Loading...
HomeMy WebLinkAboutKotzebue Electric Assoc Wind Power Project Second-Year Operating Exp 2000-2001N/A | Min Dewaw crrel Kotzebue Electric Association Wind Power Project Second-Year Operating Experience: 2000-2001 U.S. Department of Energy-EPRI Wind Turbine Verification Program Technical Report Kotzebue Electric Association Wind Power Project Second-Year Operating Experience: 2000-2001 U.S. Department of Energy-EPRI Wind Turbine Verification Program 1004040 Final Report, December 2001 Cosponsers U.S. Department of Energy Kotzebue Electric Association Alaska Energy Authority Alaska Industrial Export Authority EPRI Project Manager C. McGowin EPRI ¢ 3412 Hillview Avenue, Palo Alto, California 94304 * PO Box 10412, Palo Alto, California 94303 * USA 800.313.3774 * 650.855.2121 * askepri@epri.com * www.epri.com DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM: (A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (Ill) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR (B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT. ORGANIZATION(S) THAT PREPARED THIS DOCUMENT Global Energy Concept, LLC ORDERING INFORMATION Requests for copies of this report should be directed to EPRI Customer Fulfillment, 1355 Willow Way, Suite 278, Concord, CA 94520, (800) 313-3774, press 2. Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power Research Institute, Inc. Copyright © 2001 Electric Power Research Institute, Inc. All rights reserved. CITATIONS This report was prepared by Global Energy Concepts, LLC 5729 Lakeview Drive, Street 100 Kirkland, WA 98033 Principal Investigator R. Vilhauer This report describes research sponsored by EPRI, the U.S. Department of Energy, Kotzebue Electric Association, and the Alaska Energy Authority, Alaska Industrial Export Authority. The report is a corporate document that should be cited in the literature in the following manner: Kotzebue Electric Association Wind Power Project Second-Year Operating Experience: 2000- 2001: U.S. Department of Energy-EPRI Wind Turbine Verification Program, EPRI, Palo Alto, CA and the U.S. Department of Energy, Washington, DC and Kotzebue Electric Association, Kotzebue, Alaska Energy Authority and Alaska Industrial Export Authority: 2001. 1004040. iii REPORT SUMMARY This report describes the second-year operating experience at the 0.66-MW Kotzebue Electric Association (KEA) wind power project near Kotzebue, Alaska. Lessons learned in the project will be valuable to other utilities planning similar wind power projects. Background In 1992, EPRI and the U.S. Department of Energy (DOE) initiated the Wind Turbine Verification Program (TVP). Program goals are to help electric utility companies gain field experience with wind power, evaluate early commercial wind turbines at several U.S. sites, and transfer the experience to the wind power and utility communities. The TVP program includes four projects selected through a series of competitive solicitations and three other projects that joined the program as “associate projects.” The associate projects receive limited funding from the sponsors but benefit from the information exchange and technical assistance. KEA joined the Wind Turbine Verification Program in 1997 as the first associate project. The 0.66-MW wind turbine project is owned and operated by KEA at a site near Kotzebue in northwest Alaska. It consists of ten Atlantic Orient Corporation (AOC) 15/50 wind turbines. Each turbine has a 15-m diameter three-bladed rotor and a constant-speed turbine-generator mounted on top of a 24.4-m (80-ft) lattice tower. The first three turbines were installed in 1997, the remaining seven turbines were installed in 1999, and the project was commissioned in June 1999. Development and first-year operating experiences of the project are described in companion reports, EPRI TR-113918 (December 1999) and EPRI 1000957 (December 2000). Future reports will describe additional years of operating experience. Objective To document KEA’s second year of operating experience, to describe experiences gained and problems encountered, and to transfer lessons learned to other utilities planning similar projects. Approach Project investigators documented the second-year operating experience at the KEA project from July 2000 through June 2001. Their report describes the project’s annual performance, its operation and maintenance activities, and KEA’s continuing wind research and outreach activities. Results During the 12-month period, July 2000 through June 2001, the Kotzebue wind facility delivered 1,201,846 kWh of electricity to the Kotzebue distribution system. It operated at a 20.8% average capacity factor based on 660-kW rated capacity. Overall TVP system availability was 95.9%, allowing for all scheduled and forced wind turbine outages. Some of the turbine downtime was due to research and other activities conducted by KEA. Individual turbine availabilities ranged from 79.1% to 99.2%. The second year of operation for the full ten-turbine project at the KEA site was marked by a significant increase in energy output due to higher winds. While winds increased, providing more energy, the project experienced a 44% increase in turbine downtime relative to the first year of operation. The majority of downtime increase can be attributed to an extensive outage of Turbine 3 related to troubleshooting and repairs of an electrical problem. EPRI Perspective Through 2001, EPRI has issued 21 reports on project development and operation for seven DOE- EPRI TVP wind projects located in Alaska, Iowa, Nebraska, Texas, Vermont, and Wisconsin. An important goal of the program is to transfer experiences gained in TVP projects to utilities, wind power developers and turbine vendors, government agencies, and other interested parties so that lessons learned can be incorporated into future projects. This report will be useful because it describes actual operation experiences, which should help others avoid or reduce the impact of similar problems. Future EPRI reports will describe Kotzebue’s third year of operation and initial experiences with additional wind turbines to be installed in early 2002, as well as experiences of other TVP-funded projects. Keywords Wind power Wind resource Performance Availability Operations and maintenance vi ABSTRACT The Wind Turbine Verification Program (TVP) is a collaborative effort of the U.S. Department of Energy, the Electric Power Research Institute, and host utilities to develop, construct, and operate wind power plants. Through their involvement as an associate TVP project, Kotzebue Electric Associate (KEA) has developed, constructed, and is now operating a 0.66 MW wind power plant. The project consists of ten commercial 66 kW wind turbines manufactured by Atlantic Orient Corporation (AOC) of Norwich, Vermont. The AOC 15/50 wind turbines are installed on 24.4-m (80-ft) lattice towers at a site near the town of Kotzebue in northwest Alaska. The first phase of the project, commissioned in September 1997, consists of three turbines. The remaining seven turbines were commissioned in June 1999 and a dedication ceremony was held on August 14, 1999. This report discusses the activities and experience during the second year of operation of the full ten-turbine project. It includes summaries of the wind resource data, actual and projected energy production, and availability at the site during the second year of operation. The report discusses the operation and maintenance activities and categorizes the downtime experienced by the turbines during the period from July 2000 through June 2001. KEA is preparing for the spring 2002 installation of a Northwind 100 kW wind turbine and of two additional AOC 15/50s, bringing the project capacity to 0.89 MW. KEA may eventually increase the installed capacity of the project to as much as 2 to 4 MW. Because KEA is expanding their wind project in 2001-2002 to include additional AOC turbines and a Northern Power Systems wind turbine, additional performance evaluation under the TVP is planned for the future. vii ACKNOWLEDGMENTS A number of individuals provided information and contributed to the production of this report. Valuable input and comments were received from representatives of Kotzebue Electric Association, Atlantic Orient Corporation, the Electric Power Research Institute, the U.S. Department of Energy, and the National Renewable Energy Laboratory. Matt Bergan and Brad Reeve of KEA and Craig Thompson of Thompson Engineering were particularly helpful in providing details and clarification. Rana Vilhauer, Mark Young, Lory Widmer, and Betsie McLain of Global Energy Concepts made significant contributions to the final analysis and format of the document. Other GEC staff members also assisted with various sections of the report. ix LIST OF ABBREVIATIONS AOC ASOS EPRI GEC IEC KEA Met NREL O&M PCE SCADA TVP UWIG WECTEC Atlantic Orient Corporation Automated Surface Observing System Electric Power Research Institute Global Energy Concepts International Electrotechnical Commission Kotzebue Electric Association Meteorological National Renewable Energy Laboratory Operation and Maintenance Power Cost Equalization Supervisory Control and Data Acquisition Turbine Verification Program Utility Wind Interest Group Wind Economics and Technology, Inc. xi CONTENTS 1 INTRODUCTION 1.1 Project Background 1.2 Background on the Wind Turbine Verification Program .............::ccccesceecessessesseeeeeeeeseees 1-3 1.3) Report: Objectives!and !/SCope@ ecsccesccveccseccecscceseccsecesve-ccevececcconcecesevsnseyeseescersetsstacesnsessees 1-4 4.4 Report Organization )tccccccsseccreccssntecescsssetececncsacesssensesesaecteery srecsnsercvarseccecsecransesenescncoweren 1-4 2 WIND RESOURCE CHARACTERISTICS ........ccssssssssssssssssesseeseseessessneseeeseenseeseeneessseseseess 2-1 21) Data Comectionniteccrs-seccterccesescscvcvevoccecstsacecevasesss suease-scosvdcsssserrsnsseanascesesrarcusessestescstaren 2-1 DLO WING!SPOGG Mrcertecressercreccsscecssncerersecccsteccessoncersrencesessartsseanrevsncesionccrersvaaceteatorersoccnecseuss 2-4 QBS WING DITO CHOM cerecceterccesseccccevcccceceaccstenncccscececcecccsscesencsesnecesecanneressnces sesuacsssaueresuatecesacectes 2-8 2.4 Turbulence and Shear.. SB PROUEC TIPEREFORMANGE lcnccescrssscscscccsectaccesereccaseresrescescessvesscecessessencecastaseeserasesssaserecenece 3-1 SH AValADINY cot neccxcesatescrsecssssucserecenaeterecscestatestoresaresasectatersicseacrsscrecascecsececarsesseseasseseresecees 3-1 SSi2 EMONGY PTOGUCHION cesceccesceccsesssosccncescnsecscesenassssersecersacvevessvaesuctesmtssnerescrssacetecsssaceseetseoeses 3-3 3.2.1 Seasonal and Inter-Annual Performance Variations..............::ccsceeeseeeeeeeeeeeeeeeeeeeee 3-4 3.2.2 Utility Meter Readings and On-Site Energy Losses............c::csscesseeseeseseseeeseeeeeeseees 3-6 3:2: 3] PI OJOCLOG) ENGI Gy cersseasnarscassestessnsasaaseassscsncsanssuasueststcatsnsssetsestatsnsaessetsantssrssnterasscd 3-6 3:2'4) Lost Energy Due to|DOWntiMeste.c.cce.cescorccetosseccestcereccrenssa-asscersccestesesccsesacseasesssaees 3-9 3.2.5 Percent Time Generating So lUtiity Demandiand Project ENCrgy circccesscorscucsccsccssccesscocsccssccesacrossccoseresassencsscurssenesancers 3-12 S:4:High: Wind Penetration Study cit icccsraccevsccsecesasescsncoccecsuesotesscasstcseseatevatcosscecactsesacesancers 3-14 3.4:1 Diurnal|LoadsiandiWindiPenetration::.<---.-.0.1.-..-vscescecsseoscececesecesvesssecevecseseacerecaes- 3-14 3.4.2 Variation in Power Quality Parameters..............c:ccceccceseeeseeeseeeeeeeneeeeeeeeneesseeeeaees 3-16 4 PROJECT OPERATIONS AND MAINTENANCE .............cccscessesssessescssseessnessesesenseassenenaseneees 4-1 AMAKEA'S O&M Strategy cccccecresseccrccecccersecerersvecercessccreccocsrsdrecteontecebesece eotneeevenetavtsreneresstses 4-1 4.2 Maintenance Activities and Other Downtime Event................c:ccccccceseeceseceeeeeeeeeseeeeeee 4-1 4.2.1 Downtime Categories 4.2.2 Downtime Due To O&M Activities .. 4233 DOWNTIME DUCT OM GAUNSstetccatccatatscetacccccotererecscscreccsssectscesvcassececceceocecsacescvescrevaresas 4.3 SCADA System Experience -......:<.<....c0-ccscsenseccocrenesercocesencscccesccssacesaneccersorssesesusesunenss 4.4 Potential Performance Improvements and Turbine Testing .............::ccceeseseeeeeeeeees 4-16 AAV Slow-Starting | TUMDINCS<-..-.sr.-ccne-sor=c-c-ees-useaeeesercceseesacsecsuccazcecseacecasatcsceseronecescesesar 4-16 ALALONT UNDING!| OVEN= Fe lOGUCTION sestccersectssansecteseest crete cssarecatcuctorcestescnetersccswccseccesnesserearence 4-16 4.4.3 Turbine Power Performance Tet .............cscscsssssssessressseesceseeeseeeeeenseeseeeeese 5 5 OUTREACH ACTIVITIES AND FUTURE PLANG............cs0008 5.1 Community Education and Outreach Activities 5.2 Technology Transfer and Information Dissemination .............:::sceseeesesseeseeseeeeeseeeeeeeeees 5-1 5.3 Wind Project Planning and Development...............:ccsccssessessesssesseeeseeeeeeeeeeeeeeeeeteeeees 5-2 6 CONCLUSIONS. ....ccscsscsssssssnssssssosensscsesssoscessonsscsecsconcssssesessecsecssannasccscecccsssssvecsenessesnesaesssseess 6-1 A TVP-RELATED DOCUMENTG...........ssscesssssssccscesssssssssssssesssscssssssesnssassnsesseneesssensensenasnasneeens A-1 RPP Reports iercceccecterctetetetetenctonccecscestcostenstatenscoctesccseeceteessaseetseesercccacecectsateeceaesecucceescotocs A-1 NREL/AWEA WindPower Published Papers .............::ccsscsssssceesseeeesecesseeeeeeeeeeeeeesenaeeeeaeeees A-2 thera PROSOUNCOS ersteteccesrcretececcrarectcns ecto ccccncessstotanotetctarsctstatccanecortrssessererenessesttenserncerass A-4 B MONTHLY AVAILABILITY AND PRODUCTION BY TURBINE ..........:c:s:cssssseessereenteneeees B-1 CTVP AVAILABILITY DESCRIPTION.........:c:sccsssssssssssssesssessenstessessessensensesneeseesnesaeseeseeenesees C-1 DSPECIFIC DOWNTIME CAUSES BY TURBINE.............::scsssssessessesssesessenseeseesensnaseneeeeseesees D-1 Xiv LIST OF FIGURES Figure 1-1) Alaska State! Map reccescccsecccu-e-ccssscesaces-ceoenseesosscenaseesunessnasoscovstsvecossacoscursscassesesstarce 1-2 Figure 1-2 Photograph of the KEA Wind Power Plant .............:cs:csscsssessessessessseeeseseeseeaseneeeseees 1-3 Figure '2-1| Location of the Current KEA Met ToWe%)..2<:.2<..0-<-2ces0ceoo-cenncestenece+cccoccceoecseonscossacese 2-2 Figure 2-2 Monthly Wind Pattern at Kotzebue (26.5 M).........c:ccscsscesessesseseeeseeseeseseseseesenseeeeees 2-5 Figure 2-3 Diurnal Wind Pattern at the Kotzebue Wind Project (26.5 Mm) ...........::sscescceseeseeeees 2-6 Figure 2-4 KEA Wind Speed Frequency Distribution — July 2000 to June 2001...............0:00 2-7 Figure 2-5 Annual Wind and Energy Rose — July 2000 to JUN€ 2001 .........ceeceeeeeeseseeseeeeeeeneeee 2-8 Figure 3-1 Average Wind Speed, Availability, and Energy Production by Month..................... 3-5 Figure 3-2 Actual, Calculated, and Long-term Projected Energy — July 2000 to June 2001..... 3-9 Figure 3-3 Actual and Lost Energy by Month — July 2000 to June 2001 ............:ccseeseeeeeeeeeeees 3-11 Figure 3-4 Actual and Lost Energy by Turbine — July 2000 to June 2001 .............scceseeseeeeeeee 3-11 Figure 3-5 Percent Time Generating Power — July 1999 to June 2001.............:cecsceeeseeeeeeees 3-12 Figure 3-6 Projected and Actual Second-Year Wind Energy Contribution to KEA Energy DOMAIN ivcccscsccerseseocsascacctersrersessousevastesessonseccccecesesocersccessseasceesarsvaesuessenassuertescsssenessecsssceed Figure 4-1 Monthly Availability and Wind Speed — July 2000 to June 2001 . Figure 4-2 Total Project Downtime by Cause — July 2000 to June 2001.............cccceseeseeeeeeeeees Figure 4-3 Total Lost Energy by Cause — July 2000 to JUNG 2001..........eeceeeeeeeeeeeeeteneeeeeeees Figure 4-4 Total Project Downtime by Turbine — July 2000 to June 2001 Figure 4-5 Total Project Downtime by Month — July 2000 to June 2001 .............ccceeseeseeeeseeeeeee Figure 4-6 Comparison of Project Downtime by Turbine — July 1999 to June 2001................. 4-7 Figure 4-7 O&M Downtime by Cause — July 2000 to JuNe 2001 .........cecceeceeseeeeeeeeseeeeeeeseeseees 4-7 Figure 4-8 Lost Energy Due to O&M by Cause — July 2000 to June 2001 ...........cceeeeeeeeeeeees 4-8 Figure 4-9 O&M Downtime by Month — July 2000 to JUNC 2001 ...........ecceeeceeessesseeeeeeseeeeeteneees 4-8 Figure 4-10 O&M Downtime by Turbine — July 2000 to June 2001 ............cecceeeeseeeeeeeeeeeseeeeees 4-9 Figure 4-11 Breakdown of Fault Cause by Type — July 2000 to June 2001 ..............ceeeeeeeeeeee 4-12 Figure 4-12 Fault Frequency and Duration by Type — July 2000 to June 2001 ..............:cceeee 4-13 Figure 4-13 Breakdown of Fault Cause by Turbine — July 2000 to June 2001 .............. eee 4-13 Figure 4-14 Peak Output Compared to Minimum Temperatures — July 2000 to June 2001....4-17 XV LIST OF TABLES Table 2-1 Data Recovery Rates for Meteorological Data — July 2000 to June 2001 ................ Table 2-2 Mean Monthly Wind Speeds at Kotzebue (26.5 M).........::cecceeeeceeeeeseeeeeeteeneeeeeeeees Table 2-3 KEA Monthly Turbulence Intensity and Wind Shear — July 2000 to June 2001........ Table 3-1 Energy and Availability by Turbine — July 2000 to June 2001... eee eeeeereeeeeeee Table 3-2 Energy and Availability by Month — July 2000 to June 2001 .0...... eee eeceeeeceneeeetneeeees Table 3-3 Meter Readings and Sum of Turbine Readings — July 2000 to June 2001 ts Table:3=4) Estimated! Energy Losses strc secscscecseroececccerceveceverncssuascessenacesensssetesssetscnarsseccesosressss Table 3-5 Project Characteristics and Long-Term Net Energy Estimates ...............:cseeeeeeeee Table 3-6 Actual and Projected Long-term Energy — July 2000 to June 2001 . 5 Table 3-7 Downtime and Lost Energy by Month — July 2000 to June 2001 .........eeeeceeeeeeeeee Table 3-8 Actual and Projected Average Wind Energy Penetration on KEA System.............. Table 3-9 Summary of Wind Penetration Values — August 21 to September 20, 2000.......... Table 4-1 Downtime Hours by Category — July 1999 to June 2001 ooo... eee ecceeeceeteeeeeeeeeees Table 4-2 O&M Downtime Hours by Category — July 1999 to June 2001 ..0... eee ceeceeeereeeeee Table 4-3 Fault Downtime by Category — July 1999 to June 2001.00... eeeeseeeeeeeeeeeeeeeeees Table 4-4 Recovery Rates for SCADA System Data — July 2000 to June 2001........... cece xvii 7 INTRODUCTION This report is the third in a series of reports documenting the experiences of Kotzebue Electric Association (KEA) in developing, constructing and operating a 0.66 MW wind power plant near Kotzebue, Alaska. The project is part of the Wind Turbine Verification Program (TVP), a collaborative effort of the U.S. Department of Energy (DOE), the Electric Power Research Institute (EPRI), and host utilities to gain experience with utility operation of new wind turbine technology. Additional information on the KEA TVP project is contained in two previous EPRI reports, TR-113918 and 1000957. The first report was published in 1999 and documents the project’s development and initial operation of the Phase | turbines. The second report, published in 2000, documents the project’s first year of operating experience. Extensive background information, redundant to the previous reports, is not repeated in this report unless appropriate or necessary for comparison purposes. 1.1 Project Background The KEA TVP wind power plant is a 0.66 MW facility of small, commercial-scale wind turbines. The project consists of ten AOC 15/50 66 kW wind turbines manufactured by Atlantic Orient Corporation (AOC) of Norwich, Vermont. The turbines are installed on 24.4-m (80-ft) lattice towers on piling foundations, resulting in a hub height of approximately 26.5 m (87 ft). The AOC 15/50 is a three-bladed, downwind turbine with a 15-m (49-ft) rotor diameter. KEA’s project site is located on the tip of the Baldwin Peninsula approximately 42 km (26 mi) north of the Arctic Circle on the northwest coast of Alaska near the town of Kotzebue. With a population of approximately 3,000 residents, Kotzebue is the largest community in Northwest Alaska and serves as the economic, governmental, medical, communication, and transportation hub for the 11 communities in the Northwest Arctic Borough, an area roughly the size of Indiana. Figure 1-1 shows the location of Kotzebue on the Alaska state map. Kotzebue can be accessed only by air or water. The town itself has one paved street, but no roads connect Kotzebue to the surrounding villages. Daily jet service is available from Anchorage, and small aircraft carry passengers and supplies from Kotzebue to the surrounding villages. Informal networks of snow-mobile trails connect villages or remote homes to Kotzebue during the nine months that snow-mobile travel typically is feasible. 1-1 Introduction “NS Is) Canada nay os “Anchorage Bering Seas ? Sp ce i o~ ° mg ° A ¥ esa a Ye ° og wae °° Pacific Ocean Figure 1-1 Alaska State Map The climate in Kotzebue is characterized by long cold winters and short cool summers. The Kotzebue Sound and area rivers begin to freeze in early October, and spring breakup generally occurs in late May or early June. The 148-acre KEA wind project site is located approximately 7.2 km (4.5 mi) south of the town of Kotzebue. The land is owned by the Kikiktagruk Inupiat Corporation and leased to KEA for an initial period of ten years with an extension option of four additional ten-year periods. The wind project is situated on a relatively flat plain of treeless tundra that is well exposed to both the prevailing easterly winter winds and the prevailing westerly summer winds. Figure 1-2 is a photograph of the wind power project site. The project was installed in multiple phases at a single site over a period of approximately two years. The phases are defined by their funding sources. All seven of the Phase 2 and 3 turbines were installed in the spring of 1999 and commissioned in June 1999. Phase 1 consists of the first three turbines, installed July 1997 and commissioned September 1997. The equipment procurement, project design, turbine installation, and commissioning for Phases 2 and 3 occurred concurrently. This report covers the experience during the full project’s second year of operation, July 2000 through June 2001. 1-2 Introduction (photo courtesy of Brad Reeve, KEA) Figure 1-2 Photograph of the KEA Wind Power Plant 1.2 Background on the Wind Turbine Verification Program The objective of the TVP is to provide a bridge between the wind turbine development programs currently underway in the United States and utility purchases and evaluation of commercial, utility-grade wind turbines. The TVP is intended to assist utilities in learning about wind power through first-hand experience and to build, test, and operate enough new wind turbines to gain statistically significant performance data. A further objective of the TVP is to provide other utilities with information about wind technology and the operation of a wind power plant from the perspective of a utility owner and operator. EPRI manages the TVP program on behalf of the funding organizations and publishes periodic reports to document the experience of each TVP project. Appendix A lists the TVP reports published by the end of 2001. EPRI and DOE, through its National Renewable Energy Laboratory (NREL), also provide valuable technical and management assistance to the host utilities. The TVP was implemented in several phases. In 1994, Central and South West Services (CSW) and Green Mountain Power Corporation (GMP) were chosen by competitive solicitation to host the first two TVP projects. EPRI and DOE awarded contracts to cover a portion of the costs associated with the selected projects based on a number of criteria that demonstrated their ability Introduction to help commercialize state-of-the-art wind technology. The projects also were required to be a minimum of 6 MW and use turbines with a substantial U.S. manufacturing content. In 1996, TVP released a solicitation that focused on distributed wind generation projects. The selection criteria required that each project be connected directly to a distribution line, consist of at least two wind turbines, and be less than 5 MW in nameplate rating. The selected projects are each owned by a consortium of utilities. One project is located in Iowa and the other in Nebraska. In addition to the projects chosen through the TVP solicitations, three utility wind projects were incorporated into the TVP as “associate projects.” These projects receive limited funding from the program but benefit from the information exchange and technical assistance. In return, the program sponsors receive performance data and other valuable information. In addition to KEA, associate TVP projects include the Low Wind Speed Turbine Project in Wisconsin and the Big Spring Wind Power Plant in Texas. In 2001, the eighth TVP project, and fourth associate project, was selected through a TVP solicitation. It is the 2.0 MW wind project installed by the Tennessee Valley Authority on Buffalo Mountain in northeastern Tennessee. 1.3 Report Objectives and Scope This report focuses on the second year of operation of all ten of KEA’s AOC 15/50 wind turbines. The report discusses the project’s performance, operation and maintenance activities as well as KEA’s outreach activities and future plans. Additional KEA project reports are planned to describe the continuing performance in the future. The principal objective of this report is to summarize the KEA TVP project experience, including performance characteristics, wind resource data, operating strategy, maintenance activities, research projects, and other significant events that occurred during the reporting period. 1.4 Report Organization The report consists of six sections. Following the introduction, Section 2 describes the wind resource characteristics at the site. Section 3 discusses the project performance in terms of energy output and availability. Section 4 provides additional details on the operation and maintenance activities. Section 5 is an overview of KEA’s outreach activities and future plans. Section 6 summarizes the conclusions and experience gained during the project’s second year of operation. 1-4 2 WIND RESOURCE CHARACTERISTICS KEA has collected wind resource data for nine years from numerous locations in the Kotzebue area. The purpose of the initial installation of monitoring equipment was to establish the general wind characteristics of the area. More sophisticated monitoring equipment was installed as KEA progressed towards the development of its wind project. The wind resource data are currently being collected from a hub-height meteorological (met) tower through the Supervisory Control and Data Acquisition (SCADA) system installed in 1999. The following summarizes KEA’s wind monitoring program and the on-site wind resource during the reporting period from July 2000 through June 2001. 2.1 Data Collection In 1992, KEA purchased monitoring equipment from NRG Systems of Vermont and installed it on a roof-mounted tower on a transmitter building for an existing communication tower near the project site. Data collection continued at this site through 1996. In the summer of 1995, KEA installed a 33-m (110-ft) met tower on site and began data collection in August 1995 at heights of 19.5 m (65 ft) and 33 m (110 ft). This data collection effort suffered from marginal data recovery during its first few years. When the first three turbines were installed in 1997, construction activities further impacted the data. In addition, the completed turbine configuration created a wake impact which had an effect on the met data for winds from certain direction sectors. In August 1998, the met tower was relocated to a site approximately one rotor diameter upwind of the first row of turbines so that it could be used for performance evaluation purposes. Additional sensors were added to the tower and data were collected at 10 m, 20 m, and 30 m (33, 66, and 98 ft). However, when the next seven AOC turbines were installed upwind of the first three turbines, they created wake impacts and the met tower data were no longer representative of the wind conditions at the site. In the summer of 1999, a second met tower was installed upwind of the ten-turbine project layout. This met tower is approximately two rotor diameters upwind of Turbine 8 and serves as the primary source of wind resource data for this report. Figure 2-1 shows the location of the met tower relative to the turbines and the surrounding terrain. Sensors are installed at 10 m, 19 m, and 26.5 m (33, 62, and 87 ft). The met data are recorded by the Second Wind SCADA system which was commissioned in late August 1999. Data from this met tower were also used to conduct power performance tests on Turbine 8 during the ten-turbine project’s first year of operation. 2-1 Wind Resource Characteristics 3D 2000 3000 4000 5000 FEET 800 1000 METERS Figure 2-1 Location of the Current KEA Met Tower 2-2 Wind Resource Characteristics Wind data are also available from the Kotzebue Airport, which is located approximately 6.4 km (4 mi) northwest of the project site and serves as a long-term reference for the wind resource in the area. In 1999, KEA hired a consultant, Wind Economics & Technology, Inc. (WECTEC), to summarize the wind resource data collected in Kotzebue and to make long-term energy estimates for the site. WECTEC evaluated concurrent data from the original site met tower and the airport to compare the sites. WECTEC then developed a long-term estimated wind speed for the project site based on hourly data from the airport. The airport data, the long-term wind resource estimate, and related energy estimates are discussed in more detail in the two previous EPRI reports for the KEA wind project. A complete met data set, based on the Second Wind SCADA data, was compiled for the project site for the reporting period of July 2000 through June 2001. Table 2-1 presents the data recovery rates for the on-site wind speed anemometers at all sensor heights. The SCADA system collects and records wind speed, wind direction, temperature and pressure data every 10 minutes. During the second year of operation, the annual wind speed data recovery rate for the 26.5-m sensor was approximately 94% compared to a recovery rate of 97% the previous year. Table 2-1 Data Recovery Rates for Meteorological Data — July 2000 to June 2001 Month 10m 19m 26.5 m July 69.5% 69.5% 65.8% August 88.5% 88.5% 88.5% September 100.0% 100.0% 100.0% October 100.0% 100.0% 100.0% November 100.0% 100.0% 100.0% December 83.3% 83.3% 82.3% January 100.0% 100.0% 100.0% February 100.0% 100.0% 100.0% March 100.0% 100.0% 100.0% April 99.9% 99.9% 99.9% May 96.9% 96.9% 96.9% June 96.3% 96.3% 96.3% Annual Average 94.4% 94.4% 94.0% Data recovery was lower than usual in July and August due to a computer malfunction. During that time period, the site computer malfunctioned if it was not rebooted every few days. Because KEA personnel are not located on-site and were busy working on other activities during the month, they were not able to visit the site solely to reboot the computer. This problem was 2-3 Wind Resource Characteristics partially resolved in October when a Second Wind software upgrade was installed on the Supervisory computer. The functionality of the computer system was further improved in April 2001 when the computer hard disk was defragmented. Hard disk defragmenting as well as database archiving have since been incorporated as routine computer maintenance to ensure more reliable operation of the SCADA system. For the second year of operation, GEC replaced missing wind speed data for the 26.5-m sensor according to the following methods. When wind speed data at the 26.5-m height were missing and accurate data were available at a lower sensor height, data from the lower sensor were adjusted to represent the 26.5-m wind speed based on the relationships between the lower and upper sensors that were determined during periods of concurrent valid data. For short periods of missing data lasting up to one hour, an average of the 26.5-m wind speeds before and after the missing data was used. For limited time periods when accurate data were not available from any of the sensor heights, data were reconstructed for the 26.5-m wind speed based on a correlation between the on-site sensors and the daily average wind speed recorded at the Kotzebue Airport. For months when data were replaced using this method, GEC obtained correlation coefficient values ranging .88 to .95. A correlation factor of 1.0 indicates perfect correlation. Thus, these high correlation coefficients give confidence to data replacement methods using Kotzebue Airport data as they _ suggest a strong relationship between airport and site wind speeds. The following sections present an overview of the wind characteristics at the site during the second year of operation and include comparisons to the first year data and the estimated long-term wind resource. Additional information on performance trends and energy output is included in a later section of the report. 