443 91% 09/29/99 Ww 504 3433 3849 416 91% 09/30/99 R 263 3318 3800 482 92 % (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: Average for entire month: Average for weekdays only: 1031 780 3075 3055 4010 3489 3508 774 % 76.8 % 2-34 2. INTEGRATION Plant Demand, kw Figure 2-17. 1999, systems. The purpose of the database is to provide in- formation on the operation and management of these systems in order to learn how to design future systems to avoid problems identified from the evaluation of fielded systems. The database is intended to provide practical field performance data for potential customers. Four contractors were selected for the initial phase of the project, two for off-grid analysis, and two for on- grid analysis. The first task was to identify a site that would provide historical data for an energy storage system in operation that met the criteria for analytical tasks called out in the contract. Primary tasks included: e Acquiring data, in comma-separated variable (CSV) format that was generated by the existing data acquisition systems (DASs); e Evaluating data quality and making reliability decisions as appropriate; e Developing a Microsoft Access database for the data; and e Designing analytical tools to perform initial analyses of the system to determine the performance of the system. Average and Peak Plant Demand at Vernon for the Months of July, August, and September Additional tasks were to determine the adequacy of the DAS used for the fielded systems, determine the adequacy of the battery management strategy applied to the fielded system, and to make recommendations for ways to increase the reliability of DASs. Because of the diversity of the fielded systems, contracts were is- sued for both off-grid and grid-tied applications with the intent to develop a common database for all storage applications to assist in future data review and analysis. Although the four applications selected for analysis are very different from each other, it is anticipated that all performance information will ultimately reside in a single database. To develop the initial database structure, a data kernel was defined that was anticipated to apply to all battery applications. The kernel consisted of the fol- lowing time-referenced parameters: Ah lifetime accumulation, Wh (AC and DC) lifetime accumulation, State of charge, System DC voltage, Overcharge (%), System power, DC current, and Battery temperature. 2. INTEGRATION 2-35 The kernel also included the duty cycle count, and failures and/or problems, which would not be time ref- erenced. The kernel would be identical for all systems evaluated. Because of the diversity of the systems, it was anticipated that unique features for each system would also need to be included in the database, they would connect only to the specific system to which they related. The four contractors were directed to focus on the design of a database that fully supported their indi- vidual system with an eye toward the integration of the three other systems into the database structure and analytical tool application. It is anticipated a follow-on effort will be required in FY2000 to wrap up the work started in FY99. The following sections report on the results of the project for FY99. Field Test Data Management for Grid-connected Systems Status Metlakatla System Gridwise Engineering Company was contracted to develop a database suitable for analyzing one of the grid-connected systems, a 1-MW BESS, which was commissioned in February 1997 by Metlakatla Power & Light (MP&L) for the purpose of stabilizing its elec- trical grid and reducing its dependence on costly diesel generation. MP&L is a small stand-alone electric util- ity located on Annette Island in southeast Alaska. Its generation assets include four hydroelectric units, which together produce 5 MW, and a 3.3-MW diesel unit used primarily for load following and system sta- bilization. The BESS promised to eliminate the need for the diesel by following sharp load swings intro- duced by a local saw mill. Historical performance data for the MP&L system included 171 BESS data parameters and 73 utility pa- rameters that originated from two Supervisory Control and Data Acquisition (SCADA) systems. The data were available in a compressed, proprietary format on archive tapes. Data were converted to a usable format using software tools obtained from the SCADA soft- ware developer and in-house programming. The re- sulting electronic database was submitted as a project deliverable covering about two years of BESS opera- tion. A typical hour of operation for the BESS is shown in Figure 2-18. These data represent one-second sam- ples between the hours of 9:00 and 10:00 a.m. on July 1, 1998; a special comma-separated variable (CSV) file was created for this purpose using HistData from the station control log files. The total utility load is shown as the aggregate of all plant output including the BESS, Chester Lake, and two Purple Lake units (the third Purple Lake unit and the diesel were not operating). The BESS is shown charging and discharging at about the 0-kW level, and supporting the rapid load swings of the utility. BESS performance data corresponding to this hour were extracted from the BESS log files and are shown in Figure 2-19. It is of interest to note that the string voltage and SOC remain fairly constant while the power delivered to the grid fluctuates significantly. During the course of the hour, the hydro unit’s output rises and falls in such a way that the SOC remains at about 90%. Thus, while the BESS responds to instan- taneous fluctuations, serving the load and keeping the frequency in check, the hydro units ultimately carry the stable load. Likewise, the BESS is able to provide voltage sta- bility by rapidly varying the KVAR output. The BESS is near the utility’s load center and is controlled based upon voltage measurements taken at the Centennial Plant. After the final CSV file was created, it was easily loaded into MS Access to create the database. This database is a project deliverable, and was standardized as was agreed to at a project review Meeting in Walnut Creek, California, on June 24 and 25, 1999. The goal was to provide the parameters listed in Table 2-20 as a standardized format common to all four projects. However, not all of the parameters were available as shown in the table. In consultation with SNL, it was agreed that the operation of the Metlakatla BESS did not correspond to clearly defined cycles since the battery was always held at 90% SOC. Furthermore, the only equalization charge per- formed was at startup, and this data and initial months of operation were not available. Therefore, even by arbitrarily defining a cycle (for example, based on re- versals of DC current), the battery is never returned to full charge, the cycle count is unknown due to missing data, and there is no overcharge. 2-36 2. INTEGRATION 9am 10 a.m. Figure 2-18. MP&L Power Plant Output Including BESS between 9 and 10 a.m. on July 1, 1998. 1000 kW, % kKVAR or V 10 a.m Figure 2-19. BESS Operation between 9 and 10 a.m. on July 1, 1998. 2. INTEGRATION 2-37 Table 2-20. Database Parameters Parameter Included? Comments Cumulative Ah Yes (+ and -) State of charge Yes (SOC) String voltage Yes Overcharge (%) No No clear cycle definition. Duty cycle count No No clear cycle definition. Cumulative Wh No No integrating (AC & DC) transducer available. System power (AC) Yes Current (DC) Yes Sample cell Yes temperature In addition, the SCADA system did not include integrating watt-hour transducers, either on the AC or DC side. While it would be possible in theory to esti- mate cumulative watt-hours based upon amp-hour and string-voltage measurements, the results would be sub- ject to considerable uncertainty due to missing data. Missing amp-hour data present no problem since the integration was performed at the transducer level. However, there was no transducer that integrated volt- age. In addition to the missing data, a practical problem arises. The technique for creating CSV files results in instantaneous (not averaged) data in the 10-minute in- tervals. While this may be sufficient for slow-changing or time-integrated parameters, it leads to difficulties in reporting data for rapidly changing data. For example, BESS power output fluctuates on a sub-second basis; hence 10-minute snapshots are an insufficient basis for integrating watt-hours. While it is possible to create CSV files on a millisecond basis, it would be impractical to integrate values using this technique since the maximum amount of millisecond data that HistData could produce would be under five seconds. The CSV files would have to be manually created for each five-second interval for the two-year time span. In considering the errors introduced by missing data and dead band resolution, coupled with the im- practicality of manually integrating voltage over time, it was determined that watt-hour data could not be pro- vided with the database. The deliverable was provided on an accompanying disk to SNL, and included the files shown in Table 2-21. An economic analysis was performed using actual cost data and conventional planning practices. Key economic impacts include the following: ¢ Load Following. The BESS provides load fol- lowing service, nearly eliminating the operation of the diesel generator set. This saves about 450,000 gallons of fuel/year and reduces the net staffing re- quirements by 3.75 “full time equivalent” technical positions. e Reserve Generation Capacity. The BESS defers the need for additional reserve peaking power throughout its design life of 10 years. ¢ Voltage Regulation. The BESS defers investment in two 450-kVAR switched capacitor banks. ¢ Customer Claims. The BESS reduces voltage transients on the distribution lines, saving costs of damaged customer equipment. Results of the economic analysis are presented in Table 2-22. The actual benefit/cost ratio is calculated as 1.97 relative to conventional resource planning, and the payback period is three years. It is important to note that the primary benefit pro- vided by the MP&L BESS is system stability. This is possible since the total capacity of the BESS (1.6 MVA for 10 seconds) is as high as 40% of the total peak load. Even the continuous rating of the BESS (1.0 MVA) represents a significant portion (25%) of the peak load. For the economic results encountered at MP&L to be replicated elsewhere, a BESS of comparable propor- tion will have to be specified. As a practical matter, the ideal location will be small island grids. The optimal sizing of the BESS would be determined based upon utility-specific parameters. As a result of this project, several recommenda- tions were made for follow-on work, and they include: ¢ Conducting a detailed performance analysis based upon the converted historical data; 2-38 2. INTEGRATION Table 2-21. File Descriptions of Electronic Deliverable Filename metlakatla1.mdb MS Access database /ALG Files/*.ALG MetDataExtractor.xls files. Points List Rev 6.xls BESSDBDump.csv Description Time-stamped status information and alarms (text) MS Excel spreadsheet and macro used to create CSV files from 32-bit historical log Electronic version of “Appendix A—Points List,” of the final report. Comprehensive list of all InTouch tags, including alarms and measured data, alarm setpoints, and other information pertaining to the BESS SCADA system. UtiIDBDump.csv Similar file pertaining to the Station Control SCADA system. Table 2-22. Overall Cost and Benefits from BESS in Metlakatla System Conventional Generation BESS Net Benefit (Cost) Load following $3,869,811 $0 $3,869,811 Reserve generation capacity 400,000 154,217 245,783 Voltage regulation 25,605 8,095 17,510 Customer claims 30,723 0 30,723 BESS operations 0 93,705 (93,705) BESS capital 0 2,022,727 (2,022,727) Total $4,326,140 $2,278,745 $2,047,395 e Converting recent historical log files archived at Metlakatla; e Determining the feasibility of interconnecting the MP&L network to the larger Ketchikan Public Utility (KPU) grid, thereby utilizing the BESS for system stabilization and peaking power; e Characterizing the overall market for the island- grid stabilization application at rural electric cooperatives in the United States; e Designing and installing a server at MP&L to provide real-time data access and visualization over the internet; and e Developing a database server at SNL for providing Internet access to historical data. Brockway Standard System Energetics, Inc. was contracted to develop a data- base suitable for analyzing the other grid-connected system that was included in this study, a 1-MW/10- second power-quality system at the Brockway Standard Lithography Plant in Homerville, Georgia (Figure 2-20). A team from SNL, Energetics, and Gridwise Engineering visited the Brockway plant on May 5, 1999, to learn more about the operations and the per- formance of the PQ2000. The plant operates four production lines, two of them seven days a week. Labels are printed directly onto sheet metal that is formed into cans for food, chemicals, and specialty products (i.e., Folgers coffee, Eagle Snacks, Guy Foods, and Thompson Water Sealer) at Brockway can plants. Output averages 73.5 sheets per minute when production lines are operating (roughly 40% or 10 hours of every 24-hour day). Brockway is a full-service shop, capable of handling artwork (label design and plate preparation at subsidi- aries) and X-ray plates and screens processed on-site at Homerville. Between 1994 and 1995, on the advice of Oglethorpe Power Corporation, Brockway invested almost $600K to upgrade many of its motors that were near their end of life and to install adjustable-speed drivers (ASDs) and Omegapak adjustable frequency controllers to control all of the motors throughout the plant. The ASDs are very sensitive to power glitches that stop production lines. 2. INTEGRATION Figure 2-20. Brockway Standard Lithography Plant. There are approximately 16 motors per production line. In addition, there are motors driving the solvent disposal, incinerators, and other operations. Most of the equipment at the lithography plant dates back to the 1950s. Electricity, natural gas, and occasional propane purchases satisfy the energy requirements of the plant. Slash Pine, the local energy cooperative, is Brockway’s electricity provider, which is one of the 39 electric membership corporations that formed Oglethorpe Power Corporation in 1974. Installation of the PQ2000 was initially estimated to cost $500K, but total costs through start-up reached $1M. Brockway and Slash Pine were minor financial contributors ($32K each) while Oglethorpe Power Corp, Georgia Power, and EPRI shouldered most of the cost to install the PQ2000. In 1997, Oglethorpe re- structured into three companies, and EnerVision, the new customer service company, was interested in sell- ing the PQ2000 to Brockway. The lithography plant was not in a financial position to purchase the unit. In the end, Oglethorpe (EnerVision), Omnion, EPRI and Georgia Power agreed to let Brockway keep the PQ2000 unit; Omnion maintains the DAS. The system consists of a main system container that houses the batteries, a power controls cabinet that contains an electronic selector device (ESD), and an isolation transformer (Figures 2-21 and 2-22). The Data Acquisition System Omnion co-developed the DAS with Delphi En- ergy and Engine Management Systems. The system has operated since the PQ2000 was tested in the factory and continued operating when the battery was installed at Brockway in July 1996. The battery system became fully operational in mid-December 1996. The chro- nology is important because not all counters in the software were properly reset as the system entered op- eration. Brockway Lithography Plant PQ2000 power control cabinets Cc Isolation transformer Simulator PQ2000 cabinets ata acquisition Utility transformer Figure 2-21. Layout of PQ2000 Installation (not to scale). Figure 2-22. PQ2000 Located behind Brockway Plant. With the help of Omnion, Energetics remotely accessed and downloaded data from the PQ2000. The data from May 1996 (factory testing) through July 1999 were consolidated in a Microsoft Access database built by Energetics. The testing phase continued from May through December 1996. The first trigger event (charge) occurred on December 8, 1996, but the first discharge did not occur until January 1997, which was chosen as the starting date for analysis of the battery’s operation because all previous events occurred during the battery system’s testing phase. The testing phase data are included in the database, but not analyzed for the reporting process. Although July 31, 1999 was cho- sen as the cut-off date for the data analysis, Energetics continues to monitor and collect data. The remote access process included the capability to download data remotely from the PQ2000 DAS at 2-40 2. INTEGRATION Brockway. The DAS is housed in one of the two con- tainers sitting on concrete pads outside of the plant. Figure 2-23 shows the screen and keyboard sitting on the base of the cabinet, and Figure 2-24 shows the in- dustrial processor that drives the DAS. A Windows95 remote access package called ReachOut is used to con- nect to the computer and transfer data. Figure 2-23. PQ2000 Host Computer. PQ2000 Technical Performance Energetics performed some preliminary analysis on the PQ2000 database. The results are presented below. More time is needed to develop appropriate queries to answer additional questions about the technical per- formance of the PQ2000. 1004 90 — EB Recorded Events 80 [Gy Recorded + Unrecorded Events x re) a 3 Figure 2-24. PQ2000 Industrial Processor. The line disturbance event data was queried to deter- mine frequency of discharge events per season and per time of day in 1998 for recorded events only and re- corded with modified and unrecorded events (see Fig- ures 2-25 and 2-26). In 1998, there were 69 recorded discharges and 144 total discharges (recorded, modi- fied, and unrecorded events). Examining 1998 re- corded discharge events revealed that the majority of events (62%) occurred in the summer months (June- August) while 25% of the events occurred in the fall (September-November). This is expected since the majority of voltage disturbances occur in the spring and summer because of the frequent occurrence of lightning storms. In 1998, the spring was mild in the south. The situation is the same when the unrecorded events are included. 93 a ° Number of Events & 3 20 104 Dec-Feb Figure 2-25. Mar-May easonal Months Jun-Aug Sep-Nov PQ2000 Discharge Events by Season in 1998. 2. INTEGRATION 2-4] 7 a 60 Hl Recorded Events Recorded + Unrecorded Events 50 — Number of Events 12AM-6AM 6AM-12PM 12PM-6PM 6PM-12AM Time of Day Figure 2-26. The line disturbance analog data table was also queried to determine the time-of-day prevalence of dis- charge events. The majority of lightning storms in the Homerville, Georgia, area occur in the afternoon. This is reflected by a 44% frequency of recorded discharge events from 12:00 p.m. to 6:00 p.m. All other line dis- turbance events were spread out fairly equally among the other time periods. Including unrecorded events reinforces this time-of-day prevalence. The cumulative values for the PQ2000 operation at Brockway (Decem- ber 1996 through July 1999) are presented in Table 2-23. Table 2-23. Cumulative Values for PQ2000 Operation at Brockway: December 1996 through July 1999 Variable Value Total number of duty cycles 106 Recorded discharge output 12.4 Ah Modified discharge output 30.2 Ah Discharge output with modified and 58.8 Ah unrecorded events Normal charge input 1,755 Ah Equalization charge input 2,522 Ah Charge input 4,267 Ah PQ2000 Discharge Events by Time of Day. Figure 2-27 shows the recorded Ah discharged from the PQ2000 and the estimated Ah discharged over a two-month period. Since discharges shorter than 0.04 seconds are recorded as a discharge of zero amp-hours, Energetics did a second calculation estimating the Ah discharged during these events. Also, included in this graph is the estimated, modified, ampere-hours for missing events. Energetics estimated the amp-hours missed during these events and added them to make a final estimate of the amp-hours discharged by the PQ2000. As shown in Figure 2-27, the total estimated amp-hours discharged can be three times as much as the actual recorded data. Figure 2-28 shows the current versus time for the year 1998. Current fluctuated between 400 and 1,100 A during this year. This is typical for all opera- tional data. In 1998, June experienced the most current fluctuations due to the greater number of discharges in that month as compared to the rest of the year. Figure 2-29 shows the comparison of duty cycle charge input with both modified (zero Ah replaced with 0.02) and modified plus unrecorded events. Due to the magnitude of all normal charges into the PQ2000, this parameter is not displayed in this graph. However, the difference in normal versus duty cycle charges in 1998 can be viewed in Figure 2-30. Normal charge input represents PQ2000 battery charges that follow all 2-42 2. INTEGRATION 4.0 ~ ——-— =®= Recorded Amp-Hours 3.5 t= Modified Amp-Hours for Discharge Times of 0 3.0 =® Modified Amp-Hours for Missing Events 2.5 2.0 Amp-Hours 1.5 1.0 0.5 0.0 6/5/98 6/10/98 6/15/98 6/20/98 6/25/98 Date Figure 2-27. Brockway PQ2000 Discharge Ampere-hours Comparison in June 1998. 1000 8 Current (amps) 8 8 1WenSB 2Mer-9B 1Vey-98 SOurs8 BAB 28-Oct-98 27-DeoB Dete Figure 2-28. Brockway PQ2000 DC Current Versus Time in 1998. 2. INTEGRATION 2-43 [Hi Duty Cycle Charge EB Discharge with Unrecorded Events (1 Modified Discharge = 3/98-4/98 Sak 11/98-12/98 | 1/98-2/98 Amp-Hours 7/98-8/98 ss 9/98-10/98 tol. Date Figure 2-29. Brockway PQ2000 Charge and Discharge Comparison in 1998. 350 1 | 300 + Normal | Hl Duty Cycle Charges | 250 So | | $ 200 | ° | = | a i : | < 150 1 | 100 gett i | 50 | o -D Le ra = eee 1/98-2/98 3/98-4/98 5/98-6/98 7/98-8/98 9/98-10/98 — 11/98-12/98 Date Figure 2-30. Brockway PQ2000 Normal Versus Duty Cycle Charges in 1998. 2-44 2. INTEGRATION recorded and unrecorded discharge events as well as normal charges that were performed on the system. In the months of July through October, normal charges exceeded 600 Ah. These months contained the most recorded and unrecorded discharge events. Out of the total 144 recorded and unrecorded discharges, 60 events occurred in July through October. As Figure 2-29 shows, the duty cycle charge input is a better reflection of Ah into the PQ2000. It was expected that charge input would be greater than dis- charge output. However, discharge Ah that includes unrecorded events exceeded duty cycle charges in sev- eral months (January/February, May/June, and Septem- ber/October). It appears that by neglecting the unre- corded discharges, a more expected charge-to-discharge ratio is achieved. This supports the assumption that the unrecorded line disturbances occurred while the PQ2000 was already discharging. PQ2000 Economic Performance This economic performance analysis is based on empirical data from PQ2000 operation, cost estimates from Brockway Standard, and information from equip- ment suppliers. The Square D Omegapak ASDs are considerably more sensitive to power glitches than the motors they control. These ASDs will fault during out- ages of 0.2 seconds or greater. Of the 158 recorded discharges that occurred during the 32 months of PQ2000 operation, 96 events were longer than 0.2 sec- onds in duration. These events are the basis for the economic performance assessment. These ASD faults cause a variety of jams in the production lines that can result in significant downtime. They are shown in Table 2-24 and described in more detail in the final Energetics report. Table 2-24. Brockway PQ2000 Downtime Causes and Duration Types of jams Off-line time (minutes) ASD synchronization—minor 15 ASD synchronization—major 30 Press jam—minor 20 Press jam—major 30 Oven jam-—minor 45 Oven jam—major 60 Incinerator cleanup 30 The Brockway manager estimates that each hour of production that is lost costs the plant $275, including labor and materials. By applying that value to the downtime estimated for the four types of jams avoided since the PQ2000 became operational, the yield is $27K (Table 2-25). The cumulative costs avoided by the PQ2000 are estimated to be $168K over the 32 months of its opera- tion, or $63K annually. When these avoided costs are compared to the financial stake of Brockway in the in- stalled cost of the PQ2000 ($32K), the economic bene- fits far outweigh the costs. If reasonable annual main- tenance costs of $18K are assumed, the system economics are still attractive. The key missing ingredi- ent is in the cost of electricity, which is unknown. Further investigation is needed to determine the lithog- raphy plant’s share of the total Brockway electricity bill with the Slash Pine cooperative. If the full installed cost of the PQ2000 is consid- ered, approximately $1M, then the benefits do not jus- tify the installation. Energetics had hoped that an Table 2-25. Summary of Glitches Avoided by Brockway PQ2000: December 1996 through July 1999 Simple Typical Environmental Outage Total Number of events 50 36 4 6 96 Hours off-line 12.5 57 4 24 97.5 Lost production $ 3,400 16,000 1,000 7,000 27K Additional costs $ 140K Cumulative costs $ 168 K Annual costs $ 63 K 2. INTEGRATION 2-45 alternative system (e.g., UPS) could be examined to determine how the PQ2000 fares, but Brockway insists that it would never have purchased an alternative sys- tem. Field Test Data Management for Renewable Generation and Storage for Off-grid Systems Renewable generation and storage (RGS) systems for remote power applications represent a sizable, high- value market for the introduction of renewable energy into the United States and the world economies. BESSs are a key component to the reliable and effective func- tioning of RGS applications. The white paper titled Renewables and Energy Storage: Characterization of Initial Projects and Emerging Opportunities for Hybrid Power Systems provided an overview of several do- mestic and international RGS Projects using BES com- ponents. This task determined a method for archiving, summarizing, and reporting data from these and other field tests and began the process with data from two field tests. The operational characteristics of interest include, for example, battery, capacity, type, charge/ discharge strategy, operating data, life, cell failures, maintenance provisions, and safety precautions. Data collected from systems like the Arizona Public Service (APS) Solar Test and Research (STAR) Test Facility could be included in this project activity. The ESS Program and the contractors involved in the data management project selected two sites and have performed the data analysis. The two off-grid systems both have renewable generation resources. Status Wawona Point Off-Grid Power System Endecon Engineering was contracted to develop a database suitable for analyzing one of the off-grid RGS systems. This is a summary of the results of the pre- liminary database development and analysis project. Over the course of six months, Endecon developed a number of analytical tools using Microsoft Access 97 and Excel 97, and evaluated data from the Pacific Gas & Electric Company (PG&E) Off-Grid Power System (OGPS) located at Wawona Point in Yosemite National Park. These preliminary results are intended to provide direction for further database development that will lead to a more standardized data collection and analysis process for existing and future RGS systems. The approach on this project was to focus on the development of database tools and investigate how they could be used to describe the performance of a fairly well known system. Despite Endecon’s direct in- volvement in the OGPS Project over its life, there were many unexplained events and operational quirks that make it an interesting system to analyze. The OGPS is a self-contained, PV/propane hybrid, AC power system that was developed for PG&E’s Re- search and Development Department. Major system components are listed below: ¢ 1080-WSTC, DC PV array (18 each, Solarex MSX-60 Modules) e 5-kW 120-Vac propane generator (Onan SCCK) e 16-kWh VRLA battery bank (4 each, GNB 3- 75A27) e 4-kW Inverter (Trace SW4024) e 50-A PV charge regulator (Solar Engineering PPS- 50-24) The OGPS is located about 100 feet west of a helipad on a mountain peak known as Wawona Point about 10 miles east of the south entrance of Yosemite National Park (Figure 2-31). The peak overlooks the Wawona Valley from the south. A two-way radio re- peater used for emergency and operating communica- tions within the park is located about 50 ft east of the helipad. A television translator is located several hun- dred feet to the northwest of the helipad, which pro- vides TV signals to several buildings and residences in the valley below. Utility power (240-Vac) is supplied through a cable running up the slope from a utility transformer and revenue meter located about 0.5 miles Figure 2-31. OGPS at Wawona Point in Yosemite National Park, Consisting of a 1080 PV Array, 16-kWh/24-V VRLA BESS, a 5-kW Propane Generator, and a 4-kW Inverter. 2-46 2. INTEGRATION west of the helipad. While there is a road to the site, it is snow-bound during the winter, accessible only by foot or by snowcat. The radio-repeater and television translator each require about 5 kWh/day. Taken together, this is twice the design capacity of the OGPS (ignoring the reduced available irradiance at the selected site compared to the design). The radio repeater is a critical element of the Park Service emergency communications system. It includes its own battery bank; AC power is rectified to float the battery and power the radio. In the past, when there were utility line failures, Park Service personnel would install a small generator at the radio and hope that the PG&E line would be repaired before the gen- erator ran out of gas and the backup batteries were de- pleted. It was decided that the OGPS would be used to power the radio transmitter only, leaving the TV trans- lator utility connected. It was also determined to con- nect the utility to the Trace inverter to allow sellback of excess solar generation and would provide a secondary backup to the PV and propane generator because there was some uncertainty about both the load and the solar resource (especially the effect of shade from the trees). The OGPS system was monitored with a Campbell Scientific Inc. (CSI)-based DAS. A CSI 21X datalog- ger measured 14 analog channels, including two differ- ential measurements and 12 single ended measure- ments, as shown in Table 2-26. The datalogger sampled the input channels once every second and re- corded totals and averages every 10 minutes. Table 2-27 shows the raw data as it was stored in the data- base. The ambient temperature thermocouple was lo- cated behind the array. Its position, sometimes in direct sunlight, and lack of radiation shield, caused it to read higher than expected. On July 18, 1996, the thermo- couple was glued to the back of one module thus pro- viding module temperature from that point on. Recorded Data A brief review of various summary plots of the data offers insights into the operation of the system, as well as some surprising consequences of normal data processing techniques. Figure 2-32 shows daily average maximum and minimum battery currents (the average is a tick mark, and the minimum-maximum range is shown by a verti- cal line). The first winter/spring of low-battery transfer operation (use utility as last resort) showed regular high-charge currents as the generator had to kick in Table 2-26. Wawona Point OGPS CSI 21X Datalogger Input Channels Input Channel Description 1H Array current (V Diff) 1L co 2H Battery current (V Diff) aL o 3H POA irradiance 3L Generator status (Off=1, On=0) 4H Load power status (On=1, Off=0) 4L Generator current 5H Load voltage 5L Utility voltage 6H Load current 6L Utility power 7H Load power 7L Ambient or module temperature 8H Battery temperature 8L Battery voltage every few days to recharge the battery. The following summer/fall of float mode operation showed very little discharging or charging. Figure 2-33 shows the potential error between two ways of recording battery charging current. The sim- plest way is to record a single bi-directional current value, with positive data representing charging and negative data representing discharge. However, since the recorded data are average values over an interval of time, it is certainly conceivable that for a portion of an averaging interval the battery might charge, while, for the remainder of the interval, it could discharge. If the charge and discharge amp-hours were similar, the net average current would be zero. This loss of information would be significant both in loss of knowledge about extreme current levels, and regarding integrated quanti- ties such as charge and discharge amp-hours. Since this data set includes separate battery-power-out and bat- tery-power-in output fields, we can estimate average battery-discharge-current and average battery-charge current. It is straightforward to simulate a_bi- directional output using this data. This figure presents the error between computing the daily discharge amp- hours both ways. If there was any average discharge current, then when the net current is charging, the error will be 100%. While this figure demonstrates that the error can be quite large while the battery is floating, it also shows that significant errors (>10%) can show up even during periods of time when the system appears to be cycling regularly. 2. INTEGRATION 2-47 Table 2-27. Wawona Point OGPS Database Field Description Field Label Units Description 1 Array ID =101 (defines section of code that generated output) 2 Year Four digit year 3 Day Day of year (1 — 366) 4 HHMM Time of day (0 - 2400) 5 # of Samp Number of samples in the average 6 Vac Load Vac Inverter output voltage to load it Vac Util Vac Voltage supplied by utility 8 lac Load lac Inverter current supplied to load 9 Pac Util Wac Power supplied by utility 10 Pac Load Wac Inverter output power supplied to load af] lac Gen lac Generator output current 12 TUtil Number of samples during which the utility was supplying power 13 TGen Number of samples during which the generator was running 14 Vde Bat Vde Battery voltage 15 Pde BatChg Wdc Power supplied to battery 16 Pdc Bat Dis Wdc Power provided by battery 17 Pdc PV Wdc Power provided by PV array 18 Gen Status Generator status (1 = Off) 19 Load Status Load status (1=On) 20 T21X °C Datalogger front panel temperature (enclosure internal temperature) 21 Tamb/Tmod °C Ambient temperature 1/16/96 - 7/18/96 Module temperature 7/18/96 - present 22 TBat °C Battery temperature 23 Ir POA Wim? — Plane-of-array irradiance 24 Vde 21X Bat Vdc Datalogger battery voltage Figure 2-34 shows a close-up of one week of charge and discharge currents. Apparently, the gen- erator had difficulty starting and staying started. The Park Service was supposed to have sent someone to adjust the generator, but it is not clear from any re- corded communications what the conclusion from that inspection actually was. In any event, while this is not an ideal operation, it is real operating data that would have been 10% in error if a single bi-directional current value had been recorded. In addition, interpreting the odd charging current is simplified when the current charge and discharge data are separate. The SOC algorithm used to generate the SOC and overcharge results was based on an ampere-hour counting technique with manual and automatic SOC resets (points at which the algorithm concludes that the SOC is 100% and discards any accumulated error). This algorithm was developed as an attempt to account for SOC carryover from one partial charge cycle to an- other. In a controlled charging environment, a battery is considered at 100% SOC when it is at a specified float voltage level and the current into the battery is below a threshold value. The battery discharge amp-hours may then be accumulated, and counted against the amp- hours required to restore the battery to the original volt- age and current levels. Additional charging amp-hours required to restore the battery to its original state may be quantified and used as an alternative indicator of complete charging. 2-48 2. INTEGRATION YQ, 100 +- Current (A) 8 ° y | -20 | | 40 = | | 60 = | -80 “— © e928 88 85s &§ &§ FS F& & 2 @ 2 2 & B2 B €$¢é¢e:eessgegsessss ssesesesgseses sB 6s $$ $6 6 6 © 6 6 $¢ & F&F F&F FGF FSF FS FS = 6 8&8 & &6 =F F 8 & F&F & F F & &B F F&F FF Figure 2-32. Wawona Point Daily Battery Current. 100.0% | 90.0% + - | | 80.0% 70.0% 60.0% ir 50.0% 40.0% + W.0% 20.0% + 10.0% screeds =e 10/28/95 5/15/96 12/196 6/19/97 1/5/98 712498 2/9/99 8/28/99 Figure 2-33. Wawona Point Daily Discharge Amp-hour Error. 2. INTEGRATION 2-49 70 —Charge Discharge < = 1 = 9 © E 5 O 2 : (\ ( fh 0 ! Ltt | yoy A iy eM } \ \ Pi an j 40 —/ ee 0 ee aiwe7 31297-31897 = 81497 1897316971797 1873/19/97 3/20/97 Figure 2-34. Wawona Point Sample of Rapid Charge and Discharge Cycling Current. In an uncontrolled charge as well as discharge re- gime, the battery may not return to a full charge for an extended period of time. Estimating absolute SOC during such an extended period remains a challenging problem, but at least it is clear that the SOC value at any time is a composite function of the charge and dis- charge influences it has experienced since the last time at which the SOC was known. The first iteration of this algorithm counted amp- hours and assumed that if the battery voltage exceeded a preset value (temperature-compensated) then, when the voltage dropped below that value again, the SOC was reset to zero. While this algorithm appeared to provide reasonable results, it lacked a reasonable theo- retical justification. The algorithm actually used here is based on dis- cussions with a member of the Sentech team, in which the onset of a discharge that occurs after the minimum overcharge threshold has been reached causes the SOC reset. The theoretical basis of this comes from the as- sumption that a limited number of charge sites need to be recharged before the battery is fully charged. Once this is assumed, any measured charge in excess of this overcharge value is certain to be wasted. The algorithm maintains a single SOC value that is split into a pres- entation SOC and an overcharge value for reporting to the operator. In addition to the automatic SOC resets, a semi- automated method for handling extended gaps in the data was also used. The SOC estimate accumulates with each data record, and gaps of more than some maximum interval of time (1.5 hours in this case) must be handled specially. This mechanism allows the SOC to be estimated in one of several ways after an extended period of missing data: set to a specified value, main- tained the same as before the missing data, or linearly extrapolated with an analyst-specified SOC rate-of- change. Figure 2-35 shows the resulting average and range of SOC estimates for each day in the data set. The SOC appears to drift downward during the low-battery trans- fer mode (which simulates not having a utility grid by using it only as a last recourse). It is likely that the lack of temperature compensation is causing the bulk charge cycle to terminate too early. A tripped breaker with the inverter in low battery transfer (LBX) mode on May 6, 1997, caused the inverter to draw the battery down to the low-voltage- disconnect level. When the breaker was reset and the inverter switched to float mode, the inverter recharged the battery over a 16-hour period. This recharge re- quired 962 Ah. This is a little low when compared to the manufacturer’s rating of 975 Ah. Since the cumu- lative SOC estimate had dropped to —24% prior to the recharge, it appears that the algorithm had drifted low by about 24%. 2-50 2. INTEGRATION 10 State-of-Charge (%) 4 FF ESI EEESEr ws J BW Figure 2-35. A similar event on October 28, 1998, lead to a re- charge that took a few days and included a couple of brief discharges (50 Ah) with a net charge of 720 Ah. However, this event followed 19 months of float at low (temperature-compensated) voltage, so the actual charge available in the battery at this point was proba- bly somewhat reduced. It appears that the SOC algorithm, as currently im- plemented here, suffers from a downward drift during partial cycling, and shows an upward drift during float operation. In summary, the primary goal of this task was to build a tool for evaluating storage system characteris- tics and building a standard database of performance data for the OGPS system located at Wawona Point. In this regard, significant progress was made in develop- ing the basic processing scheme and streamlining the review of the resulting data. The resulting data analysis tool contains a semi- automated process of extracting data from the source- table and storing the cumulative results into the stan- dard database. In addition, tabular and graphical review of the data in daily aggregate and controlled time inter- vals is quick and efficient, using the versatile Excel spreadsheet software as a rapid evaluation tool. EE ES ew SX $F ESS wy gh Bh Ah ah A Wawona Point Average Range of Daily SOC. The secondary goal, that of analyzing the results to obtain conclusions regarding the present and continued operation of the OGPS system had somewhat more limited success. The weakly controlled cycling char- acteristics of this system, in its off-grid mode of opera- tion make comparisons with standard controlled cycling data challenging at best. Limitations in the data quality and even a missing power-flow measurement prevented generation of a complete and accurate standard data- base. The key measure of battery health, SOC, is diffi- cult to estimate from real-world data with incomplete charge cycles, and it is no easier to estimate during float operation. Missing data is both a real and significant impediment in this endeavor as well, requiring manual intervention and probably manual replacement in order to obtain reasonable results. This database implements SOC algorithm that attempts to evaluate SOC under real-world conditions, including analyst-specified ad- justments. The database does not implement algorithms for filling in missing data or for flagging such data val- ues in order to trace them back to the original data. Given the occasional periods where the data are “slightly wrong,” source data correction will in general be required to obtain accurate results. Semi-automated procedures for finding data inconsistencies and repair- ing them will need to be implemented. 