Loading...
HomeMy WebLinkAboutKotzebue Electric Assoc Wind Power Project First-Year Operating Exp 1999-2000crrel Kotzebue Electric Association Wind Power Project First-Year Operating Experience: 1999-2000 U. S. Department of Energy - EPRI Wind Turbine Verification Program Technical Report Nine i] ma Kotzebue Electric Association Wind Power Project First-Year Operating Experience: 1999-2000 U.S. Department of Energy —- EPRI Wind Turbine Verification Program 1000957 Final Report, December 2000 Cosponsors Kotzebue Electric Association National Renewable Energy Laboratory U.S. Department of Energy 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 (Il) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (lll) 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 Concepts, LLC ORDERING INFORMATION Requests for copies of this report should be directed to the EPRI Distribution Center, 207 Coggins Drive, P.O. Box 23205, Pleasant Hill, CA 94523, (800) 313-3774. 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 © 2000 Electric Power Research Institute, Inc. All rights reserved. CITATIONS This report was prepared by Global Energy Concepts, LLC 5729 Lakeview Drive NE, Ste. 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 First-Year Operating Experience: 1999- 2000: U.S. Department of Energy-EPRI Wind Turbine Verification Program, EPRI, Palo Alto, CA; the U.S. Department of Energy, Washington, DC; Kotzebue Electric Association, Kotzebue, Alaska: 2000. 1000957. iii REPORT SUMMARY Although much of western Alaska has abundant wind resources, wind energy technology has not been widely deployed in the state, and utilities rely primarily on diesel fuel for energy generation. Kotzebue Electric Association (KEA) is pioneering the application of wind energy technology in combination with the existing diesel generation in the remote communities in Northwest Alaska. This report describes the first-year operating experience at the 0.66-MW KEA wind power project near Kotzebue, Alaska, 26 miles north of the Arctic Circle. The lessons learned in the project will be valuable to other utilities planning similar wind power projects, particularly in cold, remote locations. Background In 1992, EPRI and the U.S. Department of Energy (DOE) initiated the Wind Turbine Verification Program (TVP). The goals of the program 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 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—owned and operated by KEA—consists of 10 Atlantic Orient Corporation (AOC) 15/50 wind turbines. Each turbine has a 15-meter diameter, three-bladed rotor and a constant-speed turbine-generator mounted on top of a 24.4-meter (80-ft) lattice tower. KEA installed the first three turbines in 1997 and the remaining seven turbines in 1999; the project was commissioned in June 1999. A companion report, EPRI TR-113918 (December 1999), describes the project development experience. Future reports will describe the second and third years of operating experience. Objectives To document KEA’s first year of operating experience, describe the experience gained and problems encountered in the first year of operation, and transfer the lessons learned to other utilities planning similar projects. Approach Project investigators documented the first-year operating experience at the KEA project from July 1999 through June 2000. The 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 1999 through June 2000, the Kotzebue wind facility delivered 733,071 kWh of electricity to the Kotzebue distribution system. It operated at a 12.7% average capacity factor, based on 660-kW rated capacity. The overall TVP system availability was 96.8%, allowing for all scheduled and forced outages of the wind turbines. Some of the downtime experienced by the turbines was due to research and other activities conducted by KEA. Individual turbine availabilities ranged between 91.1% and 99.8%. The first-year operating experience at the KEA wind project site was typical of similar small power projects using new technology at undeveloped sites. First-year issues included tip brake problems, turbine slow-starts, and longer than expected brake-cooling cycles. Although the turbine availability has been higher than expected, some minor turbine performance issues have been noted. KEA, the turbine vendor, and the TVP support contractor have identified several turbine configuration changes intended to improve performance, including repitching of the turbine blades. The configuration changes are under evaluation and will be implemented in 2001. EPRI Perspective Through 2000, EPRI has issued 17 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 the experience gained in the TVP projects to utilities, wind power developers, turbine vendors, government agencies, and other interested parties so that the lessons learned can be incorporated into future projects. The KEA operations report should be helpful in this regard because it describes negative as well as positive experiences, which should help others avoid or reduce the impact of many problems encountered. Future EPRI reports will describe operational experiences of other TVP-funded projects. Keywords Wind power Wind resource assessment Performance Availability Operation/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 the first 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 10 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 feet) 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. As part of the TVP, KEA will complete a three-year performance evaluation program of their wind project. This report discusses the activities and experience during the initial operation of Phase 1 and the first year of operation of the full 10-turbine project. It includes summaries of the wind resource data, actual and projected energy production, and availability at the site during the first year of operation. The report discusses the operation and maintenance activities and categorizes the downtime experienced by the turbines during the period from July 1999 through June 2000. KEA is evaluating the expansion of the facility and may increase the installed capacity of the project to as much as 2 to 4 MW. 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, Craig Thompson of Thompson Engineering, and Mark Young of AOC were particularly helpful in providing details and clarification. Karen Conover, Kirsten Brauch, and Sarah Meyer 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 AEP AOC ASOS AVEC AWEA EPRI GEC IEC ITI KEA Met MSI NRECA NREL O&M PCE SCADA TVP UWIG WECTEC Annual Energy Production Atlantic Orient Corporation Automated Surface Observing System Alaska Village Electric Cooperative American Wind Energy Association Electric Power Research Institute Global Energy Concepts International Electrotechnical Commission Island Technologies Incorporated Kotzebue Electric Association Meteorological Meteorological Standards Institute National Rural Electric Cooperative Association 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 TIN TRO DUCTION secececcsescrecrececersessescescnscoccnsrerececectecascostecsorss covasecsecocevocevacescsustecrecscaccesaresaceres) 1-1 Heel KPTOJOCtIBACKQFOUNC et cetessrcerstersrerseeuescrscccrerctaretsascspeseceaectacsecantescestsarencvsccsteastcssecsarecers 1-1 1.2 Background on the Wind Turbine Verification Program. .........:csccescsessseseessesseeseeseeneeeaees 1-3 1.3 Report Objectives and Scope ... 1.4 Report Organization 2 WIND RESOURCE CHARACTERISTICS.. 2.1 Data Collection LAIN SPOOG merercrcrtscteecerccucstececors scene cunssoresanscasstctorcsscstcessacsvactesescscessucsscccsuctecrssesucstarsezses Dan RI CaN a a ll ae tal cl tel eetea ZA TUPDUIONCE ANC ONCAN eres retro tereceunecssntenevasareuecscetsncsescvactetessursssansseccasucuaesecuaseas 3 PROJECT PERFORMANCE ........ccssscscssccscssscscssrsossscscrssenscscssssensesersnsvasssseacsassncenssocsecacssesse 3-1 ShPAvallabilityeercsrteeercccrcaisrererscecrcrcsestcescstecesstenssssastseceerstsucseseecnasesecacesesenaesucaectseccesecers 3.2 Energy Production 3.2.1 Seasonal and Inter-Annual Performance Variations . 3.2.2 Utility Meter Readings and On-site Energy Losses GBi2iG PMOIECIOD ENGIGV i cccnsccecccssceccoceccescsceccacsccescececesacesscssucesstctsecessaesicesraccrncnessurcessseces 9.2.4 Lost Energy due to DOWNtime ....<............ccesssssccessceosessscsscererscsccesersssrossronsssaesecons 3:3 Utility Demand and Project EMGray .................-cscccecssscssecsssecosssssrseccssrnsscosecssnssseseseseees 4 PROJECT OPERATIONS AND MANAGEMENT..........:sccccssessessessesseseessesnseesseseessesateneeeeeeees 4-1 AIKEA'SIORMISUALOY ceticecccrcee renee crrreterces cseutenessacancanccocescesstceacsnccareastarcsuscrsecsscrassecsesatsed 4-1 4.2 Maintenance Activities and Other Downtime Events... 4.2.1 Downtime Categories 4.2.2 Downtime Due To O&M Activities 2S Cererreiiret Dene "TO Fai ies cosas cxssicannnexsnasssusacremestanenss vaniianvedancniansnisieiannenensnitarit ASSISCADATSYSIOM)EXPENONCO sreerrtesraccceneceseteternrereseeet ea statastectatsucuesatereceaevsecusescsuenesecs 4:4; Potentiall Performance IMProveMentS..cccccscreesccrstrse-eescecoceoscrcseccsoccreracscacsesesnecsnsccs teed 4.4.1 Slow-starting Turbines... ccc cece cececcseeeeeesseceeecseeeeeeeeeaeeeeeeeeecseeesceeeeeseeeneeeenees 6A. BRR OCIING CNG vicctaxincnssenasccemmnmniniananccmncmenaNiasenn ina 4.4.3 Turbine Over-ProductiOn ..........cccccccsceceeseeeeseeseeceseeeeeseeeeeeeseeeeseseeeseeseseeneeeseeenaeens 4.5 Turbine Power Performance Test 4.5.1 Test Methodology 4.5.2 Test Results 4.5.3 Future Performance Testing 5 OUTREACH ACTIVITIES AND FUTURE PLANG..........ccccsssssssesssecenssesseesseessersessessenseeseeeees 5-1 5.1 Community Education and Outreach Activities.............c cc eescceeseeesseeeeseeeesseeeesseesseeeneees 5-1 5.2 Technology Transfer and Information Dissemination .............::cccccscessceseseseeesseseeseeseeees 5-2 5.3 Wind Project Planning and Development...........::cccccescessessesseesscessseesseseseeesaseaseneseesease 5-3 6 CONCLUSIONG.........csssssssssssssssssssseusesessesenseesensenseseseuseseuseeseseeseuseeseeseeeaeeseaseeseesessassesaeeaseass 6-1 A TVP-RELATED DOCUMENTSG........cccsssssssssssssessssssssessessesseuseessesseeseeessaeseeseseeeseesseseneeseeseeses A-1 EPA POTS crs csiscinasrenicsntsssninommannnctceniannsntencunnaicninena mune RnKiaeaNeNaNineN iiventines A-1 NREL/AWEA WindPower Published Papers Other TVP Resources BMONTHLY AVAILABILITY AND PRODUCTION BY TURBINE CTVP AVAILABILITY DESCRIPTION........scccscesseesessessesseseeseesesseeseeseessesesneesneeseeeeeesaesaeenaees C-1 DSPECIFIC DOWNTIME CAUSES BY TURBINE ..........csssssesessesseesseesneesesesnenseeseessenenseseene D-1 E TIP BRAKE ASSEMBLY.........cccsscssssssessessseseeseeessseeseesessnesaseaasaesseesnassnesnaeasassnassnesseseesnansaese E-1 xiv LIST OF FIGURES Figure 1-1 Alaska State Map Figure 1-2 Photograph of the KEA Wind Power Plant Figure 2-1 Location of the Current KEA Met Tower Figure 2-2 Monthly Wind Pattern at Kotzebue (26.5 m)............0 Figure 2-3 Diurnal Wind Pattern at the Kotzebue Wind Project (26.5 m) Figure 2-4 KEA Wind Speed Frequency Distribution — July 1999 to June 2000 Figure 2-5 Annual Wind and Energy Rose — July 1999 to June 2000 Figure 3-1 Energy Production, Availability, and Average Wind Speed by Month .... Figure 3-2 Actual, Calculated, and Long-term Projected Energy Figure 3-3 Actual and Lost Energy by Month — July 1999 to June 2000 . Figure 3-4 Actual and Lost Energy by Turbine — July 1999 to June 2000 Figure 3-5 Wind Energy Contribution to KEA Energy Demand Figure 3-6 Overall Wind Energy Penetration on KEA System Figure 4-1 Monthly Availability and Wind Speed — July 1999 to June 2000 ...........eceeeeeeseeeeeeee Figure 4-2 Total Project Downtime by Cause — July 1999 to June 2000 Figure 4-3 Total Lost Energy by Cause — July 1999 to June 2000 Figure 4-4 Total Project Downtime by Turbine — July 1999 to JUNG 2000 ........ ce eceeeeseeteeeeseeeee Figure 4-5 Total Project Downtime by Month — July 1999 to June 2000 Figure 4-6 O&M Downtime by Cause — July 1999 to June 2000 Figure 4-7 Lost Energy due to O&M by Cause — July 1999 to JUNE 2000...........eeeeseeeseeeeeeeees Figure 4-8 O&M Downtime by Month — July 1999 to June 2000 Figure 4-9 O&M Downtime by Turbine — July 1999 to June 2000 Figure 4-10 Breakdown of Fault Cause by Type — July 1999 to JUNE 2000 .........eeeeeeeeeeeeees Figure 4-11 Fault Frequency and Duration by Type — July 1999 to June 2000... : Figure 4-12 Breakdown of Fault Cause by Turbine — July 1999 to June 2000... Figure 4-13 Peak Output Compared to Minimum Temperatures-July 1999 to June 2000...... 4-17 Figure 4-14 KEA Topographic Site Map Used for Power Performance Test.. Figure 4-15 Power Curve at Sea-Level Density, 1.225 kg/m° Figure 4-16 Power Coefficient at Sea-Level Density, 1.225 kQ/M? ......c.ccseceseeseeeeteeteeeteeeteeees Figure 4-17 Maximum and Minimum Power Measured for the AOC 15/50 Wind Turbine ...... 