2.2 Wind Speed Table 2-2 compares the monthly and annual 26.5-m wind speeds for the first two years of operation, July 1999 through June 2001, to the estimated long-term site wind speed at 26.5 m. The average annual wind speed at the site during the second year of operation was 6.5 m/s (14.5 mph), approximately 7% higher than the estimated long-term average. The annual wind speed during the first year was 5.1 m/s (11.4 mph), a value 16% lower than the long-term estimated average wind speed of 6.1 m/s (13.6 mph). Figure 2-2 provides a graphical comparison of the wind resource at the site during the year prior to the full ten-turbine installation, the project’s first and second operating years, and the long- term estimated wind resource. This graph illustrates the inter-annual and seasonal variation in wind speed at the KEA wind site. Table 2-2 Mean Monthly Wind Speeds at Kotzebue (26.5 m) Wind Resource Characteristics atonthy 7/00-6/01 7/99-6/00 Long-term m/s (mph) m/s (mph) m/s (mph) July 5.1 (11.4) 5.5 (12.4) 5.8 (12.9) August 6.4 (14.2) 5.3 (11.8) 6.5 (14.4) September 6.4 (14.3) 8.1 (11.4) 6.4 (14.2) October 5.2 (11.7) 48 (10.7) 6.6 (14.7) November 78 (17.6) 45 (10.1) 7.1 (15.9) December 8.7 (19.5) 4.2 (9.3) 6.2 (13.9) January 7.0 (15.6) 5.3 (11.8) 6.3 (14.1) February 9.5 (21.2) Ta (16.1) 6.7 (14.9) March 6.5 (14.5) 6.5 (14.5) 5.5 (12.3) April 6.6 (14.8) 49 (11.0) 5.3 (11.9) May 4.7 (10.4) 4.4 (9.9) 5.3 (11.8) June 4.3 (9.7) 3.7 (8.2) 6.0 (13.3) Annual Average 6.5 (14.6) 5.1 (11.4) 6.1 (13.6) 10 io & — co 6 3 a 3 4 & > 2 0 J Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun 7/00-6/01 —O— 7/99-6/00 —®—7/98-6/99 = = ™ Long-term Figure 2-2 Monthly Wind Pattern at Kotzebue (26.5 m) 2-5 Wind Resource Characteristics As the figure indicates, the site experienced more variation in average monthly wind speed during the project’s second operating year than it did during the two previous years. The long- term estimated monthly wind speed pattern predicts significantly less monthly variation than that which actually occurred between July 1998 and June 2001. As shown in Table 2-2 and Figure 2-2, the average monthly wind speeds from November 2000 through April 2001 were significantly higher than the long-term wind speeds for those same months. These higher-than- average monthly wind speeds were largely due to an increase in the number of cyclonic storms. The figure also confirms that site wind speeds during the previous year (July 1999 to June 2000) were unusually low. Figure 2-3 compares the diurnal wind resource pattern during the project’s second year of operation to the long-term pattern at the project site. The wind speeds were higher during this reporting period than the estimated long-term wind speeds, but their diurnal patterns are fairly similar. However, during the second year the annual diurnal pattern peaked in the evening, whereas the long-term estimated pattern peaks earlier, at around 1:00 p.m. In general, the KEA site does not exhibit a significant variation in diurnal wind speed compared to other sites. 8 a —E 7 | 3 a ag aa Qe eee 222 swe + no = = 5] | 4+— SS 123 45 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour of the Day = 7/00-6/01 * “ _™ Estimated Long-term Figure 2-3 Diurnal Wind Pattern at the Kotzebue Wind Project (26.5 m) Figure 2-4 presents the wind speed frequency distribution at Kotzebue for July 2000 through June 2001 in tabular and graphical form. The plot shows the measured wind speed frequency distribution together with the Rayleigh distribution, which is often used to provide a simple estimate of the wind speed distribution. Using the Rayleigh distribution and the average wind speed to predict the energy generation for the second year of operation results in a fairly accurate prediction of energy generation. (It over-predicts the actual distribution by only 2%.) While the actual and Rayleigh frequency distributions appear dissimilar for wind speeds below 5 m/s (11.1 mph), the Rayleigh distribution more successfully approximates the second-year wind speed frequency distribution for operable wind speeds. The degree to which the Rayleigh distribution represents the actual wind speed frequency distribution varies by site and year. For the first year of the KEA wind power project, the Rayleigh distribution under-predicted the actual energy generation by approximately 9%. 2-6 Wind Resource Characteristics 800 600 Annual Hours L Oo oO 200 0 2 4 6 8 10 12 14 16 18 20 Wind Speed (m/s) =e Measured " " ™ Rayleigh Bin m/s Measured Hours Bin m/s__| Measured Hours 0.0 257 12.0 125 0.5 214 12.5 121 1.0 82 13.0 109 1.5 80 13.5 93 2.0 110 14.0 99 2.5 219 14.5 70 3.0 449 15.0 55 3.5 572 155 45 4.0 662 16.0 44 45 621 16.5 58 5.0 531 17.0 27 5.5 497 17.5 22 6.0 506 18.0 16 6.5 431 18.5 10 7.0 385 19.0 10 72a 368 19.5 6 8.0 366 20.0 4 8.5 275 20.5 3 9.0 265 21.0 2 9.5 234 21.5 4 10.0 216 22.0 2 10.5 201 22.5 1 11.0 157 23.0 0 11.5 138 Total 8760 Figure 2-4 KEA Wind Speed Frequency Distribution — July 2000 to June 2001 Wind Resource Characteristics 2.3 Wind Direction The KEA site achieved a directional data recovery rate of 94.6% for July 2000 through June 2001 from the Second Wind SCADA system. Figure 2-5 shows the annual wind rose at KEA for the second year of operation, based on the hours of occurrence and the energy available in each direction sector. Prevailing winds were from the east, a trend that is consistent with previous wind data at the project site. As shown in the wind rose, a direction sector width of approximately +23° from due east encompasses the majority of the energy-producing winds, which is consistent with previous years at the site. Westerly winds during the second year were negligible, whereas in previous annual wind roses, a westerly component was more apparent. Annual Wind Rose July 2000 - June 2001 S ——Percent of Total Energy + Percent of Total Time Figure 2-5 Annual Wind and Energy Rose — July 2000 to June 2001 2-8 Wind Resource Characteristics Table 2-3 KEA Monthly Turbulence Intensity and Wind Shear — July 2000 to June 2001 Month Turbulence Intensity | 10-26.5 m Wind Shear July 0.14 0.05 August 0.12 0.19 September 0.11 [ 0.20 October 0.11 0.20 November 0.09 0.23 December 0.09 0.25 January 0.09 0.26 February 0.07 0.23 March 0.08 0.24 April 0.08 0.17 May 0.09 0.09 | June 0.11 0.12 Second Year 07/00-06/01 0.10 | 0.19 First Year 07/99-06/00 0.09 0.20 2.4 Turbulence and Shear Table 2-3 summarizes the monthly turbulence intensity and wind shear during the reporting period. Turbulence intensity is a relative indicator of the turbulence characteristics of the wind. At the KEA site, the average turbulence intensity at hub height for the second year of operation was 0.10 at wind speeds above 4.0 m/s (8.9 mph). Although slightly higher than the 0.09 experienced last year, this turbulence intensity level is considered to be fairly low and unlikely to contribute to any operational problems. The wind shear factor (a) between 10 and 26.5 m was estimated to be 0.19 during the second year of operation compared to 0.20 the first year. The shear was calculated based on the power law formula! using wind speed data above 4.0 m/s (8.9 mph) from all directions. The wind shear value obtained for July is considerably lower than for other months. This anomaly is due to several occurrences of negative wind shear caused by local ground winds during the month. These ground winds are relatively common at the KEA wind project, particularly during summer when the low westerly winds occur. ' (H,/H,)"=(v,/v2) where H, and H, are measurement heights and v, and v2 are wind speeds. 2-9 3 PROJECT PERFORMANCE This section addresses the availability and energy production from the KEA TVP project during the second year of operation from July 2000 through June 2001.? The total energy produced during this period was approximately 1,200.5 MWh. The average TVP system availability, which takes into account all downtime, was 95.9% during the second year. The energy output was slightly higher than the long-term projected energy primarily due to higher-than-normal wind speeds. The monthly data presented in this section are based on the TVP reporting periods which begin at midday on the 20" of the previous month and continue until midday on the 20" of the current month. For example, the month of August includes data from July 20 to August 20. The TVP reporting periods were adjusted to coincide with KEA’s internal reporting periods. These reporting periods were also used for analysis and reporting of the first-year operating experience. 3.1 Availability Table 3-1 summarizes the energy production and availability for each turbine during the second year of operation. Table 3-2 shows the project totals by month and both tables compare the results to the first year of operation of the ten turbines. The average capacity factor for the second year of operation based on 0.66 MW of installed capacity was 20.8%. This significant improvement over the first year’s capacity factor of 12.6% is primarily due to higher wind speeds. The TVP availability for the reporting period was 95.9% based on the recovered data, with the highest turbine availability of 99.2% for Turbines | and 7, and the lowest single turbine availability at 79.1% for Turbine 3. In May, the project had the lowest availability of 82.9%, while the project was 100% available during the reporting periods of July and October 2000. Appendix B presents the monthly availability by turbine. Specific reasons for turbine downtime are discussed in Section 4. There are a number of different ways to define and track availability for individual wind turbines and wind power plants. To ensure consistency in data reporting for all projects involved in the program, the TVP has developed a definition of availability to be used for reporting performance statistics. The TVP definition of availability accounts for all downtime experienced by the individual wind turbines in a project and divides the available hours by the total hours in the > The first three AOC turbines, Phase 1, were installed in 1997. The remaining seven turbines, Phases 2 and 3 were installed concurrently and commissioned in June 1999. This report focuses on the project performance for July 2000 through June 2001, the second year of operation for the full 10-turbine project. For comparison purposes, some references are made to the full project’s first year of operation, the period from July 1999 to June 2000. 3-1 Project Performance reporting period. For example, if during a 100-hour period, a turbine is shut down for 5 hours for a site tour, 5 hours for repairs, and 5 hours due to a line outage, the TVP downtime would be 15 hours and the TVP availability would be (100% - (15/100) x 100%) or 85% for that turbine. Appendix C presents the TVP availability definition. The TVP availability values presented in Tables 3-1 and 3-2 take into account downtime hours associated with a number of different events at the Kotzebue project including research activities conducted by KEA; delays in responding to faults (the site is largely unattended); turbine reliability problems; scheduled maintenance and routine inspections; troubleshooting: delays in obtaining parts; and project-wide shutdowns in response to utility line outages, equipment upgrades, and safety concerns. Table 3-1 Energy and Availability by Turbine — July 2000 to June 2001 Wind Turbine Energy (kWh) | Capacity Factor [1] TVP Availability 1 125,682 21.7% 99.2% 2 113,271 19.6% 96.2% 3 73,730 12.8% 79.1% 4 129,616 22.4% 97.2% 5 115,475 20.0% 95.7% 6 128,478 22.2% 99.1% a 128,840 22.3% 99.2% 8 123,714 21.4% 98.7% 9 129,803 22.5% 98.4% 10 131,905 22.8% 96.3% Total Project — 1,200,514 20.8% 95.9% 2nd Year (7/00-6/01) Total Project — 733,071 12.6% 96.8% 1st Year (7/99-6/00) [1] Based on the TVP-rated turbine capacity of 66 kW. Project Performance Table 3-2 Energy and Availability by Month — July 2000 to June 2001 Month Annual Energy (kWh) | Capacity Factor [1] | TVP Availability July 2000 67,065 14.1% 100.0% August 78,823 16.1% 95.6% September 93,187 19.0% 97.4% October 48,177 10.1% 100.0% November 152,783 31.1% 98.6% December 168,620 35.5% 98.7% January 2001 103,454 21.1% 98.2% February 216,926 44.2% 98.5% March 89,447 20.2% 97.7% April 114,032 23.2% 94.6% May 46,571 9.8% 82.9% June 21,430 44% 89.0% Total Project — 2nd Year (7/00-6/01) 1,200,514 20.8% | 95.9% Total Project — 1st Year (7/99-6/00) Ty 733,071 12.6% 96.8% [1] Based on the TVP-rated turbine capacity of 66 kW. Commercial wind projects generally expect an annual turbine availability of 97% to 98%. However, due to the remote location and harsh environment of the KEA wind project an expected availability of approximately 95% was considered more reasonable and is the basis of the long-term energy estimates discussed later in this section. As shown in Tables 3-1 and 3-2, the project achieved 96.8% TVP availability during its first year and 95.9% TVP availability during the second year of operation. KEA’s higher-than-expected availability during both years of the full ten-turbine project can be attributed to the reliability of the AOC turbines and the responsiveness of the KEA site personnel in resetting faults and addressing maintenance issues. 3.2 Energy Production Tables 3-1 and 3-2 also summarize the energy production for the second year of operation. As the tables indicate, the KEA project produced 1,200.5 MWh of energy during the second year of operation and 733.1 MWh during the first year. This 64% increase in energy production is due to higher-than-projected wind speeds during the project’s second year versus lower-than-usual wind speeds during the project’s first year. 3-3 Project Performance Turbines 9 and 10 produced more energy than any other turbines during both the first and second operating years of the project. During the second year, Turbines 9 and 10 achieved a capacity factor of 22.5% and 22.8%, respectively. These turbines experience the best exposure to the prevailing winds. Seven out of the ten turbines achieved a capacity factor greater than the estimated long-term capacity factor of 20.5%. As Table 3-1 illustrates, Turbine 3 generated the least power during the second year of operation, at 73.7 MWh, caused in part to its low turbine availability for the reporting period. Turbine 3 also experienced tip brake deployment problems that frequently resulted in lower-than-normal energy production when the turbine did operate. Nonetheless, even Turbine 3’s low energy production level in the second year exceeded the average turbine production during the project’s first year. As stated previously, this notable increase in production is due to an increase in the wind resource available at the site during the second year of operation. Turbines 2 and 5 were the next lowest producers. This result is likely due to their location, which is significantly impacted by the wake effects of neighboring turbines. Appendix B presents the monthly energy production by turbine. Following is a discussion of issues related to the energy production at KEA, including seasonal and inter-annual performance variations, energy losses, projected energy, lost energy, and energy penetration. 3.2.1 Seasonal and Inter-Annual Performance Variations Figure 3-1 shows the monthly variation in wind speed, availability, and energy production at the KEA project during the first and second years of operation. The seasonal energy pattern is impacted by both the wind resource and the availability of the turbines. During the second operating year, the highest energy production occurred during February, when the project produced over 216.9 MWh or 7.7 MWh per day, yielding a 44.2% capacity factor. December was the second highest energy-producing month during the second year with over 5.6 MWh per day, a 35.5% capacity factor. The high energy-producing months correspond to the months with the highest wind speeds. The lowest energy-producing month during the second year was June, with an average of 0.7 MWh per day and a 4.4% capacity factor. During June, the average monthly wind speed was 4.3 m/s (9.7 mph), the lowest month during the second year. Figure 3-1 accentuates the differences in wind speed and energy production between the project’s first and second years of operation. During months where the second-year energy output significantly exceeds first-year production, second-year wind speeds also greatly exceed those of the first year. The project experienced abnormally low wind speeds during its first year of operation, and considerably fewer winter storms than usual. Winter blizzards typically supply a significant portion of annual high winds. 3-4 Project Performance Wind Speed (m/s) Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun | = = ™ 7/99-6/00 7100-6/01 40% TVP Availability 20% 0% Ee Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun = = ™ 7/99-6/00 7/00-6/01 250 TF N 5 o Actual Energy (MWh) a o ° oO a oS . 