2. INTEGRATION 2-51 The battery charging voltage in this system was clearly too low during float mode, and was possibly failing to complete generator-driven recharge cycles as well, due to a lack of temperature compensation. The consequence of this poor float mode operation appears to be a reduction in capacity of as much as 25% over a 19-month period of float operation. If no change in operational settings occurs, this battery may reach end- of-life within the next two years. Changing the charge control to include temperature compensation should extend this life by several years, but it is not clear at this point how the system will behave when it is con- verted back to off-grid operation, since the SOC esti- mates computed so far for that type of operation are questionable. Given the fact that this system requires regular generator run-time even during sunny periods when in LBX mode, this system is already undersized for the radio repeater load. If the existing system were to be effectively used to replace the utility entirely, the inef- ficient DC-AC-DC power conversion process should be eliminated and the PV array increased to handle low winter insolation. Eliminating the inverter will also eliminate the non-temperature-corrected charge control regime. System analysis needs attentive system operators and good operational records. Because of the distribu- tion of responsibilities and expertise between organiza- tions and the minimal site-specific system design for this application, a number of operational quirks and system problems were accepted and ignored for ex- tended periods of time. Grasmere System Sentech, Inc. was contracted to develop a database suitable for analyzing the other off-grid RGS system in this study. The site is located near Grasmere, Idaho, about 125 miles south of Boise (Figure 2-36). Major components are a PV array, battery storage, control systems, and backup diesel generators. The Grasmere RGS installation supplies power for an Air Force training unit that provides electronic emis- sion simulations for military aircraft. A crew from nearby Mountain Home Air Force Base staffs it. The station is situated in mountainous terrain at 5800 ft and contains an office/communications build- ing, a barracks, several storage/equipment buildings, high power radars and other emitters, water storage tanks, and a battery storage building. The power needs include housekeeping loads, for example, water tank freeze-prevention equipment, radars, and _ other Figure 2-36. Location of RGS at Grasmere. electronic equipment. The peaks required for the elec- tronic equipment are many times greater than those required for the basic housekeeping loads. While Air Force personnel operate the station, the power system is provided under contract with Idaho Power Company (IPC), which is responsible for acqui- sition and maintenance of this equipment. The site is approximately 40 miles from the nearest power lines, and line extension was not a viable option. From the customer’s viewpoint, the equipment currently installed has a good performance record; power shortages have not been a problem. The Air Force appeared to be sat- isfied with the cost and reliability of the power supply and the service provided by IPC. However, IPC ex- pressed the opinion that the storage system has been “overworked” and that the replacement system should be significantly larger in storage capacity. The PV array consists of 648 Solarex MSX120 panels, each rated to deliver 7-Ah at 17.1-V, or 120-W (peak output). Eighteen panels are electrically con- nected in series to provide a 308-V string, and 36 such strings are connected in parallel. The total field is rated at 252 Ah/308 V, or slightly over 77-kW peak output. Given the nominal battery voltage of 240-V, the 308-V PV output avoids the requirement for a DC/DC con- verter. The system has four Caterpillar diesel generators, two rated at 160 kW and two rated at 60 kW. The larger units are configured for automatic operation and are capable of assuming station loads or battery charg- ing duty without operator input. Since the electronic equipment is operated only sporadically, prior to the 2-52 2. INTEGRATION addition of the PV/battery hybrid the diesels ran for long periods at a small fraction of rated loads serving only the modest housekeeping loads. This resulted in cylinder glazing and increased maintenance costs (they have already been overhauled once). There was also increased incidence of diesel fuel spills. The addition of the RGS has markedly reduced diesel-run time, fuel requirements, and maintenance. During our visit, only the larger “automatic” units were in a ready-for-service mode. The inverter for the DC output from the battery or PV array was fabricated by Advanced Energy Systems (AES) and is rated at 100 kW. The direct current (DC) is converted to 120/208-V, 3-phase electrical service, which is then distributed throughout the installation. The battery was installed in January 1995. It was expected that relatively deep discharges would occur on nearly a daily basis. At this rate of usage, some 360 deep cycles would occur each year and 5 or more years would elapse before the rated life of 2000 cycles would be reached. However, the IPC representative stated that there were recent indications of two battery cycles/day, thus potentially surpassing the 2000 cycle expected life in a shorter time period. These clues suggested that the batteries were at, or approaching, end of life. Another indicator that end of life was iminent was the appearance of widespread positive-plate swelling. Deep cycling causes morphology changes in the porous active material of the positive electrode that, over time, results in plate swelling and lengthwise plate growth. The lengthwise growth initially expands into the open space between the bottom of the electrode and the cell casing. When the case bottom is contacted, continued growth pushes the electrodes and the associated posi- tive terminal upward. Most cells at the Grasmere site showed nearly a quarter of an inch of positive post up- ward movement.” Idaho Power has been aware for some time that weaknesses were becoming apparent in some cells. Accordingly, arrangements were made to acquire re- placement cells to extend the life of the overall system. During the week before Sentech’s site visit, 24 new cells in a 12x2_ series/parallel arrangement were “patched” into the 120-cell battery, and the weakest of the 12 original cells were bypassed with jumper cables. * The Hoppecke design for the seal between the positive terminal and the case resembles a sleeve that traverses upward with the post while retaining, due to its length, firm contact with the case. In the Gras- mere cells, the design was clearly effective in that, in the storage system’s 120 cells, no evidence of electrolyte leakage or acid misting was evident despite plate swelling and seal movement. The new cells are C&D type CP. They are not identical to the originals, but with two cells in parallel, they are larger in capacity and are presumed, therefore, to be “safe” for the task. The goal of the “patch” is to extend operation of the battery into the year 2000 when monthly contributions to fund battery replacement will have been increased to their required levels to purchase a new battery. Storage System Performance Analysis Analysis of the Grasmere battery storage system follows the pattern used by Sentech in prior analyses. Operational parameters important to the lead-acid bat- tery were identified and the data examined to evaluate battery exposure or performance in relation to these parameters. For Grasmere, the parameters were ad- dressed in two groups: (1) abuse checks, and (2) critical measures. Abuse checks refer to those operational parameters that may be relatively benign in moderation but would prove harmful if allowed to persist. Elements exhibit- ing the potential for abuse in lead-acid cells include the following: High rate of discharge, Deep DOD, High rate of charge, Extended periods of self discharge, and Physical mistreatment. Each of the above factors tends not to affect battery lifetime or capacity on a single occurrence or even on occasional occurrences. However, where extreme lev- els are sustained or repeated, damage to individual cells may result. The analytical approach was to identify the worst- case example for each element and to assess the fre- quency of occurrence if the worst-case value is poten- tially harmful. The Grasmere data were examined to determine worst-case occurrences for each of the pa- rameters listed, and the results are shown in Table 2-28. The four-hour discharge rate could be repeated for each discharge without harmful effect. The 72% dis- charge is below the manufacturer’s 80% recommended limit (the next deepest discharge was 53%). The seven- hour charge rate was found to be a momentary peak and did not persist for even a single charge cycle. Daily cycling nullified opportunities for self-discharge (and potential sulfation damage). No evidence was uncov- ered to suggest that physical abuse had occurred. 2. INTEGRATION 2-53 Table 2-28. Grasmere Battery Parameters Criterion Value Discharge rate (highest) C/4 Discharge depth (deepest) 72% Charge rate (highest) Cc/7 Longest idle time <1 day Physical abuse none The battery does not appear to have been damaged by any of the operating parameters addressed. Several parameters are more critical to battery health and require evaluation on a continuing basis throughout a battery’s operating lifetime. For lead-acid batteries, these are: e SOc, Degree of overcharge, and Temperature of operation. Lead-acid systems are vulnerable to electrode sul- fation if subjected to extended periods of low states of charge. Consequently, assessment of the lead-acid BES SOC can provide vital clues regarding the probable lifetime and available capacity of an operating battery. Lead-acid systems also require overcharge if full charge is to be reached on each cycle. Failure to overcharge can result in the gradual decline in SOC and slow onset of sulfation. In addition, higher temperatures exacer- bate electrode and grid corrosion and shorten cycle life and battery operating lifetime. The analytic approach was to evaluate a battery’s condition for each of these parameters on a continuing basis. For SOC and degree of overcharge, this requires algebraic manipulation of ampere-hour values. For temperature assessment, summation of operating hours at unfavorable levels is required. Normally SOC is evaluated on a cycle-to-cycle ba- sis. Initial evaluation of Grasmere data, where each current reversal begins with a new cycle, is shown in Figure 2-37. Figure 2-37 shows more than 2000 cycles, many of which appear to be rapid shifts from discharge to charge involving small percentages of battery capac- ity and occurring within a single day. Since one of the major purposes of the SOC analysis is to infer a time of exposure to low discharge states, Figure 2-37 may be misleading. To gain a clearer understanding of actual SOC, several artificial constraints were imposed on the data before plotting a graph. Nearly all related to the defini- tion of a cycle. The most realistic result was obtained by defining a cycle as a minimum of six hours of cur- rent flow in one direction before a current reversal occurred. Under this constraint, the graphic represen- tation shown in Figure 2-38 resulted. The Figure 2-38 representation is considered closer to the actual occurrence since it results in about 340 cycles, more in agreement with the approximately 420 days of the 14-month period under evaluation. The conclusion to be drawn from this data is that, in the recent three- to five-month period, the battery has spent considerable time at a partial SOC, much of it below 80%, and significant portions below 60%. If true, this is counter to expectations of the analysis team and un- doubtedly a surprise to the BESS operators. The analysts recognize that Figure 2-38, based on a six-hour continuous current flow, may not be the most precise representation of the storage system cycling during the 14-month analysis period. Other definitions of a cycle are possible. Nevertheless, the general trends evidenced in this figure regarding SOC are undoubtedly accurate. Lead-acid cells require overcharge to achieve full SOC following each discharge. This varies with cell design and manufacturer but tends to fall between 4 and 10%. Accordingly, for each cycle shown in Figure 2-38, the algebraic sum of the charge and discharge ampere-hours was calculated, and the results are pre- sented in Figure 2-39. The data support the indications from Figure 2-37 in that adequate recharges were not occurring after nu- merous discharges. As in the SOC information, the undercharge is most pronounced during the most recent three- to six-month time period. However, it is also important to observe that throughout most of the 14- month period of analysis, required overcharges were not being provided. A major cause of lead-acid life reduction is ele- vated temperature. While temperature specifications for the Hoppecke cells installed at Grasmere were not available at the time of this report, manufacturers of similar lead-acid cells have issued guidelines indicating 30% cycle life reduction for operation over 30°C and 50% reduction over 35° C. The Grasmere data system recorded cell temperatures at two locations in the bat- tery string. The percentage of time — by month — that these sensors recorded temperatures in excess of 30°C and 35°C is shown in Figures 2-40 and 2-41 respec- tively. 2-54 2. INTEGRATION 140 120 100 ve i \ be 1 | 80 Q l 8 60 +i” | 40 20 0 SO oO he oO tT M KF Dao RnR HoH tT ANH DBD DOD Oo WwW xn Oo © K ®D o ~a ovr 6 FBO §F§F& FFEYRSR BSE LS - rr rr rr rer wen Cycle Figure 2-37. Grasmere Battery SOC Based on Raw Data. 140 12017 am ik | i TH ola W Li i , i nA he Y SOc 40 a TTI INIT iim | | | 20} I | 0 | -~- rR OONHMWO rR MODNW rr MOWnrrke ODN rT B& - MO wrTOWOoOornwriawsTOR OOAT NON KR WON MY rrr r Tr TFT NNNNNHNO NO OYDCUDLCUM Cycle 3/98 6/98 9/98 1/99 3/99 Figure 2-38. Grasmere Battery SOC: Minimum Six Hours without Reversal. 2. INTEGRATION 9-55 1007 8°89 8 8 @HseYdBAO % vee LOE 062 eL2 92 6&2 ccc S02 88l LLL vSt ZEl Oct €0l 98 69 es sé 8 Cycle Grasmere Battery Degree of Overcharge and Undercharge. Figure 2-39. 100 80 2 © So vt 2 yUusd10g ‘0 0 66-1dy 66-12 66-994 66-uer 86-99 86-AON 86-190 g6-des Month, Year | Battery Temp Cell1 HBattery g6-6ny g6-Inr g6-unr g6-Aew g6-dy 86-48 Temp Cell 10 Grasmere Battery Operating Temperatures over 30°C. Figure 2-40. 2. INTEGRATION 2-56 100 7 Percent foo} o @o eo eo © ofc @¢ e¢ 7 © - a > ce Ss aD oO a oa Ss Ej S =>cts 5 7° € © © © oo a a a a Tr Ff oe ee | 7 > °o ¢ 2 = = & 8 8 38 os o© &S© Q@ on O2aq707 us £ Month, Year Battery Temp Cell 1 Figure 2-41. The graphs indicate that temperatures are excessive during the summer months. In July and August, the 30°C level is exceeded virtually 100% of the time and the 35°C level is reached nearly 60% of the time. From an annual perspective, the total high temperature hours are in the 20% range, a factor that has undoubtedly af- fected battery life. The previous analysis can be summarized to in- clude favorable and unfavorable elements. The favorable elements include: Mid-to-low-level of discharge, and Low-to-moderate charge and discharge rates. The unfavorable elements include: e Irregular SOC maintenance, © Operation at elevated temperatures, and e Insufficient overcharge. It is Sentech’s view that the unfavorable elements outlined above outweigh the favorable. This suggests that the Grasmere battery system is likely to provide somewhat less service than that suggested by the manu- facturer for the duty cycle being experienced. As indicated earlier, the Grasmere site is approxi- mately 40 miles from the nearest electric network con- Battery Temp Cell 10 Grasmere Battery Operating Temperatures over 35°C. nection. The IPC representative estimated that the cost of grid extension would have been in the range of $3 to $5M. The actual RGS installation cost was approxi- mately $1.2M, of which $250K was for the battery and $330K for the PV system, $140K for the inverter, and $100K for the building. From IPC’s viewpoint, the PV battery system was the lower capital cost option for providing electricity to the site. PREPA System Lessons Learned The 20-megawatt, 14-megawatt-hour BESS (Figure 2-42) was installed in June 1994 at the Sabana Llana substation near San Juan, Puerto Rico, primarily to mitigate under-frequency load shedding. The owner- operator, Puerto Rico Electric Power Authority (PREPA), and the architect/engineer and supplier firms learned many valuable lessons, from planning through the first four years of operations, about the largest “commercial” battery system in operation in an electric utility application. PREPA collaborated with the ESS Program to compile a “lessons learned” report for the facility. The report, Lessons Learned from the Puerto Rico Battery Energy Storage System in English and Spanish versions (SAND99-2232), was published in September 1999. Since it first began operations in 1994, the BESS has responded to dozens of load shedding events and 2. INTEGRATION 2-57 Figure 2-42. Battery Storage Building in Puerto Rico. continuous demand for frequency regulation. Almost every key component experienced some problem during initial start-up and operations. PREPA and their ven- dors and contractors have solved many of these prob- lems. Still, other challenges remain. Despite these problems, however, the BESS performed admirably in the aftermath of the worst hurricane to hit the island this century (Hurricane Georges in October 1998). The plant was able to maintain voltage support on the only transmission line from San Juan to the northeastern region that was still operating after the hurricane. Status The ESS supports PREPA generation operations with rapid reserve and frequency control, T&D with voltage regulation, and customers with more reliable service. The lessons learned report was published in Sep- tember 1999 in both English and Spanish. The follow- ing is a summary of that report: e Spinning reserve (SR) is defined as the unused generation capacity that is synchronized to the network and can respond within ten minutes to prevent interruption of service to customers (load shedding) in the event of a failure of an operating power plant. e Rapid reserve is a portion of SR that is available almost instantaneously to prevent automatic load shedding. e Frequency control is the regulation of frequency of the electricity that utilities produce within a narrow band around 60 Hz (standard in the U.S.). ¢ Voltage regulation is the ability of a power source to maintain constant output voltage with changes in load. PREPA designed its BESS to function primarily in the rapid reserve mode, preventing load shedding in its island network. The BESS experienced its first rapid discharge on November 23, 1994. A 410-MW unit of PREPA’s South Coast Steam Plant was lost, resulting in a 21% system overload. Load shedding was neces- sary, but the magnitude of the event was reduced be- cause of the availability of the BESS. PREPA’s experience identified many pitfalls as well as optimum processes for planning, design, pro- curement, construction, operation, and maintenance of a large, integrated energy storage facility. The lessons learned document will be useful to the other utilities considering construction of similar facilities. The les- sons learned are presented by project phase (see Figure 2-43) to indicate the type of challenges that can arise during each phase of a project. This project was initiated in 1989, with internal planning documents prepared by the PREPA Planning Division with assistance from its architect/engineer, United Engineers & Constructors (UE&C). The utility evaluated alternative technologies, including batteries, flywheels, and gas turbines to combat frequency control 2-58 2. INTEGRATION PROCUREMENT I CONSTRUCTION OPERATIONS PLANNING DESIGN START-UP | | | | | | | | | | | | [ ; | | | 1989 1990 1991 1992 1993 1994 995 1996 1997 1998 Resolution for Ground = 1st Battery Plant FCS PCS Humcane 2265 to End Battenes Breaking Shipment Inauguration] Optimized Acceptance Georges Load Shedding Ist Battery Response Figure 2-43. Project Timeline and Key Milestones. issues that caused load shedding. PREPA was able to draw upon the experience of two large BESS facilities already in operation: Southern California Edison’s (SCE’s) 10-MW/4-hour Chino facility operating since 1988 and Berliner Kraft und Licht’s (BEWAG’s) 17- MW/30-minute battery plant operating since 1987. PREPA determined that a BESS would provide the quickest response to power fluctuations and achieve the fastest payback. During the planning phase, the fol- lowing lessons were learned: e Establish a team to follow the project through to completion. e Identify up front who is responsible for the system within the organization throughout all project phases. e Determine project responsibilities for all participants before going to procurement. e Coordinate meetings with licensing boards to facilitate the permitting process. e Gather and evaluate as much performance data on other facilities as possible. The PREPA staff had considerable communica- tions with UE&C, who directed the design/engineering phase. The utility wanted to make certain that the BESS was not a first-of-a-kind demonstration like the Chino and BEWAG plants. They insisted that all sub- systems consist of commercially available components. However, very few components were in fact purchased off the shelf. In addition, PREPA did not want a turn- key project in which they would have minimal in- volvement; the utility wanted to be the system integra- tor. These two conditions significantly impacted the design/engineering phase. PREPA learned that the system integrator must: e Select the architect/engineer with the most relevant experience. e¢ Bean active participant in system design. e Accept that project phases are intertwined in large competitive solicitations (e.g., design phase overlapping procurement and construction phases). e Verify available infrastructure before preparing an appropriate design (e.g., roads and the supply of electricity and water). e Identify and design for site-specific environmental and climatic conditions. e Design the building to facilitate regular mainte- nance and major overhauls. e Consider the performance and cost implications of subsystem design and configuration. PREPA coordinated the procurement processes. Having a public power authority control procurement extended a complicated process, with seven sequential solicitations, into a two-year effort. In addition, PREPA had not built a major facility for 15 years. 2. INTEGRATION 2-59 Nonetheless, PREPA staff undertook a number of initiatives that paved the way for construction: e Dedicated a purchaser in the organization to ensure good communications and project continuity. e Budgeted for cost variations (see Table 2-29). e Ensured that all potential bidders are invited. e Designed specifications — that interpretations. avoid multiple e Developed contingency plans for off-schedule equipment deliveries. During the construction phase, PREPA worked closely with the general contractor, who had ample ex- perience in the construction of reinforced concrete buildings. However, the contractor’s lack of experience with electric power installations prolonged the con- struction period even though PREPA O&M personnel were present throughout construction. The participa- tion of key team members enabled several beneficial changes to building design during construction. None- theless, more proactive participation by the battery ven- dor and architect/engineer would have facilitated con- struction and acceptance testing. e Bring all suppliers together to review system impacts of subsystem designs. e Build time into the schedule for weather-related delays and construction errors. e Ensure that general contractors use qualified subcontractors for specialized work. After inauguration of the BESS in July 1994, PREPA experienced a frustrating start-up period. The system failed frequently, and it took months for the team to pinpoint all the causes of failure (such as corro- sion on the gold-plated pins of the PCS). Likewise, domino effects further complicated the isolation of problems in component design and operation. PREPA staff were able to resolve these initial start-up problems because of their involvement in all phases of the proj- ect. The major lessons learned in this phase were to: e Verify control software for battery management (original charging algorithm did not result in full state-of-charge). e Monitor all systems for unanticipated problems. e Install appropriate ground detection equipment. e Interface the BESS with central utility dispatch operations. e Access design and construction staff as needed to resolve start-up issues. e Allow for a long start-up period for one-of-a-kind projects. Table 2-29. PREPA BESS Costs by Components ($M) Contractor Principal Components ae ae United Engineers & Constructors Design/Engineering 0.95 1.49 C&D Charter Power Batteries, Racks, Watering System 4.60 4.84 General Electric PCS 5.40 5.40 Pauwels Transformers & 115 kV Interface 0.36 0.70 ABB AC Switchgear 0.19 0.19 PACS Industries DC Switchgear 0.50 0.72 Leeds & Northrup/ Facility Control System & Monitoring 0.80 1.33 Applied Control Systems Aireko Construction & Balance of Plant 4.00 4.85 PREPA Project Mgmt, Training, Testing —_ 0.80 TOTAL ais aa 16.80 20.32 2-60 2. INTEGRATION Continuity in PREPA’s BESS team facilitated the transition from initial start-up to full operations. One team member, who had been involved with the facility since the construction phase, continued as the plant manager through 1997. Unique problems in system operations were solved during the first two years of operation. PREPA staff made many modifications that improved BESS performance (a typical BESS rapid reserve response on May 2, 1997, is shown in Figure 2-44). Areas of concern and lessons learned in this phase include: e¢ Coordinate smooth turnovers with all participants. e Maintain trend data on BESS responsiveness. e Improve operations by addressing engineering details. e Choose upgradable, nonproprietary electronics and data systems. e Implement fully debugged data tracking. e Identify root cause of excessive cell failure. 23 20 15 = = 10. o = 52 Oo 5 Time (seconds) Figure 2-44. BESS Rapid Reserve Response on May 2, 1997. Battery cell failures and replacement have been the biggest maintenance issues. Building logistics aggra- vate this situation (rack design, no elevator to the sec- ond floor, and limited heavy equipment handling capa- bility). The smallest BESS outage possible to replace a few localized cells is one-third of one floor (one string is comprised of three rows). Typically, an entire floor must be taken out of service, leaving only 10 MW available for load shedding avoidance. The time and effort required to change out cells is significant and costly. Three men can remove and replace a maximum of 16 cells in a day. The entire battery contains 6000 cells. Figures 2-45 through 2-47 show one part of the removal and replacement effort. One of the biggest problems is rack design, which provides no clearance between the rack and the cell jars to permit the use of a forklift. The BESS battery technician designed a sling that was used successfully to remove good cells. A V- shaped forklift was also designed to permit removal of good cells from the bottom rack. The primary lesson learned here was to design the system for ease of maintenance: e Acquire tools appropriate for maintenance. e Install equipment to facilitate maintenance and safeguard personnel. e Have realistic staffing expectations. e Establish appropriate warranty conditions. The BESS is contributing in a way no other gen- eration asset can. The facility has successfully achieved its goal of providing rapid SR, frequency control, and/or voltage regulation. This was particularly the case in the aftermath of Hurricane Georges in 1998. Rapid discharge from the BESS continues to provide a reduction in network load shedding and in required system-wide SR that result in economic benefits for the utility. In frequency regulation mode, the BESS pro- vides a fast response that reduces frequency deviations and provides operational flexibility during generation shortages. The BESS in voltage regulation mode helps sustain system voltage levels, especially during peak hours, by augmenting the reactive power load. PREPA is in the process of deciding how to proceed with cell replacements, and is determined to improve state-of- charge measurement and data acquisition flaws that hampered operations. The utility is now beginning the process of reviewing cell specifications for a second BESS facility located at the same substation near San Juan. Despite the best efforts of those involved in the project, the PREPA BESS was still a first-of-a-kind system with its share of unanticipated problems. The lessons learned from this project should help future large BESS projects achieve full operational status in a minimum amount of time. 2. INTEGRATION 2-61 Peru System Field Test Monitoring In July 1997, the International Lead-Zinc Organization (ILZRO), the ESA, the Solar Energy Industries Association, and the country of Peru signed a memorandum of understanding (MOU) for a collaborative project. The objective of the first phase of the project was to perform a feasibility study for photovoltaics in combination with storage, power electronics, and controls for use in remote villages in the Amazon River valley. The feasibility study included a discussion of system and component sizing, performance benefits, and economic benefits. The cost of the study was shared by the MOU signatories and the ESS Program. Results of the approximately six-month study are documented in an ILZRO report. The MOU signatories have now started the hardware demonstration phase. ESS-funded activities in FY99 were planned for providing DASs in up to four demonstrations once system hardware was purchased. Figure 2-47. Installing Good Cell. Status Several meetings were held with ILZRO to discuss the status of their hardware demonstration. Activities focused on identifying funding sources for the hardware. The potential funding sources were approached, but no funding was secured in FY99. No further action will be taken on the DAS until the funding sources for the demonstration have been finalized. PV/Battery Hybrid Controller Field Test The initial evaluation of the first PV/battery hybrid controller at the SNL PV System Evaluation Laboratory (PSEL) was completed in FY95. A project plan was developed that called for the installation of the second prototype control unit at a site that could provide at least a 15-kW PV array, a 200-kWh battery, 30-kW genset, and a variable load that would simulate the electric power consumption of a small village. A multiyear operational test plan was _ successfully negotiated with APS to perform a complete system field test of the controller at the APS STAR Center in Tempe, Arizona. An MOU and a loan agreement were 2-62 2. INTEGRATION finalized in FY97. The MOU provided for the loan of a state-of-the-art inverter/controller and a DAS for up to three years. APS acquired a tubular gel VRLA battery through a special agreement with Yuasa, Inc., to expand the test results to determine whether the gel technology battery was best suited for the hybrid operational environment. A special hybrid test facility was constructed at STAR to house the battery, inverter/controller, and natural gas genset. The STAR Center with nominal loads of 2 to 35 KW would provide the desired loads for the test program. Two multiyear contracts were placed in FY98 to provide operation and maintenance for the DAS and in-depth analytical support for data generated by the system. Results from the analysis of the operational data generated at STAR are expected to provide better understanding of how to manage the components in the off-grid hybrid system operational environment. Status The Trace 30-kW inverter remained idle through- out the first quarter as APS continued a high priority test program on another inverter. Prior to placing the 30-kW inverter in mothballs in early October 1998, the Yuasa DGX battery was fully charged and placed on open circuit. The inverter was placed back in service in January 1999. Three meetings were held during the second quar- ter. The first and second, held at SNL and at STAR, respectively, had as their primary purpose an evaluation of the status of the Yuasa cells at STAR. The third meeting, held at Yuasa’s offices in Reading PA, was needed to discuss the results of the STAR and Yuasa lab testing and to then determine future courses of ac- tion. The results of the analyses performed by EECI were used as presentation materials at these meetings. Selected graphics from these presentations will be in- cluded in the final report for the present phase of the work. The most important conclusions that resulted from the analyses and from the meetings were: e The STAR battery with the Yuasa gel cells is performing quite well in the solar hybrid system, although the depth to which the battery is discharged is somewhat deeper than ultimately desirable. e Capacity testing of the Yuasa gel VRLA battery at STAR revealed that there were two cells, one in each string, that had a capacity that was ~30% less than rated. These two weak cells determine the capacity of the entire battery because of the way this test was managed. Yuasa initiated the process of obtaining some replacement cells for the weakest ones in the STAR battery. e Since the batteries in solar hybrid systems will be cycled only to ~50% DOD, continued operation with the two weak cells should not be too problematic. e The test procedures being used for Yuasa testing at SNL appear acceptable. The data show that the capacity returned during a battery charge is slightly greater than that removed in a battery discharge, but is probably insufficient to make up for coulombic inefficiencies in battery cycling. e The apparent loss of capacity of the gel VRLA battery at SNL has been traced to the one very weak cell. It was concluded that this cell should be removed from test, and that cycling should then be resumed. The data from the battery cycling should again be reviewed after the next charge cycle has been completed. e In future Yuasa lab testing, consideration should be given to performing full equalization charges before conducting capacity tests. The STAR hybrid facility was inactive during the first half of April while the 30-kW Trace inverter un- derwent extensive upgrades. During this period, Cell No. 68 in Bank A was replaced with a spare cell that had been at open circuit for 18 months (identified as Cell 68-S, S for spare). Cell 68-S was equalized at 2.35 Vde for 80 hours from April 5 through 8, 1999, and it recovered to full capacity as indicated by the manual cell voltage readings of April 26 (Table 2-30). A significant conclusion is that the Yuasa Dynacel DGX85-11 can survive 18 months at open circuit and be restored to full capacity by standard equalization. Data show that it is desirable to equalize a battery that sits on the shelf for 18 months before it is placed in service. The hybrid test did not resume normal operation until May 6, 1999, when a capacity test was performed (see Figures 2-48 and 2-49). The results of this test showed that the battery bank could only deliver 58% of rated capacity (495-Ah actual vs. 850-Ah rated). Two weak cells that were rapidly approaching cell reversal (Pilot Cells [PC] No. 2 and No. 4) primarily caused the poor performance. The capacity test was terminated to prevent cell reversal (rather than allow the battery bank to discharge to 210 V, ie., 1.75 Vpc average), which 2. INTEGRATION 2-63 t9-T NOILVYDFLNI 7 — Table 2-30. Comparison of Yuasa Battery Manual Cell Readings: April to June 1999 6/23/99 5/26/99 4/26/99 Bank A (East) Bank B (West) Bank A (East) Bank B (West) Bank A (East) Bank B (West) Pc PC PC PC PC Pc Cell 73 1 2.209 Cell 28 2.208 | Cell 73 1 2.205 Cell 28 2.204 | Cell 73 1 2.135 Cell 28 2.126 Cell 98 2.205 Cell 45 2.207 | Cell 98 2.200 Cell 45 2.202 | Cell 23 2.128 Cell 56 3 2.126 Cell 13 2.203 Cell 75 2.206 | Cell 13 2.198 Cell 75 2.201 | Cell 98 2.128 Cell 64 2.126 Cell 23 2.203 Cell 76 2.206 | Cell 23 2.198 Cell 76 2.201 | Cell 13 2.127 Cell 75 2.126 Cell 33 2.201 Cell 59 2.204 | Cell 65 2.197 Cell 48 2.200 | Cell8 2.124 Cell 76 2.126 Cell 65 2.201 Cell 64 2.204 | Cell 33 2.196 Cell 64 2.200 | Cell 33 2.124 Cell 45 2.124 Cell 103 2.201 Cell 48 2.203 | Cell 55 2.196 Cell 56 3 2.199 | Cell 65 2.124 Cell 59 2.123 Cell 8 2.199 Cell 56 3 2.202 | Cell 56 2.196 Cell 59 2.199 | Cell 28 2.123 Cell 54 2.122 Cell 56 2.199 Cell 57 2.202 | Cell 57 2.196 Cell 57 2.197 | Cell 78 2.123 Cell 55 2.122 Cell 57 2.199 Cell7 2.201 | Cell 68 2.196 Cell 106 2.197 | Cell 25 2.122 Cell 57 2.122 Cell 78 2.199 Cell 106 2.201 | Cell 103 2.196 Cell7 2.196 | Cell 43 2.122 Cell 96 lei Cell 105 2.199 Cell 54 2.200 | Cell 105 2.195 Cell 54 2.195 | Cell 48 2.122 Cell 106 2.121 Cell 18 2.198 Cell 66 2.200 | Cell5 2.194 Cell 66 2.195 | Cell 68 2.122 Cell7 ene Cell 28 2.198 Cell 81 2.200 | Cell8 2.194 Cell 81 2.194 | Cell 103 2.122 Cell 81 2.12 Cell 46 2.198 Cell3 2.198 | Cell 28 2.194 Cell 96 2.194 | Cell 18 2.121 Cell 69 2.119 Cell 74 2.186 Cell 34 2.186 | Cell 102 2.183 Cell 100 2.183 | Cell 92 2.11 Cell 87 2.108 Cell 99 2.186 Cell 58 2.186 | Cell2 2.182 Cell 29 2.182 | Cell 104 2.11 Cell 98 2.108 Cell 114 2.186 Cell 61 2.186 | Cell 4 2.181 Cell 49 2.182 | Cell 67 2.109 Cell 109 2.108 Cell 4 2.185 Cell 17 2.185 | Cell 20 2.181 Cell 63 2.182 | Cell 84 2.109 Cell 58 2.107 Cell 37 2.185 Cell 26 2.185 | Cell 37 2.181 Cell 87 2.182 | Cell 11 2.108 Cell5 2.106 Cell 102 2.185 Cell 63 2.185 | Cell 47 2.181 Cell 26 2.181 | Cell 64 2.108 Cell 26 2.106 Cell 2 2.184 Cell 77 2.185 | Cell 81 2.181 Cell 77 2.181 | Cell 99 2.108 Cell 29 2.106 Cell 10 2.184 Cell 99 2.185 | Cell 114 2.181 Cell 17 2.180 | Cell 108 2.108 Cell 35 2.106 Cell 14 2.184 Cell 22 2.184 | Cell 10 2.18 Cell 22 2.180 | Cell 34 2.107 Cell 99 2.106 Cell 108 2.184 Cell 49 2.184 Cell 108 2.180 Cell 94 2.180 Cell 44 2.107 Cell 115 2.106 ES Note: PC = Pilot Cell NOILVYDALINI 7 SOC a Table 2-30. Comparison of Yuasa Battery Manual Cell Readings: April to June 1999 (Continued) a 6/23/99 5/26/99 4/26/99 Bank A (East) Bank B (West) Bank A (East) Bank B (West) Bank A (East) Bank B (West) PC Pc Pc PC Pc PC Cell 42 2.183 Cell 94 2.183 | Cell 14 2.179 Cell 98 2.180 | Cell 47 2.107 Cell 17 2.105 Cell 81 2.183 Cell 109 2.183 | Cell 52 2.177 Cell 35 2.179 | Cell 37 2.106 Cell 92 2.104 Cell 59 2.181 Cell 35 2.182 | Cell 11 2.176 Cell 99 2.179 | Cell 114 2.106 Cell 102 2.104 Cell 11 2.179 Cell 98 2.182 | Cell 42 2.173 Cell 109 2.179 | Cell 42 2.102 Cell 49 2.103 Cell 52 2.179 Cell 102 2.182 | Cell 59 2.170 Cell 102 2.176 | Cell 59 2 2.102 Cell 22 2.102 Cell 9 2.178 Cell 43 2.176 | Cell 9 2.168 Cell 43 2.173 | Cell 12 2.1 Cell 43 2.102 Cell 12 2.178 Cell 112 2.170 | Cell 12 2.167 Cell 103 2.164 |Cell9 2.099 Cell 103 2.098 Cell 39 2.173 Cell 103 2.166 | Cell 39 2.162 Cell 112 2.162 | Cell 39 2.097 Cell 112 2.094 Cell 49 2 2.156 Cell108 4 2.155 | Cell 49 2 2.151 Cell108 4 2.147 | Cell 49 2.097 Cell108 4 2.089 a Note: PC = Pilot Cell |Battery String Voltage aes 572-899 (Julian L Bank A Voltage ——~ Bank B Voltage 288 — 2.4 SUN MON] TUE| wep) | [THU FRI) SAT V_EC = 282V - : | gee | 276+ -- ~~ - 4 tee yee ee fe dee ee ee epee ee eee es + 2.3 V_BC=276v7- | [Capacity test begun at 08:00, but load bank problems . - mwat------4-- ‘Tdelayed test until 08:58. Test | =f Eemalicaion saned ates 11 2» terminated at 13:50 when pilot Vde y:\ pv hybrid\dat 252 4 216 + 204 + eo icell 4 reached 0.4V. ES ns eid ee om ms fin eefeare ot ws nL cal fe afin a fe fon at ol 124 | yg | a 2 oe NG py V_BEOD =234v | 7° TL V_BEOD = | inverter shutdown at 1730intokeep |] foo | ....---dag batteries fully charged. Battery | V_BDOD = 232V . disconnects opened. STAR switched | back to grid [Battery banks at Voc for cell | BATTERY VOLTAGE=ZERO. | -}. - -+18 | V_BDOD,V_BEOD = 210V PY + 17 J so co sh5°2855- 85 7 tacharts\weekly\hybMay02-8_99.zip |Total amphours discharged from [The Yuasa Battery Banks and the 30-kWh |_\Trace Inverter are now the active components iin the Hybrid Facility. to 1350 were 494.5 . 2e ze =s Time (24 hr clock) Figure 2-49. Battery String Amp-hours and % DOD vs. Time, May 2 through 8, 1999 (Julian 122-128). 2-66 2. INTEGRATION limited the capacity of the entire bank to the weakest cell. Had these two cells been allowed to experience voltage reversal, the actual capacity would have been significantly higher. PC No. 4 recovered to 1.99 Vdc after being at open circuit for 1.5 hours. Normal operations continued until May 17, at which time the 30-kW Kohler generator began experi- encing over-speed fault conditions (it could not provide the STAR Center air conditioning starting current). Operation began again on June 7, when the maximum generator-power set point for the inverter was modified from 30 kW to 25 kW and the factory-modified inverter power “signal conditioner” was reinstalled. Further difficulties with the generator created the need for more modifications to this set point during June. A summary of inverter set points is shown in Table 2-31. A pre- liminary conclusion drawn from these generator diffi- culties is that it is undersized for this application. A 90- kW generator is currently awaiting connection. There was also suspicion that the inverter was not functioning properly and may be a cause for the generator fault. By the end of the fourth quarter of FY99, after 26 months of service at the STAR Center, the consensus is that the Yuasa DGX85-11 VRLA-gelled electrolyte batteries have performed well. Capacity tests have been performed periodically to measure overall per- formance as shown in Table 2-32. As was noted ear- lier, measured battery capacity decreased during the period of September 10, 1998 to May 6, 1999, due to several weak cells. A few weeks prior to a capacity test in early August, the five weakest cells were replaced with five new cells. During the capacity test (two 120- cell parallel strings—Banks A and B), it became appar- ent that Bank A was reaching end of charge faster than Bank B due to two weak cells in Bank A. Bank A’s discharge was terminated when the two weak cells ap- proached reversal (Bank A measured capacity = 293 Ah; 69% of rated capacity). Bank B was discharged until the average cell voltage was 1.75 Vdc (measured capacity = 441 Ah; 104% of rated capacity), and no cell approached reversal. The total capacity of the two banks was 733 Ah; 86% of rated capacity. Data show that the weakest cell in the string limits the actual capacity of a string of batteries. If the weak- est cell becomes reversed biased, it also creates a safety hazard, since the reverse-biased cell can become very hot. Culling out weak cells is an effective method to maximize the capacity of a string of batteries. Data Analysis from PV/Hybrid Controller Testing Following a competitive bidding process, which began late in the first quarter of FY99, a contract for analytical support was awarded in early February to EECI of Morgan Hill, California. EECI was given the task of providing analyses of data collected at the APS STAR Center for the Hybrid Controller Field Test Program. The overall objective of the DOE/SNL project at STAR is to develop an operational strategy for solar hybrid systems, which includes energy storage that will maximize performance and minimize the life cycle cost of these systems. The secondary objectives for the data analysis that is under way at EECI are as follows: 1. Assist SNL and APS personnel in fine-tuning the Trace controller so that the electric loads are supplied reliably and efficiently. N Ensure that the Yuasa cells in the energy storage battery at STAR are being maintained, repaired, and operated in a way that will maximize life and performance. 