4-21 XV LIST OF TABLES Table 2-1 Data Recovery Rates for Meteorological Data — July 1999 to June 2000 2-4 Table 2-2 Mean Monthly Wind Speeds at Kotzebue (26.5 M).........:cccccessseseeeseesseeeeeeeeeeeeeeeeees 2-5 Table 2-3 KEA Monthly Turbulence Intensity and Wind Shear - July 1999 to June 2000....... 2-10 Table 3-1 Energy and Availability by Turbine — July 1999 to June 2000... Table 3-2 Energy and Availability by Month — July 1999 to June 2000 Table 3-3 Meter Readings and Sum of Turbine Readings — July 1999 to June 2000............... 3-7 “Fae ae meme PUN TAINO ccc se dpa en fn inn tmpmmtnttntmtntmnoin ne OB Table 3-5 Project Characteristics and Long-term Net Energy Estimates..............::cc:cesseeseeeeee 3-8 Table 3-6 Actual and Projected Long-term Energy Table 3-7 Downtime and Lost Energy by Month — July 1999 to June 2000 Table 3-8 Wind Energy Penetration — September 2000.............:csscesssesseesteesseesseesseeseseeenees Table 4-1 Recovery Rates for SCADA System Data — July 1999 to June 2000... Table 4-2 Annual Energy Production at Sea-Level Density, 1.225 kg/m° xvii 7 INTRODUCTION This report is the second 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 a previous EPRI report, TR-113918. The first report was published in 1999 and documents the project’s development and initial operation of the Phase 1 turbines. Extensive background information, redundant to the previous report, 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 10 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 snowmobile trails connect villages or remote homes to Kotzebue during the nine months that snowmobile travel typically is feasible. 1-1 Introduction Arctic Circle =_- airbanks Canada ‘ fem —_s we # aS 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 10 years with an extension option of four additional 10-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 | 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 first year of operation, July 1999 through June 2000. 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 2000. 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 1-3 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 Project in Texas. 1.3 Report Objectives and Scope This report focuses on the first year of operation of all 10 of KEA’s AOC 15/50 wind turbines. The report discusses the project’s performance, operation and maintenance activities, and KEA’s outreach activities and future plans. Two additional reports on the KEA project are planned as part of its three-year TVP evaluation period. The principal objectives of this report are to summarize the KEA TVP project experience, including performance characteristics, wind resource data, operating strategy, maintenance activities, research projects, and other events of significance 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 performance of the project 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 first year of operation. 2 WIND RESOURCE CHARACTERISTICS KEA has collected wind resource data for eight 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 new Supervisory Control and Data Acquisition (SCADA) system installed in late summer, 1999. The following summarizes KEA’s wind monitoring program and the on-site wind resource during the reporting period from July 1999 through June 2000. 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 and affected the met data in certain direction sectors. In August 1998, the met tower was relocated to a site approximately one rotor diameter upwind of the first turbine string 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). Unfortunately, the next seven AOC turbines also were slated to be installed upwind of the first three turbines. As a result of wake impacts from the surrounding turbines, the met tower was no longer representative of the wind conditions at the site. In the summer of 1999, a second met tower was installed at a location upwind of the 10-turbine project layout. This met tower is positioned approximately two rotor diameters upwind of Turbine 8 and is the primary source of 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 Supervisory Control and Data Acquisition (SCADA) system, which was commissioned in late August 1999. Data from this new met tower were also used for power performance testing on Turbine 8. This met tower location meets the requirements of IEC 61400-12 for power performance testing, discussed below in Section 4.5. 2-1 Wind Resource Characteristics Prior to the commissioning of the Second Wind SCADA system, wind speed data were recorded by the Campbell SCADA system installed in early 1998 to support Phase 1 of the wind project. In addition to several turbine operating parameters, the Campbell system recorded wind speed from six anemometers; two anemometers located on the tower of each of the three Phase 1 turbines. ‘4 Met Tower —— 35 2000 3000 4000 5000 FEET eee eee eee 800 1000 METERS be! Figure 2-1 Location of the Current KEA Met Tower Wind data also are 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. 2-2 €ox Tlond Wind Resource Characteristics 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. The calculated correlation coefficient was 0.92, which indicates a good statistical relationship between the wind resource at the airport and the project site. WECTEC then developed a long-term estimated wind speed for the project site based on hourly data from the airport. From 1995 through 1997, the National Weather Service (NWS) implemented a modernization of } ) 05 meteorological monitoring equipment at sites throughout the U.S. The new equipment, designated as Automated Surface Observing System (ASOS), includes a change in the averaging methodology. The ASOS was installed at the Kotzebue Airport on December 1, 1997. While the new equipment is more accurate than the old, this change created a discontinuity in the long-term wind speed record, which introduces uncertainty in the interpretation of current wind speed records from the airport. | Wind chmak? dab donkauty Shdy NWS sponsored a study conducted by Tom Lockhart of the Meteorological Standards Institute (MSD to quantify the impact of the change from the old system to the ASOS. MSI performed a linear regression analysis to compare the wind speed from the old system to the new ASOS at eighteen stations across the country. \e char Lo hechorne +b cow WECTEC reviewed the findings of the NWS study to estimate the appropriate adjustments for the Kotzebue Airport data. From the linear regression analysis, the x-coefficients and constants are used to approximate the error between the old data collection system and the ASOS measurement methodology. WECTEC chose the x-coefficient and constant derived from the analysis at a cold-weather site, Bismarck, North Dakota, as the basis for adjusting the historical Kotzebue Airport wind speed data. A complete met data set was compiled for the project site for the reporting period of July 1999 through June 2000 based primarily on the Second Wind SCADA data. Data from the Campbell system were used for July and August 1999 data, prior to the commissioning of the Second Wind system. Table 2-1 presents the data recovery from these on-site data sources at all sensor heights. As seen in Table 2-1, data replacement was required only for July 1999, June 2000, and a few hours in December 1999. For these limited periods when accurate data were not available from either system, 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. GEC performed a linear regression analysis for representative months using the average daily wind speed data from the airport and concurrent data from the site. The resulting correlation coefficients were 0.97 for June and July and 0.88 for December. A correlation factor of one indicates perfect correlation. Thus, these high correlation coefficient factors show a strong correlation between the wind speeds at the site and those at the airport. 2-3 Wind Resource Characteristics Table 2-1 Data Recovery Rates for Meteorological Data — July 1999 to June 2000 Month 10m 19m 26.5 m July 1999 N/A N/A 80% August N/A N/A 100% September 100% 100% 100% October 89% 89% 100% November 73% 73% 100% December 91% 91% 99% January 2000 92% 92% 100% February 91% 91% 100% March 100% 100% 100% April 98% 98% 100% May 99% 100% 100% June 65% 28% 88% Annual Average 90% 86% 97% In addition to replacing missing data, GEC used data from a lower level sensor to replace the 26.5 m wind speed data that appeared to be erroneous, either due to icing or wake effects. Data from the lower sensor were adjusted to represent the 26.5 m wind speed based on relationships between the lower and upper sensors determined during periods of concurrent valid data. The following sections present an overview of the wind characteristics at the site during the first year of operation and include comparisons to 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 shows the monthly and annual mean wind speeds for July 1999 through June 2000 at the 26.5 m level compared to the estimated long-term site wind speed at 26.5 m. The average annual wind speed at the site during the first year of operation was approximately 16% lower than the estimated long-term average. An inter-annual variation of this magnitude is unusual and has a significant impact on energy production. Section 3 addresses the unusually low winds and their impact on the project performance. 2-4 Wind Resource Characteristics Table 2-2 Mean Monthly Wind Speeds at Kotzebue (26.5 m) Month 7/99-6/00 Long-term m/s (mph) m/s (mph) July 5.5 (12.4) 5.8 (12.9) August 5.3 (11.8) 6.5 (14.4) September 5.1 (11.4) 6.4 (14.2) October 48 (10.7) 6.6 (14.7) November 4.5 (10.1) 7A (15.9) December 4.2 (9.3) 6.2 (13.9) January 5.3 (11.8) 6.3 (14.1) February te. (16.1) 6.7 (14.9) March 6.5 (14.5) 5.5 (12.3) April 49 (11.0) 5.3 (11.9) May 4.4 (9.9) 5.3 (11.8) June 3.7 (8.2) 6.0 (13.3) Annual Average 5.1 (11.4) 6.1 (13.6) Figure 2-2 is a graphical comparison of the wind resource at the site during the first year of operation, the previous year, and the long-term estimated wind resource. Although there is some monthly variation between the long-term average monthly wind speeds and the monthly average wind speeds for July 1998 through June 1999, the general pattern is consistent. However, there is a significant variation between the expected long-term monthly averages and the wind speeds for the reporting period July 1999 through June 2000. As seen in the graph, the long-term pattern shows that the winds are weakest from March through July and generally begin to increase in August. The strongest winds are in October, November, and February. During the first reporting period, the monthly average wind speed was lower than normal every month except February and March. This is consistent with lower than normal wind speeds recorded at the Kotzebue Airport. Wind Resource Characteristics - : 10 | Wind Speed (m/s) 0 | Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun | —e— Jul 98-Jun 99 Jul 99-Jun oo --- Long-term. Figure 2-2 Monthly Wind Pattern at Kotzebue (26.5 m) The adjusted Kotzebue Airport data indicate that the airport’s annual average 10 m wind speed for July 1999 through June 2000 was approximately 18% lower than the long-term annual average. Although the discontinuity of the airport data, discussed earlier in this section, creates some uncertainty in estimating the long-term representativeness of the current year’s wind speed, an indication of lower than normal winds is consistent with anecdotal reports from site personnel. For example, KEA General Manager, Brad Reeve, observed that last winter there were significantly fewer blizzards (high-wind storms) than normal. A significant portion of the annual high winds typically come from the dozen or more winter blizzards. Last winter Kotzebue only experienced four or five blizzards. Figure 2-3 compares the diurnal pattern of the site’s wind resource during the project’s first year of operation to the long-term pattern. The wind speeds were significantly lower during the reporting period than the estimated long-term wind speeds; however, the diurnal pattern was essentially the same. Although the KEA site does not exhibit a significant variation in diurnal wind speed, the winds are generally slightly stronger in the afternoon. Although most months are similar to the annual pattern illustrated in Figure 2-3, a review of the monthly diurnal pattern for the long-term resource indicates there is some variation. Again, the values recorded during the first reporting period were consistently lower than the expected long-term average for the KEA project site. 2-6 Wind Resource Characteristics — 8) | an | B- a : a 3 a + é 6 = = 5+ = | 0 ee 123 4 5 6 7 8 9 1011 1213 1415 16 17 18 19 20 21 22 23 24 Hour of the Day G -- Long-term ————7/99-6/00 Figure 2-3 Diurnal Wind Pattern at the Kotzebue Wind Project (26.5 m) The peak ten-minute average wind speed recorded at the 26.5-m level during the first year of the operation was 17.4 m/s (39 mph) recorded on March 6, 2000. The peak 1999 5-second wind speed at the airport, reported in the Annual Summary distributed by the National Climatic Data Center, was 23.7 m/s (53 mph) and occurred on January 21. The 5-second peak wind speed for 2000 is not yet available. There are no long-term peak wind speed data as this parameter was not recorded at the airport prior to the installation of the ASOS system in December 1997. Figure 2-4 presents the wind speed frequency distribution at Kotzebue for July 1999 through June 2000 in tabular and graphical form. The plot shows the measured wind speed frequency distribution together with the Rayleigh distribution which is frequently used to provide a simple estimate of the wind speed distribution. While the Rayleigh distribution is a good representation of the actual wind speed distribution at some sites, this is not the case at Kotzebue and several other sites. Using the Rayleigh distribution and the average wind speed to predict the energy generation shows that it under-predicts the energy generation by approximately 9%. 