04 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun ©7/99-6/00 ™7/00/6/01 Figure 3-1 Average Wind Speed, Availability, and Energy Production by Month 3-5 Project Performance 3.2.2 Utility Meter Readings and On-Site Energy Losses Table 3-3 compares the sum of the energy generation reported for each turbine and the energy delivered to the grid as indicated by the KEA primary meter at the project interconnection point. The difference between the meter and the sum of the turbines represents the energy losses in the on-site electrical collection equipment, the energy consumption by the facility, and the differences in measurement. As the KEA meter is located at the project site, it does not account for the distribution line losses between the project and the utility load. In general the highest losses are expected to occur during the coldest months due to increased power consumption by the on-site facility and the energy used to operate the turbine transmission heaters. The coldest months typically are December, January and February. Table 3-3 Meter Readings and Sum of Turbine Readings — July 2000 to June 2001 Month KEA Site Sum of Turbine | Energy Loss| Percent Meter (kWh) Meters (kWh) (kWh) Loss July 66,600 67,065 465 0.7% August 75,600 78,823 3,223 4.1% September 90,000 93,187 3,187 3.4% October 46,800 48,177 1,377 2.9% November 148,230 152,783 4,553 3.0% December 164,033 168,620 4,587 2.7% January 99,590 103,454 3,864 3.7% February 208,634 216,926 8,292 3.8% March 86,479 89,447 2,968 3.3% April 109,980 114,032 4,052 3.6% May 45,562 46,571 1,009 2.2% June 21,124 21,430 306 14% 2nd Year - 7/00-6/01 1,162,631 1,200,514 37,883 3.2% 1st Year - 7/99-6/00 718,344 733,071 14,727 2.1% The average energy loss during the reporting period was 3.2% compared to the previous year’s loss of 2.1%. The actual energy loss is considerably higher during the second year. The monthly variation between the meters may be due to the use of multipliers, the use of different time stamps, the variation in on-site energy use, and data processing or analysis errors. 3.2.3 Projected Energy Energy projections for the site were calculated in the report, Wind Resource and Theoretical Energy Estimates for Kotzebue, Alaska and the Northwest Coast, prepared by WECTEC, March 1999. WECTEC developed an estimated long-term data set that represents the wind resource at the Kotzebue project site. This data set reflects a thorough review of all available data 3-6 Project Performance for the site as well as the long-term records from the Kotzebue Airport. As discussed in Section 2.1, the WECTEC analysis indicates a strong correlation between the airport and the project site consistent with the flat terrain and proximity of the airport to the project site. A strong correlation between the long-term reference station and the project site increases the confidence in the estimated long-term wind resource at the project site. Based on the annual distribution of winds at the KEA site and the AOC 15/50 published power curve, WECTEC estimated a gross annual energy of 131,400 kWh per turbine. Expected energy losses summarized in Table 3-4 are based on site conditions and industry experience. Table 3-4 Estimated Energy Losses Loss Factors Estimated Loss Cumulative Losses Availability 5.0% 5.0% Transformer/Line Losses 1.0% 6.0% Control System 1.0% 6.9% Blade Soiling 1.0% 7.8% Wake/Off-axis 2.0% 9.7% Total Cumulative Losses 9.7% L Despite the cold temperatures, Kotzebue has a very dry climate and rarely experiences the rime icing problems that occur in Vermont and other milder winter climates. The nearby radio tower has been in operation for over ten years and has not experienced any icing problems. As a result, no energy losses were considered for icing. Based on the expected losses in Table 3-4, the net annual energy is estimated to be approximately 118,700 kWh per turbine. This represents a capacity factor of 20.5% based on a rated turbine capacity of 66 kW. The energy estimates are summarized in Table 3-5. Table 3-5 Project Characteristics and Long-Term Net Energy Estimates Annual Energy Estimate Gross Energy per Turbine 131.4 MWh Number of Turbines 10 Gross Project Energy 1,314 MWh Estimated Energy Losses 9.7% Net Project Energy 1,187 MWh Capacity Factor * 20.5% * Capacity factor is based on 66 kW rating of the turbines. 3-7 Project Performance Table 3-6 compares the actual and calculated energy for the second year to the projected long- term monthly energy. Figure 3-2 further illustrates the relationship between the various energy measures. During its second operating year, the project produced approximately 1% more energy than the long-term energy estimate. The actual energy generated was 93% of the calculated energy for the second year. This 7% shortfall is attributed in part to turbine anomalies, such as instances when turbines were technically available but not actually producing power or producing power at lower-than-warranted levels. For example, if a brake tip is erroneously deployed, the turbine may continue to be available but will require higher winds to operate and will have significantly lower energy output. The effect of an erroneously deployed tip brake is apparent in Figure 3-2 for November and December, when there were a high number of tip brake events. This issue is discussed further in Section 4. The calculated energy estimate assumes that if the turbine is available, it is producing power at the warranted rate for the measured wind speed. Table 3-6 Actual and Projected Long-term Energy — July 2000 to June 2001 = Month Actual Calculated Long-term Actual vs. | Actual vs. Energy (MWh) | Energy (MWh) | Energy (MWh) | Calculated | Long-term July 67 52 70 128% | 95% August 79 87 109 91% 72% September 93 106 99 88% 95% October 48 58 117 83% 41% November 153 169 136 90% 112% December 169 186 110 90% 154% January 103 129 125 80% 83% February CANE 224 120 97% 181% March 89 102 87 88% 103% April 114 112 68 102% 169% May 47 39 63 120% 74% June 21 27 83 79% 26% Annual 1,201 1,292 1,187 93% 101% * Calculated energy is based on the actual wind speeds at the site met tower and the manufacturer’s power curve. While this energy estimate is based on actual wind speed and provides a valuable comparison, not all turbines experience the same winds. Although the calculated energy is adjusted for actual monthly availability and estimated parasitic losses, it is not based on the actual operating hours of the turbines. 3-8 Project Performance 250 200 150 100 Energy (MWh) 50 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun ] Actual ™ ™ ™ Calculated Long-term | Figure 3-2 Actual, Calculated, and Long-term Projected Energy — July 2000 to June 2001 3.2.4 Lost Energy Due to Downtime Table 3-7 shows the estimated energy lost during the downtime periods in each month. This calculated energy value is based on the manufacturer’s power curve at site air density and the met tower wind speed at hub height during the downtime periods and it is termed “lost energy” in this report.* The project lost approximately 59.1 MWh to downtime during its second year of operation, more than double the 28.4 MWh lost energy reported for the project’s first year. This marked increase in lost energy is due to several factors. Some of the lost energy during the first year was not accounted for because specific downtime documentation was not available for Turbines 4 to 10 for the first several months of the year. During the second year of operation the wind speeds were much higher, resulting in greater energy losses when the turbines were down. The turbine availability also decreased somewhat during the second year from 96.8% to 95.9%. Specific causes of turbine downtime during the second year are discussed in Section 4. Table 3-7 also shows an increase from the first year of the lost energy per hour of downtime. This increase is generally due to higher wind speeds. The met tower wind speed used for estimating lost energy is undoubtedly more representative of some turbines than it is of others. The most accurate estimate of lost energy would be based on the actual wind experienced at the nacelle of each turbine. However, the turbine-mounted anemometers are on the tower at approximately 18 m (59 ft), rather than the 26.5-m (87-ft) nacelle height, and the data are not representative of the turbine nacelle wind speeds. In addition to the difference between wind speeds at the two heights, the turbine-mounted anemometers are affected by wake turbulence from the turbine blades and tower. “In reality, the energy is not “lost,” but rather the opportunity to capture the wind energy is lost during periods when the turbines are unavailable to operate. 3-9 Project Performance Table 3-7 Downtime and Lost Energy by Month — July 2000 to June 2001 Month Downtime | Lost Energy Lost Energy/ Hours (kWh) Downtime Hour July 0.0 0 0.0 August 327.3 5,510 16.8 September 196.0 2,485 12.7 October 0.0 0 0.0 November 106.8 6,706 62.8 December 92.3 5,885 63.7 January 134.0 4,434 33.1 February 115.0 5,211 45.3 March 155.7 3,302 21.2 April 401.8 11,539 28.7 May 1,234.0 8,982 7.3 June 815.5 5,048 6.2 |__ 2nd Year-7/00-6/01 3,578.5 59,102 16.5 1st Year-7/99-6/00 1,975.5 28,447 14.4 The periods with the most lost energy do not necessarily coincide with the periods of the most downtime. For example, 57% of the downtime in the second year occurred during May and June; however, due to low wind speeds, energy lost during those months accounted for only 24% of the total annual lost energy. Table 3-7 also shows the lost energy per downtime hour by month. December had the highest rate of energy loss per downtime hour, at 63.7 kW lost per hour of downtime. In general, months with high energy loss per downtime hour correspond to months with high average wind speeds. Figure 3-3 shows the actual energy production, lost energy due to downtime, and average wind speed for each month. The figure confirms that the potential output of the project (actual plus lost energy) follows the wind speed pattern. Figure 3-4 shows the actual and lost energy for each turbine during the second year of operation. Assuming the turbines produce power at or near the warranted power curve values when they operate, the sum of the actual and lost energy should be nearly equivalent for all turbines, with only minor variations that account for individual turbine losses. As shown in the figure, however, Turbine 3 generated significantly lower energy than the other turbines. The sum of energy produced and lost for Turbine 2 was also lower than for the other turbines. Both turbines experienced problems with their tip brakes. Although a turbine is considered available and generally continues to operate when a tip brake deploys erroneously, the turbine output is significantly reduced. Consequently the actual energy generated during these events is very low and because the turbine is considered available, lost energy is not estimated. Section 4 addresses the tip brake events and other causes of turbine downtime. 3-10 Project Performance | 250 10 | 225 4 ro | 200 4 Ls 175 7 pe gs z 150 4 6 £& \n284 5 3 e 3 S § 100 4: = 75 3 | 50 2 i il. | 0 0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun | energy Produced MMMMEstimated Lost Energy “""""Wind Speed (m/s) Figure 3-3 Actual and Lost Energy by Month — July 2000 to June 2001 160 140 120 = o o 80 60 Energy (MWh) 40 20 GAnnual Produced Energy @Annual Lost Energy | Figure 3-4 Actual and Lost Energy by Turbine — July 2000 to June 2001 3.2.5 Percent Time Generating Figure 3-5 shows the monthly percentage of time the KEA wind power plant produced electricity during its first and second years of operation. The turbines produced power 3,965 hours (45% of the time) during the second year of operation, and 2,988 hours during the first year. During February 2001, the project’s percent time generating was 68.3%, the highest in the first two years 3-11 Project Performance of operation. While the project experienced more downtime hours during the second year than the first, it is important to note this increase in downtime hours was accompanied by an increase in the number of hours the power plant produced electricity. As previously discussed, this was due to higher winds during the second year. 100% 7 90% | 80% - 70% a - | 60% + | 50% +— | | 40% 30% | 20% + : 4 10% | - 0% Jul Aug ea Oct Nov Dec Jan Feb Mar Apr May Jun 017/99-6/00 M7/00-6/01 Figure 3-5 Percent Time Generating Power — July 1999 to June 2001 3.3 Utility Demand and Project Energy The analyses presented in this section are intended to illustrate several key points to utilities contemplating wind energy. Clearly, a wind project produces varying amounts of energy at different times of the year because of variations in the wind resource. Energy has a different value to KEA during different periods of time because of variations in the local demand. The energy is more valuable to the utility during the peak demand periods than during the periods of low demand. Figure 3-6 compares the typical seasonal energy demand at KEA with the actual demand from July 2000 to June 2001. It also shows the long-term estimated and actual second- year wind energy from the project. The long-term estimated wind energy production provides a good match to the historic utility demand with the highest wind energy output matching the highest demand, which generally occurs between October and February. The actual wind energy produced during most months of the second operating year reflected trends in the actual KEA demand for the year, although during October and January, when demand peaked, wind energy production decreased. Although KEA’s primary focus is to ensure the reliable operation of the diesel plant, they have used information such as lost energy and revenue analyses to adjust the O&M strategy at the wind project to most effectively allocate resources and schedule O&M activities. The process and the lessons learned at this site should be beneficial to KEA in its involvement with future commercial wind power plants. 3-12 500 2,500 | 450 = = = 400 2,000 = = 350 3 = 5 = 300 1,500 & 5 250 a ¢ > Ww 200 7 r 1000 & 2 iso 2 = 151 = = 100 r 500 g ° 50 a 0 ~ 0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun EEE | ong-term Projected Wind Energy MEE Actual Wind Energy | (= = _™ Typical Energy Demand em" KEA Actual Demand <_ Figure 3-6 Project Performance Projected and Actual Second-Year Wind Energy Contribution to KEA Energy Demand Table 3-8 shows the average monthly penetration, or total percentage of KEA energy demand met by actual energy produced at the wind power plant, during its second operating year. The highest average monthly penetration (11.0%) occurred in February, while the lowest penetration (1.3%) occurred in June. Table 3-8 Actual and Projected Average Wind Energy Penetration on KEA System Average Monthly Penetration Month Jul 00-Jun 01 | Long-term Projected July 4.2% 4.4% August 4.6% 6.5% September 5.6% 5.8% October 2.5% 6.3% November 8.4% 7.3% December 8.7% 5.3% January 4.7% 5.8% February 11.0% 6.3% March 5.0% 4.3% April 6.0% 3.7% May 2.6% 3.6% June 1.3% 5.2% Overall 5.5% 5.4% 3-13 Project Performance 3.4 High Wind Penetration Study’ Utilities have historically assumed that wind penetration levels of more than 25-30% would result in system instability. Higher levels of wind penetration were expected to cause serious stability and reliability problems in both the generation and distribution systems. Because wind penetration levels in the United States have generally been much lower, little data have been available to determine if such problems actually occur. With a peak load of approximately 4 MW and a minimum load of approximately 1.6 MW, the wind penetration on the KEA system is significant. On an hourly basis, KEA is currently experiencing greater than 35% wind penetration, sometimes for several consecutive hours. KEA is in the process of installing an additional 0.23 MW of wind capacity which will increase total wind capacity to 0.89 MW. During high winds and extremely cold temperatures, the wind project also operates in excess of rated power, further increasing the wind penetration on the KEA system. In order to investigate this issue further, GEC and Thompson Engineering analyzed the available data from the project during a one-month period of high wind penetration. Data for the wind project were generated using the Second Wind SCADA system. Concurrent KEA meter readings indicating total system load and production by the diesel generating facility were also available. Wind penetration estimates were calculated by dividing the measured power output from the wind facility by the total KEA system load. 3.4.1 Diurnal Loads and Wind Penetration Figure 3-7 presents the diurnal pattern of grid loads during the period evaluated. The solid line indicates the average load for each hour over the month. The bars indicate minimum and maximum values observed during the month. The approximate capacity of the wind farm (i.e., approximately 660 kW) is also indicated. Grid loads are lowest during early morning hours, usually falling below 2 MW from approximately 1:00 a.m. to 6:00 a.m. The minimum load observed during the month was approximately 1.67 MW. Loads are higher throughout the day, although loads did not exceed 3 MW during the month. The minimum system load occurred at approximately 4:00 a.m. on August 29, 2000. The highest overall wind penetration of approximately 35% was also measured at this time. August 29 was an unusually warm morning in Kotzebue, with temperatures exceeding 13° C during the overnight hours. These temperatures were over 3° warmer than measured values for the same time period on other days during the month. In addition, the winds were moderate to strong throughout the early morning hours, exceeding 15 m/s (33.5 mph) at 4:00 a.m. This combination of high winds and low demand on the KEA grid resulted in the unusually high penetration levels. 