3. Make a preliminary determination of the usefulness of gel VRLAs in the solar hybrid application. The solar hybrid test facility at STAR (Figure 2-50 shows a block diagram) consists of the following: e A 20-kW (approximately) PV array; e A 30-kW gas-fired engine-generator; e A 200-kWh battery with two strings each with 120 Yuasa DGX gel VRLA cells; and e A 30-kW PCS together with a controller to dispatch the other components as necessitated by the battery SOC, the load, and the availability of solar power. The PCS was loaned to the project by SNL. A DAS based on a Campbell Scientific CR9000 hardware and software package was also loaned to the project by SNL. The DAS allows measurement and recording of many parameters, including: e Meteorological data and temperatures in the battery and the converter room; 2. INTEGRATION 2-67 Table 2-31. STAR Hybrid Test Facility 30-kW Trace Inverter Set Point Log As of 09:15am As of 12:00pm As of 7:00am SET POINT NOTES UNITS 6/14/99 6/28/99 6/30/99 NAME Value vPCc Value vec Value vPc Goal State na Max AC Volt Vv 226 1.883 226 1.883 226 1.883 Min AC Volt Vv 185 1.542 185 1.542 185 1.542 Max AC freq Hz 62 62 62 Min AC freq Hz 58 58 58 Max AC delay s 5 5 5 Min AC delay s 5 5 5 Max Freq delay s 5 5 5 Min Freq delay s 5 5 5 Max Gnd Fault | A 10 10 10 EC = Equaliza- V_EC tion Charge Vv 282 2.350 282 2.350 282 2.350 LEC A 100 100 100 TBE apvinstta min 900 900 900 Time Interval T_Between Equ Between Equali- hr 168 168 168 zations BC=Battery V_BC Charge Vv 276 2.300 276 2.300 276 2.300 |_BC A 100 100 100 CCO=Charge cco Cut-Out A 10 10 10 BEOD=Battery V_BEOD End of Dis- Vv 234 1.950 234 1.950 234 1.950 charge BDOD=Battery V_BDOD Depth of Dis- Vv 228 1.900 228 1.900 228 1.900 charge Bat_Volt_Comp =-1Imv/A 0 oO 0 Bat_Temp_Cmp 1=.01V/degC 0 0 0 PV Start Voltage Vv 300 2.500 300 2.500 300 2.500 PV Stop Power kw 1 1 1 PV Start Delay min 1 1 1 PV Stop Delay min 6 6 6 Gen Warm Time s 60 60 60 Gen Max Power kw 25 20 15 Gen Enable 24hr Clock na 3:00 3:00 3:00 Gen Disable* 24hr Clock n/a 7:00 7:00 7:00 Pwr Track Rate Hz 1 1 1 Pwr Track Istep A 1 1 1 Gen Fail Delay s 60 60 60 Var Current A -15 -15 -15 Min Vbat Tcomp Vv 200 1.667 200 1.667 200 1.667 Max Vbat Tcomp Vv 300 2.500 300 2.500 300 2.500 BAT DI Current A 4 4 4 Date Wa Time na *This value needs to be revised during the winter months as the time for sunrise becomes later in the day. 2-68 2. INTEGRATION = Table 2-32. STAR Yuasa Battery Bank Capacity Test Comparison Parameter 9/10/98 Test 10/7/98 Test 5/6/99 Test 8/5/99 Test Amp-hour capacity 607 575 495 733 Percent of rated capacity 71.00% 68.00% 58.00% 86.00% Load power (kW) 25 25 20 25 Average battery current (Adc) 66.3 55.3 42.42 51.67 Test duration (hrs) 4.60 5.25 5.87 7.82 Min battery voltage (Vdc) 214.6 226.6 232.70 206.10 Min battery voltage (Vpc) 1.79 1.89 1.94 1.72 Min Pilot Cell #3 (V) 1.84 1.93 1.97 1.83 Min Pilot Cell #4 (V) 1.84 0.13 0.50 1.79 Figure 2-50. Loads: STAR Facility and/or Hybrid Bldg PcUu ~25kW Controls Max Pwr Tracker T2 strings T' ‘of 120 Yuasa * DGX cells * cmon flair eed Block Diagram of STAR Solar Hybrid Facility. AC and DC currents to and from the PCS; Real and reactive power flows between and among the components and the load; String voltages and currents; Voltage of each group of 12 cells. The 12-cell groups are called modules although there is no physical packaging that identifies a group as a module; Temperature of 20 of the cells in the battery, one in each of the modules; and e Voltages of four pilot cells. These data are collected, averaged over a two- minute period, and saved on a PC hard drive that is permanently attached to the solar hybrid system. The data are collated and graphed by personnel from Ari- zona State University (ASU) under a separate contract between SNL and ASU. The collated data are then sent to EECI for further analysis on a daily and/or weekly basis; more detailed data are sent when the analyses indicates that more data are required. Status The work performed by EECI, beginning on Feb- ruary 4, 1999, has met all three of the secondary objec- tives for this Data Analysis Project. For the first two secondary objectives, this progress is indicated by the results from the testing of the solar hybrid system at STAR. The results of work toward the third secondary objective will not be apparent until much more testing has been performed, and will therefore not be addressed in the this report. Progress toward the first secondary objective, fine- tuning the Trace controller for reliable and efficient operation, is best exemplified with the results shown in Figure 2-51. Figure 2-51 also shows the power flowing among the various components of the solar hybrid system for two days in August 1999. The power flowing to the load, which at the time was the STAR facility itself, is not shown in Figure 2-51, for the sake of clarity. The power to the load is the same as the inverter power when the generator is not running, and is the difference between the generator power and the inverter power when the generator is running. During a summer day, the average power to run the STAR facility is approxi- mately 16 kW, with somewhat lower levels at night. 2. . INTEGRATION 2-69 Generator Power (AC and DC kW) Inverter 300 600 900 1200 1500 1800 g ES 0 300 600 901 1200 1500 1800 2100 Tim e (24 Hour Clock) Figure 2-51. Power Flows for Components in STAR Hybrid Tests. The wide variation in each of the power levels, except for what was generated by the PV arrays, is due to large loads (primarily air conditioning units) switching on and off. The loads during August are among the high- est of the entire year for the Phoenix area where the STAR facility is located. Figure 2-51 does not show the full variation of the loads that occur, since the data are averaged over two- minute periods. Peak loads of more that 35 kW for a few 60-Hz cycles, and of 25 kW for up to a minute, must frequently be supplied by the solar hybrid system. The generator cannot respond fast enough to these step changes in load, so the battery and inverter are used to provide some of the peak power demands during peri- ods of bulk charging. This enables reliable operation of the system with a generator that is rated at less than the peak load requirements. The smooth integration of the power from several sources is apparent in Figure 2-51, even at times when the PV arrays are delivering variable amounts of power. This is obvious from examination of the data for August 16, when intermittent cloud cover was quite clearly impacting power from the PV arrays. On sunny days, for example August 17, the battery actually charges for short periods of time, even on the very hot days with high air conditioning loads that are experi- enced at STAR in August. Much of the work during FY99 dealt with proper charging and maintenance of the Yuasa cells, was also focussed on the inverter and controller. In the middle of the reporting period, some minor modifications to the equalization charge scheme programmed into the controller were first specified and then implemented. Although the firmware changes required for this were checked out in the factory and at STAR, no long-term testing was possible after their implementation, because of the development work that was proceeding on the rest of the system. Several capacity tests were run during FY99, and as a result of these, about 12 of the 240 cells in the battery were replaced. The capacity of the battery in a test at the end of FY99 was found to be 104% of Yuasa’s rated amp-hours. All of the data being acquired in the testing of the STAR system will ultimately be useful in determining the advantages and disadvantages of using gel VRLAs in solar hybrid systems. In particular, the longer life that is expected from this type of cell when used in an ISOC mode, as compared to other lead-acid battery types, will be the subject of longer term testing at STAR. As yet, however, the focus of the work at STAR has been to attain reliable and unattended opera- tion of such systems, so long-term ISOC cycle testing has not been possible. Long-term ISOC cycle testing remains a primary goal of the work at STAR, and while progress has been 2-70 2. INTEGRATION made toward this goal, it has not yet been possible to achieve it. The main issue that still needs resolution is the long-term reliability of the components and the controls of the system. Work for the coming fiscal year will therefore be focussed on finding ways to enable long-term reliable operation of the solar hybrid system and then using the system for ISOC testing of the Yuasa gel VRLA battery. Integration and Testing of Energy Storage with Flexible AC Transmission System Devices Flexible AC transmission system (FACTS) devices offer increased flexibility in decentralized control of transmission systems. As the vertically integrated util- ity structure is phased out, centralized control of the bulk power system will no longer be possible. Trans- mission providers will be forced to seek a means to get local control to address a number of potential problems such as the following: 1. Uneven power flow through the system (loop flows), 2. Transient and dynamic stability, 3. Subsynchronous oscillations, and 4. Dynamic overvoltages and undervoltages. Several FACTS topologies have been proposed to mitigate these potential problems, but transmission service providers have been reluctant to install them, usually because of cost. The integration of an ESS into FACTS devices, however, may lead to a more economically feasible and flexible transmission controller, which would have greater appeal to transmission service providers. The University of Missouri-Rolla (UMR) team has been designing topologies necessary to integrate a BESS into a static synchronous compensator (Stat- Com). The StatCom is a well known shunt-connected, commercially available FACTS device. UMR has de- signed and built a StatCom, which is currently being used to evaluate new design and control strategies. The first step in this project will be to develop interface- power-electronic topologies to incorporate energy stor- age into the StatCom. This includes potentially elimi- nating the DC capacitor bank typically used for short- term energy storage. Eliminating the capacitor bank reduces the cost and increases the reliability. After suitable topologies have been identified, both global and local control strategies will be developed to accomplish multi-purpose trans- mission-level objectives requiring both active and reac- tive power control. The project will culminate with experimental verification of the identified designs. Although the UMR StatCom is rated at 150 kVA, all topologies will be designed so that they are scaleable to transmission-level applications (that is, ratings in the 1- to 10-MVA range). A suitable battery will be supplied by SNL to UMR. Status Several significant advances have been made in the last quarter. The most important progress has been the incorporation of the BESS into the StatCom set-up. This step was delayed due to problems with infant mortality of several of the battery modules. This has been corrected and the BESS is now fully functional. The StatCom/BESS system has been successfully oper- ated in synchronism with an external system AC source. Figure 2-52 shows the main components of the integrated StatCom/BESS system layout in the UMR laboratory. ‘StatCom Controier and ‘Signal intertace Figure 2-52. UMR Statcom/BESS System Layout. 2. INTEGRATION 2-71 The battery set shown in Figure 2-53 was con- structed by Omnion and consists of 17 Trojan VRLA super-gel batteries in two strings supplying 204-V DC to the StatCom. The DAS shown in Figure 2-53 was constructed to monitor the battery voltage and string currents. Two current sensors are used to measure the string currents. A signal interface board converts the current signals into voltage signals and filters the high- frequency noise. A thermometer with two thermal sen- sors is used to measure internal battery rack and ambi- ent temperatures. Figure 2-54 shows the front panel of the StatCom. The top rack houses the StatCom sensor rack that monitors the StatCom voltages and currents, including the three phase voltages and currents, the DC voltage and the DC current. The sensor rack is shown in detail in Figure 2-55. The middle two racks house the con- verters. Only the upper converter rack is being used in this project. This rack, shown in Figure 2-56, contains six insulated gate bipolar transistors (IGBTs), the DC link capacitors, heat sinks, and control interface. The bottom rack houses the three-winding transformer (not shown). Along with the above components, a bank of three- phase 150uF capacitors is placed between the StatCom and the monitoring and control system to improve the integrity of the waveform on the system side. The monitoring and control system consist of two M5000 boards: one for data acquisition and pre-processing and the other for sine-triangle pulse width modulation (SPWM) signal generation. The signal interface board is designed to provide digital and analog signal isola- tion for the PC as well as analog filters for the analog signals. It generates the synchronous signals for the SPWM generator and the data acquisition board. The most important result has been the experi- mental verification of the proposed control methods with the StatCom/BESS connected to the external AC system. This required one modification to the local Figure 2-53. Battery Rack (left) and DAS (right) at UMR. Figure 2-54. StatCom Front Panel at UMR. Figure 2-55. UMR StatCom Sensor Rack. Figure 2-56. UMR StatCom Converter Rack. 2-72 2. INTEGRATION control method and the development of a global control scheme. The control of the StatCom/BESS may be divided into local and global control. The global con- trol algorithm is responsible for maintaining system performance according to some pre-set or user defined scheme. The global controller’s function is to provide the local controller with the desired commands to achieve the system objectives. This aspect of global control will be discussed in greater detail in the fol- lowing section. The local control objective is to interface with the global controller and to provide the switching com- mands to the inverter. A modified SPWM was devel- oped to switch the IGBTs to synchronize the Stat- Com/BESS with the external system, and control the output voltage phase angle and magnitude of the Stat- Com/BESS. Typically, a true SPWM would be ade- quate for control. However, special circumstances re- quired a modified SPWM in this project. In the experimental set-up described previously, the battery voltage is 204 V. In order to connect directly to the 230-V AC system, it was necessary to use the StatCom to boost the output voltage to this level. This was ac- complished by boosting the range of the modulation index k. The modulation index k is in the linear region between approximately 0.5 and 1.0. As k increases beyond 1.0, the maximum value of the reference signal is greater than the magnitude of the carrier signal (the triangle waveform) and the voltage magnitude output begins to level off and the voltage output cannot be further increased. The range of the modulation index k was boosted by roughly 15.5% by using a new refer- ence signal other than a sinusoid waveform. The new reference waveform is a sinusoidal signal augmented by a scaled 3rd harmonic waveform. By using the combi- nation fundamental/third harmonic waveform, the modulation index of the fundamental component was increased beyond 1.0, while the maximum value of the reference signal remained below the maximum value of the carrier signal. The third harmonic signal is elimi- nated when it passes through the interconnecting trans- former so that the output is not corrupted. Figure 2-57 shows a typical sine-triangle waveform, and Figure 2-58 shows the modified SPWM modulation signals with the third harmonic waveform imposed upon the fundamental sinusoidal waveform. The validity of the modified SPWM was verified experimentally. Figure 2-59 shows the increased linear operation range using the modified SPWM. Note that the linear region for k is increased from 1.0 to approxi- mately 1.15, and the output voltage is increased to 135 V (line-to-neutral) or roughly 235 V (line to line), which is sufficient to connect into the laboratory AC system. 1s —_ reference) Garner 1 = Ko \ bsp e/4 NY 4 \ | / \ | = of \ | =" \ / | 2 | \ / z | \, a | \ / os} \: if | \ / | \ 7 | th “WL } | 8 wo 100 180~S~C~OSCHSCCOSC SC phase enge (degrees) Figure 2-57. Typical Sine-triangle PWM. PWM signals 18 Figure 2-58. 13, 100 150 200 250 xo 350 400 phase angie (degrees) Modified SPWM. 12 w ‘StatCon/BESS output voltage in per unit 1st harmonic modulation 1st+3rd harmonics modulation 04 06 Figure 2-59. 08 1 12 14 16 18 2 modulation index k Relationship between Modulation index k and StatCom/BESS Output Voltage. 2. INTEGRATION 2-73 Figure 2-60 shows the resultant experimental waveforms from the commanded active and reactive root mean square (RMS) power changes, much as would be commanded from the global controller. In Figure 2-61, the commanded active power is changed from a steady state-value of 0 W to a discharge of 500 W. Similarly the reactive power is instantaneously changed from 0 to 500 Var discharge. The Stat- Com/BESS responds rapidly to changes in commanded values. The effect of the proportional integral (PI) controller is evident in the slight overshoot in the initial response, but the signal decays rapidly. The Stat- Com/BESS is then commanded to rapidly charge by the same amount of active and reactive power and then return to zero. The experimental measured responses show good agreement with the predicted dynamic re- sponses from previous simulation results already re- ported. 00, VARA", Pintransient-state (rand | | | ° ci 02 O3 Oo O05 O06 o7 O08 O® 1 times) Figure 2-60. Measured Results for Active Power The requirements for a StatCom/BESS global con- troller are designed to stabilize the power system. Pos- sible applications of the StatCom/BESS include voltage and steady-state power flow control, frequency regula- tion, oscillation damping, and transient stability im- provement. These requirements may change based on the size and placement of the StatCom/BESS within the power system. In this project, two applications of the StatCom/BESS were explored: transient stability im- provement and oscillation damping. The global con- troller was developed based on a _ single-machine- infinite-bus representation of the power system with the StatCom/BESS at the mid-point between the generator and the infinite bus as shown in Figure 2-62. @ Pl Va P2 Nw AW x . P satCom vi Q satcom R i 4%, | One-Machine-Infinite-Bus Power System with StatCom/BESS. Figure 2-62. A power system case is simulated by PSCAD/ EMTDC software to verify the effectiveness of the pro- posed global control strategy. In this example, a three- phase to ground fault occurs at the StatCom/BESS con- nected bus. The fault is cleared after 0.1 seconds. The global control objectives are to damp generator rotor angle oscillation, and maintain voltage magnitude and active power P2. The global control results are shown Control. in Figures 2-63 through 2-66. After the fault is cleared, (aay «00}- | | | | | | i i oh MA V VAY : c \ | 3 200 a) \ | yoyenal 6005 Ot 02 o3 O4 Os 06 O7 08 09 1 time(s) Figure 2-61. Measured Results for Reactive Power Figure 2-63. Generator Rotor Angle. Control. 2-74 2. INTEGRATION z uo8 g 8 8 StetCom/BESS Bus Voltage (V) § 3 time (seconds) Figure 2-64. StatCom/BESS Terminal Bus Voltage. $6. ___—__—__.—__.—__._——— 10 z , a § \ 2 s g + 2 -10F 15 0 ‘ime seconds) Figure 2-65. Active Power in Line 2. 05 1 15 2 25 Q 35 4 45 5 een Without StatCom/BESS —— With StatCom/BESS Reactive Power Transmission (kVar) 10) | 12 ees 0 05 1 15 2 25 3 35 4 45 5 time (seconds) Figure 2-66. Reactive Power in Line 2. the StatCom/BESS is able to effectively damp the os- cillations in both voltage and angle and hold P, Q, and V at the desired values, resulting in enhanced transient stability of the power system. This project firmly established the viability of us- ing a StatCom/BESS to enhance bulk power system operation. Several local and global controls were pro- posed that have been shown, via simulation and ex- perimental verification, to be effective in transient sta- bility improvement and oscillation damping. To this point, all local and global proposed control techniques have been based on decoupled PI-control. These controllers should be extended to include more sophisticated control methodologies such as state feed- back or nonlinear control to enhance performance. While the StatCom/BESS has been shown to be ef- fective in oscillation damping, it is desirable to develop a controller that will have a greater damping effect during the “first-swing” of the post-fault oscillation. The dynamic performance of the StatCom/BESS must also be compared to the dynamic performance of other FACTS devices including a stand-alone StatCom and the unified power flow controller (UPFC) to fully as- sess the benefits of this topology. Also, the sizing and placement of the StatCom/BESS must be explored be- fore this technology can be utilized effectively by in- dustry. RAPS Testing Methods Development The ESS Program continues to coordinate with staff from ILZRO on several collaborative projects. Each project is being co-funded by ESS and by ILZRO at varying levels of cost sharing ranging from 50-50 to 80-20, with ILZRO taking the lead in some projects and ESS in others. A project to define an international standard test-cycle regimes for RAPS has begun sched- uling goals for drafting test procedures and reviewing the procedures. Status During FY99, the RAPS testing methods develop- ment project evolved significantly and began to cement relationships with standards organizations. The goal is to prepare documents that will lead to RAPS design and certification guidelines. With major involvement of the project’s co-sponsor, the ILZRO, and Energetics, con- tractor to both SNL and ILZRO on this project, sub- stantial planning was done, documents were prepared, and standards organization interfaces were established. A major decision was made to integrate this project 2. INTEGRATION 2-75 with the existing infrastructure of the Institute of Elec- trical and Electronics Engineers (IEEE) Standards Co- ordinating Committee (SCC) 21, Fuel Cells, Photovol- taics, Dispersed Generation, and Energy Storage. This committee has several standards documents already in preparation or approved that will contribute to RAPS design, testing, and certification guidelines. The docu- ments in process all relate to PV energy systems with battery storage used in a stand-alone mode (no utility connection). SNL, ILZRO, and Energetics will con- tribute to the PV/battery documents, and will prepare new guidelines for RAPS systems-those that use a re- newable generation resource, energy storage, and a fos- sil-fueled generator in a stand-alone mode. Additionally, contact was made with IEEE SCC 29, for Stationary Batteries. While this group does not prepare standards for RAPS systems, its heavy in- volvement with battery technology requires coordina- tion on terminology and other battery issues. In order to build the necessary relationships with the IEEE SCC 21, two meetings of the Energy Storage Subsystems Working Group were attended during the year. The first was in June 1999, and the second in late September 1999. The project is providing a secretary to take notes and publish minutes of each meeting. Based on the results of these meetings, a plan was prepared for integrating the RAPS test project with ex- isting SCC 21 standards, and for preparing new stan- dards within the auspices of SCC 21. The plan divides the systems by application: PV/battery stand-alone systems and hybrid (PV-battery-diesel generator) RAPS systems. The following list describes the existing stan- dards and proposes several new ones to complete the needed documents. PV-Battery Stand-alone System Standards 1. Std. 937-1987: IEEE Recommended Practice for Installation and Maintenance of Lead-Acid Batter- ies for Photovoltaic Systems. 2. P1013/D6: Draft Recommended Practice for Sizing Lead-Acid Batteries for Photovoltaic Systems. 3. P1144: Recommended Practice for Sizing of In- dustrial Nickel-Cadmium Batteries for Photovol- taic Systems. 4. Std. 1145-1990: IEEE Recommended Practice for Installation and Maintenance of Nickel-Cadmium Batteries for Phototvoltaic Systems. 5. P1361: Guide for Selection, Test and Evaluation of Lead-Acid Batteries Used in Stand-alone Photo- voltaic Systems. 6. P1526: Procedures for Determining the Perform- ance of Stand-alone Photovoltaic Systems. 7. New Project Authorization Request (PAR): Guide for Array and Battery Sizing in Stand-alone PV Systems. Hybrid Stand-alone System Standards 1. New PAR: Guide for Sizing Hybrid Stand-alone Energy Systems. 2. Concept for PAR: Guide for Test and Evaluation of Hybrid Stand-alone Energy Systems. 3. Concept for PAR: Recommended Practice for In- stallation and Maintenance of Hybrid Stand-alone Energy Systems. During the SCC 21 Energy Storage Subsystems meetings, the working group discussed and reviewed these existing documents and proposed the two new PAR documents. The existing document 1013 does not assist in sizing the PV part of the system. Neither does 1013 address battery charging with PV. A new ap- proach is needed for sizing batteries in stand-alone PV and RAPS systems. A design document that also con- siders cost trade-offs is needed. At the suggestion of personnel from the RAPS project, two new IEEE standards will be proposed: one for sizing stand-alone remote PV systems, and one for hybrid stand-alone systems. The idea will be to create one document that guides the design of systems with a renewable source (e.g., PV, wind, etc.), an ESS (e.g., batteries, flywheels, high energy density capacitors, etc), a power conversion subsystem (e.g., inverter, con- verter), and controllers (e.g., charge controllers). The other document would also consider a fossil-fueled generator such as a diesel, propane, or natural gas gen- erator. Both documents will be prepared in a modular manner with the first versions only including PV and lead-acid batteries; in the case of the RAPS document, diesel generators will be included. Other technologies will be added later. The group recommended that the documents focus on specific sizes and applications and then add other size ranges and applications later. For stand-alone PV systems, the size should be no larger than 2 to 4 kW. A study is in progress as part of this project to better define the application requirements and size ranges for RAPS systems. The results of this study will be used to focus the RAPS design document. 2-76 2. INTEGRATION In addition, a draft document prepared by ILZRO, “Elements of a Recommended Design Practice for Community Scale RAPS with Only PV Generation,” will be used to contribute to both the PV/battery docu- ment and to the RAPS document. The ILZRO docu- ment contains significant engineering details on sizing components and design trade-offs necessary for com- plete system guidelines. These elements are lacking in existing draft documents and will be a valuable contri- bution. In addition, the ILZRO RAPS testing commit- tee formed two years ago has reaffirmed their commit- ment to reviewing and contributing to the RAPS standard. This interface will be critically important for properly developing the design guidelines for hybrid systems. Applications for the new IEEE standards will be made in early FY2000. Once approved, work on the first draft of each standard will begin. SNL, ILZRO, and Energetics will team to develop the RAPS design guidelines during FY2000. 2. INTEGRATION 2-77 This page intentionally left blank. 2-78 2. INTEGRATION 3. Components Introduction Work in the Components element focuses on im- proving the subsystems that make up energy storage systems: improved devices are developed and evalu- ated, i.e., storage (for example, battery), electrical equipment (power conversion), and controls. The ESS Program is developing and evaluating storage and other components that cost less, have higher performance, and are better integrated with other parts of the system than those now available. As part of its technical mission in support of the ESS Program, SNL performs in-house battery evalua- tion tasks. These tasks use specialized and unique fa- cilities and capabilities established during many years of program activities in battery and other storage tech- nologies. Controlled laboratory tests provide the best method of determining capacity degradation rates for discovering mechanisms to understand the relevant cause-and-effect relationships. These independent, objective tests use computer-controlled testers capable of simulating application-specific test regimes, and they provide critical data for the assessment of the status and probable success of these technologies. Development Advanced Energy Storage Development In FY97, the ESS Program started to analyze ad- vanced storage technologies (SMES, flywheel, double- layer capacitors [DLCs], etc.) and determine the true state of the art of each technology for both capability and cost. R&D is contingent upon the conclusions of these studies. The overall approach to R&D for all ad- vanced technologies will be to (1) identify the industry drivers for each technology, and (2) estimate the com- patibility and value of these storage technologies for each of the 13 storage applications (identified in the Opportunities Analysis performed earlier)" as well as for the combined applications. Based on the results of these R&D studies, plans were made for one or more contracts to be competi- *P.C. Butler, Energy Storage for Utility Applications: Phase L-Opportunities Analysis, SAND94-2605, October 1994. tively placed with suppliers of the technologies that show the most potential. Some of these contracts will be to supply SNL with hardware for testing. Other contracts may fund small demonstrations with potential users in applications that show the greatest potential for benefits. The goal of the advanced energy storage (AES) component development project is to support the im- provement of AES components. Commercial battery technologies are not included in the scope of this proj- ect. Status In FY99, the ESS Program issued an RFP and funded three companies (Saft, U.S. Flywheel, and Boeing) to conduct the first phase of a possible multi- phase research project to develop and test components that are large enough to be used in field demonstrations. In Phase I, contractors identified target applications and characterized the applications in terms of benefits and requirements. The contractors then developed storage component specifications and conceptual designs for the most beneficial applications based on state-of-the- art technology. Both prototype and production-scale costs for the technology were then estimated. Finally, contractors created a component development plan and schedule to develop the technology to the prototype field-testing stage. Work on Phase I was completed at the end of FY99. Saft presented results on an advanced battery technology and U.S. Flywheel and Boeing each pre- sented results on flywheel technology. The final re- ports documenting this work will be published at a later date. Below is a brief summary of each contractor’s Phase I work. Saft Lithium-ion (Li-ion) battery technology has re- ceived commercial recognition by the portable battery industry; however, larger cells and batteries are still in the development stage. During Phase I of this project, Saft demonstrated the feasibility of Li-ion technology for ESS applications, identified a suitable target appli- cation for the technology, and then developed a test plan and schedule for demonstrating a prototype sys- tem, as required by the RFP. 3. COMPONENTS Feasibility of Li-ion Technology for Energy Storage System Applications The main advantages of a Li-ion system over con- ventional batteries are their high-energy density and good cycle life. Table 3-1 shows the energy density and specific energy of Li-ion batteries compared to other types of conventional and advanced batteries. Li- ion batteries also provide good deep-discharge cycling capability. This technology’s shallow cycling perform- ance is shown in Figure 3-1. It shows results of 40°C cycling smaller (700-mAh) Saft cells at 10% DOD at a 2C rate. The cells are given a 100% diagnostic dis- charge every 10,000 cycles. After an initial increase in impedance, the cell characteristics stabilized, and there was almost no capacity fading. Finally, compared to conventional battery systems, Li-ion systems are rela- tively maintenance-free. The Li-ion cells are hermeti- cally sealed and are non-aqueous. Therefore, there is no electrolysis, no gassing, and no losses from the sys- tem. The cells require electronic controls to prevent overcharging. Because these controls monitor the volt- age and temperature of the cells continuously, the sur- veillance is automated. Consequently, while some monitoring of the system is required, “maintenance” is not. The two major concerns with the Li-ion technology are cost and safety. For the most part, none of the ma- terials used in these batteries has an intrinsically high Table 3-1. Energy Density of Li-ion Batteries Compared to Other Batteries Energy Specific Density Energy (Whiliter) (Wh/kg) Valve-regulated lead-acid 70-86 33 Nickel-cadmium 85 42 Nickel-metal hydride 135 65 Lithium-ion 230 130 cost. However, many of the materials are unique to this technology and have no other industrial application. The result is that the cost of these batteries is quite sen- sitive to volume. While portable Li-ion batteries re- main expensive compared to other technologies, the production cost has fallen about 30% in the last two years. This is expected to continue, particularly as pro- duction volume increases for larger cells. Eventually, cost studies predict that Li-ion batteries will have an equivalent life cycle cost to lead-acid. In the past, the main safety concern with recharge- able lithium batteries was associated with metallic lith- ium negative electrodes. Charge/discharge cycling 1 0,1 0,9 + 0,09 = 0,8 t 0,08 - 07 0,07 = | 2 9 0,6 0,06 § 2 05 005 8 oO 3 0,4 eee acacia ne s 2 03 003 £ 3 —e— 100% DOD Capacity mS 02 oo2 5 8 01 amEeaaee 0,01 0 a - + 0 0 50000 100000 150000 200000 10% DOD Cycle Number Figure 3-1. Capacity Test (100% DOD) Results for Small Saft Cells Cycling at 10% DOD at 40°C. 3-2 3. COMPONENTS caused dendrite formation, and in some cases, highly reactive powdered lithium, which in some cases led to fires. This problem has been overcome in the lithium- ion technology by the adoption of carbon-based nega- tive electrode materials. Overcharging presents another safety concern. If these batteries are overcharged at high rates (greater than C/5) and the temperature exceeds about 150°C, a fire will result. The organic electrolyte, negative active material, and separators are combustible. This potential problem can be avoided by charge-control electronics. These controls are also required to prevent deterioration in the active materials as a result of overcharging. Saft employs at least two levels of controls: e The first level of control operates on each cell or grouping of parallel cells. It either switches the cell out of circuit or shunts excess charging current once the cell reaches the maximum charge voltage. This ensures that all cells reach the same state of charge without overcharging. e The second level limits the overall charge current to a maximum level (C/5; in the case of Saft’s design). Although this level of charging will damage the active materials, it will not allow the temperature to rise to a dangerous level. Other types and levels of control circuitry are pos- sible. The main point is that redundant levels of control are used. Other protective devices include fusing to prevent damage due to short circuits, and a safety valve to prevent cell case bursting. Apart from the end-of- charge control deployed at the cell level (which is more an integral part of normal operation than a safety fea- ture), it should be noted that all of these safety devices prevent damage caused by external means, such as a charger failure. Target Applications Based on an evaluation of the data presented in the Opportunities Analysis report,” Saft identified the fol- lowing target applications for the Li-ion technology: © Power quality, e¢ Customer demand peak shaving, and e Very short-duration peak shaving. There is a definite need to address power quality issues where an energy storage system could provide * PC. Butler, Energy Storage for Utility Applications: Phase 1- Opportunities Analysis, SAND94-2605, October 1994. ride-through during voltage sags of less than one sec- ond. Previous utility experience has been that it is dif- ficult to sell storage for this application, particularly where users have not quantified a direct cost for their power quality problems. Nevertheless, with the in- creasing sensitivity to power fluctuations of new sys- tems and operations, this application will increase in importance. Demand peaks have a demonstrable effect on a customer’s utility bill. On their own, such charges are not of sufficient magnitude to provide a reasonable payback on the system cost. However, demand peak shaving can provide an attractive additional benefit in a system installed primarily for power quality. The last possible application would be to mitigate the effects of very short-duration demand peaks, such as motor inrush currents. Where such load spikes have the potential for causing power problems in the vicinity, storage can be a mitigating factor. This application contains elements of both peak shaving and distribution facility deferral. Its importance in a demonstration system would depend on the load characteristics of the chosen customer. An energy storage system that combines the capa- bility to ride through voltage sags with demand peak shaving can provide real and immediate savings for the consumer. An additional capability to handle load spikes would be beneficial. The capability of Saft’s proposed system to handle various combinations of these loads is discussed below. General Specifications To design and develop the demonstration project, Saft identified two potential industry partners. One partner would supply the system equipment (power electronics, etc.), and the other would act as the system integrator. At the conclusion of Phase I, a candidate demonstration site had not yet been identified. Without a specific customer/demonstration site, the final specifications of the proposed system are subject to change. General specifications are that the proposed system will have an output of 400 to 500 kW, and a battery storage capacity of 67 kWh. The system will operate in the on-line mode, so the output will be sized for the entire facility load (or that portion of the load that is sensitive to power problems). Output is at 480 Vac and the system will be located on the cus- tomer's premises. Saft has two specific designs in mind to meet these specifications should further work on this project be funded. 3. COMPONENTS 3-3 The development of these designs would occur in two stages. The first would comprise battery develop- ment and testing at Saft. The second stage would in- volve system integration and on-site testing at the cus- tomer's location. U.S. Flywheel Systems, Inc. U.S. Flywheel Systems (USFS) proposed a fly- wheel energy storage (FES) system for this contract. Their design, however, is fundamentally different from the one described by Boeing. Boeing proposed a sys- tem based on a single, large flywheel; the USFS sys- tem, which is described in more detail below, is based on combinations of many smaller flywheels. Target Applications Flywheels have certain operating and performance characteristics that are important when identifying tar- get applications. Flywheel systems are compatible with the full range of expected environmental conditions, and shipping, orientation, or reasonable shock and vi- bration do not generally affect their operation. USFS has developed the concept of fiber rotor benign failure, which is designed into the rotor fabrication to produce a harmless failure. Consequently, one of the main safety concerns for FES, explosion containment, is not ex- pected to be an issue, even with mass usage. Another interesting characteristic of FES is the ability to sepa- rate specific power and specific energy as design enti- ties. This ability is significant when designing a storage system for a specific application. The above characteristics lend themselves to a va- riety of applications including space (satellites), hybrid electric vehicles (HEVs), utility energy storage and UPSs, and marine and oceanic applications. Require- ments for energy storage for these applications are summarized below. e Space requirements: highest reliability, highest efficiency, highest cycle life, highest specific energy and specific power, and requires magnetic bearings with costs being less important. e Ground hybrid vehicular requirements: high human safety, very high specific power, very high cycle life, shock and vibration resistant, insensitive to temperature extremes, low cost, possibly magnetic bearings, high production. e Ground utility/UPS requirements: high human safety, highest production, designs considering ranges of very high specific power and/or specific energy, must be lowest maintenance, must be resistant to the broadest range of environmental extremes, must be fully automatic to grid support, must be price competitive, and must be easily shippable. Large production volumes may justify special magnetic bearings. ¢ Marine and oceanic requirements: very high human safety, insensitive to temperature extremes, lowest maintenance, very high specific power and/or specific energy, and may justify magnetic bearings. Basic Design Considerations for Stationary Energy Storage It was described above that there is a wide range of applications for FES systems. However, the objective of this study was to determine FES compatibility with stationary energy storage applications. A drawing of a typical USFS module is shown in Figure 3-2; the major components of the module are outlined briefly below. Rotors: These are solid isotropic materials made from steel and aluminum. Also, fiber-composite rotors (some multi-rim) comprised of glass and many different graphite materials. A great percentage of these were dissected for destructive testing; others were spin-tested to destruction. Still others were integrated into func- tional operational modules. Hubs: Isotropic materials on which the rotors were mounted. Numerous designs were employed (some were multi-piece) and were connected to, or were a part of, the flywheel shaft assembly. Shaft: This assembly protrudes through both ends of the rotor/hub assembly. Each end accepts the bear- ings (magnetic or mechanical) and the lower end ac- cepts the motor/generator rotor. Touch down Bearings: Alternate mechanical supports for the magnetic bearings. They are used during shutdown, storage, or as a safety feature for magnetic bearing (MB) malfunction. Thus far all MB design flywheels are vertical axis machines. Some of the mechanical bearing machines are horizontal axis configurations. Motor/Generator: This unit has a stationary sta- tor with the rotor turning and with the flywheel rotating assembly. The motor/generator (MG) is the unit that determines the specific power of the whole system and is in effect independent of the rest of the assembly. 3. COMPONENTS Improved Touchdown —, Bearing Spider Assembly Cooling Line at Upper >. Touchdown Bearing Se Combination Magnetic Bearing Coolant Jacket Rotor Rim —~~ Cooling Line at Lower —— Touchdown Bearing Lower Housing Cover ——~~ _-7~ Internal Sensor 7 Conditioning Electronics Board _--— Upper Housing Cover _-~ Typical Electrical “ Connector Interface to Module _-— Rotor Hub _--~ Modified Outer Housing ” Structure ~— Radial Magnetic Bearing Coolant Jacket ~.