2.3 Wind Direction The KEA site achieved a directional data recovery rate of 93.7% for July 1999 through June 2000 by combining wind vane data from the Campbell system and the Second Wind SCADA direction data. Figure 2-5 shows the annual wind rose at KEA for the first year of operation, based on the hours of occurrence and the energy available in each direction sector. Prevailing winds were generally from the east with some light westerly winds during the early summer. As shown in the graph, a direction sector width of approximately +/-45 degrees from due east encompasses the majority of the energy-producing winds. 2-7 Wind Resource Characteristics nee July 1999 - June 2000 Mean Wind Speed = 5.1 m/s 2 600 — 3 | = |= 400 oO = § 200 0 - 0 2 4 6 8 10 12 14 16 18 | Wind Speed (m/s) Measured =~ = Rayleigh | L — OO Bin Measured Bin Measured m/s Hours m/s Hours 0.0 585 10.5 134 0.5 413 11.0 115 1.0 155 11.5 99 1.5 154 12.0 77 2.0 205 12.5 54 2.5 338 13.0 35 3.0 523 13.5 31 3.5 667 14.0 24 4.0 694 14.5 23 4.5 664 15.0 16 5.0 639 15.5 9 5.5 532 16.0 6 6.0 490 16.5 2 6.5 395 17.0 1 7.0 349 17.5 0 7.5 298 18.0 0 8.0 282 18.5 0 8.5 243 19.0 0 9.0 201 19.5 0 9.5 181 20.0 0 10.0 156 8,784 Figure 2-4 KEA Wind Speed Frequency Distribution — July 1999 to June 2000 2-8 Wind Resource Characteristics 2.4 Turbulence and Shear Turbulence intensity is a relative indicator of the turbulence characteristics of the wind. At the KEA site, the turbulence intensity at hub height is generally below 0.10 at wind speeds above 4.0 m/s (8.9 mph). This turbulence intensity level is considered to be fairly low and unlikely to contribute to any operational problems. The wind shear factor () between the 10-m and 26.5-m heights was estimated to be 0.20 during the first year of operation. 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. Available data were not adequate to estimate turbulence intensity or wind shear for July or August. Table 2-3 summarizes the monthly turbulence intensity and wind shear during the reporting period. eaucany Turbulence Intensity and Wind Shear - July 1999 to June 2000 Month Turbulence 10-26.5m Intensity Wind Shear July 1999 N/A N/A August N/A N/A September 0.11 0.13 October 0.10 0.19 November 0.09 0.26 December 0.09 0.22 January 2000 0.10 0.14 February 0.07 0.20 March 0.07 0.23 April 0.08 0.23 May 0.09 0.17 June 0.13 0.20 Annual 0.09 0.20 ' (H\/H2)*=(v\/v2) where H; and H, are measurement heights and v, and v> are wind speeds. 2-10 3 PROJECT PERFORMANCE This section discusses the availability and energy production from the KEA TVP project during the first year of operation from July 1999 through June 2000.” The total energy produced during this period was approximately 733.1 MWh. The average TVP system availability, which takes into account all downtime, was 96.8%. The energy output is lower than expected primarily due to lower than normal wind speeds. The monthly data presented in this section are based on the TVP reporting periods which are from midday on the 20" of the previous month to 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 period was changed to coincide with KEA’s internal reporting periods. For this reason, monthly wind speeds may not be consistent with monthly values included in the Wind Resource section of this report, which are based on calendar months. Information in this section is based on a variety of data collection devices installed on site. Although Turbines 4-10 were fully operational by the end of June 1999, they operated without a SCADA system until late August when the Second Wind SCADA was installed. While the Second Wind SCADA operated satisfactorily, there were problems with KEA’s communication and computer interface which caused data loss. During the commissioning and early months of operating the Second Wind system, KEA relied on the Campbell SCADA which was installed in 1997 for the Phase | turbines. Although production data were collected for Turbines 4-10 during this period, turbine downtime data were not available until December. The SCADA systems are discussed in more detail in Section 4.3. 3.1 Availability Table 3-1 summarizes the energy production and availability for each turbine during the first year of operation. Table 3-2 shows the project totals by month and compares the results to the first two years of operation of the Phase | turbines. The average capacity factor for the first year of operation based on 0.66 MW of installed capacity was 12.6%. The TVP availability for the reporting period was 96.8% based on the recovered data, with the highest single turbine availability of 99.8% for Turbine 4 and the lowest single turbine availability at 91.1% for Turbine 2. The month of lowest project availability was July 1999 while the project was 100% available during March and May 2000. Appendix B presents the monthly availability by turbine. ? The first three AOC turbines, Phase 1, were installed in 1997. The remaining seven turbines, Phase 2 and 3 were installed concurrently and commissioned in June 1999. This report focuses on the project performance for July 1999 through June 2000, the first year of operation for the full 10-turbine project. For comparison purposes, some references are made to earlier performance of the Phase | turbines. 3-1 Project Performance Table 3-1 Energy and Availability by Turbine — July 1999 to June 2000 Wind Turbine Annual Capacity TVP # Energy Factor [1] Availability (kWh) [2, 3] 1 75,398 13.0% 99.1% 2 48,358 8.3% 91.1% 3 62,433 10.8% 94.4% 4 78,020 13.5% 99.8% 5 78,434 13.5% 98.7% 6 69,873 12.1% 96.8% 7 76,399 13.2% 98.0% 8 77,754 13.4% 97.5% 9 83,228 14.4% 99.1% 10 83,174 14.3% 97.8% Project 733,071 12.6% 96.8% 1] Based on the turbines’ TVP-rated capacity of 66 kW. 2] Availability for Turbines 4-10 is based on data from December 1999 through June 2000 only. 3] Annual availability is a weighted average based on the available data. Although availability data were not recorded for Turbines 4-10 until December 1999, GEC estimated turbine availability for Turbines 4-10 from July through November based on anecdotal data and comparison of turbine-to-turbine production. The resulting estimated project availability was 97.1%, slightly higher than the 96.8% determined by the recovered SCADA data. There are a number of different ways to define and track availability for individual wind turbines and wind power plants. To ensure consistency between the projects involved in the program, the TVP has developed a definition of availability to be used for reporting performance statistics throughout the program. The TVP definition of availability accounts all downtime experienced by the individual wind turbines in a project and divides the available hours by the total hours in the reporting period. For example, if during a 100-hour period, a turbine is shut down for 5 hours because a site tour is in progress, 5 hours to repair a component under warranty, 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%. Appendix C presents the TVP availability definition. 3-2 Project Performance The TVP availability values presented in Tables 3-1 and 3-2 include hours associated with a number of different activities 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 3-2 Energy and Availability by Month — July 1999 to June 2000 Month Annual Capacity TVP Energy Factor [1] Availability [2] (kWh) July 1999 31,448 6.6% 91.5% August 73,824 15.0% 94.8% September 21,709 4.4% 95.6% October 79,630 16.8% 77.6% November 39,189 8.0% 96.4% December 33,533 7.1% 97.9% January 2000 63,542 12.9% 95.5% February 139,990 28.5% 98.7% March 143,071 31.1% 100.0% April 48,000 9.8% 96.9% May 29,760 6.3% 100.0% June 29,374 6.0% 97.3% Total Project —1st Year (7/99-6/00) 733,071 12.6% 96.8% Phase 1-2nd Year (7/98-6/99) 293,579 16.9% 94.3% Phase 1-1st Year (7/97-6/98) [4] 168,182 10.0% N/A 1] Based on the TVP-rated turbine capacity of 66 kW. 2] Turbine availability for July through November includes the Phase 1 turbines only. 3] Annual availability is a weighted average based on the available data. 4] Turbine availability data are not available for 1997. 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 thought to be 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, during its first year of operation, the project achieved 96.8% TVP availability which is higher 3-3 Project Performance than the long-term expected availability. Due to startup issues, new projects commonly do not achieve long-term expected availability the first year of operation. KEA’s high first-year availability for the 10-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. The knowledge and experience gained with the Phase | turbines was also a factor in achieving high availability during the project’s first year of operation. 3.2 Energy Production Tables 3-1 and 3-2 also summarize the energy production for the first year of operation. Turbines 9 and 10 produced the most energy during the reporting period with 13.5% more energy than the average per-turbine production. These highest producing turbines achieved a capacity factor of approximately 14% compared to the estimated long-term capacity factor of 20.5%. Turbine 2 was the lowest producing turbine with a capacity factor of only 8.3%. Without Turbine 2, the project had a capacity factor of 13.2%. Although Turbine 2 availability was the lowest at 91.1%, a significant factor in the low production occurred from the turbine operating for many hours with a deployed tip brake. While this condition reduces output significantly, the turbine is technically available and operates part of the time. This issue is discussed further in Section 4. In addition to this reporting period, Table 3-2 presents the annual energy production for Phase 1’s first two years of operation. During the year July 1998 through June 1999, the Phase 1 project generated an average of nearly 97.9 MWh per turbine (16.9% capacity factor) compared to the 73.3 MWh per turbine (12.6% capacity factor) generated from July 1999 through June 2000. While turbine availability was higher than projected, unusually low wind speeds restricted the project output. Appendix B presents the monthly energy production by turbine. A discussion follows 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 energy production, availability, and wind speed at the KEA project during the reporting period. The seasonal energy pattern is impacted by both the wind resource and the availability of the turbines. On a monthly basis, the highest energy production occurred during March, with over 143.1 MWh or 4.9 MWh per day, yielding a 31.1% capacity factor. February was also a high energy month with over 4.5 MWh per day, a 28.5% capacity factor. The high energy-producing months correspond to the months with the highest wind speeds. 3-4 Project Performance a a Wind Speed (m/s) np > Oo Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun 100% 80% 7 - : 60% 40% — 20% TVP Availability 0% Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun 160 140 120 100 80 60 40 + 20 Actual Energy (MWh) oO Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun | Figure 3-1 Energy Production, Availability, and Average Wind Speed by Month 3-5 Project Performance The lowest energy month was September with only 21.7 MWh, an average of 0.7 MWh per day a 4.4% capacity factor. Although the turbines experienced an estimated 230 hours of downtime, the low production is primarily attributed to low wind speeds. In spite of 100% turbine availability, May yielded a low energy output of 29.8 MWh due to the low monthly wind speed of 4.6 m/s (10 mph). The May energy production was approximately 21% of that produced during the March reporting period. As shown in Figure 3-1, while August and October exhibited similar wind speeds, August had higher availability but generated less energy. This anomaly was caused by a performance problem with Turbine 3 during August. Although Turbine 3 has a reported availability of 85.3% for August, it produced only about 25% of the average energy produced by the other turbines during the month due to a deployed tip brake. Turbine performance and downtime are discussed later in Section 4. As discussed in Section 2, the annual average wind speed for the reporting period was only 5.1 m/s, approximately 16% lower than the estimated long-term wind speed of 6.1 m/s. The low wind speed is largely responsible for the lower than projected energy and is discussed in Section 3.2.3. 