1G) haractering the Effects of High Wind Penetration on a Small Isolated Grid in Arctic Alaska, was presented at the American Wind Energy Association’s annual conference, Windpower 2001 in June 2001. 3-14 Project Performance 3,500 3.000 | 2,500 | = 2.000 sc § 1.500 ~>—_—___—- Approximate Capacty ——————— | al of Wind Farm 1.000 500 0 rr ey - o 2 4 8 10 12 14 16 18 20 22 Time of Day Figure 3-7 Diurnal Distribution of KEA Load — August 21-September 20, 2000 Table 3-9 Summary of Wind Penetration Values — August 21 to September 20, 2000 Parameter Overall Early Morning Day Evening (12 a.m. - 8 a.m.) | (8 a.m. -5 p.m.) | (5 p.m. -12 a.m.) Maximum penetration L 35.3% 35.3% 24.5% 24.2% Average penetration 5.6% 6.0% 5.4% 5.3% Median penetration 2.0% 0.7% 2.0% 3.0% Time with less than 1% 46.2% 50.5% 46.8% 40.5% penetration Time with less than 10% 77.1% 74.6% 76.9% 80.4% penetration Time with less than 20% 92.9% 91.0% 93.4% 95.0% penetration 4 In addition to the morning of August 29, penetration values in excess of 25% were observed during the early morning hours for several days following the 29". A summary of the observed penetration values is presented in Table 3-9. Despite a few periods of high penetration, the overall average penetration for the month was approximately 5.6%. This average is highly influenced by the few significantly higher values. The median penetration for the month was 2.0%. Penetration values less than 1% occurred approximately 46% of the time; and penetration values less than 10% occurred approximately 77% of the time. Overall, penetration varied only slightly with the time of day. During the early morning hours (from midnight to 8:00 a.m.), the median penetration was 0.71%, which is somewhat below the 3-15 Project Performance overall median value. This reflects the generally lower wind speeds during this period, with the exception of August 29 and the days immediately thereafter. However, the early morning average penetration was 6.0%, reflecting the lower energy demand during this period. The opposite trend occurred during evening hours, with a higher 3.1% median penetration due to higher winds but a lower 5.3% average penetration because of higher energy demand. 3.4.2 Variation in Power Quality Parameters A variety of power quality parameters are measured by the Phasers™ and recorded in the SCADA system or can be calculated from the measured values. These parameters include (among others) line voltage, voltage imbalance, total demand distortion, and frequency deviation. This section presents an overview of how each of these parameters varied as the wind penetration increased. Power quality measurements described in this section were recorded by the Phaser™ at Turbine 8. This turbine was used in power performance testing conducted at the site, and its Phaser™ recorded a wider range of parameters than those at the other turbines. The measurements at this turbine are believed to be representative of the rest of the wind farm. The nominal voltage for the turbines is 480 V. The measured voltage exceeded the nominal value by up to approximately 20 V during the time period evaluated. However, there appeared to be no relationship between voltage and wind penetration. At the highest penetration levels, voltage was closer to nominal than at some lower penetration values. Consequently, it appears that any effect on line voltage caused by high penetration is dwarfed by factors external to the wind farm, and possibly by factors caused by the wind farm that are independent of the penetration level. Voltage imbalance was compared to the wind penetration values on a 10-minute basis. There appeared to be no relationship between wind penetration and voltage imbalance. Total demand distortion, a calculated measure of harmonic distortion, was compared to the wind penetration values on a 10-minute basis. There appears to be a slight inverse relationship between wind penetration and total demand distortion; however, the relationship is not strong and may be more related to other factors. Regardless, no adverse effect on total demand distortion is observed as wind penetration increases. Frequency deviation was compared to the wind penetration values on a 10-minute basis. The frequency deviation appears to remain relatively constant as wind penetration levels increase. Based on measurements collected between August 21 and September 20, 2000, there is no apparent adverse effect on power quality on the KEA grid as wind penetration increases. During this period, wind penetration reached a maximum level of approximately 35%. It is unlikely that wind penetration will significantly exceed 35% at KEA at the current wind turbine capacity, as the highest observed penetration levels occurred during times when a combination of high winds and low system load occurred. The planned increase from 0.66 MW to 0.89 MW installed capacity is expected to increase the wind penetration levels proportionately. This could result in peak wind penetration levels greater than 45%. Since the diesel generating equipment at KEA is updated to include independent 3-16 Project Performance generator excitation support and modern, fast-reacting electronic governors, the increase in penetration is not expected to cause a problem. In addition, 0.10 MW of the new capacity is from a wind turbine that uses variable-speed power electronics. This system, as opposed to the induction generators on the AOC turbines, is expected to have minimal impact on the power quality issue that most concerns small utilities, namely VAR supply. Depending on the actual capabilities of the new turbine’s power electronics, it may be able to supply VARs. Although KEA doesn’t expect the increase in capacity to significantly affect other power quality parameters such as frequency, voltage, and harmonics, they will continue to monitor the level of wind penetration on the grid and evaluate the power quality parameters. 3-17 4 PROJECT OPERATIONS AND MAINTENANCE 4.1 KEA’s O&M Strategy The ongoing operations and maintenance requirements of the KEA wind project are handled by KEA personnel with occasional support from their electrical support contractor, Thompson Engineering of Anchorage, Alaska. AOC travels to Kotzebue periodically to provide additional maintenance support as necessary. However, it is not logistically practical for AOC technicians to travel to Kotzebue to perform turbine repairs on a regular basis. In March 1998, KEA hired Matt Bergan, a wind energy engineer, who was formerly employed at AOC. Mr. Bergan’s immediate responsibilities included the ongoing O&M for Phase 1 and the construction oversight for the 7-turbine expansion of the project. Mr. Bergan continues to perform the majority of the O&M duties at the KEA wind farm, and he also provides some O&M services to KEA’s Wales, Alaska, wind project that consists of two AOC 15/50 turbines. KEA’s other activities significantly influence the O&M strategy at the KEA wind project. Although the wind energy engineer was hired specifically to operate the wind project, he continues to perform other duties unrelated to the wind project. The reliable operation of the diesel generators takes precedence over the operation of the wind project. In addition, KEA’s involvement in the construction of the small wind project in Wales, Alaska, has required several months of the wind energy engineer’s time during the past two years. KEA is in the process of expanding their wind project and has been trying to recruit a full-time wind technician to work with the wind energy engineer. It is difficult to find a qualified technician who is willing to live in this remote location and harsh environment. However, during the second year of operation, the wind energy engineer has provided some training to another KEA employee to perform basic operating and maintenance tasks. In spite of the limitations inherent in operating a wind project in this small remote community, the turbines experienced higher-than-expected availability during the first two years of operation. 4.2 Maintenance Activities and Other Downtime Events The KEA project experienced 3,578 total downtime hours during the second year of project operation. Although the project only reported 1,957 hours of downtime during the first year, downtime hours for Turbines 4 to 10 were not recorded for July through November while SCADA communication issues were resolved. The actual turbine availability for the first year was estimated at 97%, approximately 2,500 hours of downtime. Based on this estimated first- 4-1 Project Operations and Maintenance year downtime, the downtime hours increased 44% in the second year. Specific reasons for the increased downtime are discussed later in this section. Although availability is commonly used as a performance measurement in the wind energy industry, it is important to consider the time of occurrence and the cause of the downtime along with the actual number of downtime hours. Figure 4-1 compares the monthly availability to the average monthly wind speeds from July 2000 through June 2001. 10 100% a 8 80% > E : Zz 6 60% & 2 g n 4 40% < | a £ zc S 2 20% 0 0% Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun ——*— Wind Speed (m/s) *_* ™ TVP Availability Figure 4-1 Monthly Availability and Wind Speed — July 2000 to June 2001 Availability was high during the first ten months of the reporting period. The lowest availability months, May and June, coincide with the lowest wind months of the year. KEA understands the benefit of scheduling activities such as preventive maintenance and inspections during periods of low wind whenever possible. Fortunately the low-wind months often occur during the months of greatest daylight hours, providing adequate opportunities to perform scheduled maintenance activities. In addition to the time of occurrence, the cause of downtime and the cost to return a turbine to service are also important considerations. For example, assuming the winds are comparable, 10 hours of downtime due to a fault that is reset without additional action has less impact on the project than 10 hours of downtime due to a repair that requires significant labor, equipment, and parts replacement. Following is a discussion of the cause and impact of the project downtime during the second year. The downtime is categorized and compared to the estimated first-year downtime. The impact of events is discussed in terms of hours and lost energy, if appropriate. As discussed in the previous section, the lost energy for each downtime period is calculated for each turbine on an event-by-event basis that considers the actual wind conditions at the site during the time of the event. Because the actual time that some downtime events occurred was not documented during 4-2 Project Operations and Maintenance the first year, lost energy could not be reasonably estimated. Therefore, lost energy comparisons were not made between the first and second years. Actual O&M cost data for the project are not readily available. The majority of the spare parts used to maintain the turbines have been provided by AOC under the conditions of the turbine warranty. KEA estimates that, on an annual basis, approximately two-thirds of the wind energy engineer’s time is spent performing activities related to the operation and maintenance of the wind project. 4.2.1 Downtime Categories Figure 4-2 summarizes the total downtime by category at the KEA project during the second year of operation. The categories are: O&M-This category includes all troubleshooting, adjustments, retrofits, and repairs performed on the turbines. It includes downtime that accumulates while waiting for parts, instructions, or outside services not available on site but required to place a turbine back on line. Downtime associated with the SCADA system is not included in this category if the turbines continued to operate. O&M downtime accounts for approximately 2,729 hours or 76% of the total downtime during the second year of operation. This represents a 60% increase in O&M downtime from the first year. However, over half of the second year downtime is attributed to one electrical event. A detailed discussion of this event is provided in Section 4.2.2. Faults—This category includes only those faults that require a reset and no other action. If a maintenance activity immediately follows a fault, the downtime hours associated with the fault are combined with the repair hours and the event is included in the O&M category. In some cases, faults are not cleared until after a repair is made. In these instances, the fault time is re-classified as an O&M event if sufficient information is available to make that determination. When faults occur in the evening or on weekends, they are often reset in the morning of the next business day. Occasionally harsh weather limits access to the site, in which case a turbine reset may be delayed. The response time before the fault is reset is included in the fault category as long as the fault is not followed by maintenance. Faults account for approximately 462 hours or 13% of the total downtime during the second year of operation. Although fault downtime doubled from the first to second year, it is still a relatively low number of downtime hours due to faults. Line Outages—Specific, identifiable line outages are included in this category. While KEA tracks total feeder outages, they do not track some partial outages. Therefore, several brief undetected line outages at the wind site are likely to have occurred during the year. The line outages included in this category were documented and reported by the site personnel. During the second year of operation, line outages account for about 373 hours, or 10% of the total downtime. While this is twice the line outage downtime experienced during the first year, 269 hours are attributed to one event. On May 3 KEA disconnected the utility line to the project in order to replace a pole damaged in an automobile accident. The project was reenergized the following day. Project Operations and Maintenance ¢ Other-This category includes downtime for several brief turbine outages that cannot be easily categorized. During the second year of operation, the other downtime category accounts for 14.5 hours, or approximately 0.4% of the total downtime. The majority of this downtime, 13.8 hours, occurred during a period of missing data and no anecdotal information was available. Total Downtime: 3,578 hours Other Fault 0.4% 12.9% Line Outage 10.4% O&M 76.3% Figure 4-2 Total Project Downtime by Cause — July 2000 to June 2001 The remainder of the section describes additional analysis of the downtime associated with faults and O&M activities. Figure 4-3 shows the percentage of total lost energy by category at the KEA project during the second year of operation. O&M downtime accounts for about 47,921 kWh of lost energy; faults account for 3,328 kWh; line outages account for 6,450 kWh, and other downtime accounts for 843 kWh. Figure 4-4 shows the downtime for each wind turbine by category. Turbine 3 had significantly more O&M downtime than any other turbine. Figure 4-4 also illustrates the variability of the total downtime experienced by individual turbines. Turbine 3 experienced over 1,800 hours of downtime, more than half of the total project downtime in the second year. In contrast, Turbines 1, 6, 7, 8 and 9 combined contributed only 13% of the total project downtime. Appendix D presents additional information on the specific causes of downtime for each turbine. Project Operations and Maintenance Total Lost Energy: 59,102 kWh Other Fault 1.4% 6.6% Line Outage 10.9% O&M 81.1% Figure 4-3 Total Lost Energy by Cause — July 2000 to June 2001 Hours Turbine Boam Line Outage OFault Oother Figure 4-4 Total Project Downtime by Turbine — July 2000 to June 2001 Figure 4-5 shows the downtime for each month by category and the monthly mean wind speeds during the second year of operation. The “Other” category accounted for 14.5 hours, only 0.4% of the total downtime, and is not included in the figure. Fortunately, the months of greatest downtime, May and June, correspond with the lowest wind months, resulting in relatively low lost energy. 