___ Improved Touchdown Bearing Spider Assembly Cutaway View of Latest Developmental Hardware Figure 3-2. Typical USFS Flywheel-Cutaway View. Magnetic Bearing An efficient FES system requires exceedingly high rotational speeds. Studies show the higher the speed, the higher the overall cost (up to a point) and system efficiency. It is also shown that mechanical bearing cost can quickly reach a practical limit due to the bear- ing manufacturing tolerance limits with load and life considerations. USFS MBs have the load-centering power to shift rotor rotation about either its geometric or mass center axes, or anywhere in-between. Me- chanical bearings, on the other hand, have no choice in shifting the rotational axis to a more favorable position, nor do they have holding accuracy by an order of mag- nitude as related to either axis. For these reasons it is reasonable to predict that the MB will eventually be the choice for FES in the distant future. Power Electronics This subsystem has two major functions in the case of the magnetic bearing FES and a single function for the mechanical bearing version. The common function is to provide two- or three-phase (or DC) from the module to the nearest grid bus (at the proper voltage) and to provide an acceptable link to the motor/ generator. If the FES(s) is less than fully charged and the grid has cheap electricity, energy is accepted from the grid and is parceled to the FES(s) until they are fully charged. If the grid is near maximum capacity, the process is reversed, and energy is automatically discharged to the grid until the FES(s) is down to a se- lected 90 or 95% DOD. This process can be repeated as often as desired without limitation or harm to the FES as long as the MG is kept within its upper tem- perature limits. A heat pipe cooling subsystem keeps this limit within bounds. The second function of the power electronics for magnetic bearings is to supply power for the bearings and the FES internal instrumen- tation. External Housing/Vacuum Enclosure As mentioned earlier, with proper design, FES burst containment is not a significant issue. This opin- 3. COMPONENTS 3-5 ion is supported by work on several National Aeronau- tics and Space Administration (NASA) projects, in- cluding the International Space Station, but is generally contradicted in the automotive industry, where every internal combustion engine vehicle produced in the world today has the least safe design—an isotropic metal flywheel. For safety considerations, the ideal housing (and the one USFS recommends) should have the following characteristics: e Long-duration, high-vacuum integrity and electri- cally nonconductive, e Ability to serve as partial vibration damper, e Low heat transfer, ease of installing internal and external brackets and fittings, e High resistance to external corrosion and abrasion, e Nonbiodegradable and resistant to soil chemistry, and e Structurally tough, lightweight, high-strength, and inexpensive. An examination of the desired characteristics listed above points toward composite-fiber, wound FES rotor materials. A recommended fiber candidate is E-glass, which has the strength-to-weight ratio of five-and-a- half times that of type 4340 steel. In addition, its cost is roughly half that of steel; therefore, the strength per unit cost is about one tenth that of steel with far better char- acteristics. Housings are tubular and extra costs would ensue for stock tube shapes. For fiber winding these nonrecurring charges would disappear. Prototype Design for an Flywheel Energy Storage System for Stationary Energy Storage The design considerations discussed attempted to describe the various subsystems of a FES system, and to consider the unique possibilities of design variations that can be selected for any specific application. These opinions may not agree with those of other FES manu- facturers. At USFS, the current state of the art would suggest basing system design first on the rotor, a single or dual fiber composite rim with E-glass as the inner material and perhaps capped with a limited thickness of graphite fiber. The materials and processing used for the rotor would also be used for the housing. The other major component that must be consid- ered is the bearings. A key decision is whether to use magnetic bearings. USFS recommends magnetic bear- ings over mechanical bearings for the following rea- sons: ¢ Magnetic bearings can be made to control the rotor to rotate about either its geometric or its mass center, or in between these two limits. Presume also that some microscopic but unimportant yield occurs somewhere in the rotating mass, thus minutely shifting one or both of these centers. If a mechanical bearing assembly is used, in time this shift could compound the bearing loads and unknowingly affect the bearing’s life. Presuming that a failure does not occur, the maintenance costs could then skyrocket at some future date. e Very high-quality, high-speed, high-load, long-life mechanical bearings are also high cost. Further, these high-speed bearings are sometimes known to fail under operating loads comparable to FES operation. Magnetic bearings never wear out and are primarily dependent on the reliability of the system’s electronic components. e USFS believes that low production costs for magnetic bearings for FES can be achieved and continues to pursue efforts toward cost reduction. To design a prototype system specifically for a utility energy storage application, USFS first selected a maximum level of energy storage per module—about 1.4 kWh. This level was selected to minimize the fly- wheel’s momentum for safety and containment consid- erations. Increasing this level would probably rapidly increase developmental cost, possibly exponentially. Also, designing a single large module may eliminate many users with smaller storage needs. Using many small modules, instead of one large one may increase flywheel use and thus could contribute to cost reduc- tions based on high volume production. The lower limit of energy storage would be about 0.3 or 0.4 kWh per module. Another major design consideration was the maximum power at top design rpm. USFS selected arange of 5 to 20 kW. Figure 3-3 shows a small utility storage building comprising 1.15 MWh of energy storage and lists some additional specifications. The enclosure has a footprint of 15 ft’, is 21 ft high, and contains 1,152 modules. This modular concept will allow any individual fly- wheel module to be automatically removed and re- plugged into its required position by an “x-y-z” tram 3. COMPONENTS 21 ft t 15 ft Sample Specifications * FW module size: 1.7 ft dia x 2 ft high 2SSS6S9SSSSe5 6SSSSSSSS8SESS 6S982888888s8 SSSSSSSSssss BSSsees8S8Sss8 BNSSSSSSSsss 2SSSS98SSse28 6NSSSSSSSSSS 15 ft * FW energy storage: 1.0 kWhrs useable per module * Above complex has storage of 1.15 MW hrs. + FW operates to 95% DO! * FW eauipped with mag bearin: * Cycle life estimated at 20 to 30 years (4:1 speed ratio) + Every module accessible by auto remove/replace * Each module plugs into special sockets ¢ Every module has its own separate power electronics * Every module protected from “crow-bar short” + Above complex immune to sand, dust, temperature & water * Only maintenance is feeding “auto load/unioad system” * The above typical facility comprises 1,152 modules Figure 3-3. system. Such removals may be necessary for occa- sional maintenance. The same technique would be ap- plied to nearby power electronics shed(s). The capacity of this concept could be expanded by multiples of 15- foot cubes, or by fewer layers modules in any single cube. The power electronics for such a system should be located in an adjacent enclosure. One or more modules would be serviced by separate power electronics sys- tem(s), each of which should be arranged for ease of maintenance and/or replacement. If it should turn out that magnetic bearing controllers are cost-effective, then the bearing controllers would be in the same en- closure as the power electronics. The plot in Figure 3-4 presents major expansions of this modular concept. In this plot, the energy for a USFS Concept for Utility FES Building with 1.15-MWh Energy Storage Capacity. complex of flywheels assumed to be located in a square area can be seen up to 50 ft on a side and by two-layer increments up to the full 8-plane (8 layers) as a function of maximum energy storage. What is shown is that 15 MWh of storage can be placed in a relatively small space. Boeing The broad goal of this project is to support the im- provement of advanced energy storage components in cooperative efforts involving the government, technol- ogy developers and manufacturers, and electric power providers. There are still many gaps in the technology supporting advanced storage methods. In the case of flywheels, these range from fundamental material is- sues such as composites durability to supporting tech- nologies such as bearings, cryogenics, and safety 3. COMPONENTS 3-7 16 - MWhii8 Plane | MWhr/6 Plane Mwhr/4 Plane Mwhr/2 Plane MWhr/t Piane| oo} 0 Di £ z | Note: “8 Plane” Denotes #FW e | Modules In Vertical Stacking. a | = s- 2s i a > 2 Eo Los «| | & 2 | a a — 4- —— oS . 7 at Pl o- t + + rr 20 25 30 35 40 45 50 55 Square Building Footprint: ft? Figure 3-4. systems. One of the goals of the component develop- ment activity is to support projects that will, through a multi-phased approach, result in the development and testing of components that are large enough to be used in field demonstrations—not just “beaker-scale” R&D. Flywheel research is at a stage where the transition to such systems can be made. To do so will require many skills and the coordinated efforts of many participants. The objectives of Boeing’s Phase I effort included completing a survey of potential applications for fly- wheel storage, selecting a target application for future development, determining which technologies were most in need of further effort, and laying out a rough plan for carrying out development through a successful field demonstration. Boeing worked with two firms with experience in identification and analysis of power markets to rapidly assist with fact-finding, both for potential target appli- cations and host sites. Boeing then developed concep- tual designs for two sizes of flywheel systems and examined manufacturing costs for those systems as projected with learning curves on major components. The specific costs (on a $/W and $/Wh) basis were compared with existing solutions now on the market, primarily battery systems. A further narrowing down of the selections was performed to achieve the class of applications of interest. Some of the issues important to USFS Utility Flywheel Multi-MW Footprint Parameter Study. development of flywheels for the selected application were then identified. Target Applications Boeing used the results of two studies and its own research to evaluate possible target applications. The first study, conducted by Energetics, Inc., evaluated data from the World Bank and other studies. This work primarily addressed off-grid power systems, which in- clude both diesel and renewable generation. The Ener- getics study shows the enormous potential for storage technologies to improve the capabilities of PV, wind, hydro, and diesel generation sites. Applications range from very small but numerous household and village systems to larger communities, businesses, and military bases, which must be able to operate independent of a power grid. Most of the data in this study were based on PV systems but the storage requirements for other renewable and hybrid systems will be similar. The second study was conducted by Jack Wood Associates and included a site visit to Kotzebue, Alaska. This study focused on gathering information on several potential applications and host sites for dem- onstrations. The study also provided a variety of sup- plementary information on power generation in Alaska and opportunities for potential state agency involve- ment. 3. COMPONENTS In 1994 and 1995, Boeing’s Defense & Space Systems Group undertook a detailed study of MWh- class flywheels for a centralized load-leveling applica- tion. Such wheels would postpone or avoid the need for additional generating capacity and allow current generators to operate more efficiently. The size of the wheel was driven by the economies of scale required for this application. The tremendous leap forward in the size of flywheels, to 15,000 Ib/11-ft-diameter rims, was made feasible by two Boeing innovations: — the high-critical temperature (Tc) superconducting suspen- sion, and a non-autoclave process for composite ring fabrication. Boeing and a major utility worked together to de- fine not only the requirements but a detailed plan to obtain needed information and to reduce technical and cost risks. Advanced gas-fired turbines provided the benchmark for pricing. Ultimately it was determined that the market could use several thousand flywheels per year at a delivered price of $400 to $600/kWh. When comparing this target with reasonably aggressive manufacturing cost projections, it was determined that serving large grid-connected applications such as load leveling (LL) may be economically feasible even in the near term. The margin for error was large, however, so it was difficult to make a clear-cut case for immediate investment. Power quality (PQ) and UPS systems are also in- teresting applications for flywheels. A capable PQ system is already being marketed by Active Power of Austin, Texas. The steel wheel, essentially a high- inertia motor on conventional bearings, can deliver high powers for several seconds. Boeing’s current focus is toward the substantially increased stored energies at- tainable with composite rotors, and low-loss/wear-free suspensions based on magnetic technologies. The Boeing/DOE Superconducting Partnership Initiative (SPI) project is ultimately targeting UPS applications with a large emphasis on suitable motor technologies. After reviewing the reports of Energetics and Jack Wood Associates and previously acquired data from a variety of sources, and Boeing’s production cost esti- mates for different classes of wheels, Boeing was able to determine a more specific application. A summary of the potential market to 2010 is shown in Table 3-2. Comparable domestic United States applications can be found primarily in Alaska, Hawaii, and some mainland sites such as Indian reservations. While small as a proportion of United States power production, the off-grid component in Alaska comprises roughly a third of the state’s total power bill. The cost of power in these areas is now quite high—anywhere from $0.30 to $0.60/kWh. This leads to an acute need for power effi- ciency and for local means of adjusting to variations in demand as well as generation. Perhaps more important for the future of storage technologies, domestic devel- opment and demonstration can serve as a springboard to reach the greater worldwide markets, and can move toward eventually serving grid-connected applications. The city of Kotzebue, Alaska, has a total load of 1.5 to 3 MW that is primarily diesel-generated but in- cludes several hundred kW of wind generation. Kotze- bue falls somewhere between Applications IV and V in Table 3-2. There is a move toward more wind genera- tion in Alaska, with more “high-penetration” installa- tions that will use diesel as little as possible. The success of these installations will rely heavily on stor- age technologies being available as well as efficient backup strategies employing diesel. Without storage, a reliable system would require nearly constant operation of a diesel generator at some low fraction of its rated power. This is highly inefficient, as is shown in Figure 3-5. A preferred strategy is to use storage to serve the load in all but prolonged drops in available wind en- ergy. A flywheel system meeting the requirements of a high-penetration wind-generation site thus presents itself as an excellent target application on which to base future development efforts. Power System and Storage Unit Description The general specifications for a wind/storage/diesel system are shown in Table 3-3 below. This represents a small but practical system for a cluster of several homes or a small business center. It would use commercially available wind and diesel components. The minimum stored energy requirement is about 3 kWh. If up to 6 kWh is available, then several operating options will be available. The system could sustain longer drops in wind speed (for at least 30 minutes), provide higher peak power capability, or operate in conjunction with a second wind turbine in a larger system. The conceptual design of a flywheel storage unit for field testing is described below and shown in Figure 3-6. Major components include the rotor, the bearing systems, the containment and vacuum shells, a motor/ generator with its controller unit, and the charge/ discharge unit. Composite Rotor: While it may seem that an en- ergy storage ring should be made of dense materials, in fact the opposite is usually true. This is because stored energy increases as the square of the velocity but only in proportion to mass. A wheel made of dense materials 3. COMPONENTS 3-9 Table 3-2. Off-grid Storage Applications, Their Requirements, and Potential Markets to 2010 According to Boeing I. Single Home: Il. Developing __ Ill. Developing - dates ale! V. Advanced Application Developing Community: Community: Mo: Meee Community or Community No Industry Light Industry Industry Military Base Storage System 0.5 kW 8 kW 40 kw 400 kw 1 MW Attributes 3 kWh 45 kWh 240 kWh 3600 kWh 1.5 MWh Power Base: 0.5 kW Base: 5 kW Base: 10 kW Base: 100 kW Base: 100 kW Peak: <8 kW Peak: <40 kW Peak: <400 kw Peak: <1 MW Discharge 5 to 72 hr 5 to 72 hr 5 to 24 hr 5 to 24 hr 0.5to1hr Duration Total Projected 47 Million 137,000 40,000 84,000 131,000 Number of Systems Fraction of > 50 > 50 = 30 =10 <5 Market Captured by Storage (%) Total Number of > 24 Million 69,000 = 12,000 8,000 < 7,000 Storage Systems to Capture Market Share 14 1 12 — 10 5 8 os = | = | Ze = | | 4 | 2 () r 7 + — - : - - + 0 0.1 0.2 0.3 04 0.5 0.6 0.7 0.8 0.9 1 Fraction of Rated Power Figure 3-5. | Energy Output per Gallon of Diesel Fuel vs. Fractional Utilization of Generator Capacity. 3-10 3. COMPONENTS Table 3-3. Boeing Preliminary Specifications for Storage Component of a 12-kW Wind-Generation System with Backup and Peaking Capabilities Attribute Requirements Goals Configuration 12 kW wind gen. w/storage, part- same time diesel backup Net power output 12 kW 12 kW continuous 18 KW intermittent Discharge time 15 min 15-30 min Usable stored energy 3 kWh 3-6 kWh Voltage Varies; typ. 120 VAC, 3-phase same 240 V or 480 V Round-trip efficiency > 70% > 80% Idling loss Probably < 1 kW 500 W Footprint Not critical < 100 sq ft Routine maintenance Typically < 8 hr/month <2 hr/month Reliability Extremely high < 3 unscheduled events, 20 yrs. Site requirements Aboveground, could be atalocal § Aboveground, operable in un- power station heated wind farm shed. Operating environment 60 — 80°F(approx.) —100°F — 80°F Installation time Not critical 3 days System life 10 yr 20 yr Mobility/mass Transportable by small truck, < 1000 Ib. excluding electronics moveable with hoist 39 Motor Controller |pc Charge/Discharge ps ae (Variable Speed) |Link| Unit: DC to babel du 60 Hz AC and Loads Composite Containment Motor Generator \_ Shell a 7 | \ | Composite Rotor — Bdaring|Magnet Asbembly \ i mil | Pump — = = LS Ik “Superconducting Bearing Stator' T T Closed—Cycle \ Figure 3-6. | Boeing Conceptual Design of Flywheel System for Target Application. ‘— Touchdown Bearings Cooler ~—«———_ Rotor Diameter appx. 3’ 3. COMPONENTS 3-11 will also be subject to high centripetal forces that will tear it apart at relatively low speeds, while a low- density material can be spun to much higher speeds. The weight of the rim (Figure 3-7) is important for two reasons. First, it is a major part of system mass in itself. There is approximately a factor of six difference between the mass of a steel wheel and the mass of a high-strength composite wheel storing the same energy. As system capacities become larger, the mass of steel systems will quickly become impractical to transport or to install at most sites. The second reason has to do with the requirements imposed on the rest of the me- chanical system, especially the bearings. Very heavy wheels must be supported by mechanical bearings, which are subject to wear, will be difficult to replace, and dissipate substantial amounts of the stored energy. A light wheel, on the other hand, may be supported by a magnetic suspension. Both active magnetic bearings (which use solenoid-like actuators to control the rotor shaft position) and superconducting versions (which act passively) have been demonstrated with composite ro- tors. The combination of composite rotor and magnetic suspension can almost be taken as the definition of an “advanced flywheel” offering high energy capability, reduced system mass, long life, and minimal mainte- nance. Power consumption data in magnetic bearings are presented in Figure 3-8. Hybrid: Superconducting/Permanent-Magnet Bearing: In this application, as with batteries, energy is continuously stored in the flywheel. Therefore, any power drains will have to be made up on a near- continuous basis and will directly impact the sizing of the primary generation source. Until the recent develop- ment of bulk superconducting, self-centering high- temperature superconducting HTS bearings, the energy loss associated with both mechanical and electromechani- cal bearings has been prohibitively high. These losses are often more than 5% of the stored energy per hour. With hybrid superconducting bearings, it is now possible to obtain losses that are as low as 0.1% per hour. Boeing has fabricated superconducting magnetic bearings and is currently working to improve their quality and reduce their cost. Nevertheless, revolution- ary breakthroughs are not required. For example, good- quality materials are already available, and key proper- ties such as critical current density are stable and re- peatable. Low drag has been demonstrated using small, single-piece permanent magnet rings. However, more work is required to improve the high-speed capabilities of the magnet assemblies and to understand reliability issues affecting all parts of the bearing. Motor/Generator: The motor/generator is re- sponsible for providing and extracting energy to and from the flywheel. A variable-speed motor controller transfers power to a DC link and then to a charge/discharge unit to the 60-Hz generators and loads. Key requirements on a motor for this application are the efficiency and ease of control at the speeds re- quired. This will usually result in a motor with a small number of poles. Another important characteristic is the magnitude of destabilizing forces present that can arise due to interactions between permanent magnets and steel in the motor. Types of motors that are suit- able for the application include halbach magnet arrays, axial gap DC motors, and some types of radial gap DC motors. Power Control System: The power control sys- tem is the interface between the motor/generator, the generation source, and the system load. The controller uses active rectification to convert the motor/generator's output to a DC level, from which point it can be in- verted to supply either single- or multiple-phase 60 Hz power at any desired voltage. The DC link isolates the motor/generator from the grid, letting the controller’s firing circuits handle the job of maintaining the correct phase and power factor at the output. The signal must also be clean enough to comply with IEEE Std. 519 for harmonic power distortion. Reliability and Safety Long-term reliability requires that the flywheel op- erate at something less than the maximum stress al- lowed for the fibers in the rotor, even after taking into account fatigue and creep; thus a rotor failure should not occur within the design life. As a safeguard against failure of the primary bearings, there should also be a touchdown bearing system. These bearings or surfaces also support the wheel during transportation and start- up of the bearing. Even with these safeguards, high- energy wheels should have a backup containment structure to provide protection against any unforeseen release of energy. The energy absorption characteris- tics of such containment structures and materials are established for steels and some other materials but it is important to keep the weight to a minimum. Innovative designs and approaches are needed in this area. The designs for major structural components in- cluding the rotor, bearing housing, and containment vessel need to be based on proven finite-element com- putational codes, design-allowable methodologies, and a comprehensive database of metals and composite materials properties. In some cases the data are insuffi- cient, and more materials testing is required. 3-12 3. COMPONENTS 1400 Weight of Rim for 6 kWh Useable Energy (pounds) Figure 3-7. 1200 1000 + R bosom Steel \ \ \ Segoe te Fatigue and Factor of Safety. Figure 3-8. Bearing Dissipation (W/kg) = ° ° a 2. % 3 4 Specific Strength (x 10° inches) 5 6 TTT THIN TTT TTT Active Magnetic HTSC 1 (Iron) HTSC 2 (Ferrite) } ott Ltitt 1 )oi titi RPM 10,000 100,000 Rim Weight of Thin-Rim Flywheel vs. Specific Strength (yield strength divided by density) for Steel and Composite Flywheel Rims Delivering 6 kWh of Useable Energy. Useable Energy Accounts for 75% DOD, 20% Power Electronic Losses, and Strength Derating of 2.0 for Power Consumption in Magnetic Bearings. Boeing HTS Material Was Tested in Bearings at Argonne National Laboratory; Comparison is to Active Magnetic Bearing Data. 3. COMPONENTS 3-13 The Boeing study identified applications suitable for near-term development and demonstration of ad- vanced FES systems. Off-grid power systems have been established at a number of sites in the United States and abroad, and deregulated power markets may encourage greater use of locally generated power. Effi- cient storage technologies can substantially improve the capabilities and reliability of such systems employing diesel, renewable generation, or both. A number of host sites have been identified that have suitable appli- cations and are interested in participating in a demon- stration project. The overall conceptual design for a flywheel en- ergy system for remote applications is reasonably well defined, although more work remains to be done. The economics of flywheel storage have been studied, be- ginning with up-front costs, and appear to be favorable, provided that demand will justify production at moder- ately high quantities. An accompanying study by Ener- getics, Inc. which will be included with Boeing’s final report, looked at worldwide markets for storage and concludes that the demand will in fact be very substan- tial for favorably-priced units. Battery Simulator Development and Validation The ESS Program initiated a collaborative project with the National Rural Electric Cooperative Associa- tion (NRECA) in FY97 to develop, validate, and dem- onstrate simulators of power-quality and peak-shaving systems. The project will provide technical and eco- nomic data about peak-shaving and power-quality im- provements at electric membership cooperatives (EMCs). More important, the project will introduce a technology assessment that is more exact and no more expensive than a traditional “paper feasibility study.” The ESS portion of the project supports the devel- opment and validation of energy storage simulators that will mimic the operation of two BESSs: one that Brockway Standard operates for power quality (PQ2000) and one that Energy United (EU) Coopera- tive operates for peak shaving. The NRECA portion of the project supports field demonstrations of the energy storage simulators and the development, validation, and field demonstration of a diesel-generator simulator. The validation and demonstration activities will be hosted at sites to be selected from NRECA members. The project team of Energetics, Inc. and Orion Energy Corporation is developing the simulators, conducting the simulations, and analyzing the output of the simu- lators. For validation, analysis will compare the simu- lator behavior with the real energy storage systems. For demonstrations, analysis will determine the financial feasibility of the utility hardware being mimicked. Status Development and validation of the battery peak shaving simulator at EU’s 500-kWh ESS at its States- ville, North Carolina facility was completed. A draft report was prepared and is under review. During the validation period of August 1998, to April 1999, the utility dispatched the battery system 25 times. The simulator correctly dispatched the virtual battery 13 times. The simulator was off-line for most of the other events because of unanticipated failures of the laptop computer used to run the simulation. When the simu- lator did operate, the power level was consistently about 30% higher than the actual power level. The computer problem was resolved late in the year when the simulator was redesigned with a rugged industrial processor in place of the laptop. The power level issue was resolved with a software adjustment. In April 1999, the EU battery was specifically monitored to compare the actual battery output during a peak shave event with the simulator prediction. The simulator battery voltage and current during the dis- charge were within 5% of the actual values during the two-hour run. The system was also monitored during recharge, and while the voltage predicted was very close (within about 2%), the current (and therefore the power) varied considerably. Again, a software adjust- ment that takes into account the age of the battery will be made to resolve the discrepancy. A rough evaluation of the economic benefits of the battery indicated the simulator predicted between 10 to 30% excess benefits. This is due to the reduced actual battery capacity com- pared to the simulation. Software changes to account for the actual battery capacity will correct this offset. The power quality simulator at Brockway Standard was inoperable until late in the fiscal year. At that time, a separate backup power supply was installed and the validation process continued. The simulator output will be compared with the PQ2000 system data recently obtained for the data management project to complete the validation. Both simulators will be relocated to demonstration sites for the next stage of the project. An intensive ef- fort is under way to identify and gain the commitment of utility cooperatives to host the demonstration activi- ties. Close cooperation with the NRECA will be neces- sary for the demonstrations to be successful. 3-14 3. COMPONENTS In the diesel generator simulator development and validation part of this project funded by the NRECA, the two EU 500-kW diesel generators at Statesville were used to validate the simulator. These diesels are used to assist in load control and peak shaving. A pre- liminary report indicates that the diesel simulator is working properly and will be through the validation process in the near future. This simulator will also be relocated to a demonstration site once the validation is complete. Finally, late in the fiscal year, the NRECA ex- tended its part of the project to include the development of fuel cell and microturbine simulators. The ESS Pro- gram also extended its part of the project to include the redesign of the simulator computers (mentioned above), development of a single simulator capable of operating as any of the technologies (battery, diesel, fuel cell, or microturbine) simultaneously, and demonstration of the “universal” simulator at suitable cooperative sites. This work will take place in FY2000. Valve-Regulated Lead-Acid Reliability Improvement Project VRLA batteries have been commercially available for more than ten years and have been enthusiastically embraced by users of uninterruptible power supplies because of anticipated reduction in maintenance costs and the smaller footprint available with this technology. As field experience has accumulated, it has become more widely appreciated that VRLA batteries are more sensitive to their operating conditions than flooded lead-acid batteries. This is particularly true under con- ditions such as elevated temperature or overcharging, which can lead to battery dry-out in starved-electrolyte designs. The result is a shorter than accustomed battery life. Although some VRLA failures may be attributed to abusive environments or improper float-charging conditions, users lack confidence that all of the possible life-limiting conditions for VRLA batteries have been identified. The information on recent failures has made potential utility battery customers, including users of standby power systems, more reluctant to adopt BES technology, particularly if WRLA designs are being proposed. Because SNL believes that VRLA battery technol- ogy offers real advantages in utility and renewable en- ergy applications, a VRLA reliability improvement project was formulated. The primary objective of the project is to determine VRLA cycle and calendar life under typical utility battery operating conditions and use modes. The ESS Program and ILZRO have estab- lished a collaboration that addresses VRLA reliability issues. The two organizations generally agree that to address VRLA battery performance issues, the follow- ing areas would have to be advanced: optimizing VRLA batteries for stationary applications, establishing the best charge control system, and performing a VRLA reliability assessment. Another important issue is the fact that no method exists to rapidly characterize VRLA battery life. During FY97, the ESS Program finalized plans to develop a VRLA Industry Users Survey in collaboration with ILZRO to begin gathering informa- tion from battery manufacturers and users, which is believed to be the necessary starting point in evaluating field reliability information. A three-phase project was designed to identify and resolve these battery life issues. Phase 1 involves a survey of the industry, in cooperation with VRLA manufacturers and users, to determine objectively and accurately the status of the technology. Phase 2 inves- tigates the critical issues identified in Phase 1 and sug- gests improvements to the charging methods or other aspects of the technology. Because it is expected that charging protocols will be one of the most critical areas identified for optimum operation, Phase 2 is likely to include a charging study that focuses on those issues. Phase 3 will then attempt to correlate and match the various types of VRLA technologies to the numerous applications now using lead-acid batteries. This will assist users and battery suppliers in selecting a design for an intended application, so that the battery will be appropriately specified. The Phase 1 study consists of three tasks: Task 1: Identify VRLA manufacturers and char- acterize their share of the market by design type and application. Invite each to participate in the study and involve representative users of the products. Task 2: Develop a detailed list of data needed to characterize the VRLA technology and problems identi- fied by the users. Recover field monitoring data on the systems wherever possible. Organize a database to receive the information. Task 3: Collect the data, analyze for trend infor- mation, and summarize the results in a final report. During FY98, two contracts were placed by IL- ZRO to initiate the three tasks in the Phase 1 study. Questionnaires targeting both VRLA manufacturers and users were designed by the contractor and distributed. This analysis will commence as soon as the surveys are received. The survey questionnaires solicit information on battery physical characteristics, electrical ratings, performance and life characteristics, application and 3. COMPONENTS 3-15 operational requirements, R&D database availability, market and sales, and known instances and modes of failure in the field. Status Efforts continued in FY99 to encourage return of the surveys, especially by the user group. Manufactur- ers were also reminded to submit data when they were encountered at professional meetings and during vendor visits. Five manufacturers of VRLA batteries com- pleted surveys that discussed the materials, manufac- turing processes, and service-life expectancies of 13 different technologies that serve in stationary applica- tions. The technologies included both gel and AGM designs and the applications included both float and cycling service. The manufacturers also identified customers who have field experience with their largest-volume product, their best-performing product, and their least-favorable product. The data from the manufacturers’ surveys are recorded in an Access database. The customers who were identified by the manu- facturers and those who were identified independently received surveys that are designed to collect data on the field performance of the specific technologies ad- dressed by the manufacturers’ surveys. The applica- tions in which the end users employ the VRLA batteries include schools, federal facilities, telecommunication relays and substations, electric utility substations, elec- tric utility emergency start-up/shutdown power, and renewable/hybrid power supplies. One end user de- clined to participate, citing lack of human resources as the reason. Another user with approximately 400 AGM cells in cycling service has completed and returned a survey. In the second quarter, project analysts populated the end-user portion of the database with temporary “dummy data” to enable the analysts to test the efficacy of the database queries. The testing process will also help to refine the queries of the database as well. The responses from end users will replace the dummy data after testing of the database is complete. The time to complete the survey was extended in the third quarter because it was felt that a more asser- tive approach was required in order to gather sufficient data for the second half of the project. Manufacturers, consumer groups, and other research organizations were contacted and an expanded list of users, including names of individuals to contact, was developed. Based on feedback from recipients, the user survey was re- viewed for technical content and revised to make it easier to complete. Revised surveys were distributed to the expanded list of users on this list, and each contact was informed personally of the survey and its purpose. A transmittal letter that was much more emphatic about nondisclosure was included. Follow-up calls were made to each contact, and attempts were made to gather the required data via telephone. Where necessary, visits to sites were made to obtain the data. By the end of FY99, over 50 end users had been contacted about the survey, and ten had agreed to com- plete it. This should generate a large enough data set to carry out the trend analysis. Emitter Turn-Off Switch Development for Power Conversion System The development of a high-power semiconductor switch, an emitter turn-off (ETO) thyristor, began in the third quarter of FY98. This project entailed the devel- opment and testing of a prototype ETO. ETOs could greatly enhance the power conversion system (PCS), a vital part of the energy storage system used to interface between the storage component (for example, batteries) and the generation and T&D equipment. At the heart of the PCS are the topological connections of high-power semiconductor switches. To meet the demand of these devices, efforts have been made to improve the gate turn-off (GTO) device in the past few years for applications rated in the megawatts. The ETO is a type of GTO device developed at Virginia Tech that could substantially advance high-power ESS applications. Based on the mature technology of cur- rent semiconductor devices, the ETO could provide a low-cost, superior solution for megawatt applications. Significant advantages of the ETO compared to the conventional GTO applications include low driving power, high switching speed, elimination of the dv/dt limiting snubbers, and device-level overcurrent protec- tion. The IGCT also features the removal of turn-off snubbers and high switching speed, but the ETO has several additional features. The ETO has much lower driving power than the IGCT as well as much simpler and lower-cost driver circuitry. The ETO also has the capability to include the on-driver overcurrent protec- tion that the IGCT does not have. In FY99, this project was extended to the next phase of ETO development. The objective of this phase is to build and demonstrate an ETO in a high-power converter with thermal, electric, control, and reliability all being demonstrated and tested in one unit. The pre- liminary version started at 100 kVA and is anticipated 3-16 3. COMPONENTS to gradually move up to high power (1 MVA and higher). Status The conventional solution to high-voltage PCSs has been the GTO thyristor. However, these systems suffer several problems because of the GTO’s charac- teristics. First, because of the poor safe operating area (SOA), each GTO requires a large turn-off dv/dt (the change in voltage per the change in time) snubber ca- pacitor that is typically between 2 and 6 uF. This ex- pensive snubber causes large losses that are in proportion to the switching frequency and DC link volt- age. Second, a conventional GTO has very long turn- off transient times, including both storage time and minimum off-time. These long switching transient times combined with the above-mentioned snubber loss limit the switching frequency of a GTO application to a very low value, typically around 300 Hz. Third, device protection is very difficult for the GTO system. Even with very high power ratings, the GTO has a relatively low maximum controllable current; any turn-off opera- tion conducted at higher current will destroy the device. Complex protection strategies are required for GTO- based converters at the system level. With the introduction of high-power, snubberless turn-off thyristor devices such as the integrated gate- commutated thyristor (IGCT) and the ETO, the overall performance of a high-power system can be improved significantly. The dv/dt snubber for the switch can be scaled down or eliminated because of the improved turn-off capability. The switching frequency can be increased because of the removal of the losses in the dv/dt snubber as well as the greatly decreased turn-off time and improved minimum on/off time. Also, the improved maximum turn-off current or SOA combined with the very fast turn-off speed ensures the turn-off operation to protect the device from overcurrent dam- age. The first phase of the ETO project at Virginia Tech was the development of a power semiconductor switch, ETO4060, with high voltage (6.0 kV) and current (4.0 kA) ratings. This device was successfully devel- oped, and a sample was tested to its full rating on a pulse-testing circuit. Although the pulse tester can show some useful data, any new device must be tested in a real power converter in order to verify that the de- vice can work with continuous current where thermal issues as well as repetitive stress issues become impor- tant. The current phase of the ETO project is investi- gating the performance and the impact of a power sys- tem by using the ETO as the main switch. A smaller ETO, a 1.0-kA/4.5-kV ETO1045HS, was used to im- plement a standard phase leg and was used in a DC/DC converter system. Emitter Turn-Off Snubberless Turn-Off The device used in the system, the ETO1045HS, is shown in Figure 3-9. The snubberless turn-off capabil- ity of ETO1045HS has been verified at up to 3.0 MW of instant power through the ETO during the turn-off transient. The snubberless switching shows only a slight increase in turn-off switching loss compared to that with a 2 uF dv/dt GTO snubber, and shows a dra- matic decrease in turn-on loss due to the lack of dis- charge current from the dv/dt snubber during the voltage fall-time. The overall switching loss is thus reduced when operating in the snubberless condition. Although the ETO can safely turn off without a snub- ber, the diodes cannot. Therefore a di/dt (the change in current per the change in time) snubber is employed to limit the recovery of the diodes as they turn off. The diodes turn off as the ETO turns on, making the turn-on loss for the ETOs negligible in the presence of the di/dt snubber. Figure 3-9. Picture of ETO1045HS Model. System Configuration Three-phase inverters are typical of high-power in- dustry applications. A phase leg, composed of two switches and their anti-parallel diodes, is a standard building block for almost all high-power (voltage-fed) topologies. As the initial step of building an ETO ap- plication system, a simple configuration of the ETO system that makes use of a phase leg is shown in Figure 3-10. The boost topology was selected in order to achieve a high-voltage stress on the ETO without using a high-voltage power supply. 