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. The average energy loss during the reporting period was just over 2%. Although this annual value is consistent with experience at other wind projects, the range varies significantly from month to month. The 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 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 3-6 Project Performance 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. Table 3-3 Meter Readings and Sum of Turbine Readings — July 1999 to June 2000 Month KEA Site Sum of Percent Meter Turbine Loss (kWh) Meters (kWh) July 1999 31,388 31,448 0.2% August 73,630 73,824 0.3% September 21,695 21,709 0.1% October 79,407 79,630 0.3% November 38,838 39,189 0.9% December 33,286 33,533 0.7% January 2000 61,200 63,542 3.8% February 135,900 139,990 3.0% March 141,300 143,071 1.3% April 46,800 48,000 2.6% May 27,000 29,760 10.2% June 27,900 29,374 5.3% Annual 718,344 733,071 2.1% 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. 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 10 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. 3-7 Project Performance Table 3-4 Estimated Energy Losses Estimated Cumulative Loss Factors Loss 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% 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. Table 3-6 compares the actual, calculated, and projected long-term monthly energy for the reporting period. The actual energy production was 68% of the long-term projected energy, a shortfall of 38% with the actual energy lagging the long-term estimate during 10 of the 12 months. This energy shortfall is attributed to lower than expected wind speeds, discussed in Section 2, and turbine performance anomalies, discussed in Section 4. The shortfall attributed to turbine anomalies can be estimated by comparing the actual energy to the calculated energy.* The actual energy generated by the project was 87% of the calculated energy, a shortfall of 13%. The remaining 25% shortfall can be attributed to the lower than expected wind speeds. Figure 3- 2 illustrates the relationship between actual, calculated, and long-term projected energy. AOC and KEA are in the process of refining the configuration of the turbine to address performance anomalies and wind speeds will return to normal levels. Energy output is expected to improve significantly, in fact, the project generated 234 MWh during the first three months of the next annual reporting period, July through September 2000, compared to 127 MWh > 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 wind 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 generated during the same period in 1999. Section 3.2.5 provides additional discussion on turbine downtime and estimated energy losses. Table 3-6 Actual and Projected Long-term Energy Actual Calculated | Long-term Energy Energy Energy Actual to Actual to Month (MWh) (MWh) (MWh) Calculated | Long-term July 31 35 70 89% 45% August 74 88 109 84% 68% September 22 26 99 83% 22% October 80 84 117 95% 68% November 39 49 136 80% 29% December 34 43 110 79% 31% January 64 80 125 80% 51% February 140 154 120 91% 117% March 143 153 87 94% 165% April 48 64 68 75% 71% May 30 48 63 62% 47% June 29 20 83 144% 35% Annual 733 843 1,187 87% 62% L 200 | | | | 150 = = = > 100 > oa c wi 50 0 | 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 3-9 Project Performance 3.2.4 Lost Energy due to Downtime Table 3-7 shows the estimated amount of energy lost during the downtime periods in each month. This calculated energy value, based on the manufacturer’s power curve at site air density and the met tower wind speed at hub height during the downtime periods, is termed “lost energy” in this report.* Approximately 28.4 MWh were lost during downtime periods in the first year of operation. Due to missing data this lost energy estimate does not include Turbines 4-10 for July through November 1999. However, every indication is that Turbines 4-10 had high availability and performed well during the period of missing data. Table 3-7 Downtime and Lost Energy by Month — July 1999 to June 2000 Downtime Lost Energy Lost Energy/ Month Hours [1] (MWh) [1] Downtime Hour [1] July 1999 182.7 1.1 5.9 August 117.2 3.1 26.3 September 99.2 0.5 46 October 484.5 6.7 13.8 November 80 1.6 20.0 December 154.3 1.0 6.7 January 2000 334.8 7.9 23.7 February 86.8 1.5 17.2 March 0.0 0.0 0.0 April 233.8 4.4 18.8 May 0.0 0.0 0.0 June 202.2 0.7 3.3 Annual 1,975.5 28.4 14.4 1] Data for July through November includes the Phase 1 turbines only. 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 ‘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. 3-10 Project Performance anemometers are on the tower at approximately 18 m (59 ft) 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. The periods with the most lost energy do not necessarily coincide with the periods of the most downtime. Table 3-7 also shows the lost energy per downtime hour by month. Although November downtime was 80 hours compared to 99 hours in September, the estimated lost energy is three times higher for November than September. Similarly, August downtime was lower than July but the energy lost in August was significantly higher. Figure 3-3 shows the actual energy production and 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. Energy (MWh) Wind Speed (m/s) Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun [30 Actual MWh ELost MWh —— Wind Speed (m/s) Figure 3-3 Actual and Lost Energy by Month — July 1999 to June 2000 Figure 3-4 shows the actual and lost energy for each turbine. The sum of the actual energy produced and the estimated energy lost during downtime events should approximate the total energy possible. Consequently, the sum of the actual and lost energy should be nearly the same for all turbines with only minor variations that account for individual turbine losses. As shown in the figure, Turbine 2 generated significantly lower energy than the other turbines. Turbines 3 and 6 also generated somewhat lower energy. All three turbines experienced problems with their tip brakes. Although a turbine generally continues to operate when a tip brake deploys prematurely, the turbine output is significantly reduced. Consequently the turbine is considered to be available and the lost energy is not estimated, but the actual energy generated during these events is very low. Section 4 addresses the tip brake events and other causes of turbine downtime. 3-11 Project Performance 100 = —— 80 ~ E 70 | 60 50 | 40 | 30 | 20 | 10 | Energy (MWh) DActual MWh Lost MWh | Figure 3-4 Actual and Lost Energy by Turbine — July 1999 to June 2000 3.3 Utility Demand and Project Energy 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. 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. In addition, the output from a project has a different value to a utility at different times of the year due to variations of the utility load profile. Figure 3-5 compares the typical seasonal energy demand at KEA with the long-term estimated wind energy from the project. As illustrated, the long-term wind resource provides a good match to the utility demand with the highest wind energy output matching the highest demand which generally occurs between October and February. 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 Project Performance = Se oot | 250 2,500 = e | = 200 2,000 z > 2 2 & 150 1,500 $ : i To © 100 1,000 & a $ 8 50 500 Ww 3 3s Z g 2 Fe (went) 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec | (MM Projected Wind Energy —Total Energy Demand Figure 3-5 Wind Energy Contribution to KEA Energy Demand Energy production data from KEA’s diesel power generators were available for the September 2000 reporting period (August 20-Septemeber 20). These data were combined with wind energy production data from the SCADA system for the same time period to calculate wind energy penetration values for the site. These values are shown graphically in Figure 3-6 and are summarized in Table 3-8. During the month, the maximum hourly penetration of approximately 35% was achieved during the hour from 5 AM to 6 AM on August 29. During this time period, energy demand was extremely low due to the time of day and favorable weather conditions (temperatures were over 3°C higher than the same time period on any other day during the month), while wind speeds were high resulting in the site operating at close to 100% of capacity. No power quality problems were observed during the periods of high penetration; both the frequency and voltage were close to normal throughout the month. 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 of less than 1% were calculated for approximately 46% of the month; penetration values of less than 10% were calculated for approximately 77% of the month. Penetration varied only slightly with the time of day. During the nighttime hours (from midnight to 8 AM), the median penetration was 0.71%, which is somewhat below the overall median value. This reflects the generally lower wind speeds during this period. However, the nighttime average penetration was 6.0%, reflecting the lower energy demand during this period. The opposite trend was seen during evening hours (5 PM to midnight), with a higher median penetration (3.0%) due to higher winds but a lower average penetration (5.3%) due to higher energy demand. 3-13 Project Performance It is worth noting that the high penetration seen during some time periods in this month are not likely to be seen during other times during the year, when overnight temperatures are much lower and energy demand would therefore be higher. Consequently, it is extremely unlikely that penetrations in excess of 35% would be seen, and overall average penetration would likely be somewhat lower if averaged over the entire year. 40% | 35% | 30% 5 25% | Ss o & 20% | o & 15% | é fo 10% 4 5% | 0% | 0 20 40 60 80 Percentile Figure 3-6 Overall Wind Energy Penetration on KEA System Table 3-8 Wind Energy Penetration —- September 2000 Night Day Evening Parameter Overall (12 AM - 8 AM) (8 AM - 5 PM) (5 PM - 12 AM) Maximum penetration 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% crime with: Isssithaii-176 46.2% 50.5% 46.8% 40.5% penetration ; : 7 hime with: less'tnan 1072 77.1% 74.6% 76.9% 80.4% penetration ; ; . Time with less than’ 20% 92.9% 91.0% 93.4% 95.0% penetration 3-14 4 PROJECT OPERATIONS AND MANAGEMENT 4.1 KEA’s O&M Strategy KEA personnel handle all of the ongoing operations and maintenance requirements of the KEA wind project with occasional support from their electrical support contractor, Thompson Engineering of Anchorage, Alaska. Since the first installation of turbines, AOC has traveled to Kotzebue several times 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. After the installation of the Phase 1 turbines, KEA recognized that they needed to hire a full-time engineer to work on their wind energy program. In March 1998, KEA hired Matt Bergan, a wind energy engineer, who was formerly employed at AOC. Matt’s immediate responsibilities included the ongoing O&M for Phase | and the construction oversight for Phases 2 and 3. He also is working on KEA’s wind activities in Wales, Alaska. His experience at AOC was a significant benefit to the KEA wind program. Other full-time KEA employees also help out with the wind project tasks and additional local labor is hired as needed. The O&M strategy at the KEA wind project is largely influenced by the other activities of KEA. KEA is a small utility with approximately 15 full-time employees. Although the wind energy engineer was hired specifically to operate the wind project, he performs many other duties unrelated to the wind project. The reliable operation and timely maintenance of the diesel generators takes precedence over the operation of the wind project. In addition, KEA’s involvement in the construction of a small wind project in the remote community of Wales, Alaska has required several months of the wind energy engineer’s time during the past year. As KEA begins project expansion activities they will be looking for a second full-time wind technician to work with the wind energy engineer. KEA expects to be involved in wind project development in other Northwest Alaskan communities which will require this additional labor support. In spite of the limitations inherent in operating a wind project in this small remote community, the turbines have experienced higher than expected availability during the first year of operation. 