4-5 Project Operations and Maintenance 1,400 10 1,200 Fila 1,000 7 £ 2 8007 6 3 3 5 ov ° a. x= 600 4” Tv 400 Sie Pi 200 1 0 0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun | | G2 0aM MMM Line Outage Fault "Wind Speed Figure 4-5 Total Project Downtime by Month — July 2000 to June 2001 Table 4-1 summarizes the total downtime hours by category for the first and second years of the KEA wind project. As the table indicates, total downtime hours increased by 44% from the first to the second year of the project. Downtime attributed to O&M faults, line outages and other causes also increased from the first to the second year. Table 4-1 Downtime Hours by Category — July 1999 to June 2001 Downtime Hours Downtime Category Jul 99-Jun 00 Jul 00-Jun 01 Fault 149.7 462.0 Line Outage 123.3 372.5 O&M 1,702.7 2,729.5 Other 83.2 14.5 WT 2 Storm Damage 432.8 NA Total 2,491.7 3,578.5 During the first year of operation, an unusually gusty wind storm in September 1999 caused the maintenance rope on Turbine 2 to unravel and brake free from the leg of the tower. During the storm, the rope got caught in the rotating blades of the turbine, became entangled in the tip brake and caused considerable damage. This resulted in 433 hours of downtime for repairs, and the creation of a unique downtime category for the first year of the project. On a turbine-to-turbine basis, Figure 4-6 illustrates the distribution of downtime for the first and second year. During the two years of operations, 38% of the total project downtime is attributed to Turbine 3 and 18% to Turbine 2. 4-6 Project Operations and Maintenance 2,000 1,800 1,600 1,400 1,200 1,000 800 4 600 400 +— 200 Hours Turbine O7/99-6/00 ™17/00-6/01 | Figure 4-6 Comparison of Project Downtime by Turbine — July 1999 to June 2001 4.2.2 Downtime Due To O&M Activities As discussed in Section 3, the TVP availability was approximately 96% during the second year of operation, and total turbine downtime was 3,578 hours. The majority of the downtime, 2,729 hours, was attributed to O&M activities. Electrical repairs accounted for 2,178 hours, or 80% of O&M downtime during the second year. Figures 4-7 and 4-8 categorize the O&M downtime and related energy losses by major turbine component. Figures 4-9 and 4-10 show the O&M downtime distribution by month and by turbine. Some general observations on the turbine failures and repairs are provided below. Lost Energy Due to O&M Downtime: 47,921 kWh Tip Brake ae 10.8% Gearbox 91% 10.3% | Electrical | 78.9% Figure 4-7 O&M Downtime by Cause — July 2000 to June 2001 Ang Project Operations and Maintenance Lost Energy Due to O&M Downtime: 47,921 kWh Rot Tip Brake ate 10.8% Gearbox ~'” 10.3% Electrical 78.9% Figure 4-8 Lost Energy Due to O&M by Cause — July 2000 to June 2001 O&M Downtime: 2,729 hours 800 700 600 500 400 300 200 100 Hours Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Delectrical BGearbox MRotor HTip Brake System Figure 4-9 O&M Downtime by Month — July 2000 to June 2001 Project Operations and Maintenance O&M Downtime: 2,729 hours Hours Turbine OElectrical Gearbox Tip brake system Rotor Figure 4-10 O&M Downtime by Turbine — July 2000 to June 2001 Electrical. During the second year of operation, electrical events accounted for 2,178 hours of O&M downtime. Electrical O&M accounts for 80% of O&M downtime and approximately 61% of the project’s total downtime during the second year. During the first year, O&M downtime due to electrical issues accounted for 677 hours. While there was a dramatic increase in electrical O&M hours from the first to second year, the increase can be attributed to a single event at Turbine 3. Turbine 3 experienced roughly 1,650 hours of O&M downtime for troubleshooting and repairs due to an overspeed problem. This single electrical problem represents approximately 46% of the total project downtime for the year and about 60% of the downtime attributed to O&M. In April 2001, when Mark Young from AOC was on site, he noticed Turbine 3 was overspeeding and — determined that the cause was a failed sensor. Upon inspection, Matt Bergan noticed that the gap on the speed sensor had opened up, and he observed parking brake material on the gear teeth and speed sensor. KEA replaced the parking brake undercurrent sensor and speed sensor, and installed new contacts in the main contactor. When the turbine continued to experience overspeed problems, KEA personnel removed, tested and reinstalled the frequency/voltage unit. However, the turbine continued to experience problems and was offline for most of May and all of June 2001. More troubleshooting in June led to the discovery that the shield around the cables of the speed sensing unit was not properly grounded. KEA corrected this problem and, in early July, calibrated and installed a new shaft speed sensor and frequency-to-voltage converter on Turbine 3. This resolved Turbine 3’s overspeed problem and its speed sensing unit has been functioning properly since that time. Excluding this event, the electrical O&M downtime decreased by 22% from the first year. Miscellaneous fuses and sensors that required troubleshooting and replacement caused the remaining electrical O&M downtime during the second operating year. Project Operations and Maintenance Gearbox. Gearbox problems resulted in a total of 342 hours of O&M downtime during the second year of operation. During August 2000, KEA personnel noticed an oil leak on Turbine 2. As a result, the turbine was down for a total of 194 hours while technicians completed maintenance on the turbine, replaced the gearbox oil filtration unit with a drain/fill valve, and took oil samples. In March 2001, KEA personnel discovered an oil leak on Turbine 10. The turbine was shut down for 139 hours, until maintenance personnel were available to inspect the turbine and replace its oil vent. A few hours of downtime were also attributed to the gearbox while site personnel replaced the outer seal and cleaned oil off the rotor caused by the leaking seals. During the second year, gearbox repairs accounted for 13% of total O&M downtime. The first year of the project recorded no O&M downtime hours associated with gearbox issues. Tip brakes. Tip brake problems resulted in a total of 202 hours of O&M downtime during the second year of operation. These downtime hours include troubleshooting and repair time, and represent 14 turbine events. Turbines 2, 3, and 4 were the only turbines to experience tip brake problems during the second year of the project. In August 2000, Turbine 3 tip brake problems accounted for 101 hours of O&M downtime. KEA personnel first replaced a missing hinge block on Turbine 3, and then when the turbine continued to experience brake problems, they replaced the entire faulty tip brake. In September, Turbine 2 was down for a total of 48 hours for tip brake repairs. A blade cable had worked its way into the damper area of the tip brake, preventing the tip brake from closing properly. An important aspect of tip brake problems is reduced energy output when a tip brake is erroneously deployed. When a tip brake deploys erroneously, the turbine is considered available and continues to operate but produces significantly less power than if it were operating properly. These energy losses are in addition to the energy lost during recorded downtime events and are more difficult to accurately quantify. For example, in the January 2001 reporting period, Turbine 3 had a TVP availability of 100%, but only produced 30% of the average turbine production for that month. Similarly, during the February reporting period, while Turbine 3 was available 99% of the month, it produced 53% of the average turbine production for that month. O&M downtime for tip brake troubleshooting and repairs decreased significantly from the first to the second year. During the first year of the project, tip brake problems contributed 664 hours, or 39% of total O&M downtime. During year two, tip brake downtime accounted for just 7% of the O&M downtime. This decrease is due primarily to improvements in the AOC tip brake assembly. As discussed in the previous EPRI report documenting KEA’s first year of operation, AOC has worked to improve the reliability of the tip brakes by upgrading several small components. Upgrades have included stronger magnets, improved alignment, a more rugged damper bracket- to-hinge block joint, and improved catch plates and catch plate grommets. Since the initial turbine installations, many tip brake problems were discovered and reviewed by AOC. During the first year of operation, AOC established upgrades for several of the tip brake components including upgraded hinge eyes, catch plate washers and cap screws, and a new strain relief for the blade cables. The new hinge eyes allow for a more rugged joint to the tip brake plate. AOC expects the new hinge eyes to significantly reduce the chance of loose tip brake 4-10 Project Operations and Maintenance plates. Strain relief devices secure the blade cable at the blade root, which prevents the cable from working its way down into the tip brake mechanism. Blade cable interference with the tip brake preventing it from closing properly caused a majority of tip brake downtime during the first year. While the actual repair time was not extensive, troubleshooting required considerable time and numerous trips to the site. The strain relief device is expected to reduce the occurrence of this problem. In addition, AOC is covering the opening in the damper bracket so that if the blade cable does slide, it cannot interfere with the spring/damper mechanism. Several of these improvements have been incorporated on some of the turbines. These upgrades are being incorporated on the remaining turbines when tip brake inspection or repair is required. Rotor. During the second year of operation, rotor maintenance accounted for 7 hours of O&M downtime. Rotor O&M hours included repitching the blades on Turbine 8. Rotor maintenance accounted for 0.3% of O&M downtime during the second year. During the first year of the project, there was no O&M downtime associated with rotors. Unknown O&M. Unknown O&M consisted of only 1.2 hours during the second year of operation. This amounted to such a small percentage (0.04%) of downtime that it does not appear in the figures for O&M downtime. Unknown O&M consists of undefined maintenance events for which there is no documentation. Table 4-2 summarizes O&M downtime hours by category for both years of the project’s ten- turbine operation. As the table indicates, during the first year of operation, control system downtime accounted for 131 hours of O&M downtime, and failed anemometers accounted for 66 hours of total O&M downtime. During the second year of operation, the project experienced no O&M downtime associated with these categories. During the second year tip brake problems accounted for less than one-third of the downtime they experienced during the first year. With the exception of the long electrical O&M event on Turbine 3, the project performed well, resolving several issues that caused downtime during the first year. Table 4-2 - O&M Downtime Hours by Category — July 1999 to June 2001 O&M Downtime Category au << - a Electrical 676.5 2,177.8 Gearbox 0.0 342.0 Rotor 0.0 7.0 Tip brake system 664.3 201.5 Anemometer 65.8 0.0 Controller 131.3 0.0 Other/Unknown 164.7 12 Total 1,702.7 2,729.5 4-1) | Project Operations and Maintenance 4.2.3 Downtime Due To Faults Figure 4-11 shows a breakdown of fault downtime by type for the KEA turbines during the second year of operation. Faults accounted for 462 hours of downtime and an estimated energy loss of 3,888 kWh. One communication event that affected all ten turbines accounted for 275 hours of fault downtime, while two controller faults accounted for 78 hours. Seven faults resulting from unknown causes accounted for 75 hours of downtime. Besides total fault downtime hours, another important factor is the frequency and duration of fault events. For example, a project may experience relatively few faults but the faults may not be reset for several hours or days due to the operating strategy of the utility. On the other hand, a project may experience numerous shorter-duration faults. This is more likely to occur during the first year of operation as the turbine configuration is adjusted to suit the specific site conditions and the owner’s operating strategy. Incorporating a SCADA call-out feature that notifies a remote operator when a turbine is off-line could reduce fault downtime as well as O&M downtime. Because KEA often does not have personnel on site, they have requested that Second Wind add the call-out feature to their SCADA system. The call-out function will be programmed to call the KEA office in town during business hours when a turbine goes off line. On the weekends, the wind energy engineer or other KEA personnel would be notified in the event of a turbine outage. By the end of 2001, Second Wind expects to complete the SCADA software programming update that will allow incorporation of the call-out feature. Fault Downtime: 462 hours Electrical . 5% Tip Brake Controller 1% 17% Unknown 16% Communications 61% Figure 4-11 Breakdown of Fault Cause by Type — July 2000 to June 2001 Figure 4-12 demonstrates the duration and number of faults by month for the second year. Most of the second year’s fault downtime occurred during May 2001 and is attributed to a project- wide communication fault that was not immediately reset because KEA personnel were not available to go to the site. During the September reporting period, the project experienced the same number of faults as in May, when ten faults occurred between September 7 and September 11, but lasted a total of only 11 hours. 4-12 | ~~ Project Operations and Maintenance Total Hours 150 50 100 200 250 Jun May Apr Mar Feb Jan Nov Oct Sep Aug Jul 4 6 8 _—_____Nurmher of Failte Number of Faults Bi Total Hours 10 300 12 Figure 4-12 Fault Frequency and Duration by Type — July 2000 to June 2001 Figure 4-13 illustrates the distribution and cause of fault downtime among the turbines. While Turbines 4 and 6 both experienced only four faults, Turbine 4 had the greatest number of fault hours (104.8 hours) and Turbine 6 had the least fault downtime hours (29.5 hours). Of the 462 hours of fault downtime, 280 hours (61%) are attributed to communication faults that occurred in September 2000 and May 2001. The majority of unknown fault downtime occurred in June 2001, when four turbines faulted for a total of 71.5 fault downtime hours. Fault Downtime: 462 hours 110 100 90 80 70 50 40 30 Hours 20 10 sanfaoalh Turbine 10 Ocommunications Controller OeElectrical MW! Tip Brake Unknown Figure 4-13 Breakdown of Fault Cause by Turbine — July 2000 to June 2001 4-13 Project Operations and Maintenance Table 4-3 summarizes the fault downtime by cause for the first and second years of operation. While fault downtime more than doubled in the second year, the number of fault downtime hours is still relatively low. KEA expects to reduce fault downtime further with incorporation of the SCADA call-out feature. Table 4-3 Fault Downtime by Category — July 1999 to June 2001 Fault Category Jul 99-Jun 00 | Jul 00-Jun 01 Communications 0.0 280.5 Controller 0.0 77.8 Electrical 6.7 22.2 Tip Brake System 0.0 6.3 Unknown 109.5 75.2 Yaw 33.5 0.0 Total 149.7 462.0 4.3 SCADA System Experience The AOC turbine was initially designed as a stand-alone system and AOC did not offer any type of SCADA system at the time of KEA’s original purchase order. KEA contracted with Island Technologies Incorporated (ITI) to design and install a system that would enable remote monitoring and recording of wind turbine operation performance data and parameters and some simple remote control of wind turbine functions. The system was installed in early 1998 for control and monitoring of the Phase 1 turbines. NREL provided a Second Wind SCADA system to the KEA project as part of KEA’s involvement with the TVP. Second Wind and GEC personnel provided on-site support to KEA for the hardware installation. Second Wind personnel performed preliminary commissioning of the SCADA system in September 1999. The Second Wind SCADA system includes a Supervisor Computer which allows real-time monitoring and remote control of the turbines and interconnections with the turbine controllers. In addition to turbine data, the Second Wind SCADA collects concurrent data from the met tower sensors and redundant power quality transducers (Phaser™) also manufactured by Second Wind. The Phasers™ measure numerous power quality parameters including real and reactive power, voltage