3. COMPONENTS 3-17 Figure 3-10. Boost Converter Based on a Standard Phase Leg. System Construction In order to build a high-power converter with an ETO, it was necessary to develop a phase leg (half-bridge) with two ETOs, two diodes, the associated gate drivers, and suitable heat sinks. The ETOs and the diodes are in assembled press-pack (“hockey-puck”) packages, so mechanical clamping is necessary. The phase leg uses five liquid-cooled heat sinks so each device is cooled on both sides. Because the switches dissipate more power than the diodes, the two ETOs are separated from each other. Another requirement for the design of the phase leg is that the path length from each ETO to the opposite diode should be the same in order to equalize the voltage overshoot on the switches. The completed phase leg is shown with gate drivers in Figure 3-11. Using a laminated bus and the ETO phase leg, a DC/DC boost converter (Figure 3-12) was built. It uses the bottom switch of the phase leg, and the top switch is always gated off. Figure 3-11. Photograph of Completed Phase Leg. Figure 3-12. An Overall View of the Converter. System Performance The operating waveforms for the 100-kW test are shown in Figure 3-13. The stress on the ETO was 120-A switching current and 2000-Vde (with 2250-V peak spike). The overvoltage due to the di/dt snubber is about 250 V, which is reasonable compared to the DC link. 29-Jul-39 15:37:18 2 ms | a.s0Kv : Average (2) .2 ms 19.8nw =2814 sups Inductor Current (20 A/Div) Poe ‘Emitter Switch Gate Voltage (10 V/ div) -2 ms t Figure 3-13. Waveforms of ETO System Running at 100 kW. Major Accomplishments and Conclusion An ETO-based 500-kW single-phase-leg converter has been built. It was tested at the 100-kW level, which is the power capability of the Center for Power Elec- tronics Systems (CPES) lab, with a switch stress of 120 A/2250 V. The system exhibits a high operating fre- quency for thyristors at 1 kHz, good thermal manage- ment, and snubberless turn-off capability as well as ease of control. The gate driver features on-device overcurrent protection that has been verified on both the ETO1045 and the ETO4060. This protection acts when 3-18 3. COMPONENTS the device current has reached half of the maximum controllable current, thus preventing the current from rising out of the controllable range. This is allowed by the fast switching speed shown by the ETO in conjunc- tion with the di/dt limiting snubber that was developed in this phase of the project. Future Plan The next step for this converter is to increase the DC voltage to 3000 V. Increasing the current level without a large power supply will be accomplished by using reactive power transfer in an inductor-loaded full bridge. A three-phase multilevel inverter would be the typical application of a device such as the ETO, so the testing of such a system is the eventual goal of this converter. Additionally, soft switching techniques that can increase the operating frequency will be examined. Development of Intelligent Controls and Control Strategies for Renewable Generation and Storage Both the ESS and Photovoltaic Programs are inter- ested in the development, design, and testing of an im- proved, more versatile, and better integrated PCS. Power electronics and the supporting circuitry and software are critical to the eventual wide-scale use of renewables and energy storage technologies for many diverse applications. As documented in the recent SNL report, Renewable Generation and Storage Project— Industry and Laboratory Recommendations (SAND98- 0591), the DOE ESS Program is embarking on a major initiative to address system integration and components of renewable generation and storage (RGS) systems. This project includes a competitive procurement for an industry partner to develop an intelligent system con- troller and evaluate advanced control strategies for PCSs used in RGS applications, and will include tasks for design, applications analysis, fabrication, and com- ponent and system testing. Status In FY99, the ESS Program gathered information from the energy storage and PV industries and nonin- dustry organizations such as the Florida Solar Energy Center and the New Mexico State University Technol- ogy Development Institute to clearly define needs, di- rections, and scope of an improved RGS controller project before initiating a request for information (RFI). The objective was to prepare a clear taxonomy of RGS controls, explore how these controls should be organ- ized, and explore which functions are more natural to implement within subsystems and which functions are better suited to more advanced control principles. A statement of work (SOW) for the controller proj- ect was written. An RFI was sent to interested con- tractors to review the SOW and solicit information on the nature and the extent of the demand for an improved RGS controller. The purpose of the RFI was to achieve industry “buy-in” on the controller concept. Once the RFI has been reviewed, and if the ESS determines that it is practical to proceed with this project, a request for quote (RFQ) will be issued in the first quarter of FY2000, with a contract award anticipated later in the year. Evaluation Valve-Regulated Lead-Acid Battery Evaluation Controlled laboratory tests are the best method to determine battery capacity degradation rates and mechanisms. While batteries in field tests have the same problems, the often uncontrolled variability of the test environment slows the collection of data and makes it difficult to distinguish cause and effect. Life cycle testing continued in FY99 at SNL on VRLA batteries, which were currently under test for utility applications, specifically the GNB ABSOLYTE II and ABSOLYTE IIP AGM batteries. The GNB batteries were deliverables from a development contract with GNB. GNB is actively consulting on these tests. ABSOLYTE IIP Testing Status Testing continued in FY99 using the H-test regime shown in Table 3-4. Testing halted at the end of FY98 at Cycle 610 on September 16, 1998, to diagnose erratic Cell 1 voltage readings. At this point, the capacity had decreased to 972 Ah, near the defined end of life (EOL) of 960 Ah. Cell 5 end of discharge (EOD) voltage, which had been low throughout life cycle testing, began to decline more rapidly through Cycles 500 through 600. Testing resumed on November 30, 1998, with nine diagnostic C/8 discharge cycles that used the H- charge regime and was performed during December 1998. Testing was suspended for the annual year-end plant shutdown, and was resumed on January 27, 1999. ABSOLYTE IIP capacities are shown in Figures 3-14 and 3-15, and EOD cell voltages are shown in Figure 3-16. 3. COMPONENTS 3-19 Table 3-4. ABSOLYTE IIP and ABSOLYTE II Test Regimes H-Test Regime ABSOLYTE IIP | ABSOLYTE II Discharge at 150 A to 14.0 V Discharge at 150 A to 15.75 V 5-min rest 5-min rest Charge at 300 A to 19.2 V Charge at 300 A to 21.6 V 1-min rest 1-min rest Charge at constant 19.2 V, with a maximum 300 A, tapering to 24 A or to 7% overcharge Charge at constant 21.6 V, with a maximum 300 A, tapering to 24 A or to 7% overcharge 1-min rest 1-min rest Charge at 24 A to 7% overcharge Charge at 24 A to 7% overcharge Wait 8 hr Wait 8 hr 1200 | LO i @ Ah Out 2 Ahin EOL + = | 1 \ a 1100 + i -- —- ——— — 1050 t - = <= 2 1000 °o s a s § 950 900 850 | | 800 J = 660 670 680 690 700 710 Figure 3-14. ABSOLYTE IIP Capacity vs. Cycle Number for FY99. The erratic Cell 1 voltage problem was traced to an input voltage that exceeded common mode voltage limits on a tester input multiplex board. The board was changed to a relay-based type. Cycles 622-625 were performed using a 20% overcharge in place of the 7% overcharge of the H-test regime, which raised capacity to 1005 Ah at Cycle 625. Testing was suspended at this point for database upgrades. Testing was resumed on March 17, 1999, with capacities only slightly above the 960-Ah EOL level. The capacity continued to decline on succeeding cycles. By Cycle 693, capacity had de- clined to 914 Ah, well past the end of life. An 18% boost overcharge was performed on Cycle 694, with slight improvement in capacity. Testing was halted on May 23, 1999, at Cycle 699, pending a decision on post-service-life tests. A decision was made to attempt to recover some battery capacity. A boost charge of 1600-Ah input was performed on July 27, 1999, followed by two C/8 dis- charge capacity cycles. These three cycles (700, 701, and 702) are shown in Figure 3-17. The H-test regime was used on Cycles 701 and 702. Capacity improved slightly to 984 Ah on Cycle 701, but dropped to 938 Ah on Cycle 702. 3-20 3. COMPONENTS 1150 1100 8 capacity (Ah) 8 g 900 850 800 at | 1 a 11/1/98 12/1/98 12/31/98 1/31/99 32/99 4/2/99 9/2/99 6/1/99 date Figure 3-15. .ABSOLYTE IIP Capacity vs. Date, Showing Periods of Open Circuit. 1.9 1.85 |}; i+ + 1.8 1.75 cell voltage 5 B o 1.55 i * Cell1 = Cell2 = Cell4 -x-CellS © Cell6 + Cell7 = Cell8 ~ Cell9 T T T 7 7 ] 7 T ] 15 | i 1 I i L L I i | 600 610 620 630 640 650 660 670 680 690 700 710 cycle Figure 3-16. EOD Cell Voltages for the ABSOLYTE IIP—Cell 3 is Bypassed. 3. COMPONENTS 3-2] 1800 7 1600 > —— | | Ah out | @ AHRSRET 1400 oo : | | 1200 ao g cycle 702 © 1000 — = = o . 938 Ah 2 800 ==] E < 600 Figure 3-17. GNB suggested that useful information on the cause of the decline of capacity could be derived from a refer- ence electrode inserted into one cell. The vent was re- moved from Cell 7, and a cadmium wire reference electrode was inserted for the discharge phase of Cycle 703. The battery was discharged at a constant 150 A to 14.0 V, producing 921 Ah capacity. Results of the voltage measurements are shown in Figure 3-18. The results were sent to GNB, which replied that the cell appeared to be positive-limited, but that the negative was starting to polarize. GNB suggested that this is typical of acid starvation, and that the cell may be sulfated. This polarization could also be associated with dryout. It was suggested that a very aggressive charge profile be tried to determine if some capacity could be recovered. An aggressive charge program will be planned and tried in the first quarter of FY 2000. ABSOLYTE II Testing Status Life cycle testing of the ABSOLYTE II battery continued in the latter half of FY98 and into FY99, us- ing the ABSOLYTE II H-test regime shown in Table 3-4. This test regime is identical to the ABSO- LYTE IIP test regime in Table 3-4 except that the charge and discharge voltages of the ABSOLYTE IIP Soot eee 100 120 140 160 180 hours ABSOLYTE IIP Boost Charge and Capacity Cycles, After EOL Reached at Cycle 699. are adjusted for eight cells and those of the ABSO- LYTE II are adjusted for nine cells. A discharge cur- rent of 150 A was used for the ABSOLYTE II through most of the life testing instead of a C/8 current of 130 A, so direct comparisons could be made with the ABSOLYTE IIP. The capacity had been steadily decreasing, as re- ported in the ESS Program Report for FY98, SAND98- 0883. As was also reported, there were ambient tem- perature fluctuations in the building where the ABSO- LYTE II was being tested that caused fluctuations in measured capacity and contributed to tester malfunc- tions. In October 1998, the testing was halted for air conditioning and tester repairs. Twenty-three cycles (376-398) using the H-test regime were run during De- cember to evaluate tester modifications. The battery was then on open circuit from December 22, 1998, to February 13, 1999, when testing resumed. Capacity vs. cycle number is shown in Figure 3-19, and capacity versus date is shown in Figure 3-20. Cell 4 EOD voltage, which had been consistently below that of other cells, began to decline sharply shortly before the testing was halted in October 1998 (Figure 3-21) and stayed low. Capacity also declined, reaching the EOL value of 832 Ah at Cycle 440. At Cycle 441, a boost overcharge of 148% returned 1232 Ah was performed, with a slight improvement in 3-22 3. COMPONENTS cadmium wire reference electrode T i T 4 Positive reference aaa 0.4 negative reference (calc. V) = re 206 mv |_| 93 8 < 2 02 £ o - @ 19 0.1 = = 3 18 0 gt & 3 17 ae 0.1 16 calculated negative reference f 0.2 15 0.3 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 time (hours) Figure 3-18. Voltages Measured and Calculated Using a Cadmium Reference Electrode in Cell 7 of the AB- SOLYTE IIP. 1300 1200 1100 = < 1000 2 8 a 900 S ° 800 700 600 340 Figure 3-19. ABSOLYTE II Capacity vs. Cycle Number for FY99. 3. COMPONENTS 3-23 1300 1200 1100 1000 900 capacity (Ah) 800 700 boost charge | = — Cell 4 bypassed 4 AhnOut O AhiIn — EOL 600 i i 9/1 Figure 3-20. 198 10/1/98 11/1/98 12/1/98 1/1/99 1/31/99 date 3/3/99 4/2/99 5/3/99 «6/2/99 7/3/99 ABSOLYTE II Capacity vs. Date, Showing Periods of Open Circuit. = 16 pO - [ - | o , | 2 ‘ | = 1.5 -— OO ———— RK i = > x | x - Cell 4 bypassed }———__» a4—— 7, _f a 7 * | | 1.3 po et * _| — Z | “ | x gk | EP x | | T | | | | | | | | 1.1. -——-— a | | @Cellt =Cell2 «Cell3 xCell4 aCells @Cell6 +Cell7 -Cell8 ~Cell9 1 | i I I 340 360 380 400 420 440 460 480 500 520 cycle Figure 3-21. EOD Cell Voltages for the ABSOLYTE II. Cell 4 was Bypassed after Cycle 446. 3-24 3. COMPONENTS capacity and in Cell 4 EOD voltage on subsequent cy- cles. Cell 4 EOD voltage was still significantly lower than other cells, and at Cycle 446 Cell 4 was bypassed. The ABSOLYTE II H-test regime was modified to be appropriate for eight cells, making it identical to the ABSOLYTE IIP H-test regime. Life cycle tests continued with declining battery capacity. By Cycle 506 (June 15, 1999), capacity was consistently below the 832 Ah that corresponded to 80% of rated capacity, and testing was halted. The battery was on open circuit for the remainder of FY99. A decision will be made on further testing or disposi- tion of the battery. Yuasa Laboratory Testing for Hybrid Environment In early 1998, a 24-Vde string of 12 Yuasa Exide, Dynacel DGX tubular gel VRLA cells were received at SNL. A test plan was developed to test these special cycling batteries in a laboratory. The test program emulated the off-grid hybrid operational environment and was run in parallel with the actual field testing of the DGX batteries at the APS STAR Center Hybrid Test Facility in Tempe, Arizona. This test program provided a data point to compare to the Dynacel DGX battery at STAR, which is being tested in an actual hy- brid operating environment. The battery is exposed to a cycling regime denoted as a complete capacity cycle (CCC), in which the battery is operated above 50% and 100.00% penne 90.00% +> below 80% SOC for extended periods of time. It is currently believed that the tubular gel battery can oper- ate at intermediate states of charge for extended periods of time with little if any sulfation and no stratification effects. If this is true, the use of this technology can possibly lead to the development of an optimal operat- ing strategy for off-grid RGS applications. Cycle test- ing began in October 1998 and continued throughout FY99. All testing was done in-house at SNL. Status The first cycle of FY99 (CCC 1) ended on October 24, 1998. The nameplate rating for the DGX battery is 405 Ah (all charges and discharges were con- ducted at a 5-hour rate). At the end of CCC 1, the ca- pacity test indicated that the battery was at 88% of its rated capacity. At the end of CCC 2, the battery capac- ity had risen to 91%. From that point on, capacity de- clined for the next six CCCs. During the first half of the fiscal year, the battery exhibited a severe drop in ca- pacity, to a rating of 77% by CCC 8. At this time, the battery was operated for about 500 cycles at 30% ca- pacity. As indicated in Figure 3-22 for CCCs 3 through 8, string capacity fell by more than 10% over a period of about six months. Several meetings and telephone con- ferences were conducted during the period in an attempt to determine why the DGX cells were exhibiting pre- mature capacity loss (PCL). 80.00% + 70.00% 60.00% + — 2 (Capacity 8 50.00% a $ oO 40.00% +—— — 30.00% —— - | 20.00% —— — _ 10.00% — ~ | 0.00% 1 3 5 7 9 "1 13 15 Complete Capacity Cycle Figure 3-22. Yuasa Exide DGX Capacity History. 3. COMPONENTS 3-25 Testing of the DGX battery was temporarily halted in late February following CCC 8 as test results indicated that the battery was not being properly recharged following each complete cycle. In Cycle 509, which was completed in mid- February, capacity had fallen to 77% of the name- plate rating after completing the equivalent of 163 cycles at 100% DOD. A meeting was held in mid- February with Yuasa engineers to review the test results. A second meeting was held in mid-March at the Yuasa offices in Reading, Pennsylvania. EECI represented SNL at the meeting. The find- ings from the meeting were that the battery had not been charged correctly while charging from 50 to 80% SOC. It was recommended that the charge voltage be raised to 2.40 Vpc. This increase in charge voltage appeared to solve the problem, and over the next several CCCs capacity increased to near 86% rated capacity. Following CCC 11, however, the DGX cells exhibited a severe drop in capacity after the com- pletion of approximately 380 operational cycles. During the period of June to July following CCC 12, string capacity fell dramatically by more than 10%. Several meetings and telephone conferences were conducted during this period in an attempt to determine why the cells were again exhibiting PCL after the recovery in late February. In early June, Yuasa engineers provided a new charge algorithm. The new algorithm proposed a constant voltage, constant current (CVCI) charge regime, which is considerably more aggressive than the original constant voltage charge algorithm. In the constant current charge, the cell voltage was allowed to rise to 2.55 Vpc while the current was limited to 12 A. As noted in Figure 3-19, capacity did not continue to fall but appears to stabilize around the 75% ca- pacity point. However, according to the test plan, the test was to be terminated when the capacity of the battery fell below 80% rated capacity. At the end of FY99, the test was terminated after com- pleting 1,082 operational cycles at 50 to 80% ca- pacity. Table 3-5 shows the complete history of the test program. The battery was expected to deliver 800 cycles at 100% DOD for an expected energy life of 324,000 Ah at the 5-hour rate of the tests. Using accumulated ampere hours delivered (130,603 Ah), the battery was at 75% nominal nameplate rated capacity after having delivered 40% of its rated life in ampere hours. Another interesting item in Table 3-5 is found in the percent overcharge column. It appears there was no signifi- cant overcharge delivered (accumulated Ah re- turned/removed), which was not the case. During the 30% charge and discharge cycles, the exact Table 3-5. Yuasa Exide Dynacell DGX Life History Table eae NOMINAL ; Accumulated Amp Hours ccc Cycles Capacity Capacity Ah Removed a a —— . 1 64 357.18 88.19% 283.45 8,834 8,933 101.12% 2 128 368.40 90.96% 309.36 16,999 17,158 100.94% 3 191 367.25 90.68% 300.52 25,150 25,359 100.83% 4 261 358.40 88.49% 293.37 34,109 34,378 100.79% 5 323 348.90 86.15% 289.21 42,118 42,430 100.74% 6 385 333.73 82.40% 265.70 50,086 50,445 100.72% 7 447 324.93 80.23% 268.31 58,048 58,443 100.68% 8 509 312.72 77.21% 251.47 65,979 66,412 100.66% 9 573 331.23 81.79% 296.03 74,287 74,795 100.68% 10 636 345.20 85.23% 271.52 82,261 82,833 100.70% 11 698 343.93 84.92% 275.02 90,245 90,894 100.72% 12 763 302.48 74.69% 265.22 98,177 98,825 100.66% 13 826 307.57 75.94% 266.85 106,459 107,268 100.76% 14 889 300.33 74.16% 260.28 114,398 115,276 100.77% 15 952 310.39 76.64% 254.79 122,339 123,262 100.75% _'6 1,017 _ 306.93 _ 75.79% _ 254.75 130,603 131,706 100.84% 3-26 3. COMPONENTS number of ampere hours removed were returned during the charge portion of each 30% cycle. This strategy is very reasonable as the battery charge acceptance effi- ciency is nearly 100% over this operating regime. Further analysis yielded information that verified that the proper manufacturer’s recommended overcharge of 10% was applied during the normal full recharge peri- ods, which occurred every 60 cycles at 80 to 50%. Following each 110% complete charge, the battery was fully equalized, again using the manufacturer’s sug- gested procedure for equalization. The question as to why the battery did not recover under the aggressive CVCI charge program needed to be answered. The most plausible explanation for this may be that the battery had been severely, if not perma- nently, damaged during the preceeding 12 CCCs and required even more controlled overcharge to recover the cells. Another question that needed to be answered was why the cells were experiencing PCL. After many discussions with Yuasa battery engineers, it appeared that another mechanism besides the typical sulfation buildup was driving the PCL. Yuasa engineers will attempt to determine the characteristics of this PCL mechanism. In early FY2000, the SNL DGX battery will be dismantled and returned to Yuasa for further testing and tear-down. Intermediate State-of-Charge Testing The ISOC test project was initiated in early FY98, and testing of batteries from several participating manufac- turers began in the first quarter of FY99 at the SNL test laboratories. Two VRLA gel electrolyte batteries and one AGM battery from three manufacturers were tested as single units and in a 60-V string to determine whether the batteries can operate in an ISOC environ- ment without damage. The ISOC environment allows for better energy management in off-grid applications by not requiring a full 100% recharge following dis- charge operations. During FY99, the test project re- quired internal technical support to ensure continued and reliable operation of the test equipment and data management system. No external contracts were needed to support this project. Because of initial agreements with the battery manufacturers who provided batteries for this test pro- gram, information on test results are reported for Bat- teries A, B, and C with no specific reference to the manufacturer or battery model. Following completion of the test program, information for test results from specific manufacturers and models may be made avail- able if manufacturers give their permission to release the information. Status Testing continued throughout FY99. Tables 3-6 through 3-8 summarize the results through the end of the period. Batteries A and B are flat-plate gel technol- ogy, and Battery C is an AGM. All are exposed to the ISOC operational regime where the battery is operated between 50% and 80% SOC with periodic capacity tests followed by full charges and equalization charges that occur every 60 ISOC cycles. Following each of the three tables are discussions of the particulars of each of the three batteries. Columns 1 and 2 (in Tables 3-6 through 3-8) give the data on the 12-V modules under test. Column 3 in each of the tables gives the data on the 60-V string consisting of five modules. Because of the way energy is removed from bat- teries in a hybrid cycling environment, the term “cycle” is somewhat misleading when one considers what really constitutes a cycle. Depth and rate of discharge both influence battery cycling performance. However, all batteries are designed to deliver a fixed number of am- pere hours based on the amount of active material that is manufactured into a battery. For fixed cycling appli- cations where loads and depth of discharge are always the same, a cycle is obvious. On the other hand, when loads and depth of discharge vary, defining a cycle is not so obvious. Consequently, the use of ampere hours delivered is a more accurate definition of battery per- formance in a variable cycle environment. All capacity measurements reported for ISOC testing are based on the total expected ampere hours that are available from the battery given the number of 100% cycles at the five- hour rate. Battery A is rated to deliver 20,700 Ah over a period of 300 cycles at the five-hour rate. Battery B is rated to deliver 24,150 Ah over a period of 300 cy- cles at the five-hour rate. Battery C is rated to deliver 28,000 Ah over a period of 400 cycles at the five-hour rate. All batteries undergoing ISOC testing have ex- ceeded their rated cycle life as of the end of the FY. Table 3-6. Module and String Data for Batteries on Test-Type A A-1 A-2 A-3 Battery A 12-V 12-V 60-V Module Module String ISOC Cycle Sets 16 16 16 Cycles Completed 1,012 1,010 1,009 Rated Capacity Ah 69 69 69 Current Capacity 53.7 57.6 42.7 % Rated Capacity 77.8 83.5 61.9 Equivalent 100% Cycles 322 314 312 Completed Equivalent 30% Cycles 1,072 1,046 1,041 Completed 3. COMPONENTS Battery A was not rated for a specific number of cycles in the manufacturer’s specification sheet. Typi- cally, this type of battery is expected capacities of 300 complete 100% cycles. Nominal capacity has fallen below the 75% capacity target for A-1 and A-3; how- ever, the batteries continue to support their loads with no apparent effects from the reduced capacity. The batteries will continue on test for several more CCCs in FY2000 in order to determine how the batteries behave in a cycling environment as they approach the end of their functional life. Table 3-7. Module and String Data for Batteries on Test-Type B B-1 B-2 B-3 Battery B 12-V 12-V 60-V Module Module String ISOC Cycle Sets 16 16 16 Cycles Completed 1,012 1,010 1,009 Rated Capacity Ah 69 69 69 Current Capacity 53.7 57.6 42.7 % Rated Capacity 77.8 83.5 61.9 Equivalent 100% Cycles 322 314 312 Completed Equivalent 30% Cycles © 1,072 1,046 = 1,041 Completed Battery B is rated for 300 cycles at 100% DOD in the manufacturer’s specification sheet. Nominal ca- pacity has held in the 80% range for Module 2 while Module 1 and the string capacity has fallen below the test capacity termination threshold of 75%. Testing will be terminated for all three units in early FY2000. Note that all three B batteries have greatly exceeded the manufacturer’s 100% cycle projection of 300 cycles and all units remain functional despite their decline in measured capacity. Table 3-8. Module and String Data for Batteries on Test-Type C C-1 Cc-2 c-3 Battery C 12-V 12-V 60-V Module Module String ISOC Cycle Sets 20 20 20 Cycles Completed 1,280 1,269 1,274 Rated Capacity Ah 80.5 80.5 80.8 Current Capacity Die 64.5 49.5 % Rated Capacity a2. 81 62 Equivalent 100% Cycles 412 410 382 Completed Equivalent 30% Cycles 1,374 1,312 1,314 Completed Battery C is rated for 400 cycles at 100% in the manufacturer’s specification sheet. Nominal capacity has held in the 80 to 85% range for the last several CCCs. There has been one module failure in the string (C-3), which resulted in the removal of the defective module and the resumption of cycling with only four modules in the string. All units are near or have ex- ceeded the manufacturer’s life cycle rating and all units are holding up well. Testing will continue until capac- ity falls below the 75% rated capacity. It is anticipated that the ISOC test program will meet its expected goals and testing will be terminated during the first quarter of FY2000. An interesting ob- servation as to how capacity declines and how to im- prove management strategies for a cycling environment will be published as a SAND report sometime in FY2000. Details of the performance of each battery will also be published in FY2000. If permission is granted by the manufacturers, details of each model and part numbers will be included in the report. System Evaluation of a 200-kWh Zinc/Bromine Battery at “The Barns” In FY97, after completing a competitive procure- ment process, the ESS Program initiated a contract with Powercell Corporation to perform in-house tests on Powercell’s Zinc-Flow™ battery. The project was ex- panded in FY99 to include the field characterization of 100-kWh units that Powercell calls the PowerBlock. Two 100-kWh PowerBlocks were built and will be field tested at “The Barns,” a community theatre at Franklin Park in Loudon County, Virginia. This project will entail monitoring both battery performance and energy use/generation of all the major system components. Status The location of the field test was negotiated early in FY99. The size of the battery was increased to 200 kWh for this application. It will be charged by a solar array and grid power. Its primary purpose is to store solar energy generated in the day for use at night. The electrical design for the building was com- pleted in the third quarter. The design allows for opti- mum use of the BESS. A plan was developed based on this design to measure energy use/generation of the major components. The plan enables the project team to perform energy balances for the entire system. A draft plan for implementation (Figure 3-23) was completed in the third quarter. 3-28 3. COMPONENTS ~_| interruptibie | Loads Scroll Compressor Cooling Tower Fan Condenser Water Pump Glycol Pump —™ Non-critical Loads Incandescent Lights Selected Fluorescents Heat Recover Critical Loads | I Taper All stage lights Emergency Equipment {iG Sound Fans HVAC Controls Pumps not listed above 4h { By pass contacer Tow PV Amy = ~_/) awreica | —~ AL AM ‘Wom Trace Techndoges Figure 3-23. The Barns Implementation Plan. The Barns theatre is intended to be a net producer of green power, with a 200-kWh energy storage system and a 70-kW PV array. PowerBlock performance data and its response to grid failure are presented in Figures 3-24 and 3-25. Some of the most significant features of the PowerBlock unit are: four-quadrant control system, zero switching time backup, smal] footprint, mimimum maintenance, environmentally safe and benign. The PowerBlock will function to eliminate the following events: Sags, surges, and spikes Noise and harmonic distortions Under- and overvoltages Interruptions and outages The unit also functions to reduce the following: e Demand charges e VAR charges e Generator maintenance costs The state-of-the-art power electronics unit (Figure 3-26) for the system is IGBT based, and has high effi- ciency, four-quad operation, double conversion, fiber- optic controls, and a water cooling system. VV MA VV iN Alen + i Ws Grid Voltage Figure 3-24. PowerBlock Performance Data. 3. COMPONENTS 3-29 ee PowerBlock’s Response to Grid Fail- ure. Figure 3-25. PowerBlock State-of-the-Art Power Electronics. Figure 3-26. 3. COMPONENTS 3-30 4. Analysis Introduction The analytical tasks being performed in FY99 de- rive from studies conducted in past years. The Analysis program element contains two subelements: (1) system studies and (2) technology assessments. A system study is an initial screening study per- formed to identify and evaluate the potential applica- tions and benefits of energy storage. This screening study establishes a rough estimate of the benefit-to-cost ratio of storage using a limited examination of user- specific operating and financial data as a basis. Technology assessments evaluate the state of the art of storage and other relevant devices. These evaluations include the identification of appropriate technologies, a literature search, discussions with sup- pliers, and modeling. The ultimate objectives of these assessments are to characterize technologies to deter- mine their suitability for stationary applications and to define research needed to successfully develop and field advanced technologies in working systems. System Studies Opportunities Analysis—Phase II The Opportunities Analysis was conceptualized in FY94 as a two-phase project. Phase II of the project, started in FY98, is an extension of Phase I of the study performed earlier by the ESS Program. In a prelimi- nary assessment of national benefits, SNL estimated that generation and transmission applications of storage could represent $17.2B in national benefits. In Phase I of the Opportunities Analysis, the T&D benefits were found to be significantly higher than previous estimates. Phase II of the study included a refinement of the technical and economic understanding of the role of energy storage in the utility industry. Increased under- standing will help promote appropriate development and more rapid commercialization of utility ESSs. The current information is insufficient to estimate market size with a high degree of confidence, especially from a system supplier’s perspective. Activity in Phase II pri- marily focused on the need to characterize the near- and long-term utility application requirements for energy storage. Status The first meeting of Phase II of the Opportunities Analysis study was held November 1 and 2, 1998, in Santa Fe, New Mexico. The meeting was attended by 15 representatives from the storage and related indus- tries, several Sandians, and staff from Energetics. The meeting focused on specifying application requirements for stationary energy storage and categorizing them according to a structure meaningful to electricity sup- pliers and consumers, and reaching consensus on clear terminology. On April 22 and 23, 1999, representatives from DOE/SNL and Energetics and stakeholders from vari- ous parts of the energy storage, power electronics, and electric power industries met to conduct Meeting II. The goals of the meeting were to: 1. Reassess the value of applications for energy storage in a competitive electric power industry; 2. Refine definitions of application requirements for ESSs; 3. Identify the best applications for specific ESS technologies; Participants in the meeting included representatives from the following stakeholder groups: e — Electric-service providers — Enron Energy Services, — Northern States Power, - Public Service of New Mexico, — Southern Company, e Technology developers - Active Power, - Argo Tech, — American Superconductor, - ESA, — GNB Industrial Technologies, — Omnion Power Engineering, — ZBB, e Industry and technology experts - US. DOE, — Energetics, — ILZRO, and - SNL. 4, ANALYSIS Participants reviewed the requirements for each of the following applications and modified some of the application names to better reflect the nature of the ap- plications. The following application names emerged from this review: Rapid reserve, Area control and frequency responsive reserve, Commodity storage, Transmission system stability, Transmission voltage regulation, T&D facility deferral, Distribution facility deferral, Renewable energy management, Customer energy management, and Power quality and reliability. The group of industry experts who participated in the Phase II analysis used the perceived need for the application across the nation and the potential technical and economic benefits for utilities as criteria to evaluate the applications. From this process emerged the defini- tions shown in Table 4-1 for ten individual applications of energy storage that are in general demand and have high value with electric power producers and their cus- tomers. Table 4-1 also shows an organization of the applications under headings of Generation, T&D, and Customer. Although utility functions of generation, T&D, and customer service now generally occur in separate business units, the organization of the applica- tions under those headings does not signify that a stor- age system could or should serve only one application or application type. In fact, a storage system is most valuable when it performs multiple functions in more than one of these groups of applications. Regulatory, operational, and economic influences create technical requirements for each electric power application of energy storage. The industry experts who participated in this study identified the application requirements shown in Table 4-2 as the most significant to electric power applications: power; duration of dis- charge; AC system voltage; floor-space requirements; portability; and the type, number, and distribution of duty cycles. The characteristics of the operating envi- ronment are also important to technical requirements by the application on an ESS. The information in Table 4-3 summarizes the re- quirements of each of the applications. The technical attributes of the various storage tech- nologies considered in this study are compatible with some electric power applications and incompatible with others. Table 4-4 identifies the level of compatibility between each of the ten applications and the storage technologies under consideration. The following infor- mation was included when the compatibility ratings were being considered: the technologies’ capabilities in the near- and mid-term (as presented by participants in this study who develop the technologies and as re- viewed by the other participating technologists); the information in the preceding descriptions of the storage technologies; and technical and nontechnical alterna- tives and the application requirements presented in pre- ceding tables. Table 4-4 does not address the disparate state of development and experience of the storage technologies; nor does it address the cost- competitiveness of the technologies in the various ap- plications. For example, flooded lead-acid and VRLA batteries have a D rating for area control and frequency responsive reserve, while composite rotors have an L rating. These ratings indicate that flooded lead-acid batteries can definitely serve the application and that composite flywheels are likely to be able to serve the application in the future. These ratings are based on the fact that lead-acid technology has already successfully served the application, and composite flywheel devel- opers believe that their technology will also success- fully serve the application. The ratings do not reflect which technology will address the application require- ments in a more cost-effective manner when they are at the same level of technical maturity and experience. Participants also reviewed the information on tech- nologies with which storage will have to compete with in serving these applications. From this review, the group identified a need to characterize competing tech- nologies to ensure that the list of technologies is com- plete, that the technology capabilities are accurately represented, and that the power and energy costs asso- ciated with the competing technologies are current. In particular, participants agreed that an evaluation of the market share for storage would require refined cost breakdowns for each competing technology. A need for identifying competitive drivers for each of the ap- plications was expressed. Another result of the study was that research is needed on the interface and balance of system issues associated with the differences in the storage media. For example, VRLA batteries require tightly controlled charge management to achieve reasonable service life; FES needs to stabilize a variable signal from its MG before serving a load. Participants also identified the various stages of development that each type of storage system has achieved to date. For example, flooded lead-acid batteries are fully mature and commercially available, and are integrated with the PCS. VRLA is commercially available and is integrated with the PCS 4-2 4. ANALYSIS Table 4-1. Definitions and Categories of Electric Power Applications of ESSs Definition Category Rapid Reserve Generation Generation capacity that a utility holds in reserve to meet National Energy Reliability Council (NERC) Policy 10* requirements to prevent interruption of service to customers in the event of a failure of an operating generating station. Area Control and Frequency Responsive Reserve The ability for grid-connected utilities to prevent unplanned transfer of power between them- selves and neighboring utilities (area control) and the ability of isolated utilities to instantane- ously respond to frequency deviations (frequency responsive reserve). Both applications stem from NERC Policy 10 requirements. Commodity Storage Storage of inexpensive off-peak power for dispatch during relatively expensive on-peak hours. In this report, Strategic Storage for Systems Management refers to applications that require less than four hours of storage. Transmission System Stability Transmission and Distribution Ability to keep all components on a transmission line in sync with each other and prevent sys- tem collapse. Transmission Voltage Regulation Ability to maintain the voltages at the generation and load ends of a transmission line within 5 percent of each other. Transmission Facility Deferral Ability of a utility to postpone installation of new transmission lines and transformers by supple- menting the existing facilities with another resource. Distribution Facility Deferral Ability of a utility to postpone installation of new distribution lines and transformers by supple- menting the existing facilities with another resource. Renewable Energy Management Customer Service Applications through which renewable power is available during peak utility demand (Coincident Peak) and available at a consistent level. Customer Energy Management Dispatch of energy stored during off-peak or low-cost times to manage demand on utility- sourced power. Power Quality and Reliability Ability to prevent voltage spikes, voltage sags, and power outages that last for a few cycles (less than one second) to minutes from causing data and production loss for customers with demands of less than 1 MW. —— — =<=——==—— — — * NERC Policy 10, Draft 3, is available for downloading through Word 97 and Acrobat at http://www.nerc.com/~oc/standards/ 4, ANALYSIS 4-3 = = Table 4-2. Definitions for Applications Requirements Power requirements are measured in kilo- or megawatts (kW or MW) and kilo- or mega-vars (kVAR or MVAR) if the application requires reactive power: Power requirements determine the capacity of the PCS and can influence the size of the system via the power-to-energy ratio. The ESS must be rated so that power drain does not reduce its cycle life. Power requirements impact the size and cost of the ESS, the support structure, and the building. High power levels increase the cost of the control and power-handling equipment. Electric arc furnaces, cranes, welding machines, rolling mills, and other in- duction motors typically have large and/or widely fluctuating needs for reactive power. Inadequate reac- tive power supply can cause production or operational disturbances and reduce the life of manufacturing equipment. Utilities experience higher impedance and reduced capacity on T&D lines that experience reactive loading. Therefore, power providers charge a kVA demand charge to compensate for this cost. To avoid this charge, industrial facilities often try to supply reactive power on site using generators, ca- pacitors, synchronous motors, and other technologies. Duration of discharge, measured in minutes (min), determines the energy requirements for the storage system as measured in kilo- or megawatt-hours (kWh or MWh) or as measured by the hours (hr) of storage: Energy is the amount of power delivered over a period of time. Therefore, the longer the discharge duration at any power, the greater the energy that the storage system must be able to deliver. Energy requirements typically determine the size of the system. Consideration must be given to effects of discharge depth on the service life of the system. Higher energy requirements result in in- creases in the size and cost of the shelter and storage structural requirements. AC system voltage requirements as measured by the root-mean-square of the load’s requirement for kilovolts (kVa:): The AC system voltage determines the size and cost of the transformer used be- tween the PCS and the AC source and load. Voltage requirements also influence the gauge and cost of cabling for the system as lower voltages to reduce transformer costs cause higher currents and increased cable size and cost. Floor space or footprint requirements, as measured in square feet of area, that the ESS occupied (ft?): As implied above, the application requirements influence physical size, and the physical size affects the cost of the entire system. For many applications, and for some utilities, space availability is very sig- nificant in the selection and cost of storage systems. Portability requirements, as measured by relative difficulty or cost of transporting the ESS: Some applications are temporary in nature; therefore, the ability to transfer a storage system from site to site can significantly increase its overall value. Portability varies greatly between types of systems. SMES units, battery storage systems and flywheel ESSs are all now offered commercially as pre-packed, turn- key systems that fit into trailer containers with all of their monitors, controls, and power conversion equip- ment for easy transportation and installation. For large systems requiring significant energy content, however, portable systems may not be feasible, as the size of the storage media often becomes impracti- cal or uneconomical for transport. Number and distribution of duty cycle requirements as measured by number of cycles over a spe- cific time period and the way that the cycles occur (at regular intervals, in clusters at intervals, randomly): The type and frequency of duty cycles affects the service life of storage devices. Cycling requirements influence the size and change-out interval for both the storage media and peripheral com- ponents. The nature of the duty cycle profile, distribution, and frequency also affect the efficiency of all storage systems. While frequent cycling increases the efficiency of some storage media, it decreases efficiency of others. Frequent cycling also introduces transient effects and associated system inefficien- cies that can increase cycle life cost, and can also increase the necessary size of the storage or power electronics, which will increase the cost of the system. 