HomeMy WebLinkAboutDelta Junction Area Wind Turbines Plant Impact Study - Dec 2010 - REF Grant 2195370WWW.EPSINC.COM
PHONE (425) 883-2833 4020 148th AVE NE, SUITE C, REDMOND, WASHINGTON 98052 FAX (425) 883-0464
PHONE (907) 522-1953 3305 ARCTIC BLVD., SUITE 201, ANCHORAGE, ALASKA 99503 FAX (907) 522-1182
Alaska Environmental Power
Golden Valley Electric Association
Delta Junction Wind Plant Impact Study
Draft Report
December 13, 2010
David W. Burlingame, P.E.
Dr. James W. Cote, Jr., P.E.
Michael J. Bradley, P.E.
John D.L. Hieb
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
ii
Revision Revision Date Revision Description
0 December 13, 2010 Rough Draft to AEP
1
December 14, 2010
Final Draft to AEP/GVEA
Table of Contents
1 INTRODUCTION...................................................................................................... 1
2 RAILBELT DATABASE ............................................................................................ 2
2.1 GVEA Generation Additions ................................................................................. 2
2.2 Power Flow Cases ................................................................................................ 2
2.3 BESS Configuration .............................................................................................. 3
2.4 Generation Configuration – Winter Peak .............................................................. 3
2.5 Generation Configuration – Summer Peak ........................................................... 4
2.6 Generation Configuration – Summer Valley ......................................................... 5
2.7 Wind Generation Cases ....................................................................................... 6
3 POWER FLOW CONTINGENCY ANALYSIS ........................................................ 10
4 DISTURBANCES ................................................................................................... 13
5 DYNAMIC STABILITY RESULTS .......................................................................... 14
5.1 Transmission Fault and Trip Scenarios .............................................................. 14
5.2 Generator Trip Scenarios ................................................................................... 16
6 WIND MODEL/ANALYSIS ..................................................................................... 23
6.1 Wind Data ........................................................................................................... 23
6.2 Wind Analysis ..................................................................................................... 25
7 REGULATION ........................................................................................................ 36
7.1 Ramp Rate Analysis ........................................................................................... 37
7.2 Excess Energy – Curtailment ............................................................................. 38
7.3 Voltage Regulation/Coordination ........................................................................ 40
8 WIND RAMP DYNAMIC SIMULATIONS ............................................................... 41
9 INTERCONNECTION ............................................................................................ 44
9.1 Description.......................................................................................................... 44
9.2 Protection Requirements .................................................................................... 44
9.3 Wind Turbine Options ......................................................................................... 45
9.4 Interface ............................................................................................................. 46
9.5 Voltage Regulation ............................................................................................. 46
9.6 Curtailment Procedures ...................................................................................... 46
APPENDIX A_TRANSMISSION LINE DISTURBANCES ............................................. 48
APPENDIX B – GENERATION DISTURBANCES ....................................................... 49
APPENDIX C – WIND RAMP RATES .......................................................................... 50
APPENDIX D – EPS WIND ANALYSIS ....................................................................... 51
APPENDIX E – CLARITY ANALYTICAL WIND ANALYSIS ......................................... 52
APPENDIX F – WIND LOGIC - WIND ANALYSIS ....................................................... 53
APPENDIX G – EPS PROPOSED ONE-LINE FOR AEP WIND FARM ....................... 54
List of Tables
Table 1: Winter Peak Load - No Wind Generation .......................................................... 4
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
iii
Table 2: Summer Peak Load - No Wind Generation ....................................................... 5
Table 3: Summer Valley Load - No Wind Generation ..................................................... 6
Table 4: Winter Peak Wind Generation Cases ................................................................ 7
Table 5: Summer Peak Wind Generation Cases ............................................................. 8
Table 6: Summer Valley Wind Generation Cases ........................................................... 9
Table 7: Winter Peak Contingency Results – Line Overloads ....................................... 10
Table 8: Summer Peak Contingency Analysis Results – Line Overloads ...................... 11
Table 9: Winter Peak Contingency Analysis - Voltage Violations .................................. 12
Table 10: Contingency List ............................................................................................ 13
Table 11: Dynamic Stability Results Table…………………………………………………. 22
Table 12: Summer Valley Load-Nominal Import…………………………………………… 38
Table 13………………………………………………………………………………………... 38
Table 14: Available Regulation for Wind Ramp Events…………………………………… 41
List of Figures
Figure 1: d_sp_brad_d2 Sample Plot ............................................................................ 14
Figure 3: GE LVRT Undesired Tripping Event .............................................................. 16
Figure 4: Winter Peak Maximum Import - Healy Clean Coal Trip .................................. 17
Figure 5: Winter Peak Maximum Import - HCCP Trip + Transfer Trip Healy-Cantwell .. 18
Figure 6: Winter Peak Maximum Import - HCCP trip + BESS Auto-Scheduling ............ 19
Figure 7: Winter Peak Maximum Import - NPCC1 Trip .................................................. 20