4.2 Maintenance Activities and Other Downtime Events The KEA project experienced approximately 2,000 total downtime hours during the first year of the operation. Due to on-site start-up problems with the SCADA system, the data necessary to determine the downtime events for Turbines 4 through 10 from July through November 1999 are not available. However, based on anecdotal site data and turbine-to-turbine production comparisons it was determined that these turbines experienced high availability during the period 4-1 Project Operations and Management of missing data. As discussed in Section 3, while the exact downtime numbers are slightly low, the actual overall turbine availability for the project is likely higher than presented in this report. Although availability is commonly used as a performance measure 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 1999 through June 2000. The two highest wind months, February and March, coincide with two of the highest availability months. As shown in the figure, the monthly wind speeds during the reporting period were relatively low in July 1999 and between March and June 20000, which is consistent with the long-term wind resource. 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. However, during the first year of operation, typical “scheduled maintenance” activities have been combined with unscheduled maintenance activities and additional scheduled time was not required. 8 100% | ao ate 9, | Eo _} 80% 2 | z, 60% & = | a + 40% | Bo - : . a a = + 20% F 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 1999 to June 2000 In addition to the time of occurrence, the cause of downtime and the cost to return a turbine to service also are 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. The following sections of this chapter address the cause and impact of the project downtime. The downtime is categorized and the impact of events is discussed in terms of hours and lost energy, if appropriate. As discussed in the previous chapter, the lost energy for each downtime period was calculated for each turbine on an event-by-event basis that considered the actual wind conditions at the site during the time of the event. 4-2 Project Operations and Management 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 first year of operation. As previously discussed, this analysis does not include downtime for Turbines 4 through 10 from July 1999 through November 1999. The categories are: O&M-This category includes all troubleshooting, adjustments, retrofits, and repairs performed with the exception of the downtime incurred on the Turbine 2 storm damage discussed below. It includes downtime that accumulated while waiting for parts, instruction, 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 1,322 hours or 68% of the total downtime during the first year of operation. Storm Damage-This category includes the downtime attributed to storm damage on Turbine 2 that occurred on September 30, 1999. A maintenance rope hangs down the center of the lattice tower and is used to lift parts and tools up to the crew when working up-tower. After the maintenance work is complete, the rope is secured to the leg of the tower. An unusually gusty wind storm beat the rope against the tower leg until the rope came unraveled and broke free. The wind caught the rope and swung it into the path of the rotating blades. The rope wrapped around the rotary transformer and became entangled in the tip brake. This incident accounts for approximately 433 hours of downtime or 22% of the total downtime during the first year of operation. Repair work was delayed because site personnel were not available to make repairs. Faults—This category includes only those faults that required a reset and no other action. If a maintenance activity immediately followed a fault, the downtime hours associated with the fault have been 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 was re-classified as an O&M event if sufficient information was 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 was reset is included in the fault category as long as the fault was not followed by maintenance. Faults account for approximately 101 hours or 5% of the total downtime during the first year of operation. Line Outages—Specific, identifiable line outages are included in this category. KEA does not track the date and duration of line outages. Therefore, several brief line outages are likely to have occurred during the year that went undetected. The line outages included in this category were documented and reported by the site personnel. During the first year of operation, line outages account for about 18 hours per turbine, or 1% of the total downtime. 4-3 Proj ect Operations and Management Other-This category includes downtime for several brief turbine outages where no documentation is available to determine the cause. This category also includes a few hours when turbines were taken off line during the commissioning of the Second Wind SCADA system and during the installation of the primary site meter. Finally, this category includes a few hours of downtime while KEA was verifying the configuration of the current transformers (CTs), and documenting serial numbers of the turbine generators and blades. During the first year of operation, this category accounts for 83 hours or approximately 4% of the total downtime. The following sections of this chapter describe additional analysis of the downtime associated with faults and O&M activities. — _ _—_______ Total Downtime: 1,957 hours 2 Oth | Storm er Fault . | 9, Line Outage Damage 4.3% 5.1% 0.9% 9 | Figure 4-2 Total Project Downtime by Cause — July 1999 to June 2000 Figure 4-3 shows the total lost energy by category at the KEA project during the first year of operation. O&M downtime accounts for about 18,385 kWh of lost energy; Turbine 2 repairs account for 5,279 kWh; faults account for 1,652 kWh; line outages account for 587 kWh, and other downtime accounts for 1,894 kWh. The contribution of downtime hours to lost energy is fairly consistent for each of the downtime categories. 4-4 Project Operations and Management Total Lost Energy: 27,797 kWh WT 2 Storm Other paul Line Outage Damage 6.8% 5.9% 21% 19.0% Ba O&M 66.1% Figure 4-3 Total Lost Energy by Cause — July 1999 to June 2000 Figure 4-4 shows the downtime for each wind turbine by category. The downtime associated with O&M activity is particularly apparent on Turbines 2 and 3 as well as the storm damage on Turbine 2. Besides Turbines 2 and 3, the only other turbines with more than 100 hours of downtime during the year were Turbines 6, 7 and 8. wy Hours b oO oO | | | MFault Cine Outage Figure 4-4 Total Project Downtime by Turbine — July 1999 to June 2000 4-5 Project Operations and Management Figure 4-4 also illustrates the variability of the total downtime experienced by individual turbines. Turbines 2 and 3 experienced well above the average downtime hours during the first year of operation, accounting for nearly 65% of the total project downtime. In contrast, Turbines 1, 4, 5, and 9 had significantly less downtime and together accounted for only 10% of the total project downtime. Appendix D presents additional information on the specific causes of downtime for each turbine. Figure 4-5 shows the downtime for each month by category and the monthly mean wind speeds during the first year of operation. The project experienced 100% availability for the reporting months of March and May 2000. The high wind speeds in March coincided with the 100% availability and resulted in record production of 493 kWh per turbine per day, a capacity factor of over 31%. The project experienced the most downtime during October due to the Turbine 2 storm damage discussed earlier. The second highest downtime month was January which is attributed to the replacement of the 120 V surge suppressor and two fuses on Turbines 6, 7, and 8. 8 7 -_ ”o ce g ° 3 | 3 3 = 30. 7 2 | Ss } ° : an | | Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun | = Fault Oe co Line Outage a | ZZAO&M Other | WT Maint Incident —— Wind Speed EE — _] Figure 4-5 Total Project Downtime by Month — July 1999 to June 2000 4.2.2 Downtime Due To O&M Activities As discussed in Section 3, the TVP availability was approximately 97% during the first year of operation, and total turbine downtime was 1,957 hours. The majority of the downtime (1,322 hours) was attributed to O&M activities and accounted for 68% of the downtime. The majority of the O&M downtime was attributed to tip brake problems and electrical repairs. Tip brake problems and electrical repairs account for 50% and caused 34% respectively of the O&M downtime during the year. Figures 4-6 and 4-7 show a breakdown of the downtime and related energy losses included in the O&M category by major turbine component. Figures 4-8 and 4-9 show the categorical distribution of O&M downtime by month and by turbine. Some general observations on the failures and repairs are provided below. 4-6 Project Operations and Management O&M Downtime: 1,322 hours Anemometer 5.0% Controller 9.9% Tip Brake Electrical 50.3% Other 34.3% 0.5% Figure 4-6 O&M Downtime by Cause — July 1999 to June 2000 Lost Energy during O&M: 18,385 kWh Anemometer Controller 4.2% 2.1% | Other Electrical | .O, | 34.6% 0.2% 58.9% | Figure 4-7 Lost Energy due to O&M by Cause — July 1999 to June 2000 4-7 Project Operations and Management O&M Downtime: 1,322 hours 400 | | 300 ;—— 200 oi 0+ | Jul Aug ean Oct Nov Dec Jan Feb Mar Apr May Jun Hours |MControlier Electrical eee Brake OOther | Figure 4-8 O&M Downtime by Month — July 1999 to June 2000 O&M Downtime: 1,322 hours | 500 400 — + | 2 300 = aaaseeeeeneeesneee = 3 x 200 + - — i : | | 100 +—— : T=, FY ” | YG ZY Y | > Mn ZA oo ; _a | | 1 2 3 4 5 6 7 8 9 10 | Turbine ‘mController elisa Tip Brake “Other | Figure 4-9 O&M Downtime by Turbine — July 1999 to June 2000 Tip brakes. Tip brake problems resulted in a total of 664 hours of the O&M downtime during the first year of operation. These downtime hours included time for troubleshooting and repairs and accounted for approximately 26 turbine events. These hours do not include the downtime hours for tip brake damage attributed to the Turbine 2 storm damage. The tip brake downtime accounts for 50% of the O&M downtime and 34% of the total downtime during the first year of operation. An important aspect of tip brake problems is reduced energy output when a tip brake is prematurely deployed. When a tip brake deploys prematurely, the turbine is available and continues to operate but produces significantly less power. These energy losses are in addition to the energy lost during recorded downtime events and are more difficult to quantify accurately. 4-8 Project Operations and Management This was evident in the November 1999 production and availability for Turbine 2. Although the November availability for Turbine 2 was 100%, it essentially never ran, producing only 4 kWh compared to the average turbine production of approximately 4,350 kWh. Similarly, in April 2000, Turbine 2 was available 100% of the time but generated only 2,065 kWh compared to an average turbine production of 5,120 kWh. As shown in Figure 4-8, a significant percentage of the tip brake downtime hours occurred during July 1999 and June 2000. The July downtime consisted of several events with all but one event affecting Turbine 3, whereas the June downtime consisted of one long event on Turbine 2. The remaining tip brake downtime occurred during August, September, and November 1999 and April 2000. As illustrated in Figure 4-9, Turbines 2 and 3 experienced 92% of the total tip brake downtime during the first year of operation. The longest tip brake event occurred during the June 2000 reporting month on Turbine 2 and lasted nearly 200 hours. Although data are missing for Turbines 4 through 10 for the first five months of the annual reporting period, available data indicate that the newer turbines have experienced fewer significant tip brake problems. AOC has continually reviewed tip brake problems as they have occurred on the KEA turbines as well as on other AOC turbines. AOC has improved the reliability of the tip brakes by upgrading several small components. This includes a number of modifications made to the AOC tip brake assemblies of the Phase | turbines. The improvements include stronger magnets, improved alignment, a more rugged damper bracket-to-hinge block joint, and improved catch plates and catch plate grommets. These improvements have been incorporated into Turbines 1, 2, and 3 and were standard in Turbines 4 through 10. Additional improvements that will be incorporated into Turbines 1 through 10 include 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 that the new hinge eyes will significantly reduce the chance of loose tip brake plates. The upgraded hinge eyes address the problem that caused the tip brake failure on Turbine 6 in April 2000. 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. A majority of tip brake downtime on Turbines 2 and 3 was caused by a blade cable interfering with the tip brake and preventing it from closing properly. 