4-4 4, ANALYSIS Application Rapid reserve Area control and frequency responsive reserve" Commodity storage Transmission system stability Transmission voltage regu- lation Transmission facility deferral Distribution facility deferral Customer energy management Renewable energy management Power quality and reliability Table 4-3. Summary of Significant Applications Requirements Special Power Storage aie Floor Space Portability Di Seema of Demands of the (MW) (min) (importance) (importance) Operating (kVac) Duty Cycles Environment 10'-10° 10'-10° 10'-107 Medium Low 10'/year, random, Unremarkable discharge only 10'-10° Charge — 10'-10° Low Low Random, con- Unremarkable discharge tinuous charge/ cycles of <10' discharge cycles clustered in 2- hour blocks daily 10°-10° 10°- 10° 10'-10° Medium Negligible 10*/year, regular, Harmonics are periodic, weekday more important block discharge, _ than in other increased use in generation ap- shoulder months __ plications 10'-10° 10°10" 10'-10° Medium Low 10*/year, random, Unremarkable charge and dis- (MvA) charge cycles 10°-10' = 10'- 107 10'-107 Medium High 10°/year, random Safety concerns (MVAR) charge and dis- are important charge cycles more likely on weekdays, sea- sonal by region — at least 6-7 months 410'-10° = 107 10-107 High High 10°/year, most Safety concerns likely during are important weekday peaks, charge and dis- charge 10'-10° 10° 10°-10' High High 10*/year, most Safety concerns likely during are important weekday peaks, charge and dis- charge 107-10' 10'- 107 10-10" High Varies 10° to 10”year, Safety concerns (MVA) regular periods are important 107-10? 10°- 10° TBD High High 10° to 10*year, Hostile environ- regular periods, ments including (MVA) discharge only, extreme heat unpredictable and cold, par- source ticulates and corrosive at- mospheres 107-10' 10°—10° 107-10" High Varies 10° to 10*/year, Safety’ concerns irregular periods, are important (MVARS) charge and dis- charge *NERC Policy 10, Draft 3 provides details of changes that are under way for Frequency Responsive Reserve. The document is available at http:/ www.nerc.com/~oc/standards/ 4. ANALYSIS 4-5 oF SISATVNV ‘+ Table 4-4. Capability of Technologies to Serve Electric Power Applications Electromechanical Management Renewable Energy Management Power Quality & Reliability D - definite capability, L — likely capability, P- possible capability Applications Electrochemical Batteries Flywheels Electrical Devices —— Flooded- Steel Composite Ultraca- Pb/acid VRLA NalS NIMH za znier Rotor Rotor SMES pacitor Rapid Reserve D D P P Area/Frequency D D P E L Regulation Commodity D & P P P P Storage Transmission D D P Pp P Pp Pp System Stability Transmission D D P P P P | Pp Voltage Regulation T Transmission D L P Pp P Pp P Facility Deferral | Distribution Facility D E Fe P P P P P Deferral Customer Energy D D P P P P P with some charge-control issues. Lithium/polymer is in development and has rudimentary integration (tele- communications only). String limits and PCS integra- tion remain to be established. Participants also agreed that ranking all applications as primary or secondary is important, and that developing groups of primary/ secondary combinations is also important. Value of Storage for Restructured Utility Industry The electric utility industry in the U.S. is being re- structured and is evolving from a regulated monopoly to a partially competitive, partially regulated group of electricity providers. The public’s expectation of plen- tiful, high-quality, and low-cost electricity for all con- sumers has not changed, and if anything, will grow in the coming years. Generation, transmission, distribu- tion, and use of electricity will be performed in a great variety of new and evolving ways through the imple- mentation of creative regulatory frameworks, financial instruments, and technologies. One technology that could have tremendous impact is energy storage. A study initiated last year by Distributed Utility Associ- ates (DUA) described several scenarios likely for the utility industry in the future. Status DUA completed the study, and the final report was received in June 1999. The study describes a series of scenarios in a deregulated utility industry and it dis- cusses the possible benefits from storage in each case. This report outlines a wide range of innovative ways in which storage could be advantageously used in all aspects of the electric supply system of the future, including customer-sited storage. It discusses ways to expand the envelope of possible storage applications and suggests creative uses for storage. It also presents many possibilities for communicating the value and flexibility of storage. The report begins with a summary of the assump- tions made for the study. These assumptions include continuing state-by-state restructuring with federal re- structuring directives lagging behind the states. Gen- eration will be a competitive industry, but T&D will be regulated using performance-based rate-making tech- niques. While overall electricity rates are assumed to decrease, inequities will exist from state to state and between different classes of customers. Technologies for storing energy will continue to improve, but no dramatic breakthroughs will occur. A series of scenarios that consider the use of stor- age in the restructured industry were developed as fol- lows: Core (business as usual with steady evolution) Very inexpensive and efficient storage Environmental emergency Fluctuating electricity price Demanding customer Storage packaging breakthrough Gas and electric industry convergence Energy security Extreme deregulation and competition poor anNhwWn The following themes were determined from an as- sessment of the scenarios. 1. Storage is more likely to be installed at customer sites than coupled to central power plants. 2. An increased interest in environmental issues would accelerate storage technology market entry in many ways; the expanded use of storage is com- pletely consistent with cleaner energy systems. 3. Packaging, ease of use, low initial cost, and high reliability (rather than efficiency and energy den- sity) are the key technology factors in several ma- jor market opportunities. 4. Regulatory structures that allow more freedom to solve problems with innovative approaches would be more likely to lead to increased uses of storage. The implications of each of these themes to elec- tricity providers as well as storage system suppliers and research organizations were also considered. The final report on the DUA study will be pub- lished in FY2000. Utility Operating Cost Analysis This task was initiated during FY94 through the placement of a contract with UMR to use EPRI’s DYNASTORE computer program to perform calcula- tions of utility operating costs with and without BESS. Operating cost savings are one important component of the battery system cost/benefit picture, along with the system capital cost and other projected utility benefits. In this initial study,” UMR calculated generating costs * Assessment of Costs and Benefits of Battery Energy Storage for Electric Utility Applications, Document AL-9306, revised March 6, 1995, prepared by Dr. Max D. Anderson and John T. Alt. 4. ANALYSIS 4-7 for a medium-sized utility system (Utility B) that was not interconnected with other utilities. The results of this work showed that significant production cost sav- ings could be obtained by using a battery system for SR. In FY95, a new contract was placed with UMR for a follow-on study to perform a similar operating cost analysis for a grid-connected utility system.** KCPL, which was selected as the subject for this new study, is a typical Midwestern electric utility with many interconnections and a mix of generating plants. As with the previous study, the approach was to run a unit commitment program on energy storage units along with generating units and calculate operating costs with and without energy storage, so that savings could be quantified. This study was completed at the end of the third quarter of FY96, and the greatest production cost savings were projected for frequency regulation applications in this case. A reanalysis of data from the grid-connected utility system study began in FY97 in order to take into ac- count factors not addressed in the first study, such as battery O&M costs, and to clarify the reasons for cer- tain trends observed in the previous results). UMR studied the impact of fixed and variable BESS O&M costs and found it to be minimal on operating cost sav- ings. A Monte Carlo analysis was used to investigate forced outages. While the Monte Carlo results are viewed as being the most realistic, the differences com- pared to a deterministic method that does not simulate forced outages were again small in most cases. A final report comparing the results for the island and grid- connected utility cases was submitted during FY98.*** Another task that began in FY98 was an investiga- tion of ways to include renewable energy generators in DYNASTORE. Representing the renewable generation as an hourly power purchase transaction proved to be a viable method for including renewables in the DYNASTORE calculations. UMR presented prelimi- nary findings from this task at the end of the fourth quarter of FY98. Work in this area continued under a new contract placed with UMR during the second quarter of FY99. A major part of this study added the capability to in- A Study to Quantify the Benefits of Adding Storage to a Utility System, Document AO-4841, revised June 28, 1996, prepared by Dr. Max D. Anderson and Xiangling Gao. “= Electric Utility Savings from BESS Applications, Document AU-7834, revised August 31, 1998, prepared by Dr. Max D. Ander- son and John D. Stickley. clude wind-powered generators in DYNASTORE cal- culations. The effect of renewable generation on oper- ating costs was also investigated for the KCPL system. Status The method used to represent a renewable energy generator in DYNASTORE and assumptions made re- garding the generation profile for a 200-MW PV source were discussed in the ESS Annual Report for FY98. Initial DYNASTORE calculations of operating cost savings were carried out with both a BESS and the re- newable energy PV source on the KCPL system. This was done for the BESS applications of spinning reserve (SR) only, load leveling (LL) only, and load leveling with spinning reserve (LL + SR) for the years of 1995, 1996, and 1997. Several different BESS powers (40, 100, 200, and 300 MW) and energies (1-hr, 4-hr, and 8- hr duration) were included. These results were com- pared to earlier DYNASTORE calculations of operating cost savings due only to the BESS (no renewable en- ergy source on the system). The ranges of savings for these two cases are shown in Table 4-5. The differ- ences between the two sets of numbers show the oper- ating cost savings afforded by the PV generator alone. In most cases, the savings resulting from PV alone are greater than the savings resulting from the BESS alone. Table 4-5. Operating Cost Savings Range ($K/yr) Estimates for KCPL (DYNASTORE Simulation Method: Monte Carlo, 12 Iterations) BESS Plus Renewable Energy Source Spinning reserve only 7900-—11,900 900-4800 Load leveling only 6300 — 9000 200 — 4000 Load leveling with 7500 — 13,000 700-7600 spinning reserve Application BESS Only Savings in Operating Costs Due to Battery Energy Storage System with and without Renewable Sources To determine savings in operating costs from the BESS in the presence of renewables, the savings af- forded by the renewables must be subtracted from the total savings calculated. The resulting savings are then due solely to the BESS. This subtraction was done for a PV source of 200 MW, for all of the BESS applica- tions, SR only, LL only, and LL+SR, for the years 1995 through 1997. The results were presented in Fig- ures 4-1 through 4-9 for the ESS first quarter report in 4-8 4. ANALYSIS FY99. The BESS application of frequency regulation has also been considered, and is discussed later in this section. A comparison of these results is as follows: Spinning reserve (SR) only-savings resulting from BESS are generally higher with 200 MW of PV oper- ating. Load leveling (LL) only-savings are generally lower because PV is used to shave peaks. Conse- quently, the BESS is used less. LL + SR-savings are generally nearly the same be- cause the savings are greater for SR only and less for LL only. Frequency regulation—PV will increase the need for frequency regulation; hence, BES should be more valu- able for this application. With DYNASTORE, the savings afforded by frequency regulation can be evalu- ated, but the need for frequency regulation cannot. Be- cause PV is dependent upon day/night, cloud cover, and other meteorological conditions, it cannot be considered a reliable energy source for frequency regulation (nor for SR), even though it may be available to pick up the frequency regulation requirement at certain times, thus providing some savings. This lack of PV availability will actually increase the need for frequency regulation, making BESS more valuable (and also more valuable for SR). DYNASTORE calculates savings in frequency regulation mode by calculating generating unit operat- ing costs for generating units on and off of regulation. The two cost numbers are subtracted to obtain the sav- ings. Using this approach, the benefits of PV would be subtracted out, affording no savings. Other approaches for calculating savings for frequency regulation need to be investigated. The calculated results enable the following conclu- sions: 1. Renewable sources (PV in this case) actually feed into the load, causing the BESS to be used less for the LL application. The hourly PV generation profile has a shape similar to the hourly load profile, except that the load profile may lag the PV generation profile by two or more hours. Hence, the PV generation will aid in serving part of the peak load. 2. When renewable generation sources are in use by the system, a smaller-size BESS may be adequate in some cases. The calculations show a smaller increase in savings with larger BESS size for the LL only application if the PV source is present. 3. Use of PV increases the need for frequency regulation as well as SR. This is due to the periods that PV may be unavailable or have reduced power output because of day/night times, cloud cover, or other meteorological variables. In the second quarter, a new contract was placed with UMR to continue modeling electric utility cost savings afforded by BES with renewable energy sources. A major part of this study is to develop the capability to include wind power generators in DYNASTORE. Wind speed data used in this project were obtained from the Solar and Meteorological Surface Observa- tional Network (SAMSON). The data set is available from the National Oceanographic and Atmospheric Administration’s (NOAA) National Climatic Data Center (NCDC). The set is divided geographically into regions: Eastern, Central, and Western United States. It contains hourly solar radiation data along with selected meteorological elements for the period 1961-1990. It encompasses 237 stations in the United States, plus stations in Guam and Puerto Rico. The data set in- cludes both observational and modeled data. The hourly solar elements are extraterrestrial horizontal and extraterrestrial direct normal radiation, and global, dif- fuse, and direct normal radiation. Certain meteorologi- cal elements are also recorded. These are total and opaque sky cover, temperature and dew point, relative humidity, pressure, wind direction (true north) and speed, visibility, ceiling height, present weather, pre- cipitable water, aerosol optical depth, snow depth, days since last snowfall, and hourly precipitation. The data were originally developed by a joint effort between NCDC and the National Renewable Energy Laboratory (NREL). The wind energy resource atlas of the United States shows the areas that are potentially suitable for wind energy applications are dispersed throughout much of the United States Estimates of the wind resource in this atlas are expressed in wind power classes ranging from Class 1 to Class 7, with each class representing a range of mean wind power density or equivalent mean speed at specified heights above the ground. Areas desig- nated Class 4 or greater are suitable with advanced wind turbine technology under development today. Power Class 3 areas may be suitable for future genera- tion technology (year 2000 and beyond). Class 2 areas 4. ANALYSIS 49 are marginal and Class | areas unsuitable for wind en- ergy development. The South Central region, consist- ing of Arkansas, Kansas, Louisiana, Missouri, Oklahoma, and Texas, has Class 3 or higher annual average wind power. A map of the of annual average wind power, along with wind power classifications, is presented in Figure 4-1. The average wind power in Missouri is shown in Figure 4-2. During the third quarter, a model for estimating the electrical power available from wind in the KCPL area was constructed. The three steps involved in calculat- ing the power from the wind are shown below. 1. Calculate wind speeds for 12 typical weeks from the 30 years of data available from the NCDC. 2. Translate wind speeds from recorder height to turbine hub height. 3. Calculate electrical power from wind speed at hub height. Step 1: Evaluation of wind speed for 12 typical weeks from 30-year data. Hourly wind speeds are available for 30 years from 1961 through 1990. The following procedure was used to get data from 12 typi- cal weeks, one week for each month of the year. Con- sidering one month at a time, average wind speeds were found for each hour of each day of the month. In this way, 30 months of data (the same month each year for 30 years) is replaced by a single month with average hourly wind speeds. Then, by similar averaging, a typi- cal week is found for the month. This is repeated for all the months. At the end of this procedure, we have 12 typical weeks of hourly wind speed data, one for each month of the year. Figure 4-3 shows the average wind speeds at ground level for a representative day in Janu- ary. It should be noted that Figure 4-3 is not intended to represent actual wind speeds in any given day be- cause it is derived from averaging. These average val- ues are useful for predicting mean power levels generated by the wind turbines, but in the final analysis, the variation about the mean must also be included. This will be done in follow-on work. Step 2: Calculation of wind speeds at hub height. The wind speed is measured at the height of the ane- mometer — typically ground level. Because the wind turbine blades will be at a different height from the anemometer height, the recorded wind speeds must be translated to the hub height before power can be calcu- lated. Zond Energy Systems, Inc., is one of a handful of wind turbine manufacturers in the United States that produces large turbines. Their 750-kW Z-5O turbine, which has been successfully used in the United States and abroad, was used for wind power calculations. According to machine specifications, the hub height is 65 m and the blade diameter is 50 m, yielding a total swept area of 1963 m’. The wind speed at a 65-m hub height can be calculated using the one-seventh power law. The power law (Equation 1) states that the wind speed at a height of z2 will be a 22 wa) =() * n(zl), (1) zi where: zl = the height at which wind speed was meas- ured z2 = the height at which wind speed is to be calculated u(zl) = wind speed at height z1 u(z2) = wind speed at height z2 In Equation 2, typical values of constants a and b were used to calculate a a=a-b * log (u(z1)), (2) Typical values of 0.11 and 0.061 in the daytime and 0.38 and 0.209 at night. For the months of March through September daytime is 6 a.m. to 8 p.m., and for the remaining months daytime is 7 a.m. to 5 p.m. The wind speed measurement height is 10 m. Wind speeds at the hub height were calculated for the 12 typical weeks. Step 3: Calculation of wind power. Based on the above wind speeds, wind power was calculated using the formula given by Equation 3. Using this equation, wind-generated powers are calculated for all 12 typical weeks that were required by the DYNASTORE pro- gram. Figure 4-4 shows the average power generated by the Zond Z-50 WTG for a representative day in January. P = 0.5pA CpV3NeNy (3) where: P = power in watts, p= air density (about 1.225 Kg/m’ at sea level; this value is taken even though Kansas City is 315 m above sea level) A = rotor swept area = 1963 m’ from specifi- cations C, = coefficient of performance = 0.35 Vs = wind speed in m/sec N, = generator efficiency = 0.8 N, = gearbox/bearings efficiency = 0.9 4-10 4. ANALYSIS SISATVNV & [Il-¢ UNITED STATES ANNUAL AVERAGE WIND POWER ALASKA a Meow PRINCIPAL - HAWAIIAN, ’ ISLANDS o” Lae aeons Vv. Se ae Pe aei CD woee coast esewars coca arves 190017 Figure 4-1. A Map of Annual Average Wind Power, along with Wind Power Classifications. 92° * Columbia > | , ‘ ne “St. Louis © tye ‘ ee Kilometers Figure 4-2. Average Annual Wind Power in Missouri. 8 7 6 5 Ya ZT 8 es a 4 | Zs £ Wind speed at | = 3 ee | | 2 1 | ! | | 0 0 4 8 12 16 20 24 Hour of typical day Figure 4-3. Average Wind Speeds at Anemometer and Hub Height for a Representative Day in January. 4-12 4. ANALYSIS 140 4 porn 120 100 80 60 Power (kW) 40 20 0 4 8 12 16 20 24 Hour of typical day Figure 4-4. Average Power Generated by the Z-50 Wind Turbine Generator (WTG) on a Representative Day in January. Averaging Method Was Used. The above calculations enabled production of input data for the DYNASTORE program. With these data in hand, UMR completed DYNASTORE calculations. In order to make realistic calculations of the operating cost data with wind power included in the generation mix, UMR assumed a 400-MW capacity from the wind farm. This is expected to result in 200 MW of actual wind power some of the time, which should be comparable with the PV generation. In the PV case, a peak genera- tion capacity of 200 MW _ was used. The DYNASTORE runs were being made for the BESS cases of 40-, 100-, 200-, 300-MW capacity and 1-, 4-, and 8-hr duration. The modes of BESS operation were SR only, LL only (PS), and LL + SR. The years for the study were 1995, 1996, and 1997. This is the same format used in previous studies, which will enable comparisons to be made to previous results. Savings in operating costs for each BESS application for each year are shown in Figures 4-5 through 4-13. A complete listing of the results is given in Appendix B. The results to this point have shown savings in op- erating costs when energy storage is present on a utility system. The estimated values vary by factors of two to three, depending on the particular application for the energy storage and the utility configuration with regard to interconnections. Some typical values are shown in Table 4-6 for a grid-connected electric utility. SR is generally the application showing the most savings due to BESS, while PS/LL shows the least savings. Table 4-6. Operating Cost Savings Range ($K/yr) Estimates for KCPL (DYNASTORE Simulation Method: Monte Carlo, 12 Iterations) BESS Plus Application 400-MW Wind BESS Only Power Source Spinning reserve 23,400 — 900 — 4700 only 30,300 Load leveling only 21,800 — 200 — 4000 26,000 Load leveling with 23,000 — 700 — 7600 spinning reserve 31,000 Analysis of the situation with renewables is com- plicated and is still not complete, but an indication is emerging that the operating cost benefits of storage in this case may be much larger than previously estimated. The reason for this is that renewable energy sources are not continuously available (e.g., PV at night) and there- fore cannot be allocated to certain applications, such as SR, that require 100% availability. However, the use of energy storage will broaden the range of applications to which renewables can be applied by providing the ca- pability to ride through periods of low renewable gen- erator output. Therefore, the storage not only directly 4. ANALYSIS 4-13 Savings (Thousands of $) a 8 3S T T T T T T 1 0 50 100 150 200 250 300 350 Battery size (MW) — 1 hour battery —#—4 hour battery —#— 8 hour battery | | Figure 4-5. | Operating Cost Savings with a BESS Used for Spinning Reserve Only by KCPL (1995) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). NO 88 Q 8 o Savings (Thousands of $) ora 838 2000 1000 0 T T T T t r 1 Oo 50 100 150 200 250 300 350 Battery size (MW) —*— 1 hour battery —#—4 hour battery —4&— 8 hour battery Figure 4-6. | Operating Cost Savings with a BESS Used for Spinning Reserve Only by KCPL (1996) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). 4-14 4. ANALYSIS 8000 5 | 7000 + | 6000 - 5000 4000 3000 T T T T T T 1 50 100 150 200 250 300 350 Battery size (MW) Savings (Thousands of $) Oo —@— 1 hour battery —#- 4 hour battery —#— 8 hour battery | | | Figure 4-7. Operating Cost Savings with a BESS Used for Spinning Reserve Only by KCPL (1997) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). 8000 “ Savings (Thousands of $) as T T T 7 | Battery size (MW) | —@— 1 hour battery —#- 4 hour battery —#— 8 hour battery | | | | | 0 50 100 150 200 250 300 350 | Figure 4-8. | Operating Cost Savings with a BESS Used for Load Leveling Only by KCPL (1995) after Sub- traction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). 4. ANALYSIS 4-15 Figure 4-9. 8000 + 7000 6000 5000 4000 3000 000 = 1h 0 50 Savings (Thousands of $) °° _—_—_——==EES 0 T 7 T T T 7 100 150 200 250 300 350 Battery size (MW) —¢— 1 hour battery —#-— 4 hour battery —t#— 8 hour battery Operating Cost Savings with a BESS Used for Load Leveling Only by KCPL (1996) after Sub- traction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). 8000 7000 6000 5000 4000 3000 Savings (Thousands of $) 2000 +——— 1000 : ——— 100 150 200 250 300 350 Battery size (MW) —— 1 hour battery —#- 4 hour battery —&— 8 hour battery Figure 4-10. Operating Cost Savings with a BESS Used for Load Leveling Only by KCPL (1997) after Sub- traction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). 4-16 4. ANALYSIS 8000 7000 6000 5000 4000 3000 2000 1000 0+ T 0 50 Savings (Thousands of $) T T T 100 150 200 Battery size (MW) =e T 250 300 350 —¢— 1 hour battery —#®- 4 hour battery —&— 8 hour battery Figure 4-11. Operating Cost Savings with a BESS Used for Load Leveling with Spinning Reserve by KCPL (1995) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). T 0 50 Savings (Thousands of $) > o S 3S T T T 100 150 200 Battery size (MW) 250 300 7 350 —*—1 hour battery —#- 4 hour battery —#— 8 hour battery Figure 4-12. Operating Cost Savings with a BESS Used for Load Leveling with Spinning Reserve by KCPL (1996) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). 4, ANALYSIS 4-17 8000 + 6000 4000 2000 0 T 1 Savings (Thousands of $) 0 50 100 150 200 250 300 350 Battery size (MW) T T T T 1 —@— 1 hour battery —#— 4 hour battery —#— 8 hour battery Figure 4-13. Operating Cost Savings with a BESS Used for Load Leveling with Spinning Reserve by KCPL (1997) after Subtraction of the Savings Afforded by a 200-MW Wind Generator (Simulation Method: Monte Carlo, 12 Iterations). replaces more expensive sources of generation, such as combustion turbines, it also enables renewable genera- tion to serve many of these applications. Since renew- ables generally have very low operating costs, this significantly increases operating cost savings. As shown in Table 4-6, these savings can be quite large, four to six times greater than those from adding BES alone to a utility system. The low end of the savings ranges shown in Table 4-6 reflects a BES power of 40 MW and a 1-hr duration while the high end corresponds to 300 MW of BES power with an 8-hr duration. For reference, this utility has a summer peak load of about 3,000 MW and the total generating capacity is around 5,000 MW. Annual operating costs were estimated by DYNASTORE to be $115M per year without any wind generation or BES; therefore, a savings of $30M per year represents a 25% reduction in operating cost. Although the calculation of cosvbenefit ratios is not the focus of this study, it is worth pointing out that the cost of 40 MW of 1-hr BES could be recovered within roughly two years at this savings rate. The savings figures shown assume that energy storage allows the renewable wind generation to be fully allocated to the indicated applications, and that the wind generation is always “on” for an amount calcu- lated by statistical averaging. However, we know that the output of wind and solar energy generators can vary considerably over a period of time. It is intuitive that a BESS is necessary to “take up the slack” in power for any wind or solar facility in order to make it a suitable energy source for certain applications. In fact, the vari- able nature of the renewable generator output is ex- pected to make energy storage even more valuable when this type of generation is present. However, the extent of use of BESS and the savings in the case of variable renewable generator output have not yet been quantified. Including the full effect of variability in renewable generator output will be a major focus of the next phases of this study. The strategy for use and level of acceptance of renewable generation sources by sys- tem operators is also unclear due to the relatively low penetration of renewables in the utility market. International Energy Agency Annex IX The DOE Office of Power Technologies is the United States participant for the International Energy Agency (IEA) Implementing Agreement for Research and Development on Energy Conservation through En- ergy Storage. In June 1996, the OPT made a commit- ment to participate in the recently created electrical Energy Storage Annex IX by pledging funding in sup- port of its activities. The primary benefits of United States participation are increased awareness of analyti- cal and technical developments in storage in the inter- national arena, identification of projects of mutual interest, and the ability to assess the competitive posi- tion and market opportunities for energy storage sys- tems in the overseas markets. ILZRO is also interested in these objectives and has contributed funding for par- ticipation by the United States in Annex IX. 4-18 4. ANALYSIS OPT and ILZRO designated the ESS Program at SNL to be the USA Participating Agent in the activities of Annex IX. This responsibility requires attending executive committee and experts meetings in the United States and abroad, coordinating United States repre- sentation by experts for the various storage technolo- gies, and identifying projects of common interest to the participants and supporting the implementation of the projects. Status Phase 2 of Annex IX Electrical Energy Storage Technologies for Utility Network Optimization work was concluded at the end of FY99. A final report, which was an overview of this work, was submitted in November 1999. Topical reports on Subtasks 1, 2, and 3 were submitted in the third and fourth quarters of FY99. e Subtask 1: Applications Case Studies Subtask 2: Project Definitions Subtask 3: Applications Modeling A summary of these reports follows: Subtask 1: Applications Case Studies—The Case Studies report was received in the third quarter of FY99 and represents the final output of Subtask 1. The report focuses on storage systems as solutions for specific problems. It is aimed at users who may be relatively unfamiliar with storage. Preexisting electri- cal energy storage installations are described from the perspective of the end user as distinct from the particu- lar storage technology employed. Four different broad application categories are con- sidered. e Power quality/quality of supply, e Distribution capacity deferrals/peak demand-side management (DSM), shaving/ e Integration of renewables, and e Frequency regulation/spinning reserve/stability Where possible, attention is drawn to factors deemed important in project management of the instal- lations, with a view toward learning lessons for future storage installations. The decision-making processes for a potential user of electricity storage are discussed under the following headings: e System cost, e System benefits, ¢ System performance, and e Time scales of installation and of expected system lifetime. Power Quality/Quality of Supply The following is a background and performance summary of two of the systems reviewed as part of the case studies addressing the power quality application. In 1996, Oglethorpe Power Corporation installed a 1-MW, 10-second battery storage system at a lithogra- phy plant in Homerville, Georgia. The system was de- signed to remedy voltage sags and short power outages with a 4-ms response time. The system uses flooded lead-acid batteries and features a transportable con- tainer (4.5 m long x 2.3 m wide x 3.7 m high; weight = 18.1 tons). The initial cost of this system was $1.06M, and it has an estimated discounted life cycle cost over 20 years of $1.65M. Costs of similar systems are said to have diminished since this installation. Construction began in May 1996, the containerized equipment arrived in July 1996, and operation began in December 1996. From December to July 1996 more than 50 supply disturbances were successfully dealt with. The South African National Electricity Company Eskom installed a 2.4-MJ low-temperature SMES sys- tem at a paper mill in Stranger, South Africa. The ob- jective was to remedy voltage dips due to flashovers induced by lightning, sea mists, and fires. Steady volt- age is essential for accurate motor speed control in pa- per mills. The South African installation is a relatively small “micro-SMES PQ-VR” model. High-temperature su- perconducting connectors are used in the newest in- stallations to minimize losses in the resistive parts of the system. The Stranger paper mill installation was commis- sioned in April 1997. The system had achieved 100% reliability up to February 1998, even through 68 voltage dips, about half of which would have been serious enough to stop production at the mill, were it not for the storage system. 4, ANALYSIS 4-19 Distribution Capacity Deferrals/Peak Shaving/ Demand-Side Management The following is a summary of one of the systems reviewed as part of the case studies addressing distribu- tion capacity deferrals, peak shaving and DSM. GNB Technologies is a major U.S. battery manu- facturer and operates a lead smelting and recycling center in Vernon, California. Large electric fans are crucial to the plant’s environmental protection meas- ures, and are vulnerable to power failures. The com- pany has a strong commercial interest in the development of battery storage systems and therefore installed a system to demonstrate the viability of the technology. The plant’s maximum load is 5 MVA, critical load 2.1 MVA, and typical load 3.5 MVA. The BESS peak rating is 5S MW/5 MWh and runs the whole plant for 10 seconds during non-critical load shedding and then the critical load for up to two hours. In the event of long power outages, an orderly shutdown is possible, with the critical loads continuing to be supplied by the battery. As well as its principal backup function, the BESS has been used for peak shaving to generate electricity purchase savings. Costs of the BESS were considered against a con- tinuously operating motor-generator flywheel system, coupled with a diesel engine. The former provides only limited time of operation, but has low parasitic costs, compared with the 200 kVA, $0.1M per year of the motor generator. Overall the discounted life cycle costs were deemed similar, but these estimates depend upon battery life. The BESS cost $4.2M, against $4.5M for the alternative. The system is not considered cost-effective solely on the basis of peak shaving and demand reduction; the overriding rationale is maintenance of environmental safety through power outages. The project was announced in November 1994, construction began in January 1995, and was completed in August 1995. The plant was fully commissioned in January 1996. Startup of the BESS takes less than one second. The system performs routine automatic peak lopping at a preset power demand. Fifty percent capacity is re- tained for emergency backup. Overall benefits include: Environmental safety assurance Savings in peak demand charges for utility e Peak tariff unit charges avoided Integration of Renewables The following is a background and performance summary of two systems reviewed as part of the case studies addressing the integration of renewables appli- cation. A lead-acid battery storage system consisting of three modules totaling 1.2 MW and 1.2 MWh for peak shaving and power quality improvement, was installed at Herne-Sodingen, Germany, for the local utility Stadtweke-Herne AG. The storage system at Herne operates in parallel to a PV plant, which became opera- tional in the summer of 1999. The battery itself has operated in a peak-load shaving mode since December 1997. It was deemed more cost effective to install the storage system than to build utility network extensions. The principal applications envisioned for the scheme were integration of the renewable generation resource, load leveling, and power quality improvements, as well as the deferral of new capacity. Also, it was intended to assist in the development of process automation for storage systems. Planning, preparation for construction, and con- struction took two, eight, and eleven months respec- tively. Construction ran from January to November 1997. The battery has operated in peak shaving mode since December 1997, and was deemed fully commis- sioned in February 1998. A PV plant was operational and integrated in the summer of 1999. The customer paid about half the cost, with government and interna- tional agency grants covering the rest. The storage me- dium accounted for 40% of the capital costs, and the interface plus the power conversion system for 20%. Frequency Regulation & Spinning Reserve The following is a background and performance summary of a system reviewed as part of the case stud- ies addressing frequency regulation, spinning reserve, and stability applications: Dinorwig is a pumped hydro facility on the edge of the Snowdonia National Park in North Wales and is the largest storage facility in Europe, constructed as a ma- jor national investment when UK generating capacity 4-20 4. ANALYSIS and the national grid were under monolithic state con- trol. The system has ratings of: e 1800 MW max (6 x 300 MW motor/generator) e =1380 MW x Sh = 6900 MWh Eighty percent full generating power can be achieved in 10 seconds if the generators are kept spin- ning in air in synchronization with the grid, taking 5 MW per machine. Electrical storage turnaround effi- ciency is about 67%. It is used principally for frequency control and emergency backup for major plant failures on the UK National Grid. There are smaller 360-MW pumped storage schemes nearby at Ffestiniog, and another in Scotland at Cruachan. Costing $425M in 1975, construction of the system was the largest civil engineering contract to date in the UK. Construction began in 1974 and took 10 years to complete. This project now seems out of scale with the diver- sity of the privatized UK electricity supply industry, and has to compete with many suppliers. This illustrates the potential investment pitfalls when a long-term view is taken in modern energy markets. It is unlikely that the planning of such a project in the future could overcome the environmental objec- tions. Case Studies Conclusions The major conclusions from the case studies work are as follows: e Planning takes a long time (often years), but to avoid pitfalls, careful consideration is worthwhile. e A “turnkey” approach is always desired by the end user, but a constructive partnership is essential among planners, suppliers, consulting engineers, and end users. e Construction often takes approximately a year, except in the case of trailer-mounted turnkey installations. e Careful monitoring and documentation of initial testing, subsequent performance, and problems are highly desirable because of the immaturity of the technology. Modularity and standardization of system assemblies will reduce costs, improve reliability, and diminish the demands on the system designers and engineers. It will often be difficult to make a sound cost- benefit case relative to fossil-fuelled generation options; the storage field is still in an exploratory phase and will mature in the future. Subtask 2: Project Definitions—The report was completed in the third quarter of FY99 and represents the final output of Subtask 2. The Project Definition Report focuses on two energy storage applications: power quality (Table 4-7) and primary substation appli- cations (Table 4-8). The report presents a cost break- down for the various electrical energy storage system solutions currently available. Generic information is provided without focusing on any single suppliers’ so- lution or any particular site. Table 4-7. Power Quality Installation Application Power rating 2 MVA, 2 seconds (typical) Energy storage 1 to 2 kWhor 3 to 6 Mi (typical) capacity Storage Batteries (including lead-acid and technology advanced battery technologies) SMES Flywheels Capacitors Supercapacitors = — Hybrid (i.e., combination of above) Table 4-8. Utility Substation Application Specifications Power rating 2 to 4 MVA, 15 mins to 2 hours (typical) Energy storage 1 to 4 MWH (typical) capacity Storage Batteries (including lead-acid and technology advanced battery technologies) Hybrid system (i.e., incorporating a flywheel, SMES, capacitors, super- capacitors, etc.) 4. ANALYSIS 4-21 Key system features include: Modularity of system configuration, Transportability (i.e., ease of relocation), Re-configurable functionality, and Remote on-line monitoring. Power Quality Project Definitions Reasons for the choice of the power quality appli- cation follow: Power quality and quality of supply are of growing importance as electronic devices form an increasingly large proportion of overall electrical load. Electronic motor drives are becoming widespread in manufactur- ing industry. They, and other power electronic devices and computer-related equipment can both be suscepti- ble to and themselves cause electricity supply distur- bance. The full range of practical electrical energy storage systems can be applied to many power quality applications, so this application provides an interesting project definition, with much scope for innovative de- velopment and demonstration. Based on work carried out in Phase 1 of Annex IX and subsequently, it was judged that an appropriate installation for the costing study would be sized at: e 2-MW rating for a duration of up to 2 seconds of continuous supply, i.e., 1 to 2 kWh or 3 to 6 MJ. A “quality of supply” application would ideally have a ride-through