and current on each phase, voltage and current total harmonic distortion (THD), frequency deviation, and voltage imbalance. There is a Phaser™ attached to each turbine in addition to the project interconnection point. Table 4-4 presents the data recovery rates for SCADA data downloaded from the Second Wind system during the second year of operation. The table includes recovery rates for the turbine 4-14 Project Operations and Maintenance data, the Phaser™ data and the met data. The data recovery and overall reliability of the SCADA system improved during the second year. In October, Second Wind installed a software upgrade designed for the 32-bit architecture of the Windows 98 program. While this provided some improvement in terms of software performance, the computer continued to require rebooting on a weekly basis. The low data recovery in July was before any of the improvements were integrated into the system. The low data recovery in December made KEA and Second Wind realize there was still a problem with the system performance. The SCADA program is constantly scanning the output from the turbine’s programmable logic computer (PLC) and writing to a database located on the computer hard drive. This constant hard drive accessing tends to fragment the hard drive space. Second Wind recommended running a defragmentation program. Since KEA has started running the defragmentation program regularly, the computer performance has become significantly more reliable. The KEA Site Supervisor also archives the database on a regular base to keep the database at a manageable size. Table 4-4 Recovery Rates for SCADA System Data — July 2000 to June 2001 Month Turbine Phaser™ [1] Met [2] July 68.5% 50.5% 67.9% August 97.0% 90.9% 95.0% September 100.0% 99.5% 100.0% October 100.0% 99.5% 100.0% November 100.0% 100.0% 100.0% December 82.1% 78.6% 81.7% January 98.9% 98.2% 100.0% February 99.8% 99.0% 100.0% March 100.0% 99.9% 100.0% April 98.9% 96.5% 99.9% May 96.5% 93.4% 95.8% June 100.0% 100.0% 100.0% Jul 00-Jun 01 95.2% 92.2% 95.1% Jul 99-Jun 00 88.9% 94.9% 91.2% [1] Values are based on Real Turbine Power data collected from the Second Wind Phasers™. [2] Values are based on recovery of hub-height wind speed data. 4-15 Project Operations and Maintenance 4.4 Potential Performance Improvements and Turbine Testing 4.4.1 Slow-Starting Turbines As discussed in the first-year operating report, data analysis and observations at the wind site indicate there are times when the turbines were available to operate, and the winds speeds were above turbine cut-in, but the turbine did not connect to the utility line. This slow-start condition sometimes persists for several hours and appears to affect some turbines more than others. AOC, KEA, and GEC have worked together to identify the possible causes and a solution to this issue. During the first year of operation, KEA, AOC and GEC performed data analysis and turbine testing to better understand the characteristics of slow-start events and to identify possible causes. Investigation indicated that the slow-start condition was not caused by cold temperatures and was probably related to rotor drag. AOC discussed the slow-start issue with some of their component vendors. In September 2000, AOC performed two friction tests and collected additional information to help resolve the slow-start issue. The tests included the measurement of static drive train friction (the brake) and yaw bearing friction. While preliminary testing did not indicate that drive train (brake) friction was a significant factor, in April 2001 AOC installed a free-wheeling program in the control software that allows Turbines 6 and 7 to shut down in low winds without applying the parking brake. Preliminary data analysis indicates that the disuse of the parking brake has not eliminated or decreased significantly slow-start behavior. In low winds below cut-in some improvement may be realized using freewheeling because the rotor will continue to turn. This allows better alignment to the wind and eliminates the influence of any yet to be identified drive train static friction. However, compared to turbines with the standard software, the freewheeling program tends to allow the turbine to motor more in low winds, thus consume more energy. Before incorporating the program on other turbines, KEA would like to see a slight modification to minimize turbine motoring while still allowing freewheeling. The yaw bearing friction testing indicated a higher degree of friction on turbines that were slow to start and recommended an increase of yaw bearing lubrication. Those turbines were lubricated and it appears they yaw more easily as a result. Additional time and analysis are required to confirm the effectiveness of the additional lubrication. Additional analysis of the slow-start issue will be performed in the coming year. 4.4.2 Turbine Over-Production During the first year KEA observations and data analysis confirmed that the turbines regularly generate power at levels higher than the 66 kW-rated capacity of the turbine. This is particularly true during very cold weather and is partly attributed to the resulting increased air density at the site.° Figure 4-14 shows the peak power output from the KEA turbines on a monthly basis. The ° Standard sea level air density of 1.225 kg/m’ is based on an annual average temperature of 15° C. The average annual temperature in Kotzebue is -5.8° C and during the winter months the temperature is routinely -15° to -25° C. These cold temperatures result in an average annual air density of approximately 1.32 kg/m’ and winter air density in excess of 1.40 kg/m’. The increase in maximum turbine output is proportional to the increase in air density. 4-16 Project Operations and Maintenance peak power represents the highest single turbine output at any time during the month. While the peak output is affected by wind speeds, the trend is for the turbine to generate the highest output during the coldest months. This trend is consistent with that experienced during the first year. KEA was concerned that although the cold Kotzebue climate may protect the turbines from overheating during periods of over-production, gearbox life and long-term reliability of the turbine may be reduced. Fairfield Manufacturing, the designers of the AOC gearbox, reviewed one year of 10-minute data records from Turbine 8. The Fairfield engineers concluded that the turbines are not in danger of catastrophic failure and that the 30-year design life of the gearbox will not be jeopardized. The gearbox life is affected by the number of shutdowns and startups, but not by the limited number of high output hours experienced during the cold winter months. [ 100 80 60 40 Power (kW) 20 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Peak Power ™ ™ ™ Min Temp Figure 4-14 Peak Output Compared to Minimum Temperatures — July 2000 to June 2001 The generator is largely protected by over-temperature sensors and is not believed to be jeopardized by operating at high output levels. This is particularly thought to be true because the periods of high output only occur during extremely cold temperatures. In September 2000, prior to Fairfield’s final analysis, KEA repitched the TVP test turbine, Turbine 8. The pitch angle was reduced from the original setting of 1.15° to -0.6°. This pitch adjustment reduced the peak output to approximately 78 kW compared to the >90 kW output previously experienced during high winds and cold temperatures. Based on Fairfield’s findings, KEA may increase the pitch on Turbine 8 to the original setting to allow for the higher output. 4.4.3 Turbine Power Performance Test During the first year, as part of its technical support, TVP conducted a third-party evaluation of the power performance characteristics of an AOC 15/50 wind turbine installed at Kotzebue. Wind speed and concurrent power were collected from calibrated power transducers and 4-17 Project Operations and Maintenance meteorological sensors and processed according to the International Electrotechnical Commission (IEC) Standard 61400-12, Wind Turbine Power Performance Measurement. The test provides a measured power curve for comparison against the manufacturer’s predicted power curve and enables baseline production projections for evaluating performance. The methodology and results of the power performance test were reported in the first-year operating report and presented at the annual American Wind Energy Association conference, Windpower 2000. Test results indicate that, at sea level air density, the AOC 15/50 does not meet the design cut-in wind speed of 4.6 m/s (10.3 mph) and reaches a peak power output of approximately 68 kW at about 14.5 m/s (32.3 mph). This is 36% above the manufacturer’s nominal rating of 50 kW and 3% above the manufacturer’s specified sustained peak rating of 66 kW. At site density (1.361 kg/m?) the peak output is 75.5 kW at 14.5 m/s (32.3 mph). Consistent with the slow-start problem discussed previously, the body of the measured power curve for low to moderate winds (5-11 m/s (11.1-24.5 mph)) was below the manufacturer’s estimate. The problem is severe enough and frequent enough to reduce average power levels at these wind speeds by 10-30%. As discussed in Section 4.4.2, the blades on Turbine 8 were repitched to reduce the maximum power output and KEA plans to return the turbine to its original blade pitch. KEA is also likely to incorporate the free-wheeling program on Turbine 8. NREL, AOC and GEC are still considering a second power performance test once the optimal turbine configuration is determined. When a solution to the slow-start is approved, it would be valuable to retest Turbine 8 as an “optimally” configured AOC turbine for the KEA wind site. 4-18 5 OUTREACH ACTIVITIES AND FUTURE PLANS The primary objective of KEA’s wind energy program is to bring more affordable electricity and jobs to remote Alaska communities. During its second year of operation, KEA continued its community outreach, technology transfer, and wind project planning and development activities. 5.1 Community Education and Outreach Activities KEA continues to promote and participate actively in outreach activities within the local community and at the state and national level. In May 2001 Brad Reeve, KEA’s General Manager, participated in a Wind Powering America meeting in Anchorage. The meeting focused on wind energy applications on Native American lands. The KEA website’ provides a variety of information, including an overview of KEA’s wind energy-related goals, some wind energy basics, and specific information about the AOC wind turbines. KEA provides some wind project statistics, such as energy produced and diesel fuel saved, and gives a summary of the expected benefits of their wind energy program. KEA provides site tours for interested groups within the constraints of KEA’s employee resources. Special efforts are made to accommodate local school groups, the news media, public policy officials, and utility and other technical groups with specific need for information about the project. 5.2 Technology Transfer and Information Dissemination In addition to community education, KEA’s outreach includes technical information dissemination including project performance reporting through the TVP, participation in the Utility Wind Interest Group (UWIG) program, and presentations to local, regional and state officials, and wind and utility industry groups. In 2000 and 2001, papers related to the KEA wind experience were presented at a number of industry events including the annual conference of the American Wind Energy Association, a meeting of the UWIG, and the annual TVP workshop. The majority of KEA’s technical analysis and information dissemination has been performed as part of the TVP program. This includes monthly performance reporting that tracks the production, availability, and maintenance activities of the KEA wind project as well as routine 7 http://www. kotzelectric.com 5-1 Outreach Activities and Future Plans interaction with the TVP support contractor and other utilities. The TVP provides statistics for all of the TVP projects, including KEA, in its quarterly TVP Bulletin. In addition to the TVP program, KEA continues to be actively involved with UWIG. Brad Reeve has served as UWIG President for the past three years. Participation in UWIG activities provides KEA with opportunities to interact on a regular basis with other utilities from around the country that are using or considering wind energy as part of their energy mix. Mr. Reeve continues to make presentations on the project experience at UWIG meetings and other utility and renewable energy conferences. As the President of UWIG, Mr. Reeve has been instrumental in planning the annual TVP Workshop, which is held in conjunction with the fall UWIG meeting. Last year the TVP/UWIG Workshop was held in Morgantown, West Virginia, and included a tour of the new 10.4 MW wind project in Garrett, Pennsylvania. KEA continued to work with local and state educational organizations to develop a “windsmith” training program that will teach wind technician maintenance skills to rural Alaskans. KEA believes that as the wind energy industry grows in Alaska, the experience gained by local workers and companies in Kotzebue will help them get work installing wind projects for other communities. 5.3 Wind Project Planning and Development During the second year of operations KEA continued to promote the integration of wind energy into the power systems of surrounding communities. KEA provided project engineering for the Wales wind project, a high-penetration installation completed in the fall of 2000. Wales is the first community in Alaska powered almost exclusively by wind energy. KEA owns the turbines and sells the power to the Alaska Village Electric Cooperative (AVEC) who serves the electrical needs of the community. The two AOC 15/50 turbines are expected to provide the community with 100% of their electricity during a significant portion of the year. In order to accomplish this level of wind penetration on the grid, the power plant is instrumented with a state-of-the-art control system, developed by NREL, that allows the use of excess electricity produced during periods of high winds. During this past year there have been many challenges at the Wales wind project, particularly with the dump load systems that use the excess energy. Several water and fuel lines required resizing in order to perform properly. Without the dump loads available to take excess wind energy, it was not possible to operate both turbines during the majority of the year. Adjustment to the system required coordination between KEA, AVEC, and NREL. Both turbines are currently operating. KEA continues to work towards a long-term goal of 2 to 4 MW of wind generation capacity at Kotzebue, enough to entirely meet the electrical needs of the community during peak demand. Currently Kotzebue uses about 20 million kWh of electricity and 5.7 million liters (1.5 million gallons) of diesel fuel each year. The next wind project expansion in Kotzebue will add one Northwind 100 kW turbine and two AOC 15/50 turbines and is expected to occur during spring 2002. The Northwind turbine and tower arrived in Kotzebue on a summer barge. The two AOC turbines have been shipped from Vermont and are expected to arrive in Kotzebue 5-2 Outreach Activities and Future Plans in November 2001. The AOC towers are currently being assembled and should arrive before spring. The AOC blades are also expected in the spring. The new turbines will extend the existing three rows of turbines with one additional turbine at the end of each row. All three new turbines will be incorporated into the Second Wind SCADA system. The site road was expanded farther onto the site and along the turbine rows in the summer of 2001. These site roads will provide easier access to the turbines for maintenance during the summer months when the ground is thawed. 