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. Appendix E provides a diagram of the AOC tip brake assembly. Electrical. During the first year of operation, 454 hours are attributed to electrical events. The majority of the electrical downtime, 313 hours, was for the replacement of the 120 V suppressor in mid-January on Turbines 6, 7, and 8. KEA personnel suspect that the failed suppressors came from a bad batch. There have been no problems with the suppressors on the other turbines. The remaining electrical downtime was caused by miscellaneous fuses and sensors that required troubleshooting and replacement. This category accounts for 34% of the O&M downtime and approximately 23% of the project’s total downtime during the first year. 4-9 Project Operations and Management Controller. During the first year of operation, the project experienced 131 hours of control system O&M downtime. All of the controller downtime occurred in early December on Turbines 1, 2, and 3. KEA personnel replaced and rewired the terminal strip in the controller cabinet on all three turbines. Replacement of terminal strips generally is not a difficult or time-consuming task. Many of the hours attributed to this event are likely related to the limited daylight and harsh temperatures during December. This category accounts for 10% of the O&M downtime and approximately 7% of total downtime during the first year. Anemometers. Failed anemometers accounted for 66 hours of O&M downtime. In August 1999, Turbine 3 was shut down by the turbine controller because both anemometers had failed. In April 2000, Turbine 3 had numerous slow-start events that were attributed to a failed anemometer and were therefore counted as turbine downtime. This category accounts for 5% of the O&M downtime and approximately 3% of total project downtime. Other O&M. Other O&M consisted of only seven hours during the first year of operation. This included a couple of hours a turbine was shut down for safety purposes while site personnel performed maintenance on a nearby turbine. There also were a couple of very brief inspections in this category. The remaining three hours in this category consisted of an undefined maintenance event. The seven hours in this category accounted for 0.5% of the project’s O&M downtime during the first year. 4.2.3 Downtime Due To Faults Figure 4-10 shows a breakdown of fault downtime by type for the KEA turbines during the first year of operation. The fault downtime for the project was very low with approximately 101 hours of downtime and an estimated energy loss of 1,650 kWh. This represents only 10 hours per turbine or 5% of the total project downtime during the first year of operation. The actual number is probably a little higher due to the missing data for Turbines 4 through 10. The majority of the fault downtime, 60.5 hours, was for unknown reasons. Yaw faults account for 33.5 hours of fault downtime with the remaining seven hours attributed to electrical faults. 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. For example, the Big Spring project tends to experience a higher number of faults at night due to the characteristics of the wind resource. Turbines that fault off at night are not reset until morning. Fault downtime as well as O&M downtime could be reduced by incorporating a SCADA call-out feature that notifies a remote operator when a turbine is off-line. 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 would be programmed to call the KEA office in town during business hours when a turbine went off line. On the weekends, the wind energy engineer or other KEA personnel would be notified in the event of a turbine outage. 4-10 Project Operations and Management Fault Downtime: 101 hours Electrical Yaw 6.6% | 33.3% Unknown 60.1% Figure 4-10 Breakdown of Fault Cause by Type — July 1999 to June 2000 As shown in Figure 4-11, nearly all fault downtime occurred during February and April 2000 and a few events occurred in December 1999 and January 2000. As illustrated in the graph, while there was more fault downtime in February and April, significantly fewer events occurred than in December and January. Figure 4-12 illustrates the distribution and cause of fault downtime among the turbines. There were no recorded fault events for Turbines 1 and 10. Turbine 9 experienced the highest fault downtime, most of which was for unknown reasons. The yaw faults are attributed to a one-time occurrence during the April reporting period when Turbines 6, 7, and 8 were turned upwind and required resetting by KEA personnel. Turbines 2, 3, and 4 experienced the majority of the 7 hours of electrical fault time. 4-11 Project Operations and Management Total Hours 70 Number @ Total Hours Number of Faults | 14 Figure 4-11 Fault Frequency and Duration by Type — July 1999 to June 2000 Fault Downtime: 101 hours | 80 | 40 +— |. £ 30 +——— 3 x 20 —— 10} 0 [ye 1 2 3 r ed | MYaw = CElectrical MHUnknown Figure 4-12 Breakdown of Fault Cause by Turbine — July 1999 to June 2000 4-12 Project Operations and Management 4.3 SCADA System Experience The AOC turbine initially was designed as a stand-alone system and AOC did not offer any type of Supervisory Control and Data Acquisition (SCADA) system at the time of KEA’s original purchase order. However, KEA recognized the need to collect turbine data and the value of remote monitoring and control capabilities. KEA contracted with Island Technologies Incorporated Inc. (ITI) to design and install a system that would enable remote monitoring and recording of wind turbine performance data and also to enable some simple remote control of wind turbine functions. The system was designed around the Campbell Scientific CSI CR10X Data Logger that ITI has used in several other remote and arctic wind turbine and power plant monitoring projects. The system was installed in early 1998 for control and monitoring of the Phase 1 turbines. As part of KEA’s involvement with the TVP, NREL provided a Second Wind SCADA system to the KEA project. Second Wind and GEC personnel provided on-site support to KEA for the hardware installation. Preliminary commissioning of the SCADA system was performed in September 1999 by Second Wind personnel. The Campbell system continued to collect data for the Phase | turbines for a few months after the Second Wind system was commissioned. Having a redundant SCADA system was a valuable tool in understanding the data collected through the new Second Wind system. 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. Although the Second Wind SCADA system began recording data in late August and was fully commissioned in early September, problems with KEA communication lines resulted in low data recovery for several months. Table 4-1 presents the data recovery rates for SCADA data downloaded from the Campbell system during the first few months of the reporting period, followed by the Second Wind system during the remaining months. The table includes recovery rates for the turbine data, the Phaser™ data and the met data. KEA experienced several months of low data recovery from their SCADA system. 4-13 Project Operations and Management Table 4-1 Recovery Rates for SCADA System Data — July 1999 to June 2000 Month Turbine Phaser™ [2] Met [3] July [1] 66.9% N/A 66.9% August [1] 100.0% N/A 99.6% September [1] 100.0% N/A 99.7% October [1] 100.0% N/A 100.0% November 58.0% 58.0% 99.2% December 85.0% 87.3% 100.0% January 87.3% 71.0% 87.4% February 95.6% 91.8% 96.4% March 96.8% 91.2% 96.7% April 100.0% 92.8% 93.5% May 78.0% 78.0% 78.0% June 96.8% 93.8% 74.5% Annual 88.9% 94.9% 91.2% [1] Values are based on the data collected from the Campbell SCADA system. [2] Values are based on Real Turbine Power data collected from the Second Wind Phasers™. [3] Values are based on recovery of hub height wind speed data. In September and October, KEA was debugging the communication systems that connected the Second Wind computer, at the project site, to the KEA office in Kotzebue. Consequently the data recovery was so low that KEA and the TVP continued to rely on the Campbell SCADA system for the majority of monthly data analysis. After the communication problems were resolved, the computer provided with the Second Wind system began having problems. After several weeks, as the SCADA program collected data and wrote records to the database, the computer performance became noticeably slower. The computer also locked up at times and, if KEA personnel were not available to visit the site and reset the computer, data were lost. Other TVP sites with Second Wind systems encountered similar problems. Second Wind and GEC worked with each of the TVP projects with Second Wind SCADA systems to define differences and similarities anong SCADA issues. They determined that the increasing size of the database was using too many computer resources and causing the system to become unstable. GEC developed a procedure specific to each TVP project for archiving data 4-14 Project Operations and Management and clearing out the database on a monthly basis. Regular archiving of the data improved system reliability and data recovery increased. Although data recovery improved, KEA continued to experience some computer lockups. Second Wind determined that the SCADA program used was not fully compatible with the new 32-bit architecture used by the Windows 98 operating system installed on the SCADA computers. The Second Wind software has been upgraded to 32-bit architecture and fully tested at the TVP projects in Algona, Iowa and Big Spring, Texas. A system upgrade was recently installed at KEA and are being scheduled other TVP sites. 4.4 Potential Performance Improvements 4.4.1 Slow-starting Turbines Through data analysis and observations at the wind site, KEA and GEC determined that 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 cause of slow-starting turbines. GEC analyzed several months of data to better understand the characteristics of the slow-start events. Sue Childs of AOC met with GEC in May 2000 to discuss the issue, review the preliminary analysis, and plan additional activities. Further investigation indicated that the slow- start condition was not caused by cold temperatures and was probably related to rotor drag. However, the source of the drag could not be determined through data analysis alone. AOC discussed the slow-start issue with some of their component vendors. Potential slow-start factors that were considered and discounted include main bearing and gear friction, temperature- dependent drive train friction, and blade pitch angle. During a site visit to Kotzebue 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. To determine the static drive train friction, a pull test was conducted that measured the torque required to rotate the rotor from rest with the parking brake released. Turbines 6 and 10 were tested because they are the turbines that are respectively most and least prone to slow-starts. Measurements were repeated several times on each turbine. The test yielded similar results for both turbines. For the yaw bearing pull test, in addition to Turbines 6 and 10, Turbines 5 and 7 were tested. Turbines 5 and 7 are respectively the second most and least prone to slow-starts. From the test results, AOC concluded the following: e Although testing was done under a limited number of conditions, drive train (brake) friction does not appear to be a factor. e Turbines 6 and 7, the two turbines most prone to slow-starts, had much higher yaw bearing friction than Turbines 10 and 5, the two turbines least prone to slow-starts. 4-15 Project Operations and Management e Additional data should be collected and analyzed after the yaw bearings on Turbines 6 and 7 are lubricated and retested. An indication of yaw bearing friction implies that the turbines are slow to start because they are not yawing into the prevailing wind in a timely manner. However, site personnel have observed occasions when the turbines appear to be pointed into the wind properly and do not start up, while neighboring turbines have already come online. It seems likely that there are multiple contributing factors to the slow-start problem. AOC, KEA, and GEC will continue to work together to define the cause of the slow-starts and improve the performance of the KEA turbines. 