of the 10 seconds needed for elec- tromechanical protection systems such as auto-reclosers to operate after fault clearance, when possible. How- ever, shorter durations are common in current storage installations, on grounds of storage cost, but clearly offer a lesser degree of protection. It was decided to cost a “power quality” application rather than one on “quality of supply” as defined above. The latter appli- cation would always have to have storage capacity and switchgear capable of supporting the whole load, whereas a power quality application would generally need to support only a part of the total load. This application is highly relevant for network op- erators, as it impinges upon another topic of increasing importance, i.e., asset management. In increasingly competitive and liberalized modern energy markets, effective use of assets is clearly of enormous economic significance. Substation Project Definition Storage can play a role in avoidance of network reinforcement, alongside alternatives such as localized generation. Unlike the latter, however, storage systems can perform a secondary power quality function within the same installation. Local embedded generation is, in many cases, more likely to create rather to than solve short-term power quality problems. It was judged that a suitable project of this kind would be a primary substation application with a ration of: e 2 to 4 MW for a time of up to 30 minutes of continuous supply, i.e., 500 to 2000 kWh or 1800 to 7200 MJ. Cost Estimates The estimated costs for a power quality applica- tion, as defined previously, are shown in Table 4-9. The cost of supercapacitors is not included, as no cost information is available, but the technology is included in the study because of its excellent future potential. A breakdown of the initial costs involved with the purchase, installation, and commissioning of energy storage devices for a substation application are shown for the four storage technologies in Table 4-10. As with the power quality application, the equipment to load interface cost will vary depending upon the rating of the storage device. The type of storage medium used will not normally affect the cost of the connection and is therefore valid for most applications. The connection of any electrical equipment to a network must obvi- ously meet strict safety requirements that will make certain fixed costs unavoidable irrespective of capacity. The expected cost for such an application with a 2-MW rating is in the range $40K to $60K. This cost may diminish in the future to the same degree as the power quality application, due to modularization. These figures are estimates derived from total project costs from the literature and known costs for the appropriate elements of similar applications. In the case of the advanced and conventional bat- tery storage systems, a solid state switch is fitted to the network interface to allow the rapid isolation of the storage system in the event of a network fault. The purpose of the switch is to avoid the storage system “feeding back” into the faulted network. The slow re- action time of the compressed air energy storage system is such that a conventional mechanical isolation switch can be used. 4-22 4. ANALYSIS Table 4-9. Initial Cost Breakdown for a 2-MW, 2-sec Power Quality Application Battery Flywheel Superconducting Supercapacitor (BESS) ($K) (FES) (SK) Magnet (SMES) ($K) (S-CS) ($K) Storage medium High 100 600 800 Low 80 400 600 Interface High 60 60 60 60 Low 40 40 40 40 PCS High 825 825 1000 825 Low 360 360 800 360 Other High 1000 1200 1800 1000 Low 500 700 1000 500 Total High 1985 2685 3660 Low 980 1500 2440 = Table 4-10. Initial Cost Breakdown for a 2-MW, 2-MWh Substation Application Battery (BESS) Advanced Battery, e.g., | Compressed Air ($1K) Redox Flow Cells ($1K) (CAES) ($1K) Storage medium High 600 4065 700 Low 400 1100 500 Interface High 160 160 60* Low 140 140 40* PCS High 825 825 825 Low 360 360 360 Other High 1500 1500 1500 Low 1000 1000 1000 Total High 3085 6550 3085 Low 1900 2600 1900 * No fast-acting switch for protection required because of reduced system reaction time. 4. ANALYSIS 4-23 The PCS is typically 15 to 20% of the overall cost of a substation application. Cost will be in the range $360K to $825K for a typical system. Again, these estimates are derived from literature values for total and partial costs for similar installations. Future cost reductions are likely as power elec- tronic systems advance. However, the scope for stan- dardization will be less for this application than for the power quality case, since there are likely to be fewer and more varied substation installations. The cost of the energy storage medium for this ap- plication often represents the largest single element in the total cost breakdown, at around 20 to 30% for bat- tery installations. There are some cost categories that do not fall into the initial cost categories listed above. These costs con- sist of taxes incurred, site considerations, transporta- tion, financing, and other services that add significantly to the initial costs of each project. As in the power quality application, the choice of technology for this substation application is best viewed in terms of cost versus tangible benefits. The battery energy storage solution has all the advantages and dis- advantages as pointed out above for the power quality application, but with an even greater need for a rigid and stable base with the increased battery capacity re- quired for this application. The increased maintenance and battery replacement needed for a larger scale sys- tem might also make a battery solution less attractive in the long term. Advanced battery technologies at this stage entail some technical risk. These systems offer enhanced storage densities and therefore smaller footprint than the better established lead-acid technology. In most other respects they are generally equivalent. A key difference is that redox flow type batteries have an added feature of being able to decouple their power and energy rating, since the parameters are determined by different elements within the battery system. Thus, the power rating is determined by the cell stack design, while the energy storage rating is determined by the volume of electrolyte stored in separate tanks. Redox flow cells represent a promising development in the field of advanced battery research. Each of the candidate advanced battery systems has its own potential environmental difficulties (as have conventional, established battery systems). The par- ticular chemical used will have a strong bearing on the cost of necessary safety precautions, and these have not been explored fully for many of the emergent technolo- gies. Compressed air energy storage (CAES) systems use proven industrial technologies. The low cost per kilowatt claimed by one developer in their prequalifi- cation document indicated this may be a relatively cheap storage method. Development of flywheel technology has been fo- cused on smaller capacity devices, e.g, 5 kW/ SkWh. Larger capacity devices are under develop- ment, e.g., 100 KW/S kWh, making them unlikely con- tenders for this application in the short term. One flywheel company launched their “new pirouette” fly- wheel in May 1999, which will be rated at 100 kW/3.2 kWh and will be capable of being cascaded to more useable levels of storage. Each of the storage technologies has an associated tisk; for example, the potential for lead-acid battery problems resides primarily in acid spillage, and secon- darily in lead release in the event of fire. These risks are in addition to the risk of sudden release of the stored energy, common to all storage systems. Storage of very large volumes of either toxic or relatively innocuous chemicals for a redox flow system would be subject to environmental and safety regulation. The project definitions work provides indicative costings for two applications of electrical energy stor- age. The intention is that these be used to generate pro- posals for future demonstration _ installations, incorporating innovations and advances in the technol- ogy. Subtask 3: Applications Modeling—Work on this subtask was nearing completion in the fourth quarter of FY99. A final report was received in October 1999 and represents the final output of Subtask 3. The purpose of the power quality application model is to help the user to design an energy storage system to meet the needs and requirements placed on it. The model uses a benefit/cost calculation to give the optimum system scaling for a given set of parameters. The model consists of a stand-alone executable written in Borland Delphi (32 bit) for use with MS Windows 95/NT/98 or above. It is intended that the model be used by system de- signers to help make the decision whether energy stor- age is viable, which storage technology is best suited, and to define the optimal system scaling in terms of power and energy requirements for a particular sce- nario. 4-24 4. ANALYSIS The user must input the voltage level (and number of phases) and peak load current for the application(s) to be “protected.” The user must input approximate figures for the cost of each of the levels and duration of voltage dip; this is possibly the most difficult and least accurate approximation. However, this thought process is part of the most basic economic analysis of the energy storage system application. The user should consider the following cost categories when attempting to quantify the cost of each voltage dip: Cost of lost production Cleanup cost Equipment damage Cost of waste/scrapped materials Reduced equipment life span (years) Reduced product quality oeoeoeee The model uses the above information to calculate the benefit matrix. Figure 4-14 shows the modeled benefit of mitigation at different voltage magnitudes and durations. The user then inputs cost information for an energy storage system in terms of $/kW and $/kWh. The voltage connection level and rms current are used to calculate the power and energy requirements over a range of dip durations. From this information, a cost matrix is calculated. The user can employ a variety of system costs for comparison using different storage media, and in this way cost comparisons of different storage technologies can be made. Figure 4-15 shows a three-dimensional plot of the modeled cost of a typical energy storage system. It is clear that the cost rises rapidly with both increasing dip duration and reducing voltage magnitude. Dividing the benefit matrix by the cost of the ma- trix gives the benefit/cost matrix. The optimal benefit over cost ratio will be seen as the maximum value in this matrix and will identify the best system scaling for the user’s specification. Figure 4-16 shows a three-dimensional plot of a cumulative benefit/cost. The peak value, at 75 to 80% retained voltage magnitude and a duration between 10 and 20 mains cycles, indicates that this is the most eco- nomic range of events for which to design a system. From the optimized dip duration, shown in Figure 4-16, the voltage level and the peak current, the re- quired power and energy levels for the system can now be calculated. Network Simulation Integrated Model—As there are a variety of software packages available to the network designer to simulate network performance, each with its own merits, a specification for this model has been produced that can be applied to any of the major packages that are commercially available. The specification was written in accordance with the guide- lines laid down by the IEEE Standard 830-1993-IEEE Recommended Practice for Software Requirements Specifications. The scope of the specification for the network ap- plications modeling of energy storage systems is clearly defined as covering the following areas: e A model definition suitable for use with any of the commonly used network simulation packages. e The model will cover the applications of distribu- tion capacity deferrals, peak shaving, frequency regulation, spinning reserve, and stability. In carrying out this work a number of software manufacturers were contacted and questioned about the possibilities of integrating energy storage system com- ponents into their packages. The overall response from these companies was that when they experience a sig- nificant demand, they would then act to integrate stor- age into their packages. In addition to the above requirements for model- ing/load flow analysis, this integrated energy storage model must incorporate economic factors and be able to simulate the interaction of the storage device with the network over a time frame long enough to realize the benefits the system was designed to fulfill. Transmission Power Quality Study The Transmission Power Quality (TPQ) study will monitor and define power quality levels of the electric transmission system for a defined 13 state region in the Southeastern United States. A previous study con- ducted by the Electric Power Research Institute (EPRI) characterized power quality levels of the distribution system. However, very little data now exist for cus- tomers of a transmission network. Many large indus- trial facilities get their power directly by the transmission network without a distribution system. Data captured from this study will assist the ESS Pro- gram in the design and scope of the Substation Power Quality project, which is intended to mitigate power quality disturbances from the transmission system. This project will provide utilities with baseline data for system performance, and experience in a large-scale 4. ANALYSIS 4-25 70-75 Retained Voltage Magnitude r 75-80 (%) 28 s 2 1000< Dip Duration Range (cycies) Figure 4-14. Plot Showing the Modeled Benefit of Mitigation at Different Levels. Cost($) 30000600 25000000 20000060 15000006 10000600. <60 50000090 65-70 A * Retained 75-80 Ce Voltage 2 = EY 85-90 <3 3-5 5-10 10-20 20-50 50-100 100- 1000< 1000 Dip Duration Range (cycles) Magnitude (%) Figure 4-15. The Modeled Cost of an Energy Storage System. 4-26 4. ANALYSIS Benefit/Cost 8 Dip Duration Range (cycles) Figure 4-16. Cumulative Benefit/Cost. monitoring project. Also, the resulting data may be used to validate system models for studies on the be- havior of the transmission system. This study will also provide valuable data for the DOE’s Electric Reliability Initiative. Status A study team composed of representatives from utilities, SNL, EPRI, and the Southeastern Electric Ex- change (SEE) conducted a scoping study that outlined the project justification, methodology, cost information, and schedule. Monitoring sites will be chosen through- out the Southeast to provide a statistically valid sample of the transmission system. Criteria for choosing sites are voltage class and lightning stroke density. Partici- pating utilities will procure, install, download, and maintain appropriate monitoring equipment. Data from the participating utilities will be transferred to EPRI, which will serve as a central collection point. EPRI will compile the transferred data into a composite data- base, removing all utility-specific information. The database will be transferred to SNL for further analysis and characterization. SEE will serve as a facilitator for the project, recruiting participants and hosting project review meetings and workshops. The DOE will pro- vide funding for the project, in conjunction with cost- sharing from industry. DOE will also coordinate the 70-75 Retained Voltage Magnitude (%) ESS program with other SNL programs that may bene- fit from this study. The Southeastern United States was chosen for the monitoring region for TPQ. Thirteen states have been identified, as shown in Figure 4-17, based on geo- graphic interpretations of a “Southeastern” boundary, and the location of participating utilities in the SEE. This organization is comprised of member utilities in the Southeastern U.S. for the purposes of sharing in- formation on operational practices such as safety and benefits all members. The SEE also serves as an ex- cellent central point for disseminating information to prospective participating utilities. Statistical analysis concludes that a minimum of 300 sites must be selected to provide a statistically valid sample. The duration of the monitoring period will be a minimum of 24 months. A minimum of 11 utilities will be required to participate, with five specific utilities identified as necessary participants based on the num- ber of potential monitoring sites and size of the service territory. Total project cost is estimated at $7.3M over the four-year life of the project. Cost-sharing opportu- nities exist with industry participants, including the use of existing monitoring equipment and installations, utility and SEE labor, and development of data analysis and management software. 4. ANALYSIS 4-27 Figure 4-17. The 13-State Region Under Study Defined as the Southeastern U.S. Table 4-11 lists the utilities in these 13 states that own transmission assets. Most of these utilities are SEE members; however non-SEE members will be in- vited to participate in the study. Likewise, EPRI mem- bership (or lack thereof) has no bearing on whether or not a particular utility may participate in the study. Technology Assessments Performance and Economic Analysis of Superconducting Magnetic Energy Storage, Flywheels, and Compressed Air Energy Storage Systems Project The scope of the ESS Program includes a portfolio of energy storage technologies for electric utility appli- cations. The program approach has been to apply ex- pertise gained from work with BES to the development of storage media, PCSs, peripheral devices, and ad- vanced storage systems that depend on similar compo- nents. The ESS Program initiated this analysis project with Energetics, Inc., in FY97 to identify the areas in which program expertise directly applies to this ex- panded range of technologies and where such program expertise must be developed. Status The final report, A Summary of the State of the Art of Superconducting Magnetic Energy Storage Systems, Flywheel Energy Storage Systems and Compressed Air Energy Storage Systems, was published in July 1999, and can be referenced in SNL’s Technical Library using the following number SAND99-1854. Table 4-11. Potential Participating Utilities Utility States Covered American Electric Power (AEP) VA, WV, KY (Only southern operating region considered for study) Allegheny Power WV, VA, MD Baltimore Gas & Electric (BG&E) MD Cleco Corporation LA Carolina Power & Light (CP&L) NC, SC Duke Power NC, SC Entergy LA, MS, AK Florida Power & Light (FP&L) FL Florida Power Corporation (FPC) FL Jacksonville Electric Authority FL Kentucky Utilities KY Louisville Gas & Electric (LG&E) KY Georgia Transmission Corp GA Potomac Electric Power (PEPCO) MD, VA, DC South Carolina Electric & Gas sc (SCE&G) Southern Company GA, AL, FL, MS Tampa Electric Company (TECO ) FL Tennessee Valley Authority (TVA) TN Utilicorp wv Virginia Power VA, NC * Southern Company comprised of member companies: Georgia Power, Mississippi Power, Alabama Power, Savannah Electric, and Gulf Power Long- Versus Short-Term Storage Study A study to characterize the stationary applications and technologies of short- and long-term energy storage was completed by the ESS Program. Applications of energy storage have a wide range of performance re- quirements. One important requirement is storage time or discharge duration. In this study, applications and technologies have been evaluated to determine how storage time requirements match technology character- istics. Comparisons have also been made on the basis of capital cost for various ESSs operating over a range of discharge times, categorized as short term (< 2 hr) and long term (2 to 8 hr). Special categories of very short term (< 1 min) and very long term (a day to weeks) 4-28 4. ANALYSIS were also considered. The technologies evaluated in- cluded batteries (lead-acid and advanced), flywheels (low and high speed), supercapacitors, SMES, com- pressed air energy storage (CAES), pumped hydro, and hydrogen. Status A draft copy of the report has been received and is under review. The following is a brief summary of the draft report: The areas that have long been the focus of the ESS Program’s R&D of energy storage for the utility sector are reliability, expanded use of renewables, and en- hanced utility/industry productivity. Energy storage can address and benefit these areas in a variety of ways. Applications in these areas can be more specifically categorized by the functions that are recognized by utilities and their customers. These applications in- clude: Load management—Load management includes the traditional load-leveling application of energy stor- age, in which energy is stored during off-peak hours (typically at night) and then discharged during peak hours. This not only saves money on the basis of the difference between peak and off-peak rates, but also provides a more uniform load factor for the generation, transmission, and distribution systems. Other types of load management are ramping and load-following. Remote power—In some remote locations it is not practical to bring power to a site from an established utility grid. Power may be generated from diesel or gas generators, fuel cells or renewable sources. For local load management, it may be useful to include energy storage to minimize the generation capacity. Spinning reserve—Most electric utilities operate with a requirement for SR. This generation is ready, or in “hot standby,” should an electric generating unit somewhere on the system fail. The available reserve power is determined by the configuration and mix of unit capacities on the system. Typically, the reserve power must equal the power output of the largest gen- erating unit in operation. Renewables matching—Renewable energy sources, such as wind and solar, are desirable because they are non-polluting and in plentiful supply. By their very nature, however, they are intermittent; often the profile of energy generation does not coincide with the demand cycle. Energy storage can be used to match the output of renewable sources with any load profile. Transmission enhancement—Energy storage can improve transmission capacity by providing line stabil- ity, voltage regulation, frequency regulation, and volt amp reactive (VAR) or phase angle control. Special- ized power electronic equipment must be located in suitable locations along transmission lines. The amount of energy injected is often small, but at relatively high power. Distributed resources—Distribution systems in many growing urban and suburban areas are subject to dramatic daytime peaking. It is often more cost- effective to add distributed generation resources in critical locations than to upgrade distribution wires. Energy storage can be ideal for this application because recharging can take place during off-peak periods. Power quality—Utility power sometimes suffers disturbances such as momentary voltage sags or even outages. These events, along with harmonic distortions and other imperfections, can affect sensitive processing equipment that needs extremely clean power to operate properly. ESSs are being successfully installed to pro- vide reliable, high-quality power to sensitive loads. Sometimes the systems are coupled directly to the criti- cal equipment, and sometimes to the bus feeding a fa- cility or even on a feeder line. End use—Although the primary end-use applica- tion for energy storage is for power quality, there are other customer uses. These include local PS (to avoid time-of-day charges) and process enhancements (e.g., in pulsed power processes, or other specialized indus- trial applications). Transit—Many electric transit systems could bene- fit from energy storage because of the highly variable load they create during braking and start-up. Many types of energy storage can provide regenerative brak- ing, accepting energy from the propulsion system dur- ing deceleration and then providing a boost during acceleration. These applications can also be characterized by their technical requirements, i.e., power level, energy storage capacity, and response time. The energy storage capacity is specifically determined by the time duration delivery or discharge. Applications tend to fall into time categories of very short, short, long, and very long. Cost Analysis One major objective of this study was to compare system capital costs for the various technologies in sev- eral representative applications. For those systems that consist of the energy storage unit and a single PCS that 4, ANALYSIS 4-29 operates in both the discharge and charging modes, the system cost is the sum of the component costs plus bal- ance of plant (BOP). The costs and efficiencies used in this study are listed in Table 4-12. Most of these were developed through discussions with vendors while others are found in recent literature. These values represent the latest available data and projections. Table 4-13 indi- cates the relative maturity of the technologies and cer- tainty in the cost assumptions. Performance Fit Technologies can be matched to applications in a variety of ways. Certainly cost can be a deciding fac- tor. But the performance must also meet the application requirements. The most important characteristics are power, stored energy, and response time. If a technol- ogy cannot provide all of these characteristics, it is not suited to the application. Figure 4-18 shows numerous ESSs plotted by characteristics of power delivered and energy stored. Overlaid on the chart are lines indicating discharge times of one second, one minute, and one hour. The plot is logarithmic and covers a wide range of time scales. Some general application areas are indi- cated, for example, power quality, load management, and distributed utility. After considering the performance fit and capital costs, Table 4-14 is presented to summarize the tech- nologies that best fit the various applications, and their characteristics. Some general conclusions from Figure 4-18 are: e Ultracapacitors are best suited for smaller applica- tions, i.e., end use. e Batteries and SMES cover the broadest range of applications, from less than a MW to thousands of MW. e For very large power applications, only a few tech- nologies are suitable—CAES system and pumped hydro. Figure 4-19 indicates typical response times for the various technologies. Those with solid-state power conversion interfaces can often respond at subcycle rates, assuming they are on “standby.” Those with me- chanical inertia, such as air or water turbines, require longer start-up or response time. Most fuel cell systems also require warm-up or flow time, but recent advances are making quick-response fuel cells available as well. The various energy storage technologies serve some applications better than others do. Distinctions can be made on the basis of long- or short-term storage (or discharge duration), size (power level), response time, and also on the basis of cost. Some conclusions from this study include: e Flywheels are a good match for a range of short- term applications up to a size of several MW. e Batteries currently have the broadest overall range of applications. ¢ Fuel cells should be applicable and cost-effective in a very broad range of applications in the future. e Hydrogen—fueled combustion engines are a cur- rently available technology for short-term applications including distributed utility applica- tions, renewables matching, and SR. e CAES and pumped hydro are best for load management when geology is available and response time in the order of minutes is acceptable. e SMES is a niche technology for power quality and especially high-power distribution or transmission networks. Projected costs for bulk storage, however, show it to be expensive. In this study, only capital costs were considered. It is recommended that operating costs also be considered, because the different technologies have different energy efficiencies, parasitic requirements or losses, operations and maintenance cost; some include fuel costs and dif- ferent lifetimes or replacement costs. It is also recommended that fuel storage options for hydrogen systems be considered more broadly. In power quality applications, for example, it is possible to simply deliver hydrogen to the site, rather than use an on-site electrolyzer to produce it. This may be more cost-effective. Finally, a sensitivity analysis to cost assumptions would be valuable. 4-30 4. ANALYSIS — Table 4-12. Energy Storage Technologies Costs and Efficiencies Energy <== Related a Balance of Electrolyzer Compressor Glachore Cost Plant ($/kWh) (S/kW) (S/scfm) aos (S/kWh) Cost (S/kW) Efficiency Lead-acid batteries (low) 175 200 50 0.85 (average) 225 250 50 0.85 (high) 250 300 50 0.85 Power quality batteries 100 250 40 85 Advanced batteries 245 300 40 0.7 Micro-SMES 72,000 300 10,000 0.95 Mid-SMES (HTS projected) 2000 300 1500 0.95 SMES (HTS projected) 500 300 100 0.95 Flywheels (high-s 1) 25,000 350 1000 0.93 Flywheels (low-speed) 300 280 80 0.9 Ultracapacitors 82,000 300 10,000 0.95 Compressed air energy storage 3 425 50 0.79 (CAES) Compressed air storage (CAS) in 50 517 50 0.7 vessels Pumped hydro 10 600 2 0.87 Hydrogen fuel cell/ gas storage (low) 15 500 50 300 112.5 0.59 Hydrogen fuel cell/ gas storage (high) 15 1500 50 600 112.5 0.59 Fuel cell/ underground stor- 1 500 50 300 112.5 0.59 age Hydrogen 15 350 40 300 112.5 0.44 engine/gas storage ° . 4. ANALYSIS a5) Table 4-13. Comparison of Commercial Maturity and Cost Certainty for Energy Storage Technologies Technology Commercial Maturity Cost Certainty Lead-acid batteries Power quality batteries Advanced batteries Micro-SMES Mid-size SMES Superconducting magnetic energy storage (SMES) Flywheel (high-speed) Flywheel (low-speed) Ultracapacitor Compressed air energy storage (CAES) Compressed air storage (CAS) in tanks Pumped Hydro Fuel cells (conventional) Fuel cells (dynamic response for power quality) Hydrogen combustion engine Pn ESE] EEE eee Cae Mature products, many sold Commercial products, multiple units in the field Prototype units in the field Designs available Price list available Price quotes available Costs determined each project Costs estimated 4-32 4. ANALYSIS 1,000,000.000 100,000,000 110,000,000 1,000.00 100,000 10.000 Energy, kWh 1.000 0.100 0.010 0.001 1 10 100 1,000 10,000 100,000 Power, kW Figure 4-18. Power and Energy Characteristics of Energy Storage Products. Power Quality! Transmission | Distributed Utility | Load Leveling SMES, >1 MW | Batteries on Power Level 8 8 10kW 1kW] | | | | I I I | 1/4 cycle I cycle 1 sec 1 min >1 min Response Time Figure 4-19. | Response Characteristics of ESSs. 4. ANALYSIS 4-33 = = Table 4-14. Applications, Characteristics, and Appropriate Technology = APPLICATION Very short duration Power Time Storage Energy Time kWh Response Technologies End use ride-through, power quality, Motor Starting 1<MW Transit <1MW T&D stabilization Short duration up to 100’s MW secs secs secs ~.2 < 1/4 cycle ~.2 <1cycle 20 - 50 < 1/4 cycle kWh Flywheel Ultracapacitors Micro-SMES Lead-acid battery He fuel cell Flywheel Ultracapacitors Micro-SMES Lead-acid battery He fuel cell SMES He fuel cell Lead-acid battery Distributed generation .5 to 5 MW (peaking) End use PS (to avoid demand charges) <1MW Spinning Reserve— rapid response within 3 sec to avoid auto- matic shift Conventional response ” within 10 min Telecommunications 1-2kW backup Renewable matching (intermittent) 1-100 MW up to 10 MW ~1hr ~1hr < 30 min < 30 min ~2 hrs Min - 1 hr 5000 - 50,000 <1min 1000 <1min 5000 - 500,000 <3 sec ” < 10min 2-4 <1cycle 10- 10,000 <1 cycle Flywheel Advanced batteries SMES Lead-acid batteries Fuel cell or engine CAS Flywheel Advanced batteries Lead-acid batteries SMES Fuel cell or engine CAS Flywheel Lead-acid battery Advanced battery SMES Fuel cell or engine CAS Flywheel Lead-acid battery Advanced battery SMES Fuel cell or engine CAES CAS Pumped hydro Flywheel Ultracapacitors Lead-acid battery Advanced battery He fuel cell Flywheel Lead-acid battery Advanced battery He fuel cell SMES (continued on next page) 4-34 4. ANALYSIS Table 4-14. Applications, Characteristics, and Appropriate Technology (continued) — Storage APPLICATION i Time Power Energy Response Time Technologies Uninterruptible Power Supply up to ~2 MW ~2 hrs Long duration 100-4000 sec Flywheel Lead-acid battery Advanced battery SMES CAS Hp fuel cell Hz engine MWh Generation, 100s MW Load leveling 6-10 hrs Ramping, load 100s MW following several hrs Very long duration 100 - 1000 100 - 1000 min SMES Lead-acid battery Advanced battery Pumped hydro CAES CAS He fuel cell He engine SMES Lead-acid battery Advanced battery He fuel cell < cycle MWh Emergency backup 1 MW Seasonal storage 50 - 300 MW weeks 24 hrs 24 10,000 - min Lead-acid battery H2 engine Ho fuel cell CAS Advanced battery CAES sec — min 100,000 Renewables backup 100 kW-1MW _ 7 days Power Conversion System Magnetics and Functionality Analysis A previous study” conducted by the ESS Program focussed on identifying the state of the art in PCS tech- nology and identified several promising research direc- tions whereby cost and footprint reductions could be achieved. Additionally, the study cited several areas Atcitty, S., S. J. Ranade, and A. Gray-Fenner, Summary of State- of-the-Art Power Conversion Systems for Energy Storage Applica- tion, SAND98-2019, Sandia National Laboratories, Albuquerque, September 1998. 20 - 200 Battery Advanced battery CAES CAS Pumped hydro He fuel cell/underground storage sec — min that could potentially contribute to cost reduction, but available information did not permit a thorough as- sessment. The following list contains two promising ideas that could not be adequately evaluated. 1. Simplify PCS specifications by defining the minimum functionality required for a PCS used as part of an ESS. 2. Improve the magnetics used in the PCS. The purpose of the work being undertaken at this time is to further evaluate the above areas. As such, this report is an extension of and relies heavily on the 4, ANALYSIS 4-35 previous study. The objective of Task 1 is to determine if there is a standardized approach to specifying PCS functionality that can reduce cost in ESSs. The objec- tive of Task 2 is to determine if significant savings can be achieved in cost and footprint for PCS magnetic components. Status The PCS is a vital part of all ESSs. It provides the means for transferring power and energy between dif- ferent energy sources such as AC power systems and generators, DC battery storage systems and DC PV sources. The PCS performs conversion from DC to AC, for example, to supply energy from batteries to a load. One of the ESS Program’s goals is directed at improving the power conversion process in terms of reliability and cost. The study approach consisted of developing an as- sessment of current designs, identifying benefits through discussion with vendors and through a survey of literature, and developing a cost-savings assessment. Discussions were held with five organizations, two in- volved in the RGS area, one involved in the UPS area, one expert in motor control design, and a power semi- conductor manufacturer. In order to provide a better focus, the scope of this study was limited to grid-independent RGS systems. An RGS system may consist of photovoltaics as a re- newable source, BES, and an engine generator serving a load. Such a system may include several power con- version subsystems, namely, the DC-AC inverter, a DC-DC battery charge/discharge controller, and possi- bly an array peak-power tracker. These subsystems, along with the controller, may be integrated in a single package as the PCS. This report is concerned with the inverter. PCSs use inductors and capacitors and, more gen- erally, filters for waveform control and to support basic energy conversion mechanisms. Transformers are often required to interface to the AC load. These components are collectively referred to as magnetics. In the survey described the earlier study, many vendors indicated that there is significant potential for footprint and cost reduction in the magnetics part of the PCS. Therefore it was considered worthwhile to ex- plore the issue further. The approach, again, primarily involved discussions with vendors. Some vendors suggested that in conventional PCS, magnetics represent 20 to 30% of the cost and contrib- ute 30 to 50% to footprint and weight. Some RGS ven- dors prefer to use off-the-shelf magnetics because the low product volume precludes custom design or in- house design and construction. Off-the-shelf or third- party supplied magnetics are often designed using low- frequency design approaches. When used in a PCS, where the current and voltage contain high-frequency distortion, these devices can require additional cooling and can represent a point of failure, but this problem is not unusual. In areas other than RGS, the market vol- ume appears sufficient to justify magnetics optimiza- tion. It was also suggested that significant savings can be achieved if the magnetics were optimally designed with respect to the PCS application. For example, volt- age-sourced PCS are connected to a load or grid through an inductor and then a transformer. A trans- former can be designed to perform both the voltage transformation and inductive filtering. In volume pro- duction, the net cost savings would be substantial. The potential methods of reducing magnetics cost and footprint are: e Increase in switching frequency would substan- tially reduce filtering requirements. Obtaining higher switching frequency will, with present de- vice technologies, necessitate soft-switching tech- nology. e PCS topologies exist that can eliminate transformers and filters for some applications. Typically these have not been demonstrated in conjunction with ESS or RGS; thus it is difficult to evaluate their savings potential. e Reduce cost by using use newer technologies for capacitors and inductors. Newer film capacitors are being used in DC links, with as much as a 50% reduction in weight and volume. There does not appear to be any attempt to use newer technologies such as amorphous cores and high-frequency coil design techniques in connection with inductors and transformers. Apparently the design and material costs outweigh benefits in the RGS area. The magnetics have a significant impact on foot- print and cost. However, the key issue appears to be that the cost of modifications may outweigh the benefits in small-volume applications. It is believed that these costs will continue to decline. This area is thus not seen as a major R&D priority. The objective of this task was to answer the fol- lowing questions: 4-36 4. ANALYSIS 1, What type of simplification and standardization is appropriate for PCS used in RGS systems? 2. What are the benefits of such an approach? The discussions with vendors focussed on PCS de- sign, PCS-related problems in RGS, and the need and approach for standardization. Observations from these discussions are summarized below. An initial analysis of cost savings is also developed. RGS is a low-volume application, and this makes the design and integration of RGS systems somewhat unique. Most often the PCS vendor acts as the system integrator; that is, the PCS and controls are engineered as one unit. The PCS then becomes the “heart” of the system in the sense that the PCS controller must control the power conversion process as well as control other elements in the RGS. The “true cost” of the PCS is thus masked by costs associated with upgrading the overall control system to include RGS control. The study (PCS study done previously) cites a cost range from $200-$1200/kW. Another drawback of existing systems is that although modern PCSs are quite reli- able, downtimes can be significant since a failed PCS cannot be easily replaced. This is because the PCS is not an “off-the-shelf” component and usually must be repaired or replaced by the original vendor. One conclusion, and thus a fundamental assump- tion in further work, is that in order to specify a mini- mum functionality PCS it is necessary to separate the RGS control functions and interface hardware from the control functions and hardware required for the power conversion function. This assumption drives the re- mainder of the discussion. The benefits of unitizing power electronic subsys- tems are well recognized in industry, particularly through research originating from the Power Electronic Building Blocks (PEBB) Project." Although terminol- ogy is not completely standardized, industry is moving towards the types of subsystems described below. e “Power modules,” also called “power assemblies,” are commercially available packages that are comprised of power devices, device drivers, and current/voltage/temperature sensing, cooling, and protection. Devices used and the topological connection depend on function (e.g., DC to three- phase AC conversion) and voltage current ratings. Power modules can include additional components such as DC-side capacitors. With highly optimized * http://pebb.onr.navy.mil/ design and packaging, and leveraged by markets in the motor control area, the industry target cost for these units is $25/kW. e A “power cell” (a term suggested by one of the vendors), a “power processor” or “macro- processor” (a term coined by Silicon Power Co.), would be comprised of the power module and a controller that achieves a specific conversion function. Such a controller would presumably use microprocessor and digital signal processing technology. In the context of RGS, a power cell would be a DC-AC power converter or inverter capable of supplying a specific AC current based on a reference command. Additionally, such power cells could be designed to inherently allow series/parallel connections for higher ratings. The inclusion of a controller mandates that software development tools be made available to program the controller from basic building-block software. Vendors tend to standardize such design. e A “basic inverter,” on the other hand, would consist of the power cell, associated filters, disconnect means, and a controller that would interface with the overall RGS controller. Thus the RGS controller might specify a setting for AC power; the inverter control would develop the target current command for its current regulator. It appears that, analogous to the trend in power modules, power cells can also be highly optimized and characterized in a way that any system integrator can utilize them as a subsystem in a PCS. It is also clear that vendors are already moving in this direction and away from one-of-a-kind designs. Although the cost of some of the components of the PCS are difficult to assess without detailed design, an estimate of cost of a power cell was developed. The previous study provided cost data for actual and pro- posed RGS applications. The data are broken down by power stage, controller, and balance of system cost. The power stage cost in the minimum functionality PCS was taken to be the target cost of power modules de- scribed above, with a contingency cost for assembly and integration. The cost of the controller is extremely variable and probably dominated by custom control development. It is believed that the controller for a minimum functionality PCS is a mature technology and a reasonable cost was assumed. The balance of system cost was estimated using standard installed cost for electrical equipment. The total cost was then compared with median cost of a small village RGS. 4. ANALYSIS 4-37 It is estimated that PCS cost reduction of 50% may be possible. Rather than the specific percentage cited, it should be emphasized that substantial benefits may accrue by reorganizing designs of RGS and leveraging advances in technology in the power electronics indus- try. This can be done by developing standardized specifications for a PCS at the power cell or basic inverter level. The reduction comes primarily from the use of advanced technology such as power modules, with costs dictated by a much larger market segment. It should be noted that these benefits are offset to some extent by the need for a RGS controller and possibly additional PCSs for the PV system or battery. In summary, anticipated benefits of moving to- wards a standardized power cell or basic inverter ap- proach include: e Potential reduction in PCS cost, e Interchangeability of PCS from different vendors in RGS designs, e Ability to easily replace a failed PCS in an RGS system. RGS systems continue to be one-of-a-kind designs. The development of an industry-supported specification of PCS as a RGS subsystem, rather than the RGS “con- troller,” is recommended. It is believed that this would move the RGS industry in a direction to utilize the same technology as is used in the larger market such as mo- tor-drive, UPS, and other industrial markets. The con- cept is not novel. It is noted that vendors are moving in this direction in PCS design. Some vendors already offer basic inverter modules and separate controllers. The ESS Program can help accelerate this development. EESAT Conference The ESS Program, under the sponsorship of the DOE and in cooperation with the ESA, took over the responsibility for organizing the Electrical Energy Stor- age Systems Applications and Technology (EESAT) Conference to be held September 18 to 20, 2000, in the Royal Plaza Hotel, Lake Buena Vista, Florida, at the Disney World resort. The international conference ad- dresses all aspects of ESS technologies including con- ventional and advanced batteries, ultracapacitors, SMES, flywheels, CAES, pumped hydro, power elec- tronics and control systems, and studies and economic analyses of storage systems. Individual components, system applications, and research into advanced sys- tems and components will also be included in the con- ference agenda. Status A three-person steering committee was established in the third quarter to do the initial organizing for work the conference. The steering committee consisted of the DOE/ESS program manager, the ESA board chair, and the SNL/ESS program manager. The full interna- tional conference organizing committee will be chosen early in 2000. To aid the committee in the administra- tive duties of the conference, a competitive contract was placed with Complete Meeting Concepts (CMC), Inc. in Orlando, Florida. Duties of CMC will include developing a pre-conference schedule and financial plan; preparing and mailing all conference announce- ments, registration, paper collection and publication; and acting as the prime contact with the conference hotel for the meeting logistics. The steering committee will be responsible for technical aspects of the confer- ence (organizing committee selection, paper review and approval, session arrangements, and overall conference direction). The contract for the hotel was successfully trans- ferred from the ESA to SNL in February 1999. A call for papers was issued in September 1999, with abstracts due in January 2000. Technical and eco- nomic papers are encouraged. The component areas of interest include the ESS, power electronics, control and data acquisition. The application areas of interest in- clude generation, transmission, distributed resources, distribution power quality (DPQ), energy management, and customer applications. The research areas of inter- est include advanced systems and components, identifi- cation of the potential users and needs for energy storage and the future of the deregulated utility indus- try. The conference schedule is shown in Table 4-15. Table 4-15. EESAT 2000 Conference Schedule Call for papers nr Sapionbes 1999 Abstracts due January 2000 Final papers June 2000 Final Agenda August 2000 EESAT 2000 September 18-20, 2000 Conference 4-38 4. ANALYSIS 5. References Alaskan Division of Energy. Statistical Report of the Power Cost Equalization Program, Ninth Edition, 1998. Anderson, Dr. Max D., and John T. Alt. Assessment of Costs and Benefits of Battery Energy Storage for Electric Utility Applications, Document AL-9306, revised March 6, 1995. Anderson, Dr. Max D., and Xiangling Gao. A Study to Quantify the Benefits of Adding Storage to a Utility System, Document AO-4841, revised June 28, 1996. Anderson, Dr. Max D., and John D. Stickley. Electric Utility Savings from BESS Applications, Document AU-7834, revised August 31, 1998. Atcitty, S., S. J. Ranade, and A. Gray-Fenner. Summary of State-of-the-Art Power Conversion Systems for Energy Storage Application, SAND98-2019, Sandia National Laboratories, Albuquerque, New Mexico, September 1998. Butler, P. C. Energy Storage for Utility Applications: Phase I — Opportunities Analysis, SAND94-2605, October 1994. Power Electronics Building Blocks (PEBB) Program. http://pebb.onr.navy.mil/. U.S. Navy, Office of Naval Research. Updated April 9, 1998. NERC Policy 10, Draft 3 provides details of changes that are under way for Frequency Responsive Reserve. The document is available at http:/www.nerc.com/~oc/standards/. 