5-3 6 CONCLUSIONS Through their involvement in the TVP, KEA has successfully developed, constructed, and is now operating a wind power plant in Kotzebue, Alaska. For the second year in a row the ten-turbine project performed well, with a higher-than-expected availability. The wind speeds, which had been unusually low during the first year, returned to normal and the project exceeded its projected production levels. During the second year of operation, the project produced approximately 1,200 MWh of energy with an average TVP system availability of 96.8%. The project generated approximately 1% over the estimated long-term annual energy of 1,187 MWh. This significant improvement from the previous year’s 38% energy shortfall is attributed to higher wind speeds. During the first year of operation, the average wind speed at the site was 5.1 m/s (11.4 mph), approximately 16% lower than the long-term average. During the second year, the average wind speed was 6.5 m/s (14.5 mph), approximately 7% higher than the long-term average. During both years, the turbine availability was higher than the long-term expected availability of 95%. KEA is successfully meeting the challenges of operating a commercial wind farm in the Arctic. The AOC 15/50 turbine is performing well in the cold, harsh environment and AOC continues to support the project with configuration adjustments and appropriate component upgrades. AOC has also provided KEA with turbine maintenance support. The success of the KEA Wind Power project has aided the continuing development of wind energy in Alaska, including the completion of the high-penetration wind project in Wales, Alaska. KEA is in the process of expanding their ten-turbine project. They have contracted to purchase three additional turbines for installation in 2001-2002. Current project expansion plans include the addition of one Northwind 100-kW turbine and two additional AOC 15/50s. While other Northwest Alaskan communities are still considering wind energy development, no final agreements have been reached. The KEA project has also been effective in achieving the TVP’s objectives of verifying the performance, reliability, maintainability, and cost of new wind turbine designs and system components in commercial utility environments. Consistent with TVP objectives, KEA is also providing other utilities and stakeholders with information about wind technology and the development and operation process from the perspective of utility owners and operators. With the incorporation of the larger Northwind 100-kW turbine and the installation of two additional AOC 15/50 turbines, the TVP expects to continue their support of the KEA project. The TVP plans to provide ongoing technical support as well as performance evaluation and 6-1 Conclusions reporting beyond the original three-year evaluation period. The unique environment of the KEA wind project is providing valuable lessons learned to wind industry stakeholders. The TVP is continuing to provide utilities and turbine manufacturers with valuable experience in wind power plant development, operation and maintenance, and technology transfer. The lessons learned through the TVP will be passed on to other projects in which EPRI and DOE have a management role and to the rest of the wind and utility industry through continuing outreach activities. A TVP-RELATED DOCUMENTS EPRI Reports Wind Turbine Verification Project Experience: 1999, EPRI 1000961, December 2000. Big Spring Wind Power Project Second-Year Operating Experience: 2000-2001, EPRI 1004042, December 2001. Big Spring Wind Power Project First Year Operating Experience: 1999-2000, EPRI 1000958, December 2000. Project Development Experience at the Big Spring Wind Power Project, EPRI TR-113919, December 1999. Iowa/Nebraska Distributed Wind Generation Projects First- and Second-Year Operating Experience: 1999-2001, EPRI 1004039, December 2001. Lessons Learned at the Iowa and Nebraska Public Power Wind Projects, EPRI 1000962, November 2000. Project Development Experience at the lowa and Nebraska Distributed Wind Generation Projects, EPRI TR-112835, December 1999. Kotzebue Wind Power Project Second-Year Operating Experience: 2000-2001, EPRI 1004040, December 2001. Kotzebue Wind Power Project First Year Operating Experience: 1998-2000, EPRI 1000957, December 2000. Project Development Experience at the Kotzebue Wind Power Project, EPRI TR-113918, December 1999. Wisconsin Low Wind Speed Turbine Third-Year Operating Experience: 2000-2001, EPRI 1004041, December 2001. Wisconsin Low Wind Speed Turbine First and Second Year Operating Experience: 1998-2000, EPRI 1000959, December 2000. A-1 TVP-Related Documents Wisconsin Low Wind Speed Turbine Project Development, EPRI TR-111438, December 1998. Green Mountain Power Wind Power Project Third Year Operating Experience: 1999-2000, EPRI 1000960, December 2000. Green Mountain Power Wind Power Project Second Year Operating Experience: 1998-1999, EPRI TR-113917, December 1999. Green Mountain Power Wind Power Project First Year Operating Experience: 1997-1998, EPRI TR-111437, December 1998. Green Mountain Power Wind Power Project Development, EPRI TR-109061, December 1997. Central & South West Wind Power Project Third Year Operating Experience: 1998-1999, EPRI TR-113916, December 1999. Central & South West Wind Power Project Second Year Operating Experience: 1997-1998, EPRI TR-111436, December 1998. Central & South West Wind Power Project First Year Operating Experience: 1996-1997, EPRI TR-109062, December 1997. Central and South West Wind Power Project Development, EPRI TR-107300, December 1996. DOE-EPRI Wind Turbine Verification Program TVP MI-112231 Status Report, 1998. Building Community Support for Local Renewables and Green-Pricing Projects EPRI TR- 114203, 1999. NREL/AWEA WindPower Published Papers Central & South West’s 1998 Operations and Maintenance Field Experiences. B. Givens, Central & South West Services. Presented at WindPower 1999. Characterizing the Effects of High Wind Penetration on a Small Isolated Grid in Arctic Alaska. G. Randall, R. Vilhauer, Global Energy Concepts, LLC, C. Thompson, Thompson Engineering Company. Presented at Windpower 2001. Characterizing Wind Turbine System Response to Lightning Activity: Preliminary Results. MeNiff, B.; LaWhite, N.; Muljadi, E. Collection of the 1998 ASME Wind Energy Symposium Technical Papers Presented at the 36" AIAA Aerospace Sciences Meeting and Exhibit, 12-15 January 1998, Reno, Nevada. New York: American Institute of Aeronautics and Astronautics, Inc.(AIAA) and American Society of Mechanical Engineers (ASME); pp. 147-156; NICH Report No. 25563. 1998. A-2 TVP-Related Documents Comparison of Projections to Actual Performance in the DOE-EPRI Wind Turbine Verification Program. H. Rhoads, J. VandenBosche, T. McCoy, A. Compton, Global Energy Concepts, LLC, B. Smith, National Renewable Energy Laboratory. 14 pp.; NICH Report No. CP-500-28608. _ Presented at WindPower 2000. CSW Small Wind Farm Operating Experience °96 — °98. W. Marshall, Central & South West Services, Inc. Presented at WindPower 1998. Development and Plans for the Kotzebue Wind Power Plant. B. Reeve, Kotzebue Electric Association, and E. Davis, Wind Energy Consulting & Services. Presented at WindPower 1998. Distribution Line Power Quality Experience with the Nebraska Distributed Wind Generation Project. M. Hasenkamp, Nebraska Public Power District. Presented at WindPower 2000. DOE-EPRI Distributed Wind Turbine Verification Program (TVP III). C. McGowin and E. DeMeo, Electric Power Research Institute, S. Calvert and P. Goldman, U.S. Department of Energy, B. Smith, S. Hock and R. Thresher, National Renewable Energy Laboratory. Presented at WindPower 1997. DOE-EPRI Wind Turbine Verification Program (TVP). C. McGowin, EPRI, T. Hall, U.S. Department of Energy and B. Smith, National Renewable Energy Laboratory. Presented at WindPower 1998. EPRI/DOE Wind Turbine Performance Verification Program. Calvert, S.; Goldman, P.; DeMeo, E.; McGowin, C.; Smith, B.; Tromly, K. 6 pp.; NICH Report No. CP-440-22486. Presented at Solar Energy Forum 1997. Evaluation of Lightning Protection Retrofit on the Z-750s in Springview, Nebraska. M. Hasenkamp, Nebraska Public Power District. Presented at Windpower 2001. Green Mountain Power’s 6-MW TVP Wind Project in Searsburg, Vermont. J. Zimmerman, Green Mountain Power Corporation. Presented at WindPower 1998. Green Mountain Power’s Searsburg Project. B. Ralph, Green Mountain Power Corporation. Presented at WindPower 1999. Iowa TVP III Project. T. Wind, Cedar Falls Utilities. Presented at WindPower 1999. Lightning Activities in the DOE-EPRI Turbine Verification Program. T. McCoy, H. Rhoads, T. Lisman, Global Energy Concepts, LLC, B. MeNiff, McNiff Light Industry, B. Smith, National Renewable Energy Laboratory. 14 pp.; NICH Report No. CP-500-28604. Presented at WindPower 2000. Nebraska TVP III Project. M. Hasenkamp, Nebraska Public Power District. Presented at WindPower 1999. A-3 TVP-Related Documents Power Performance Testing Activities in the DOE-EPRI Turbine Verification Program. J. VandenBosche, T. McCoy, H. Rhoads, Global Energy Concepts, LLC, B. MeNiff, McNiff Light Industry, B. Smith, National Renewable Energy Laboratory. 15 pp.; NICH Report No. CP-500- 28589. Presented at WindPower 2000. Power Performance Testing Progress in the DOE-EPRI Wind Turbine Verification Program. J. VandenBosche, G. Randall, T. McCoy, Global Energy Concepts, LLC. Presented at Windpower 2001. - Power Quality of Distributed Wind Projects in the Turbine Verification Program. J. VandenBosche, T. Lettenmaier, Global Energy Concepts, LLC, T. Wind, Wind Utility Consulting, M. Hasenkamp, Nebraska Public Power District. Presented at Windpower 2001. Program on Lightning Risk and Wind Turbine Generator Protection. Muljadi, E.; MeNiff, B. National Renewable Energy Laboratory 8 pp.; NICH Report No. CP-440-23159. 1997. Project Performance in the DOE-EPRI Wind Turbine Verification Program. J. VandenBosche, R. Vilhauer, G. Randall, Global Energy Concepts, LLC, B. Smith, J. Green, National Renewable Energy Laboratory, National Wind Technology Center. Presented at Windpower 2001. “Projects-at-a-Glance” Summaries of Projects Within the DOE-EPRI Wind Turbine Verification Program. K. Conover, S. Meyer, H. Rhoads, S. Simon, K. Smith, J. VandenBosche and R. Vilhauer, Global Energy Concepts, LLC. Presented at WindPower 2000. Review of Operation and Maintenance Experience in the DOE-EPRI Wind Turbine Verification Program. K. Conover, J. VandenBosche, H. Rhoads, Global Energy Concepts, LLS, B. Smith, National Renewable Energy Laboratory. 13 pp.; NICH Report No. CP-500-28620. Presented at WindPower 2000. TU/York Big Springs Project. L. Herrera, TU Electric. Presented at WindPower 1999. Wind Farm Generation Impact on a Small Municipal Utility System. T. Wind, Wind Utility Consulting. Presented at WindPower 2000. Wisconsin Low Speed Wind Turbine Project Development Experience. J. VanCampenhout, Wisconsin Public Service Corporation. Presented at WindPower 1998. Other TVP Resources Joint Utility Wind Interest Group/Turbine Verification Program/Wind Powering America Technical Workshop. 2000. TVP News Bulletins. Global Energy Concepts. 1999-2001. A-4 B MONTHLY AVAILABILITY AND PRODUCTION BY TURBINE KEA Turbine Availability - July 2000 through June 2001 wT Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Annual 1 100.0% 100.0% 99.9% 100.0% 99.0% 100.0% 100.0% 100.0% 100.0% 100.0% 92.4% 98.5% 99.2% 100.0% 85.8% 80.1% 100.0% 98.3% 100.0% 100.0% 100.0% 100.0% 100.0% 92.4% 98.5% 96.2% 100.0% 86.4% 100.0% 100.0% 98.2% 99.7% 100.0% 98.9% 99.6% 62.1% 5.7% 0.0% 79.1% 100.0% 89.8% 98.0% 100.0% 98.6% 100.0% 91.2% 100.0% 100.0% 100.0% 89.3% 100.0% 97.2% 100.0% 100.0% 100.0% 100.0% 98.5% 88.8% 100.0% 85.6% 99.2% 84.9% 92.3% 100.0% 95.7% 100.0% 100.0% 99.8% 100.0% 98.6% 100.0% 100.0% 100.0% 98.8% 99.6% 92.4% 100.0% 99.1% 100.0% 100.0% 99.6% 100.0% 98.8% 100.0% 100.0% 100.0% 100.0% 100.0% 92.4% 100.0% 99.2% 100.0% 99.7% 96.9% 100.0% 98.6% 100.0% 100.0% 100.0% 100.0% 100.0% 92.4% 96.6% 98.7% 9 100.0% 99.7% 99.6% 100.0% 98.3% 100.0% 90.8% 100.0% 100.0% 100.0% 92.4% 100.0% 98.4% 10 100.0% 94.6% 99.8% 100.0% 98.6% 98.6% 100.0% 100.0% 79.3% 99.5% 86.7% 96.7% 96.3% aon Oak WOW ND Project 100.0% 95.6% 97.4% 100.0% 98.6% 98.7% 98.2% 98.5% 97.7% 94.6% 82.9% 89.0% 95.9% KEA Energy Production - July 2000 through June 2001 (kWh) Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Annual 6,190 9,020 10,191 5,062 14,998 17,492 10,705 24,127 9,259 12,701 4,705 1,230 125,682 6,314 5,670 4,210 4,748 14,664 16,170 10,916 21,896 9,316 11,569 5,351 2,447 113,271 5,678 5,124 10,020 4,926 12,762 12,226 3,065 11,532 6,698 1,698 0 0 73,730 7,205 8,424 9,869 4,999 15,931 19,307 10,061 24,590 8990 13,509 4,141 2,590 129,616 6,831 8497 9,680 4,625 16,005 10,924 11,462 19,671 10,035 9,445 5444 2,855 115,475 6,480 8149 9,621 4,597 15,127 18,887 11,496 23,162 9,449 13,386 5,493 2,631 128,478 6,747 8,508 9,842 4,902 16,510 19,091 11,178 22,985 8577 12,959 5.416 2,125 128,840 7,418 8895 9,954 4,849 14,604 17,126 11,286 21,323 9,027 11,654 5,419 2,158 123,714 6,900 8267 9,682 4,232 16,005 19,388 9,909 23,791 10,326 13,553 5,282 2,470 129,803 10 7,301 8,269 10,118 5,238 16,177 18,009 13,375 23,850 7,769 13,558 5,318 2,923 131,905 oor 92H He wnH A/F Project 67,065 78,823 93,187 48,177 152,783 168,620 103,454 216,926 89,447 114,032 46,571 21,430 1,200,514 C TVP AVAILABILITY DESCRIPTION There are a number of different ways to define and track availability for individual turbines and wind power plants. To ensure consistency among the projects involved in the program, the TVP developed a definition of availability to be used for reporting on performance statistics throughout the program. The TVP definition of availability takes into account all downtime experienced by the individual wind turbines in the project and divides the available hours by the total hours in the period. For each turbine, the TVP availability is: % Turbine Availability = {[H- (Downtime Hours for Turbine) ]/H} X 100% where H is the number of hours in the period and Downtime Hours for Turbine accounts for all downtime experienced by the turbine during the period of interest (i.e., week, month, year-to- date, or 8760 hours for an annual period). For a wind power plant, the TVP availability is: % Wind Power Plant Availability = {[(H X N)-(Sum of the Downtime Hours for N Turbines) ]/(H X N)} X 100% where H is the number of hours in the period and N is the number of turbines in the project. Although the above definitions use “hours” in the calculation, it is important to collect data that shows the turbine status (i.e., available or unavailable) on a time interval of 10 minutes or less so that fractions of an hour can be included in the availability calculation. The TVP availability includes downtime caused by different events including: e research activities; e testing; e delays in responding to faults; e public relations (i.e., site tours); e turbine maintenance and retrofit activities; C-1 TVP Availability Description e scheduled maintenance and routine inspections; e troubleshooting; e delays for parts or equipment; e line outages; and e force majeure events. There are several of other availability definitions that exclude some of these events. Although these approaches are intended to serve a specific purpose, the TVP uses the 7VP Wind Power Plant Availability definition to ensure consistency among the projects. C-2 D SPECIFIC DOWNTIME CAUSES BY TURBINE Turbine 1 Downtime: 73.5 hours Line Yaw Faults Outage 37% 47% Unknow n 1% Unknow n Faults 15% Turbine 3 Downtime: 1831.7 hours Tip Brake Yaw Faults [2% Other 0% Bectrical O&M 90% Turbine 5 Downtime: 372.7 hours Line Unknown Faults Outage 11% Gearbox O&M 1% Turbine 2 Downtime: 331.5 hours Tip Brake Unknown O&M Faults 18% 3% Unknow n 8% Turbine 4 Downtime: 243.8 hours Tip Brake Unknow n O&M 11% Comm Fault 15% 1% Bectrical O&M 27% Turbine 6 Downtime: 77.2 hours Bectrical Gearbox O&M O&M 11% 3% Yaw Faults Line Outage 36% 48% D-1 Specific Downtime Causes by Turbine Turbine 7 Downtime: 66.2 hours Comm Line Outage 54% Turbine 8 Downtime: 115.3 hours Bectrical O&M Line Outage 17% 32% Rotor O&M 5% Turbine 9 Downtime: 141.2 hours Faults Unknow n Rotor O&M = 9% 20% 1% Line Outage Hectrical 28% O&M 49% Turbine 10 Downtime: 325.5 hours Hectrical O&M Bectrical Unknown 23% Faults 5% Line Outage a ~ Gearbox 12% O&M 44% D-2 Target: Wind Power Development Support About EPRI EPRI creates science and technology solutions for the global energy and energy services industry. U.S. electric utilities established the Electric Power Research Institute in 1973 as a nonprofit research consortium for the benefit of utility members, their customers, and society. Now known simply as EPRI, the company provides a wide range of innovative products and services to more than 1000 energy- related organizations in 40 countries. EPRI's multidisciplinary team of scientists and engineers draws on a worldwide network of technical and business expertise to help solve today’s toughest energy and environmental problems. EPRI. Electrify the World © 2001 Electric Power Research Institute (EPRI), Inc.All rights reserved. Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power Research Institute, Inc. @ Printed on recycled paper in the United States of America 1004040 EPRI * 3412 Hillview Avenue, Palo Alto, California 94304 * PO Box 10412, Palo Alto, California 94303 * USA 800.313.3774 * 650.855.2121 + askepri@epri.com * www.epri.com 1007-0007 :®2uatsedxg Buneiado yel!o1g 4eMog PUIAA aNgEe20y IIS 00001