4.4.2 Brake-cooling Cycles Each time the electro-dynamic and parking brakes are applied, the AOC turbine is programmed to initiate a 15-minute brake-cooling cycle during which time the turbine will not be allowed to operate. The brake-cooling cycle is necessary to prevent the possibility of multiple braking cycles in a short period of time which could cause overheating of the brakes. When the winds are hovering around cut-in (4.6 m/s) and are variable, the turbine controller will release the turbine brake, the turbine will begin to spin up or operate for a short time and then the wind will drop and the turbine will shut down. During the next 15 minutes, if the wind speed increases above cut-in speed, the turbine cannot restart because it is still in the brake-cooling cycle. The duration of the standard AOC brake-cooling cycle is based on normal operating conditions, including warm or even hot climates. GEC performed some analysis and determined that measurable energy is lost during these 15-minute cycles. AOC considered the specifications of the brakes and the cool Kotzebue climate and determined that a reduction in the duration of the brake-cooling cycle was acceptable for low-wind braking events. AOC has recently reduced the brake-cooling cycle to five minutes for low-wind braking events on all turbines except for Turbine 8, the TVP test turbine. Section 4.5 describes the test turbine configuration. 4.4.3 Turbine Over-Production 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.” As part of the monthly TVP reporting activities, GEC began tracking the maximum 10-minute peak power output from the turbines shortly after the Phase 1 turbines were installed in 1997. After the installation of the Second Wind SCADA system, GEC began tracking the instantaneous peak power of each of the 10 turbines. Figure 4-13 shows the peak power output from the KEA turbines on a monthly basis. The 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. 5 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 Management Min Temp (C) 20 -——_______________________+ -70 | Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun | Max Power = = = Min Temp | Figure 4-13 Peak Output Compared to Minimum Temperatures-July 1999 to June 2000 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. Although the Fairfield engineers concluded that the turbines are not in danger of catastrophic failure, the impact of running the turbines at the current peak output on the gearbox life was not determined. AOC’s preliminary conclusion is that the gearbox life would be negatively impacted over a period of time. AOC engineers are performing additional investigation and analysis to determine the appropriate level of peak power output. The AOC 15/50 turbine is a fixed-pitch turbine. It is common industry practice to make seasonal pitch adjustment for fixed-pitch turbines to account for changes in air density or wind speed, thereby optimizing energy production. A reduction in peak power output is accomplished by reducing the pitch of the turbine blades causing blade stall at a lower wind speed. While blade pitching is common in the industry, the AOC turbine was not designed to be repitched after turbine installation. Consequently, AOC had to design a pitch measuring device that was then used to measure the original pitch angle of the blades and then to reduce the angle of the pitch. The new pitch angle was recently incorporated on the TVP test turbine, Turbine 8. The pitch angle was reduced from the original setting of 1.15° to -0.6°. Data are currently being reviewed to determine the yield of the new pitch setting. When the appropriate blade pitch has been verified, AOC and KEA expect to repitch the remaining nine turbines. 4.5 Turbine Power Performance Test 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 meteorological 4-17 Project Operations and Management 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 test also enabled the TVP to gain experience with the IEC power performance measurement standard and to provide feedback to the IEC working group responsible for developing the Standard. As previously discussed, the Second Wind SCADA system was installed and commissioned in late 1999. This system was used to collect data for the performance testing. The test data collection period ran from November 8, 1999 to May 31, 2000. A post-test recalibration of the primary anemometer was performed on June 28, 2000. In accordance with the Standard, test data were considered valid only from a certain direction sector determined to be free from obstructions. This sector was identified through a defined site assessment process that excluded any direction sectors in which the met tower or test turbine (Turbine 8) would be affected by the wake of nearby turbines. Figure 4-14 shows the locations of the site reference met tower, and the wind turbines relative to the surrounding terrain. The valid measurement sector is 32.5° clockwise to 225°. 4.5.1 Test Methodology According to the IEC Standard, the test site must be evaluated by three criteria to determine which data must be excluded from the database: topographic variations, neighboring and operating wind turbines, and other significant obstacles. If the topography of the site meets the criteria given in the IEC Standard, then the wind speed at the meteorological mast is assumed to be identical to the wind speed at the wind turbine. Ifthe topographic variations exceed the criteria, then an experimental calibration of the test site is required to determine corrections to the wind speed. The neighboring wind turbines were evaluated to determine whether they disturb the free stream wind velocity. There were no other significant obstacles. Using the criteria set forth in the IEC Standard, direction sectors in which the free stream wind flow is disturbed by the adjacent turbines, were excluded from the database. Due to the flat terrain at the KEA site and the absence of obstacles, a site calibration was not required. 4-18 Project Operations and Management = N yes mh, ~32 5.9. ¥ : / . wrs © wr7s WIP \ wT2¢@ wree® | jw \ } wrie WTB.@ { wre eto | WT4e \ | Ip — eee nee 8L. ees a from 32.5 to 225 degfees 225° es 0 1000 2000 3000 FEET J dl 0 200 400 600 800 1000 METERS Figure 4-14 KEA Topographic Site Map Used for Power Performance Test During the test period November 8, 1999 through May 31, 2000, 2,198 hours of valid data were collected while the wind direction was within the valid measurement sector (32.5° to 225°), icing conditions were not apparent, and the turbine was available. The highest wind speed bin filled with at least three valid 10-minute average data points (and with wind speed normalized to sea- level density) was the 17.0 m/s bin. According to the Standard, data should be collected up to a wind speed equal to 1.5 times the wind speed at which the turbine reaches 85% of its rated power. Based on the TVP definition, the AOC 15/50 turbine is rated at 66 kW, and the power curve provided by the manufacturer indicates that the performance data should be collected up to 19.0 m/s. 4.5.2 Test Results 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 and reaches a peak power output of approximately 68 kW at about 14.5 m/s. 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/m3) the peak output is 75.5 kW at 14.5 m/s. 4-19 Project Operations and Management Consistent with the slow-start problem discussed previously, the body of the measured power curve for low to moderate winds (5 to 11 m/s) 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% to 30%. Figures 4-15 through 4-17 present the results of the power performance test after adjustment to sea-level. The error bars on the power curves represent the combined uncertainty calculated in accordance with IEC 61400-12. They include the error introduced by the standard deviation of the measured data points and the uncertainty in the sensors and data acquisition system. No error bars are shown for the data points with less than three 10-minute samples in the bin. 75 70 . a e¢ 65 oo ; ; cacaas 60 ‘ ' 32° g 55 ae -- 3 50 - 3 ° 45 g - 40 a 35 : 3 304----- 3 25 flees 3 20 me 3 15 r 8: 10} - - flees ied 5 =, st @.. 0 000000000007 | 5 - | 0 2 4 6 8 10 12 14 16 18 20 | Wind Speed (m/s) Figure 4-15 Power Curve at Sea-Level Density, 1.225 kg/m’ Electrical Power (kW) 0.4 | tt | 03 ,t 4+, fs | Boz ¢ The d: | 2s + mc) + +e | ‘= o1 4 % | o | 5 0.0 5 . | 5 oe | z -0.1 | mn a -0.2 + -0.3 —— : 0 2 4 6 8 10 12 14 16 18 20 Wind Speed (m/s) | aE REET naire sae) Figure 4-16 Power Coefficient at Sea-Level Density, 1.225 kg/m’ 4-20 Project Operations and Management 80 ¢ Minimum Power | | © Maximum Power | 0 2 4 6 8 10 12 14 16 18 20 Wind Speed (m/s) Figure 4-17 Maximum and Minimum Power Measured for the AOC 15/50 Wind Turbine Table 4-2 presents the gross annual energy production (AEP) results for the measured and extrapolated power curves along with the associated uncertainties. The measured AEP includes only those wind speed bins for which there are three or more data points. The extrapolated AEP assumes the power at higher wind speed is equal to the power of the last filled wind speed bin. Wherever the energy production from the measured curve is 95% or greater than the energy for the extrapolated curve, the resulting data are labeled complete. When the measured energy is less than 95% of the extrapolated energy, the data are labeled incomplete. 4.5.3 Future Performance Testing At the conclusion of the power performance test, the met tower was lowered, and the hub-height anemometers were removed and submitted for post-calibration. The original anemometers were replaced and the met tower was reinstalled. The blades on Turbine 8 recently were repitched to reduce the maximum power output. AOC is investigation potential solutions to the slow-start problem issue. NREL, AOC and GEC are discussing a second power performance test. When a solution to the slow-start is determined, it would be valuable to retest Turbine 8 as an “optimally” configured AOC turbine for the KEA site. At present, the turbine is being tested informally to determine the appropriate pitch angle for the KEA turbines. For this reason, no other configuration changes have been made to the turbine, such as the reduction in the brake-cooling cycle, which was recently performed on the other turbines. 4-21 Project Operations and Management The testing program at KEA provides the TVP with a unique opportunity to test an arctic installation in multiple configurations. AOC and KEA have been very supportive of the testing efforts and the results are likely to benefit all parties. Table 4-2 Annual Energy Production at Sea-Level Density, 1.225 kg/m* Rayleigh Mean AEP-Measured (from AEP-Extrapolated (from Hub Height measured power Uncertainty of AEP- extrapolated power Wind Speed curve) Measured curve) (m/s) (MWh/yr) Status (MWh/yr) (%) (MWh/yr) 4 18.2 Complete 2.6 14.15 18.2 5 50.9 Complete 5.1 10.15 51.0 6 94.9 Complete 7.5 7.9% 95.6 7 140.6 Complete 9.4 6.7% 144.8 8 179.7 Incomplete 10.5 5.8% 191.9 9 207.9 Incomplete 11.0 5.3% 232.7 10 224.8 Incomplete 11.1 5.0% 264.8 1 232.0 Incomplete 10.9 4.7% 287.6 4-22 5 OUTREACH ACTIVITIES AND FUTURE PLANS The primary objectives of KEA’s wind energy program are to bring more affordable electricity and jobs to remote Alaska communities. During its first year of operation, KEA continued its community outreach, technology transfer, and wind project planning and development. In March 2000, Brad Reeve received an R&D Achievement Award from the National Rural Electric Cooperative Association (NRECA) and its Cooperative Research Network. NRECA recognized Reeve for his service and leadership in the electric cooperative industry, and especially for his pioneering efforts to bring wind turbine technology to Kotzebue and the state of Alaska. In April 2000, KEA received AWEA’s utility award for its “pioneering efforts in support of wind energy applications in the Arctic.” According to AWEA Deputy Executive Director Tom Gray, “the Kotzebue Electric Association demonstrates that wind energy can be a reliable, money- saving complement to diesel as an energy source for remote communities.” 5.1 Community Education and Outreach Activities KEA continued to promote and participate actively in outreach activities within the local community and at the state and national level. KEA held a well-attended dedication ceremony for the project involving local and state dignitaries August 14, 1999. KEA expanded and updated the Wind Energy Area of its website® during the first reporting period. The 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. The website also contains a slide show that illustrates the raising of a wind turbine. KEA developed an information kiosk for the project site that explains the project to residents, tourists, and other site visitors. Site tours are provided 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. KEA continues to encourage coverage of the project by the news media through regular press releases, providing interviews, and hosting interested reporters. The project has been visited ° http://www.kotzelectric.com Outreach Activities and Future Plans numerous times by state officials, newspaper journalists, and television crews from around the state. KEA provides regular communication to other utilities of the project’s progress through Ruralite, a regional magazine distributed to rural utility membership. KEA provides regular project updates to cooperative members at its Annual Meetings. During the past year, a significant accomplishment for KEA and other electrical cooperatives in Alaska was the successful extension of the Power Cost Equalization (PCE) program. The high cost of electricity is currently subsidized by the Alaskan State Government through the PCE which helps reduce the cost of electricity in rural communities to 20 cents per kWh, an average reduction of approximately 52%. The PCE program had been slated for elimination or at least a significant reduction due to a decline in the state’s oil revenues. The elimination of the program or a change in the implementation would have had a significant impact on the cost of electricity in Kotzebue and the other northwest Alaskan villages. Fortunately, through the efforts of many, the PCE received a permanent endowment that will ensure the continuation of this vital program. 