5. REFERENCES 5-1 This page intentionally left blank. 5. REFERENCES Appendix A: Data From Vernon BESS Operations APPENDIX A A-] This page intentionally left blank. A-2 APPENDIX A == Table A-1. October 1998 Discharge Data from Vernon BESS Operations (Tables for November and December 1998 are not included because preventive maintenance required the BESS to be offline) October 1998: 3,050-kW Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period Values (at left) Period 10/1/98 Th 895 3057 3446 389 72% 10/2/98 F 870 3053 3449 396 69% 10/3/98 Sa 615 2981 3196 215 93% 10/4/98 Su 895 3030 3169 139 92% 10/5/98 M 785 3052 3455 403 64% 10/6/98 Tu 826 3053 3602 549 69% 10/7/98 WwW 753 3051 3847 796 78% 10/8/98 Th 99 2970 3160 190 91% 10/9/98 F 597 3029 3491 462 89% 10/10/98 Sa No PS 800 2983 3598 615 - 10/11/98 Su No PS 28 2931 3244 313 - 10/12/98 M 617 3036 3450 414 72% 10/13/98 Tu 676 3050 3496 446 71% 10/14/98 Ww 772 3051 3696 645 76% 10/15/98 Th Note A 1143 3076 3610 534 - 10/16/98 F No PS 1830 3373 3972 599 - 10/17/98 Sa No PS - 3002 3572 570 - 10/18/98 Su No PS : 2754 3105 351 - 10/19/98 M No PS 2017 3224 3810 586 - 10/20/98 Tu No PS 2140 3313 3975 662 - 10/21/98 WwW No PS 2100 3243 3923 680 - 10/22/98 Th No PS 1284 2011 3703 1692 - 10/23/98 F No PS 1201 3109 3775 666 - 10/24/98 Sa No PS 472 2982 3201 219 - 10/25/98 Su No PS - 2755 3106 351 - 10/26/98 M No PS 2027 3226 3778 552 - 10/27/98 Tu No PS 2147 3343 4319 976 - 10/28/98 Ww No PS 2114 3228 3768 540 - 10/29/98 Th No PS 2135 3282 3838 556 - 10/30/98 F No PS 2151 3379 4032 653 - 10/31/98 Sa No PS (Data logger interrupted) - (Note: Data acquired from BESS has 10-s sample rates on discharge and 3-min. on recharge, PS = peak shaving) (Note A: PCP #2 (inverter) tripped ~3:30 PM for unknown reasons; peak shaving nullified for the month) Largest of the month: Average for entire month: Average for weekdays only: 1185 1326 3054 3100 4319 3593 3709 78.0% 75.1% APPENDIX A A-3 So 9S ° wo Demand, kw 3000 Plant TRI-BATT-0064-0 Average and Peak Plant Demand at Vernon for the Month of October 1998. Figure A-1. APPENDIX A A-4 = Table A-2. January 1999 Data from Vernon BESS Operations January 1999: 3050-kW Peak-Shaving Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period values (at left) Period 1/1/99 F 0 2606 2681 75 92 % 1/2/99 Sa No PS - 2577 2744 167 - 1/3/99 Su No PS - 2528 2644 116 - 1/4/99 M 26 2739 3283 544 92% 1/5/99 Tu 0 2502 2861 359 92 % 1/6/99 WwW 0 2439 2577 138 92 % 1/7/99 Th 29 2699 3310 611 92 % 1/8/99 F 64 2797 3317 520 92 % 1/9/99 Sa No PS - 2922 3450 528 - 1/10/99 Su No PS : 2825 3286 461 - 1/11/99 M 75 2767 3341 574 92 % 1/12/99 Tu 196 2907 3383 476 86 % 1/13/99 WwW 116 2860 3375 515 92 % 1/14/99 Th 193 2914 3382 468 92 % 1/15/99 F 232 2880 3356 476 91% 1/16/99 Sa No PS - 2608 2736 128 - 1/17/99 Su No PS - 2561 2710 149 - 1/18/99 M 270 2948 3342 394 92 % 1/19/99 Tu 220 2908 3284 376 92 % 1/20/99 WwW 182 2889 3401 512 91% 1/21/99 Th 33 2766 3273 507 92% 1/22/99 F 132 2843 3302 459 92 % 1/23/99 Sa No PS - 2672 2823 151 - 1/24/99 Su No PS - 2640 2784 144 - 1/25/99 M 586 3039 3404 365 83 % 1/26/99 Tu 273 2938 3394 456 93 % 1/27/99 Ww 195 2891 3335 444 92 % 1/28/99 Th 112 2839 3376 537 92% 1/29/99 E 206 2882 3436 554 91% 1/30/99 Sa No PS - 2667 2801 134 - 1/31/99 Su No PS - 2657 2748 91 - (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 3450 Average for entire month: 150 2765 3134 91.2% Average for weekdays only: 151 2755 3110 91.2% APPENDIX A A-5 = = ———— Table A-3. February 1999 Data from Vernon BESS Operations — February 1999: 3000-kW Peak-Shaving Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period values (at left) Period 2/1/99 M 142 2842 3346 504 89 % 2/2/99 Tu 184 2849 3324 475 91% 2/3/99 W 100 2803 3312 509 92% 2/4/99 Th 15 2708 3220 512 92% 2/5/99 F 9 2591 3168 577 92% 2/6/99 Sa 0 2568 2708 140 92% 2/7/99 Su No PS - 2613 2689 76 - 2/8/99 M 16 2631 3113 482 92% 2/9/99 Tu 40 2739 3273 534 92% 2/10/99 Ww 62 2733 3281 548 92% 2/11/99 Th 690 2998 3357 359 83 % 2/12/99 F 761 2992 3398 406 85 % 2/13/99 Sa NoPS - 2832 2993 161 - 2/14/99 Su No PS - 2838 3015 177 - 2/15/99 M 451 2959 3306 347 89 % 2/16/99 Tu 708 2978 3450 472 87 % 2/17/99 Ww 602 2996 3356 360 80 % 2/18/99 Th 543 2985 3389 404 86 % 2/19/99 F 577 2984 3352 368 84% 2/20/99 Sa NoPS - 2788 2916 128 - 2/21/99 Su No PS - 2725 3012 287 = 2/22/99 M 104 2738 3243 505 92 % 2/23/99 Tu 241 2825 3284 459 92 % 2/24/99 WwW 336 2881 3372 491 89 % 2/25/99 Th 228 2825 3307 482 91% 2/26/99 F 661 3001 3386 385 64% 2/27/99 Sa NoPS - 2886 3107 221 - 2/28/99 Su No PS - 3044 3183 139 - (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 3450 Average for entire month: 308 2834 3209 87.9 % Average for weekdays only: 324 2853 3312 87.7 % A-6 APPENDIX A a a = Table A-4. March 1999 Data from Vernon BESS Operations March 1999: 3000-kW Peak-Shaving Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period values (at left) Period 3/1/99 M 672 2996 3385 389 76 % 3/2/99 Tu 788 3001 3501 500 73% 3/3/99 Ww 502 2988 3359 371 83 % 3/4/99 Th 617 2999 3395 396 75 % 3/5/99 F 529 3003 3460 457 80 % 3/6/99 Sa NoPS - 2956 3102 146 - 3/7/99 Su No PS - 3045 3259 214 - 3/8/99 M 690 3002 3473 471 69% 3/9/99 Tu 1101 2979 3210 231 90 % 3/10/99 Ww 1 2920 3076 156 93 % 3/11/99 Th 351 2969 3417 448 91% 3/12/99 F 584 2982 3356 374 87% 3/13/99 Sa NoPS - 3005 3216 211 - 3/14/99 Su NoPS - 2920 3073 153 - 3/15/99 M 22 2898 3095 197 93 % 3/16/99 Tu 0 2423 2602 179 92 % 3/17/99 WwW 0 2368 2492 124 92% 3/18/99 Th 0 2450 2956 506 92 % 3/19/99 IF 0 2498 3166 668 92 % 3/20/99 Sa NoPS - 2615 3342 727 - 3/21/99 Su = NoPS - 2895 3199 304 - 3/22/99 M 81 2938 3121 183 91% 3/23/99 Tu 360 2944 3137 193 82 % 3/24/99 Ww 554 2952 3159 207 93 % 3/25/99 Th 717 3000 3405 405 80% 3/26/99 F 124 2838 3060 222 93 % 3/27/99 Sa 824 2968 3084 116 92 % 3/28/99 Su No PS - 3100 3253 153 - 3/29/99 M 1007 3006 3401 395 69 % 3/30/99 Tu 722 2989 3247 258 82% 3/31/99 Ww 873 3000 3341 341 74% (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 3501 Average for entire month: 463 2892 3205 84.8 % Average for weekdays only: 448 2876 3209 84.4 % APPENDIX A A-7 4000 3500 3000 2500 Plant Demand, 2000 i ’ March Peak 1500 y~ March Ave. 1000 50 Feb. Peak ; Feb. Ave. Month =. i ~ N 28 @ Day of the Month Nom 5 Figure A-2. Average and Peak Plant Demand at Vernon for the Months of January, February, and March 1999. A-8 APPENDIX A Table A-5. April 1999 Data from Vernon BESS Operations a <== April 1999: 2975-kW Peak-Shaving Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period values (at left) Period 4/1/99 Th 596 2972 3366 394 78% 4/2/99 F 1054 2972 3112 140 91% 4/3/99 Sa 1221 2981 3120 139 91% 4/4/99 Su No PS - 3085 3242 157 - 4/5/99 M 641 2971 3713 742 71% 4/6/99 Tu 784 2978 3147 169 84% 4/7/99 Ww 779 2964 3322 358 87 % 4/8/99 Th 620 2974 3188 214 82% 4/9/99 F 36 2932 3080 148 92% 4/10/99 Sa 0 2848 2945 97 93 % 4/11/99 Su No PS 7 3089 3235 146 - 4/12/99 M 799 2974 3377 403 72% 4/13/99 Tu 542 2958 3305 347 90 % 4/14/99 WwW 1114 2978 3373 395 76 % 4/15/99 Th 835 2978 3374 396 70% 4/16/99 F 529 2962 3073 111 92 % 4/17/99 Sa No PS - 3023 3228 205 - 4/18/99 Su No PS 7 2996 3176 180 - 4/19/99 M 771 2974 3358 384 62 % 4/20/99 Tu 875 2975 3331 356 60 % 4/21/99 WwW 543 2926 3396 470 91% 4/22/99 Th 1465 3059 3521 462 95 % 4/23/99 F 95 2873 3142 269 92 % 4/24/99 Sa No PS = 3040 3183 143 - 4/25/99 Su No PS : 3024 3210 186 7 4/26/99 M 689 2976 3337 361 59 % 4/27/99 Tu 905 2981 3515 534 60 % 4/28/99 WwW 640 2974 3432 458 77 % 4/29/99 Th 898 2972 3322 350 84% 4/30/99 F 126 2892 3118 226 93 % (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 3713 Average for entire month: 690 2977 3275 80.9 % Average for weekdays only: 697 2964 3314 79.9 % APPENDIX A A-9 Table A-6. May 1999 Data from Vernon BESS Operations May 1999: 3100-kW Peak-Shaving Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period values (at left) Period 5/1/99 Sa 497 3014 3313 299 93% 5/2/99 Su 94 2979 3209 230 92% 5/3/99 M 769 3089 3571 482 80% 5/4/99 Tu 526 3045 3527 482 90 % 5/5/99 WwW 404 3005 3487 482 92% 5/6/99 Th 367 3011 3462 451 91% 5/7/99 FE 2802 2985 183 93% 5/8/99 Sa 0 2850 3056 206 93 % 5/9/99 Su No PS = 2833 2942 109 : 5/10/99 M 402 2990 3498 508 89 % 5/11/99 Tu 245 3012 3188 176 93 % 5/12/99 Ww 98 2923 3445 522 82% 5/13/99 Th 926 3095 3679 584 83 % 5/14/99 lz, 119 3070 3452 382 91% 5/15/99 Sa No PS - 3031 3860 829 - 5/16/99 Su No PS a 2930 3196 266 5/17/99 M 748 3078 3438 360 87% 5/18/99 Tu 566 3089 3540 451 74% 5/19/99 Ww 590 3082 3536 454 85% 5/20/99 Th 1063 3092 3463 371 86% 5/21/99 F 844 3092 3529 437 71% 5/22/99 Sa No PS - 3060 3504 444 - 5/23/99 Su No PS - 3147 3425 278 : 5/24/99 M 911 3099 3547 448 73 % 5/25/99 rie 901 3100 3505 405 73% 5/26/99 Ww 744 3099 3472 373 89% 5/27/99 Th 733 3080 3571 491 89 % 5/28/99 F 94 3001 3222 221 92% 5/29/99 Sa No PS . 3044 3296 252 5/30/99 Su No PS - 2992 3260 268 5/31/99 M No PS - 3063 3537 474 - (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 3860 Average for entire month: 506 3026 3410 86.1 % Average for weekdays only: 553 3044 3460 85.2 % = A-10 APPENDIX A Table A-7. June 1999 Data from Vernon BESS Operations June 1999: 3100-kW Peak-Shaving Trigger Level No. of Average kW Largest Peak Difference Lowest SOC Day of Discharge During During Between During Date the Operations Required Peak- Largest & Peak- Week During Required Demand Shaving Average kW Shaving Demand Period Period Period values (at left) Period 6/1/99 Tu 700 3084 3496 412 82% 6/2/99 Ww 471 3069 3543 474 92% 6/3/99 Th 255 2937 3469 532 91% 6/4/99 F 610 3030 3500 470 91% 6/5/99 Sa 0 2746 3029 283 94% 6/6/99 Su 102 2901 3175 274 92% 6/7/99 M 408 2992 3485 493 87 % 6/8/99 Tu 394 3024 3472 448 87% 6/9/99 Ww 496 3000 3584 584 91% 6/10/99 Th 575 3057 3571 514 85 % 6/11/99 F 533 2952 3571 619 92% 6/12/99 Sa No PS 2817 3032 215 6/13/99 Su No PS - 2943 3338 395 - 6/14/99 M 382 2991 3453 462 89% 6/15/99 Tu 228 2960 3450 490 92% 6/16/99 WwW 351 2982 3429 447 92% 6/17/99 Th 207 2945 3406 461 92% 6/18/99 Ei 572 3053 3489 436 86 % 6/19/99 Sa No PS - 2817 3129 312 : 6/20/99 Su No PS - 2817 3183 366 6/21/99 M 432 2979 3442 463 90 % 6/22/99 Tu 472 3021 3491 470 85 % 6/23/99 Ww 152 2862 3399 537 92% 6/24/99 Th 575 3044 3468 424 83 % 6/25/99 F 488 3021 3457 436 89 % 6/26/99 Sa No PS 2798 3167 369 : 6/27/99 Su No PS 2867 3319 452 - 6/28/99 M 87 2834 3307 473 93 % 6/29/99 Tu 463 3004 3461 457 89 % 6/30/99 Ww 541 3025 3605 580 87 % (Note: Data acquired from BESS has 10-sec sample rates on discharge and 3-min on recharge.) PS = peak shaving Largest of the month: 3605 Average for entire month: 396 2952 3397 89.3 % Average for weekdays only: 427 2994 3479 89.0% APPENDIX A A-1] [ i | ' V7 June Peak i/ June Ave. May Peak Lc April Peak April Ave. Day of the Month ho Figure A-3. Average and Peak Plant Demand at Vernon for the Months of April, May, and June 1999. A-12 APPENDIX A Appendix B: Savings Due to BESS Only APPENDIX B B-1 This page intentionally left blank. B-2 APPENDIX B Table B-1. Savings ($k) Due to BESS Only, When Wind Power is Installed 1995 Spinning Reserve Load Leveling Load Leveling with Spinning Reserve 1 4 8 1 4 8 1 4 8 40 1825 1825 1825 229 508 617 1727 1635 1725 100 3951 3951 3951 437 934 1238 3682 3572 3608 200 6072 6072 6072 677 1480 2211 5930 5970 6116 300 6094 6094 6094 771 1859 2324 6706 7144 7238 1996 Spinning Reserve Load Leveling Load Leveling with Spinning Reserve 1 4 8 1 4 8 1 4 8 40 1618 1618 1618 229 482 546 1680 1650 1563 100 3986 3986 3986 366 828 930 3586 3280 3266 200 5760 5760 5760 643 1196 1385 5274 5253 5328 300 6008 6008 6008 676 1297 1678 6328 6481 6440 1997 Spinning Reserve Load Leveling Load Leveling with Spinning Reserve 1 4 8 1 4 8 1 4 8 40 1896 1896 1896 402 765 854 1813 1732 1744 100 3649 3649 3649 727 1427 1695 3386 3407 3421 200 6037 6037 6037 953 1934 1843 5661 5398 5331 300 6059 6059 6059 1120 2170 2141 6723 6622 6521 APPENDIX B B-3 Table B-2. Savings ($k) Due to BESS Only (no renewables) Load Leveling 8 1456 2767 3153 3963 Load Leveling 8 690 1656 3213 3861 Load Leveling 1995 Spinning Reserve 1 4 8 1 4 40 1826 1826 1826 560 955 100 3042 3042 3042 584 1718 200 3781 3781 3781 933 2371 300 3637 3637 3637 1019 3028 1995 Spinning Reserve 1 4 8 1 4 40 945 945 945 284 516 100 3136 3136 3136 458 132 200 4838 4838 4838 793 2095 300 4744 4744 4744 850 2594 1997 Spinning Reserve 1 4 8 1 4 40 1632 1632 1632 229 610 100 2764 2764 2764 451 1222 200 4435 4435 4435 851 2161 300 4520 4520 4520 1163 2929 8 965 1997 3112 3961 Load Leveling with Spinning Reserve 1 4 8 1098 2091 2127 3297 3564 3779 4389 4773 4900 4926 5739 6381 Load Leveling with Spinning Reserve 1 4 8 674 1540 1577 3053 3054 3256 3965 4404 5010 4682 5805 6757 Load Leveling with Spinning Reserve 1 4 8 1681 1820 1838 2752 2767 2955 4936 5285 5578 5799 6764 7556 APPENDIX B Appendix C: Presentations and Publications APPENDIX C C-l This page intentionally left blank. C-2 APPENDIX C Appendix C: Presentations and Publications Presentations A. Akhil, “Battery and Diesel Concepts,” presented at the Rural Alaska Alternative Energy technical conference, Fairbanks, Alaska, August 17, 1999. S. Atcitty, “Power Conversion System Analysis,” PCS Study—presented at the DOE Executive Review meeting, Washington, DC, September 8 through 9, 1999. J. D. Boyes, P.C. Butler, “Energy Storage is Underway at Sandia National Laboratories (DOE),” presented at the ESA spring meeting, Raleigh, North Carolina, May 6, 1999. J. D. Boyes, “Substation Power Quality Project,” presented at the Power Systems World Conference, Santa Clara, California, November 1998. J. D. Boyes, “Transmission Power Quality Project,” Southeast Electric Exchange Power Quality Subcommittee meeting — New Orleans, Louisiana, April 1999. J. D. Boyes, “Status of Energy Storage Program,” presented at the ESS Annual Peer Review, Washington DC, September 1999. J. D. Boyes, “Lessons Learned in the Puerto Rico Battery Energy Storage Project,” presented at the Sixth International Batteries for Utility Energy Storage Conference, Essen, Germany, September 1999. J. D. Boyes, “Demonstrations of Utility Sized Zinc/Bromine Battery Systems,” presented at the Sixth International Batteries for Utility Energy Storage Conference, Essen, Germany, September 1999. J. D. Boyes, “A Pilot Renewable/Diesel Hybrid Village System in Mexico,” presented at the Sixth International Batteries for Utility Energy Storage Conference, Essen, Germany, September 1999. P. C. Butler, “Energy Storage Simulator Applications and Enhancements,” presented at the NRECA Power Supply Task Force meeting, Albuquerque, New Mexico, January 1999. P. C. Butler, “ILZRO/SNL VRLA Reliability Study,” presented at 111th BCI convention, Nashville, Tennessee, May 4, 1999. G. P. Corey, “Intermediate State of Charge Testing Program,” presented at the spring ESA meeting, Raleigh, North Carolina, May 7, 1999. P. DiPietro, “Analysis of Flywheels & SMES in Stationary Applications,” presented at the ESA spring meeting, Raleigh, North Carolina, May 7, 1999. M. Mclntyre, “Database for Alaska Community Electric Power Systems,” presented at the Fourth Annual Science and Technology Outreach symposium, Albuquerque, New Mexico, August 5, 1999. C. Namovicz, R. Sen, N. Miller, “Advanced Lithium Batteries for Stationary Applications,” presented at the spring ESA meeting, Raleigh, North Carolina, May 7, 1999. D. Rovang, A. A. Akhil, J. D. Boyes, “Substation Power Quality Project,” presented at the ESA spring meeting, Raleigh, North Carolina, May 7, 1999. R. Sen, “Renewable Generation and _ Storage,” presented at the ESA spring meeting, Raleigh, North Carolina, May 7, 1999. P. A. Taylor, P. T. Moseley, P. C. Butler, “Preliminary Results of an ILZRO-Sponsored Field-Data Collection and Analysis to Determine Relationships Between Service Conditions and Reliability of VRLA_ Batteries in Stationary Applications,” presented at the Intelec 98 conference, San Francisco, California, October 7, 1998. APPENDIX C C-3 W. Torres, “Four Years of Experience with the PREPA BESS,” presented by P.C. Butler at the 111th BCI convention, Nashville, Tennessee, May 4, 1999. Publications J. D. Boyes, Lessons Learned from the Puerto Rico Battery Energy Storage System, English and Spanish versions available, SAND99-2232, Sandia National Laboratories, Albuquerque, New Mexico, September 1999. J. D. Boyes, Lessons Learned in the Puerto Rico Battery Energy Storage Project, Sixth International Batteries for Utility Energy Storage Conference, Essen, Germany, September 1999. P. C. Butler, Energy Storage Systems Program Report for FY98, SAND98-0883, Sandia National Laboratories, Albuquerque, New Mexico, April 1999. N. C. Clark, P. Eidler, and P. Lex, Development of Zinc/Bromine Batteries for Load-Leveling Applications: Phase 2 Final Report, SAND99-2691, Sandia National Laboratories, Albuquerque, New Mexico, October 1999. P. Eidler, Development of Zinc/Bromine Batteries for Load-Leveling Applications: Phase I Final Report, SAND99-1853, Sandia National Laboratories, Albuquerque, New Mexico, July 1999. C. Miralles, B. Jensen, A. Flack, B. Yelin, J. Cromwell, R. Lopez, and S. Berge, Solar-Powered Systems for Environmental Remediation, SAND99-0936, Sandia National Laboratories, Albuquerque, New Mexico, April 1999. B. L. Norris and G.J. Ball, Performance and Design Analysis of a 250-kW, Grid-Connected Battery Energy Storage System, SAND99-1483, Sandia National Laboratories, Albuquerque, New Mexico, June 1999. P. Taylor, L. Johnson, K. Reichart, P. DiPietro, J. Philip, and P. Butler, A Summery of the State of the Art of Superconducting Magnetic Energy Storage Systems, Flywheel Energy Storage Systems, and Compressed Air Energy Storage Systems, SAND99- 1854, Sandia National Laboratories, Albuquerque, New Mexico, July 1999. J. Wohlgemuth, J. Miller, and L. B. Sibley, Investigation of Synergy Between Electrochemical Capacitors, Flywheels, and Batteries in Hybrid Energy Storage for PV Systems, SAND99-1477, Sandia National Laboratories, Albuquerque, New Mexico, June 1999. G. Kern, System and Battery Charge Control for PV- Powered AC Lighting Systems, SAND99-0935, April 1999. C4 APPENDIX C Bob Weaver 777 Wildwood Lane Palo Alto, CA 94303 Robert W. Fenn 6335 Coleridge Road Painesville, OH 44077 Per Danfors ABB Power T&D Co., Inc. 16250 West Glendale Drive New Berlin, WI 53151 Jim Balthazar Active Power 11525 Stonehollow Dr. Suite 255 Austin, TX 78758 David Lockard Alaska State Div. of Energy 333 West Fourth Ave. Suite 220 Anchorage, AK 99501-2341 Distribution Eric Rudd 35 Harmon Ave Painesville, OH 44077 Dutch Achenbach 2111 Buffalo Casper, WY 82604 Hans Weinerich ABB Power T&D Co., Inc. 1460 Livingston Ave. P.O. Box 6005 North Brunswick, NJ 08902-6005 Robert Wills Advanced Energy Systems, Inc Riverview Mill P.O. Box 262 Wilton, NH 03086 Percy Frisby Alaska State Div. of Energy 333 West Fourth Ave. Suite 220 Anchorage, AK 99501-2341 Peter Crimp Alaska State Div. of Energy 333 West Fourth Ave. Suite 220 Anchorage, AK 99501-2341 C. Shih American Elec. Pwr. Serv. Corp. 1 Riverside Plaza Columbus, OH 43215 Tim Ball Applied Power Corporation Solar Engineering 1210 Homann Drive SE Lacey, WA 98503 Christian St-Pierre ARGO-TECH Productions, Inc. Subsidiary of Hydro-Quebec 1580 de Coulomb Boucherville, QC J4B 7Z7 CANADA Ira Bloom Argonne National Laboratories 9700 South Cass Avenue CTD, Bldg. 205 Argonne, IL 60439-4837 Thomas Ablel American Superconductor Corp. 8371 Bunchberry Court Citrus Heights, CA 95610 Meera Kohler Anchorage Municipal Light & Pwr 1200 East 1st Avenue Anchorage, AK 99501 Ralph M. Nigro Applied Energy Group, Inc. 46 Winding Hill Drive Hockessin, DE 19707 Gary Henriksen Argonne National Laboratories 9700 South Cass Avenue CTD, Bldg. 205 Argonne, IL 60439 Alan Wolsky Argonne National Laboratories 9700 S. Cass Avenue Building362 Argonne, IL 60439-4837 Herb Hayden Arizona Public Service 400 North Fifth Street P.O. Box 53999, MS8931 Phoenix, AZ 85072-3999 Robert Hammond Arizona State University East 6001 S. Power Rd. Bldg. 539 Mesa, AZ 85206 Greg Kern Ascension Technology, Inc. 4700 Sterling Drive Unit # Boulder, CO 80301 Glenn Campbell Babcock & Wilcox P.O. Box 785 Lynchburg, VA 24505 Michael L. Bergey Bergey Windpower 2001 Priestley Avenue Norman, OK 73069 Ray Hobbs Arizona Public Service 400 North Fifth Street P.O. Box 5399, MS8931 Phoenix, AZ 85072-3999 Edward C. Kern Ascension Technology, Inc. P.O. Box 6314 Lincoln, MA 01773-6314 Rick Lawrence AVO International P.O. Box 9007 Valley Forge, PA 19485-1007 Richard L. Hockney Beacon Power Corp. 6 Gill St. Woburn Industrial Park Woburn, MA 01801-1721 Klaus Kraemer Berliner Kraft und Licht (BEWAG) Bewag AG Energy Planning & Trading D- 12432 Berlin, GERMANY Massoud Assefpour BHP Research & Tech Dev. 600 Bourke Street Melbourne Victoria, 3000 AUSTRALIA Gerald W. Braun BP Solarex 630 Solarex Court Frederick, MD 21703 Dr. Les Holden C&D Charter Pwr. Systems, Inc. Washington & Cherry Sts. Conshohocken, PA 19428 Dr. Sudhan S. Misra C&D Technologies Washington & Cherry Streets Conshohocken, PA 19428 J. Holmes California State Air Resc. Board Research Division P.O. Box 2815 Sacramento, CA 95812 Samuel B. Wright Boeing Inform., Space & Defense Sys. P.O. Box 3999 MS 82-97 Seattle, WA 98124-2499 Salim Jabbour Business Management Consulting 24704 Voorhees Drive Los Altos Hills, CA 94022 Larry S. Meisner C&D Powercom 1400 Union Meeting Road P.O. Box 3053 Blue Bell, PA 19422-0858 Jon Edwards California Energy Commission 1516 Ninth Street, MS-46 Sacramento, CA 95814 Pramod P. Kulkarni California Energy Commission Research & Dev. Office 1516 9th Street, MS43 Sacramento, CA 95814-5512 Rod Boucher Calpine Corporation 50 W. San Fernando Suite 550 San Jose, CA 95113 John Cooley Chugach Elec. Association, Inc. P.O. Box 196300 Anchorage, AK 99519-6300 N. Tai Consolidated Edison 4 Irving Place New York, NY 10003 R. B. Sloan Crescent EMC P.O. Box 1831 Statesville, NC 28687 Bob Galyen Delphi Energy. & Engine Management Systems P.O. Box 502650 Indianapolis, IN 46250 Tom Lovas Chugach Elec. Association, Inc. 5601 Minnesota Dr. P.O. Box 196300 Anchorage, AK 99519-6300 M. Lebow Consolidated Edison 4 Irving Place New York, NY 10003 R. Stack Corn Belt Electric Cooperative P.O. Box 816 Bloomington, IL 61702 J. Michael Hinga Delphi Energy & Engine Management Systems P.O. Box 502650 Indianapolis, IN 46250 Jim Daley Department of Energy 1000 Independence Ave. SW EE-12 FORSTL Washington, DC 20585 Dan T. Ton Department of Energy 1000 Independence Ave. SW EE-11 FORSTL Washington, DC 20585 Dan Sanchez Department of Energy Albuquerque Operations Office P.O. Box 5400 Albuquerque, NM 87185 Alex O. Bulawka Department of Energy 1000 Independence Ave. SW EE-11 FORSTL Washington, DC 20585 J. A. Mazer Department of Energy 1000 Independence Ave. SW EE-11 FORSTL Washington, DC 20585 Richard J. King Department of Energy 1000 Independence Ave. SW EE-11 FORSTL, SH-095 Washington, DC 20585 James E. Rannels Department of Energy 1000 Independence Ave. SW EE-11 FORSTL Washington, DC 20585-0121 Mark B. Ginsberg Department of Energy 1000 Independence Ave. SW EE-90 FORSTL SE-052 Washington, DC 20585 Tien Q. Duong Department of Energy 1000 Independence Ave. SW EE-32 FORSTL, Rm. 5G-030 Washington, DC 20585 J. P. Archibald Department of Energy 1000 Independence Ave. SW EE-90 FORSTL Washington, DC 20585 Russ Eaton Department of Energy Golden Field Office 1617 Cole Blvd., Bldg. 17 Golden, CO 80401 Kenneth L. Heitner Department of Energy 1000 Independence Ave. SW EE-32 FORSTL, Rm. 5G-030 Washington, DC 20585 Allan Jelacic Department of Energy 1000 Independence Ave. SW EE-12 FORSTL Washington, DC 20585 Alex G. Crawley Department of Energy 1000 Independence Ave. SW EE-90 FORSTL Washington, DC 20585 Benjamin A. Finzel Department of Energy 1000 Independence Ave. SW EE-1 FORSTL Washington, DC 20585 Joe Galdo Department of Energy 1000 Independence Ave. SW EE-10 FORSTL Washington, DC 20585 Neal Rossmeissl Department of Energy 1000 Independence Ave. SW EE-13 FORSTL Washington, DC 20585 R. Eynon Department of Energy 1000 Independence Ave. SW EI-821 FORSTL Washington, DC 20585 Philip N. Overholt Department of Energy 1000 Independence Ave. SW EE-11 FORSTL Washington, DC 20585-0121 Jack Cadogan Department of Energy 1000 Independence Ave. SW EE-11 FORSTL Washington, DC 20585 Paul Maupin Department of Energy 19901 Germantown Rd ER-14 E-422 Germantown, MD 20874-1290 Dr. Imre Gyuk Department of Energy 1000 Independence Ave. SW EE-14 FORSTL Washington, DC 20585 Dan W. Reicher Department of Energy 1000 Independence Ave. SW EE-1 FORSTL Washington, DC 20585 Robert K. Dixon Department of Energy 1000 Independence Ave. SW EE-10 FORSTL Washington, DC 20585F Gary D. Burch Department of Energy 1000 Independence Ave SW EE-11 FORSTL Washington, DC 20585 Sigmund Gronich Department of Energy 1000 Independence Ave. SW EE-15 FORSTL Washington, DC 20585 Tina M. Kaarsberg Department of Energy 1000 Independence Ave SW EE-10 FORSTL Washington, DC 20585 Peter R. Goldman Department of Energy 1000 Independence Ave. SW EE-12 FORSTL Washington, DC 20585 William P. Parks, Jr. Department of Energy 1000 Independence Ave. SW EE-10 FORSTL Washington, DC 20585 Lynne M. Gillette Department of Energy 1000 Independence Ave. SW EE-11 FORSTL Washington, DC 20585 Patricia A. Hoffman Department of Energy 1000 Independence Ave. SW EE-22 FORSTL Washington, DC 20585 Joseph J. lannucci Distributed Utility Associates 1062 Concannon Blvd. Livermore, CA 94550 Alan Collinson EA Technology Limited Chester CH1 6ES Capenhurst, England UNITED KINGDOM M. Stanton East Penn Manufact. Co., Inc. Deka Road Lyon Station, PA 19536 Steve Eckroad Electric Power Research Institute P.O. Box 10412 Palo Alto, CA 94303-0813 Steve Chapel Electric Power Research Institute P.O. Box 10412 Palo Alto, CA 94303-0813 Steven Whisenant Duke Power Company P.O. Box 1006 Charlotte, NC 28201 Jim DeGruson Eagle-Picher Industries. Inc. C & Porter Street Joplin, MO 64802 Daniel R. Bruck ECG Consulting Group, Inc. 55-6 Woodlake Road Albany, NY 12203 Robert Schainker Electric Power Research Institute P.O. Box 10412 Palo Alto, CA 94303-0813 Phillip C. Symons Electrochemical Engineering Consultants, Inc. 1295 Kelly Park Circle Morgan Hill, CA 95037 Dave Feder Electrochemical Energy Storage Systems, Inc. 35 Ridgedale Avenue Madison, NJ 07940 Harald Haegermark Elforsk-Swedish Elec Utilities R&D Co Elforsk AB Stockholm, S-101 53 Sweden Mindi J. Farber De Anda Energetics, Inc. 501 School Street SW Suite 500 Washington, DC 20024 Rich Scheer Energetics, Inc. 501 School Street SW Suite 500 Washington, DC 20024 Ed Skolnik Energetics, Inc. 501 School Street SW Suite 500 Washington, DC 20024 Michael Dodge Electrosource P.O. Box 7115 Loveland, CO 80537 Chuck Whitaker Endecon Engineering 2500 Old Crow Canyon Road Suite 220 San Ramon, CA 94583 Howard Lowitt Energetics, Inc. 7164 Gateway Drive Columbia, MD 21046 Laura C. Johnson Energetics, Inc. 7164 Gateway Drive Columbia, MD 21046 Greg J. Ball Energy & Env. Economics, Inc. 353 Sacramento Street Suite 1540 San Francisco, CA 94111 Amber Gray-Fenner Energy Communications Consulting 7204 Marigot Rd. NW Albuquerque, NM 87120 Robert Duval EnerVision P,O, Box 450789 Atlanta, GA 31145-0789 Erik Hennig EUS GmbH Munscheidst 14 Gelsenkirchen, D-45886 GERMANY J. Mills Firing Circuits, Inc. P.O. Box 2007 Norwalk, CT 06852-2007 Steven Kraft Frost & Sullivan 2525 Charleston Road Mountain View, CA 94043 Dale Butler EnerTec Pty. Ltd. 349 Coronation Drive PO Box 1139, Milton BC Old 4044 Auchenflower, Queensland, 4066 AUSTRALIA David H. DaCosta Ergenics, Inc. 247 Margaret King Avenue Ringwood, NJ 07456 John Breckenridge Exide Electronics 8609 Six Forks Road Raleigh, NC 27615 James P. Dunlop Florida Solar Energy Center 1679 Clearlake Road Cocoa, FL 32922-5703 Dave Coleman Frost & Sullivan 2525 Charleston Road Mountain View, CA 94043 Bob Zrebiec GE Industrial & Pwr. Services 640 Freedom Business Center King of Prussia, PA 19046 Declan Daly General Electric Drive Systems 1501 Roanoke Blvd. Salem, VA 24153 Anthony B. LaConti Giner, Inc. 14 Spring Street Waltham, MA 02451-4497 Joe Szymborski GNB Tech. Ind. Battery Co. Woodlake Corporate Park 829 Parkview Blvd. Lombard, IL 60148-3249 John Boehm GNB Tech. Ind. Battery Co. Woodlake Corporate Park 829 Parkview Blvd. Lombard, IL 60148-3249 Nick Miller General Electric Company 1 River Road Building 2, Room 605 Schenectady, NY 12345 Gerry Woolf Gerry Woolf Associates 17 Westmeston Avenue Rottingdean, East Sussex, BN2 8AL UNITED KINGDOM George Hunt GNB Tech. Ind. Battery Co. Woodlake Corporate Park 829 Parkview Blvd. Lombard, IL 60148-3249 Sanjay Deshpande’ GNB Technologies Woodlake Corporate Park 829 Parkview Blvd. Lombard, IL 60148-3249 Steven Haagensen Golden Valley Elec. Assoc., Inc. 758 Illinois Street P.O. Box 71249 Fairbanks, AK 99701 Ben Norris Gridwise Engineering Company 121 Starlight Place Danville, CA 94526 George H. Nolin HL&P Energy Services P.O. Box 4300 Houston, TX 77210-4300 Patrick Moseley ILZRO 2525 Meridian Parkway P.O. Box 12036 Research Triangle Park, NC 27709 Ron Myers Imperial Oil Research Centre 3535 Research Road NW Room 2E-123 Calgary, Alberta, T2L 2K8 CANADA Albert R. Landgrebe Int'l Electrochemical Sys & Technology B14 Sussex Lane Long Neck, DE 19966 Clyde Nagata Hawaii Electric Light Co. P.O. Box 1027 Hilo, HI 96720 Carl D. Parker ILZRO 2525 Meridian Parkway P.O. Box 12036 Research Triangle Park, NC 27709 Jerome F. Cole ILZRO 2525 Meridian Parkway PO Box 12036 Research Triangle Park, NC 27709 Ken Belfer Innovative Power Sources 1419 Via Jon Jose Road Alamo, CA 94507 David Warar Intercon Limited 6865 Lincoln Avenue Lincolnwood, IL 60646 A. Kamal Kalafala Intermagnetics General Corp. 450 Old Niskayuna Road P.O. Box 461 Latham, NY 12110-0461 Imelda G. Francis Just Do IT 21658 W. Ravine Lake Zurich, IL 60047 Elton Cairns Lawrence Berkeley Nat'l Lab University of California One Cyclotron Road Berkeley, CA 94720 Kim Kinoshita Lawrence Berkeley Nat'l Lab University of California One Cyclotron Road Berkeley, CA 94720 Susan M. Schoenung Longitude 122 West, Inc. 1010 Doyle Street Suite 10 Menlo Park, CA 94025 John Neal Internation! Business & Tech. Services, Inc. 9220 Tayloes Neck Road Nanjemoy, MD 20662 Gerard H. C. M. Thijssen KEMA T&D Power Utrechtseweg 310 P.O. Box 9035 ET, Ernhem, 6800 The Netherlands Frank McLarnon Lawrence Berkeley National Lab University of California One Cyclotron Road Berkeley, CA 94720 J. Ray Smith Lawrence Livermore Nat'l Lab University of California P.O. Box 808, L-641 Livermore, CA 94551 Joseph Morabito Lucent Technologies, Inc. 600 Mountain View Ave. P.O. Box 636 Murray Hill, NJ 07974-0636 Cecilia Y. Mak Lucent Technologies 3000 Skyline Drive Room 855 Mesquite, TX 75149-1802 Floyd Russell Metlakatla Power & Light P.O. Box 359 3.5 Mile Airport Road Metlakatla, AK 99926 Anthony Price National Power PLC Harwell Int'l Business Ctr. Harwell, Didcot, OX11 0QA UNITED KINGDOM Bill Brooks NC Solar Center Corner of Gorman & Western Box 7401 NCSU Raleigh, NC 27695-740 Bart Chezar New York Power Authority 1633 Broadway New York, NY 10019 Stephen R. Connors Massachusetts Inst of Tech The Energy Laboratory Rm E40-465 Cambridge, MA 02139-4307 D. Nowack Micron Corporation 158 Orchard Lane Winchester, TN 37398 Steven P. Lindenberg National Rural Elec Cooperative Assoc. 4301 Wilson Blvd. SSER9-207 Arlington, VA 22203-1860 Andrew L. Rosenthal New Mexico State University Southwest Tech. Dev. Institute Box 30001/Dept. 3SOL Las Cruces, NM 88003-8001 Denise Zurn Northern States Power Co. 414 Nicollet Mall Minneapolis, MN 55401 Jack Brown NPA Technology Two University Place Suite 700 Durham, NC 27707 Robert Hawsey Oak Ridge National Laboratory P.O. Box 2008 Bldg. 3025, MS-6040 Oak Ridge, TN 37831-6040 Brendan Kirby Oak Ridge National Laboratory P.O. Box 2008 Bldg. 3147, MS-6070 Oak Ridge, TN 37831-6070 John DeStreese Pacific Northwest Nat'l Lab Battelle Blvd. P.O. Box 999, K5-02 Richland, WA 99352 Thomas H. Schucan Paul Scherrer Institut CH - 5232 Villigen PSI Switzerland John Stoval Oak Ridge National Laboratory P.O. Box 2008 Bldg. 3147, MS-6070 Oak Ridge, TN 37831-6070 James VanCoevering Oak Ridge National Laboratory P.O. Box 2008 Bldg. 3147, MS-6070 Oak Ridge, TN 37831-6070 Douglas R. Danley Orion Energy Corporation 10087 Tyler Place #5 Tjamsville, MD 21754 Daryl Brown Pacific Northwest Nat'l Lab Battelle Blvd. MS K8-07 P.O. Box 999 Richland, WA 99352 Brad Johnson PEPCO 1900 Pennsylvania NW Washington, DC 20068 Hans Meyer Power Electronics Div S&C Electric 2010 Energy Drive P.O. Box 879 East Troy, WI 53120 P. Prabhakara Power Technologies, Inc. 1482 Erie Blvd. P.O. Box 1058 Schenectady, NY 12301 Reznor I. Orr PowerCell Corporation 99 South Bedford Street Suite 2 Burlington, MA 01803 Roger Flynn Public Service Co. of New Mexico Alvarado Square MS-2838 Albuquerque, NM 87158 Norman Lindsay Queensland Department of Mines and Energy G.P.O. Box 194 Brisbane, 4001 QLD. AUSTRALIA Stan Sostrom POWER Engineers, Inc. P.O. Box 777 3870 US Hwy 16 Newcastle, WY 82701 Rick Winter Powercell Corporation 6724 Preston Avenue Suite C Livermore, CA 94550 Jerry Neal Public Service Co. of New Mexico Alvarado Square MS-BA52 Albuquerque, NM 87158 Wenceslao Torres Puerto Rico Elec. Pwr. Authority P.O. Box 364267 San Juan, PR 00936-4267 J. Thompson R&D Associates 2100 Washington Blvd. Arlington, VA 22204-5706 Al Randall Raytheon Eng. & Constructors 700 South Ash Street P.O. Box 5888 Denver, CO 80217 George V. Fantozzi S&C Electric Company 6601 North Ridge Blvd. Chicago, IL 60626-3997 Guy Chagnon SAFT Research & Dev. Ctr. 107 Beaver Court Cockeysville, MD 21030 G. E. "Ernie" Palomino Salt River Project P.O. Box 52025 MS PAB 357 Phoenix, AZ 85072-2025 Robert Reeves Sentech, Inc. 9 Eaton Road Troy, NY 12180 K. Ferris RMS Company 135 Post Office Rd. South Salem, NY 10590-1106 Jim McDowall SAFT America, Inc. 3 Powdered Metal Drive North Haven, CT 06473 Michael C. Saft SAFT Research & Dev. Ctr. 107 Beaver Court Cockeysville, MD 21030 Dr. Charles Feinstein Santa Clara University Dept. of Dec. & Info. Sciences Leavey School of Bus. & Admin. Santa Clara, CA 95053 Kurt W. Klunder Sentech, Inc. 4733 Bethesda Avenue Suite 608 Bethesda, MD 20814 Rajat K. Sen Sentech, Inc. 4733 Bethesda Avenue Suite 608 Bethesda, MD 20814 Clay Aldrich Siemens Solar 4650 Adohr Lane P.O. Box 6032 Camarillo, CA 93011 Scott Sklar Solar Energy Ind. Assoc. (SEIA) 1111 North 19th St Suite 2604th Floor Arlington, VA 22209 K. Vakhshoorzadeh Southern Company Services, Inc. 600 North 18th Street P.O. Box 2625 Birmingham, AL 35202-2625 James T. Collins Southern Electric Exchange 3379 Peachtree Road NE Suite 250 Atlanta, GA 30326 Nicole Miller Sentech, Inc. 4733 Bethesda Avenue Suite 608 Bethesda, MD 20814 Deepak Divan Soft Switching Technologies 2224 Evergreen Road Suite 6 Middleton, WI 53562 Naum Pinsky Southern California Edison 2244 Walnut Grove Ave. P.O. Box 800, Room 418 Rosemead, CA 91770 Richard N. Schweinberg Southern California Edison 6070 N. Irwindale Avenue Suite I Irwindale, CA 91702 C. Seitz SRI International 333 Ravenswood Avenue Menlo Park, CA 94025 George Zink Stored Energy Engineering 7601 E. 88th Place Indianapolis, IN 46256 Jon Hurwitch Switch Technologies 4733 Bethesda Avenue Suite 608 Bethesda, MD 20814 Harold Gotschall Technology Insights 6540 Lusk Blvd. © Suite C-102 San Diego, CA 92121 Haukur Asgeirsson The Detroit Edison Company 2000 2nd Ave. 435 SB Detroit, MI 48226-1279 Michael Orians The Solar Connection P.O. Box 1138 Morro Bay, CA 93443 Bob Bish Stored Energy Engineering 7601 E. 88th Place Indianapolis, IN 46256 Terri Hensley Tampa Electric Company P.O. Box 111 Tampa, FL 33601-0111 Thomas J. Jenkin The Brattle Group 44 Brattle Street Cambridge, MA 02138-3736 Charles E. Bakis The Pennsylvania State University 227 Hammond Building University Park, PA 16802 Tom Anyos The Technology Group, Inc. 63 Linden Avenue Atherton, CA 94027-2161 Bill Roppenecker Trace Engineering Division 5916 195th Northeast Arlington, WA 98223 Michael Behnke Trace Technologies 161G South Vasco Road P.O. Box 5049 Livermore, CA 94550 Jim Drizos Trojan Battery Company 12380 Clark Street Santa Fe Springs, CA 90670 Paul C. Klimas U.S. Agency for Intn'l Development Center for Environment Washington, DC 20523-3800 Henry Chase U.S. Flywheel Systems 1125-A Business Center Circle Newbury Park, CA 91320 Bill Erdman Trace Technologies 161G South Vasco Road P.O. Box 5049 Livermore, CA 94550 Donald A. Bender Trinity Flywheel Power 6724D Preston Avenue Livermore, CA 94550 James Fangue TU Electric R&D Programs P.O. Box 970 Fort Worth, TX 76101 Dr. Gerald P. Ceasar U.S. Department of Commerce NIST/ATP Bldg 101, Room 623 Gaithersburg, MD 20899 Wayne Taylor U.S. Navy Code 83B000D, NAWS China Lake, CA 93555 Edward Beardsworth UFTO 951 Lincoln Avenue Palo Alto, CA 94301-3041 John Herbst University of Texas at Austin J.J. Pickel Research Campus Mail Code R7000 Austin, TX 78712 Mariesa L. Crow University of Missouri-Rolla 233 EECH Rolla, MO 65409-0040 Marnell Hoover US Air Force OO-ALC/LIPC 6040 Gum Hill Air Force Base, UT 84056-5825 Mike Stern Utility Power Group 21250 Califa Street Suite 111 Woodland Hills, CA 91367-5029 Richard Lemmons United States Air Force OO-ALC/LIPC 6040 Gum Lane Hill Air Force Base, UT 84056-5825 Max Anderson University of Missouri - Rolla 112 Electrical Eng. Bldg. Rolla, MO 65401-0249 G. Alan Palin Urenco (Capenhurst) Ltd. Capenhurst, Chester, CH1 6ER UNITED KINGDOM Steve Hester Utility Photo Voltaic Group 1800 M Street NW Washington, DC 20036-5802 Rick Ubaldi VEDCO Energy 12 Agatha Lane Wayne, NJ 07470 Gary Verno Virginia Power Innsbrook Technical Center 5000 Dominion Blvd. Glen Ellen, VA 23233 Randy Bevin Walt Disney World Design and Eng'g P.O. Box 10,000 Lake Buena Vista, FL 32830-1000 Howard Saunders Westinghouse STC 1310 Beulah Road Pittsburgh, PA 15235 Nicholas J. Magnani Yuasa, Inc. 2366 Bernville Road P.O. Box 14145 Reading, PA 19612-4145 R. Kristiansen Yuasa-Exide, Inc. 35 Loch Lomond Lane Middleton, NY 10941-1421 Alex Q. Huang Virginia Polytechnic Instit. & State Uni Virginia Power Electronics Center 672 Whittemore Hall Blacksburg, VA 24061 Gerald J. Keane Westinghouse Elec. Corp. Energy Management Division 4400 Alafaya Trail Orlando, FL 32826-2399 Frank Tarantino Yuasa, Inc. 2366 Bernville Road P.O. Box 14145 Reading, PA 19612-4145 Gene Cook Yuasa-Exide, Inc. 666 Edison Furlong Road Furlong, PA 18925 Henry W. Zaininger Zaininger Engineering Co., Inc. 9959 Granite Crest Court Granite Bay, CA 95746 Robert J. Parry ZBB Technologies, Inc. 11607 West Dearbourn Ave. Wauwatosa, WI 53226-3961 MS-0340/01832 MS-0953/02500 MS-0953/02501 MS-0613/02521 MS-0613/02521 MS-0613/02521 MS-0613/02522 MS-0614/02523 MS-0613/02525 MS-0613/02525 MS-0613/02525 MS-0613/02525 MS-0613/02525 MS-0613/02525 MS-0899/09616 MS-0741/06200 MS-0704/06201 MS-0455/06201 MS-0708/06214 MS-0753/06218 MS-0753/06218 MS-0753/06218 MS-0753/06218 MS-0753/06218 MS-9403/08723 MS-9018/08940-2 MS-0612/09612 Jeff W. Braithwaite William E. Alzheimer J. Thomas Cutchen Daniel H. Doughty Terry M. Unkelhaeuser Rudolph G. Jungst Paul C. Butler Robert W. Bickes, Jr. Garth P. Corey Gustavo P. Rodriguez James T. Crow Nancy H. Clark Stanley Atcitty John D. Boyes (10) Technical Library (2) Samuel G. Varnado Abbas A. Akhil Marjorie L. Tatro Henry M. Dodd Christopher P. Cameron Russell H. Bonn Thomas D. Hund John W. Stevens Ward I. Bower James C. F. Wang Central Technical Files Review & Approval For DOE/OSTI (1)