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 1999 and 2000, they presented papers 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 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. KEA’s general manager Brad Reeve has served as UWIG President for the past two 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. 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. Outreach Activities and Future Plans 5.3 Wind Project Planning and Development KEA is interested not only in reducing the future electricity costs of Kotzebue customers, but in supporting 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 consisting of two AOC 15/50 turbines connected to a small, isolated grid serving a community of 165 residents. The Wales wind project was completed during the Fall of 2000 making Wales the first community in Alaska powered almost exclusively by wind energy. KEA will own the turbines and sell the power to the Alaska Village Electric Cooperative (AVEC) who serves the electrical needs of the community. Wales has an excellent wind resource with an average annual wind speed of approximately 9.3 m/s (21 mph) at 26 m (85 ft). The two 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-art control system, developed by NREL, that allows the use of excess electricity produced during periods of high winds. KEA has also had some preliminary discussions with AOC related to the development of another project in a nearby community. This project would include one to three AOC 15/50 turbines and would be located in a community somewhat larger than Wales, with a population of approximately 650. KEA is working towards an eventual 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 19.7 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 include two additional AOC 15/50 turbines slated for installation in the Spring and Summer of 2001. 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. During the 10-turbine project’s first year of operation, the turbines performed well with a higher than expected availability. As wind speeds return to normal, the project is expected to meet its projected production levels. During the first year of operation, the project produced approximately 733 MWh of energy with an average TVP system availability of 96.8%. The project experienced an energy shortfall of approximately 38% based on an estimated long-term annual energy of 1,187 MWh. The majority of this shortfall, 66%, is attributed to the annual wind speeds which were approximately 16% lower than the long-term average. The remaining energy shortfall can be attributed to turbine performance anomalies that include slow-starting turbines and prematurely deployed tip brakes. During the reporting period, 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. KEA was responsible for the installation of two AOC 15/50 turbines in Wales, Alaska that began operation in October 2000. The Wales wind project is a collaborative effort of the Alaska Science and Technology Foundation, Alaska Energy Authority/Alaska Industrial Development Export Authority, KEA, Alaska Village Electric Cooperative, NREL, and the U.S. Environmental Protection Agency. KEA provided engineering, electrical, and management support and will ultimately own and operate the turbines. KEA will sell the energy to AVEC who provides electrical service to the residents of Wales. KEA anticipates an expansion of their 10-turbine project in the near future, possibly with a larger turbine, a different turbine type, or more AOC turbines. Project expansion will begin in the spring of 2001 with two additional AOC 15/50’s and could ultimately result in 2 to 4 MW of wind capacity for the utility. Other nearby communities are considering wind energy development with some preliminary negotiations already underway. The reliability of the SCADA system has improved and Second Wind has worked with KEA to determine system upgrades that will best serve the needs of the utility. A major upgrade of the Second Wind software was recently installed at KEA and is currently underway at the other TVP project sites. During this first year of operation of the 10-turbine project KEA installed a primary site meter that is being incorporated into the Second Wind SCADA system. 6-1 Conclusions The KEA project has also been effective in achieving 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. 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 industries 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 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. Lessons Learned at the Iowa and Nebraska Public Power Wind Projects, EPRI 1000962, November 2000. Project Development Experience at the Iowa and Nebraska Distributed Wind Generation Projects, EPRI TR-112835, December 1999. 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 First and Second Year Operating Experience: 1998-2000, EPRI 1000959, December 2000. 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. A-l TVP-Related Documents 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 Wind Turbine System Response to Lightning Activity: Preliminary Results. McNiff, 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. 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 Expereince ’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. A-2 TVP-Related Documents 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. Green Mountain Power’s Searsburg Project. B. Ralph, Green Mountain Power Corporation. Presented at WindPower 1999, Green Mountain Power’s 6-MW TVP Wind Project in Searsburg, Vermont. J. Zimmerman, Green Mountain Power Corporation. Presented at WindPower 1998. 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. McNiff, 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. Power Performance Testing Activities in the DOE-EPRI Turbine Verification Program. J. VandenBosche, T. McCoy, H. Rhoads, Global Energy Concepts, LLC, B. McNiff, McNiff Light Industry, B. Smith, National Renewable Energy Laboratory. 15 pp.; NICH Report No. CP-500- 28589. Presented at WindPower 2000. Program on Lightning Risk and Wind Turbine Generator Protection. Muljadi, E.; McNiff, B. National Renewable Energy Laboratory 8 pp.; NICH Report No. CP-440-23159. 1997. “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. A-3 TVP-Related Documents 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. 1999-2000. A-4 B MONTHLY AVAILABILITY AND PRODUCTION BY TURBINE KEA Turbine Availability - July 1999 through June 2000 WT Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr May Jun YTD 1 99.9% 99.5% 98.9% 99.1% 99.8% 92.5% 100.0% 100.0% 100.0% 100.0% 100.0% 99.8% 99.1% 2 100.0% 99.5% 99.3% 39.0% 89.7% 93.3% 99.5% 100.0% 100.0% 100.0% 100.0% 73.3% 91.1% 3 74.7% 85.3% 88.5% 94.7% 99.8% 95.7% 100.0% 100.0% 100.0% 94.4% 100.0% 100.0% 94.4% 4 N/A N/A N/A N/A N/A 100.0% 99.8% 100.0% 100.0% 98.7% 100.0% 100.0% 99.8% 5 N/A N/A N/A N/A N/A 100.0% 97.8% 93.4% 100.0% 100.0% 100.0% 99.7% 98.7% 6 N/A N/A N/A N/A N/A 100.0% 85.9% 100.0% 100.0% 91.6% 100.0% 100.0% 96.8% 7 N/A N/A N/A N/A N/A 100.0% 87.7% 100.0% 100.0% 98.5% 100.0% 100.0% 98.0% 8 N/A N/A N/A N/A N/A 99.8% 84.4% 100.0% 100.0% 98.5% 100.0% 100.0% 97.5% 9 N/A N/A N/A N/A N/A 100.0% 100.0% 93.8% 100.0% 100.0% 100.0% 100.0% 99.1% 10 N/A N/A N/A N/A N/A 97.4% 100.0% 100.0% 100.0% 87.0% 100.0% 100.0% 97.8% Project 91.5% 94.8% 95.6% 77.6% 96.4% 97.9% 95.5% 98.7% 100.0% 96.9% 100.0% 97.3% 96.8% KEA Energy Production - July 1999 through June 2000 (kWh) WT Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr May Jun Total 1 2,661 8,698 2,213 8,465 4,347 1,991 7,289 14,178 14,261 5,344 3,281 2,671 75,398 2 2,318 = 7,771 2,183 3,989 4 3,253 5,050 9,275 9,683 2,041 122 2,671 48,358 3 590 1,962 1,257 7,128 4,185 3,299 7,442 11,669 13,404 5,104 3,722 2,671 62,433 4 3,818 8604 2,213 8,217 4,393 2,706 6,135 15,071 15,008 5,879 2,963 3,013 78,020 5 3,990 7,924 2,397 8,758 3,140 3,790 5,971 14,487 15,240 6,062 3,674 3,001 78,434 6 3,826 8,068 2,105 8,591 4,267 3,450 4,310 14,336 13,494 1,702 3,055 2,669 69,873 7 3,465 7,418 2,056 7,869 4,390 3,898 6,813 14,074 14,981 5495 2,988 2,952 76,399 8 3,031 8,248 2,448 9,043 4,513 3,712 4,206 15,668 15,112 5,641 2,916 3,216 77,754 9 3,978 7,469 2,340 8,450 4,845 3,834 7,855 15,398 16,307 6,156 3,555 3,041 83,228 10 3,771 7,663 2,497, 9,119 5,106 3,600 8471 15,835 15,581 4576 3,485 3,470 83,174 Project 31,448 73,824 21,709 79,630 39,189 33,533 63,542 139,990 143,071 48,000 29,760 29,374 733,071 B-1 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; ¢ public relations (i.e., site tours); e turbine maintenance and retrofit activities; e scheduled maintenance and routine inspections; C-1 TVP Availability Description ¢ troubleshooting; e delays for parts or equipment; e line outages; and e force majuere 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: 76.8 hours Miscellaneou Other s O&M 20.0% 28% Turbine 2 Downtime: 776.0 hours Line Outage Controller Fault 1.7% O&M 11% 14.4% Other O&M 0.4% Electrical O&M 59.1% Controller Line Outag¢ Tip Breke 0am 72% O&M 701% 82.6% Turbine 3 Downtime: 490.3 hours Turbine 4 Downtime: 108 hours Fauts Line Outage i Unknown Fault rae Anemoreter ain 12% Line Outage Other 7.8% . O&M 2% 1.9% i 13.4% Controller 5% . Ba 8M 6.1% Generator Tip Brake 01% . 69.2% Bectrical 08M. Turbine 5 Downtime: 59.8 hours Turbine 6 Downtime: 167.5 hours Unknown Urknown Faults Line Outage 038M 11.1% 24.0% 0.3% Lae ea 5.6% . Electrical O&M 2.7% D-1 Specific Downtime Causes by Turbine Turbine 7 Downtime: 103.0 hours Faults Line Outage Electrical 11.0% 0.2% 038M 88.8% Turbine 8 Downtime: 129.2 hours Faults eS a a Line Outage 7 01% Electrical 90.1% Turbine 3 Downtime: 46.5 hours Line Outage 0.4% 99.6% Turbine 10 Downtime: 96.7 hours Bectrica Line Outage 99.8% 0.2% D-2 E TIP BRAKE ASSEMBLY ITEM_[ DWGNO | AOCPIN DESCRIPTION QTY 1 10146-D 10280-C [TIP BRAKE PLATE 1 2 10274 20179 [RETROFIT HINGE EYE, LEFT 1 3 10274 20179 RETROFIT HINGE EYE, RIGHT 1 4 10254 10067 |BEARING, 3/8-15/32x.7 2 5 10137-C 10282-A |cLEvis 1 6 N/A 10068 |FHSCs, 1/4-20x5/8" 18-8 SS 2 a N/A 10619 IBHSCS 1/4-20x 1 1/2" 18-8 SS 2 8 10139-C 10283-CC |CATCH PLATE 1 9 N/A 10069A |GROMMET, EPDM, 5/16 IDx13/16"OD x1/4 slot 1 10 N/A 10394 Washer, #10 13/64"ID 3/4"OD, .06 THK 1 11 N/A 10094 ISHSCS 10-32 UNF x 5/8" 18-8SS 1 12 10255 10071-B BUMPER, 1/4"X13/16"X9/16" cut into 2 1/4" pieces 1 13 10133-E 10284-E HINGE BLOCK (MFG'd W/Pin) 1 14 10135-B 10285-A HINGE PIN 1 15 10136-D 10286-D [DAMPER BRACKET 1 16 N/A 10072 |FHSCS, 8-32 UNC x 1/2" S.S. (bracket to block) 6 17 N/A 10105 IBHSCS 8-32x3/8" 18-8 4 18 N/A 10073 |SLHN, 3/8-24UNF SS Nylock (reduced height) 2 19 N/A 10074 ISLHN, 5/16-24UNF SS Nylock (reduced height) 4 20 N/A 10172 |SPRING, ext.(10.44x1.0x0.94,302SS/.04"rnd.wire) 1 21 20148-D 10287-B [SPRING RETAINER 2 22 N/A 10288 DAMPER, extension (M8 thread) 1 23 N/A 10289 jRoD END, female 5/16 bore M8x1.25 5 in-oz 2 24 N/A 10075 |cLEvis PIN, 5/16 x 1 1/8" (Outboard) 1 25 N/A 10106 HAIRPIN COTTER PIN,.0585"x 1 1/8" 2 26 N/A 10171 |CLEvis PIN, 5/16X1 7/16" HEADLESS (INBOARD) 1 27 N/A 10077 |COTTER, Hairpin, 3/8-3/4x2.5", 0.093"D wire, S.S. 1 28 10230-A 10078-A [Mortin Cup 1 29 N/A 10290 |ELECTROMAGNET, 1.75"OD (190 Ib,12VDC,16W) 1 30 N/A 10079 [BRIDGE RECTIFIER (25A, 1000V) 1 31 N/A 10080 |QUICK-CONNECT, Blade, #18-22 .25" Female Insulated 4 32 N/A 10081 |MACHNE SCREW, ROUND HEAD, 10-32X5/8"18.8 S 1 33 N/A 10082 |SLHN, 10-32 18-8 SS (for rect.) 1 34 N/A 10083 IsTup, 5/16-18 UNC X 2 SS (for E-mag) 1 35 N/A 10084 JQUICK-CONNECT, RING 18022 STUD 10 (FOR GROUND) 1 36 N/A 10085 [Rubber Cushioned Steel Loop Strap 1 37 N/A 10174 |RHCDS 8-32x1/2" 1 38 N/A 10175 |LOCKWASHER, #8 EXTERNAL TOOTH 1 39 N/A 10176 RIVET NUT #8 1 40 N/A 10087 |SLHN 5/16-18 UNC S.S. (for E-mag stud) 3 41 N/A 10088 [CABLE TIE, 5 1/2" x 0.14" 1 42 10618 |FHSCs, 1/4-20 X 1 1/2" 18-8 SS 2 43 10620 '1/4-20 SS NYLON LOCKNUT 4 E-1 ca UPRESS OTHERWISE SPECIFIED CAD GRRABATIO DRAWING. OMENSIONS ARE I RICHES 00 HOT MANUALLY UPDATE TOUBANICES ARE ee ors SEE BOM aie seem vo wo seus sare USED O11 18/90 BLADE ASS Atlantic Orient Corporation A WIND ENERGY SYSTEMS COMPANY 1c Vt goss mW ASSEMBLY "1452-0 Ayquassy ayoag di] 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 © 2000 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 1000957 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