Figure 8: Winter Peak Maximum Import - NPCC1 Trip BESS Auto-Scheduled ............. 21
Figure 9: Graph and Chart of Distribution of MW Changes in 2 MW bins……………….. 25
Figure 10: Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes
Ahead………………………………………………………………………………………...… 26
Figure 11: Scatter of Negative Ramp Distribution from Starting Power Level 1 Minutes
Ahead…………………………………………………………………………………………... 27
Figure 12: Delta Junction distribution of Hourly 10 minute Min-Max-Avg………………. 28
Figure 13: Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes
Ahead with Negative Ramp Curve………………………………………………………….. 29
Figure 14: Scatter of Negative Ramp Distribution from Starting Power Level 20 Minutes
Ahead with Negative Ramp Curve………………………………………………………….. 30
Figure 15: Scatter of Negative Ramp Distribution from Starting Power Level 30 Minutes
Ahead with Negative Ramp Curve………………………………………………………….. 31
Figure 16: Delta Junction and Eva Creek Distribution of Hourly 10 minute Min-Max-Avg
…………………………………………………………………………………………………... 32
Figure 17: Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes
Ahead with Negative Ramp Curve………………………………………………………….. 33
Figure 18: Scatter of Negative Ramp Distribution from Starting Power Level 20 Minutes
Ahead with Negative Ramp Curve………………………………………………………….. 34
Figure 19: Scatter of Negative Ramp Distribution from Starting Power Level 60 Minutes
Ahead with Negative Ramp Curve………………………………………………………….. 35
Figure 20: Scatter of Negative Distribution from Starting Power Level 1 Minute Ahead
…………………………………………………………………………………………………... 37
Figure 21: Wind Ramp with Insufficient Capacity in Islanded Condition………………... 43
Figure 22: Wind Ramp with Insufficient Capacity Connected System…………………... 44
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
iv
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 1
1 Introduction
Alaska Environmental Power (AEP) contracted with Electric Power Systems Inc. (EPS) to
perform a grid impact study for the proposed new Delta Junction Wind Plant, interconnected to
the Golden Valley Electric Association (GVEA) system near Delta Junction. The Delta Junction
wind plant consists of 16 wind turbines rated at 1.6 MW each, step-up transformers, and a
collector system connecting to the GVEA system at the Jarvis Creek substation. The study will
analyze the operational impacts of the variable wind energy. Power regulation requirements will
be recommended based on analysis of 10-minute wind rate ramp analysis. Additionally, the grid
impact study will analyze the power flow and transient stability impacts of the Delta Junction
wind plant on GVEA’s grid. The protection impacts to the GVEA system as well as the
protection requirements for the AEP wind farm are also identified. The proposed 24 MW Eva
Creek wind plant will be included in some of the impact studies, evaluated in conjunction with
the Delta Junction wind plant for excess energy issues and the impact on power flows and
transient stability in a limited number of cases. Both wind projects are assumed to consist of
GE wind turbines. The Delta Junction Wind farm is modeled with the “WindInertia” package
installed to provide an inertial response from the wind turbines during under-frequency grid
events.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 2
2 Railbelt Database
The seasonal power flow cases and dynamic database used in this study were created from the
2020 Railbelt PSS/E database being completed by EPS for the Railbelt System Studies
Subcommittee (SSS). The GVEA system loads were updated to reflect the forecast loads for
2015. The generation dispatches for the SSS cases were expanded to create various
Anchorage-Fairbanks intertie import scenarios, and wind generation scenarios.
2.1 GVEA Generation Additions
The two wind projects (Delta Junction and Eva Creek) result in three wind generation
configurations for analysis. The three scenarios are:
1) a base case with no wind,
2) cases with Delta Junction wind only, and
3) cases with Delta Junction wind and Eva Creek wind both online.
2.2 Power Flow Cases
The power flow cases were created to represent several GVEA operating situations with respect
to the Anchorage-Fairbanks import. The cases are:
Case 1 is configured with a nominal import on the Anchorage-Fairbanks Intertie,
measured along the Douglas – Stevens section of the 138 kV line. In case 1, the
generation is configured so that the intertie flow is equal to GVEA’s portion of the
Bradley Lake generation.
Case 2 is configured with an assumed maximum import of 75 MW, measured on the
Douglas – Stevens 138 kV line section looking north. The import is achieved by taking
Bradley Lake generation plus economy sales from the south.
Case 3 is configured as an islanded case with the Anchorage-Fairbanks tie line out of
service.
Each of the three cases were further augmented to include the three different wind
combinations of 1) no wind, 2) Delta Junction wind only, and 3) Delta Junction and Eva Creek
wind. Additionally, the three load seasons of winter peak, summer peak, and summer valley will
be analyzed. There are 27 base case power flows, resulting from 3 load conditions, 3 intertie
flow conditions, and 3 wind plant conditions. The power flow case naming convention is as
follows:
The first section of the case name is the wind configuration
o d_ represents the case with the Delta Junction Wind Farm
o de_ represents the case with both Delta Junction and Eva Creek Wind Farms
o Blank represents the case without wind generation
The second section of the case name is the season
o wp represents the winter peak case
o sp represents the summer peak case
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 3
o sv represents the summer valley case
The third section of the case name is the import condition
o ‘75’ represents the maximum import case
o ‘brad’ represents the nominal import case
o ‘open’ represents the open tie case
An example case name of d_wp_75.sav would correspond to a maximum import, winter peak
case with only the Delta Junction Wind farm generation.
2.3 BESS Configuration
The Wilson BESS configuration is important in the configuration of the dynamic cases. GVEA
has stated that the BESS will not be used for the regulation of the wind power. However, it is
assumed that the primary functions of SVC operation and emergency power injection for
extreme under-frequency events are still available from the BESS.
2.4 Generation Configuration – Winter Peak
Starting from the 2020 IOC / SSS Railbelt base cases, EPS created the following winter peak
load cases. The GVEA generation was changed in order to get the three Anchorage-Fairbanks
tie flow conditions discussed previously. Table 1 shows the load, losses, import, unit
commitment, and unit dispatch for three winter peak load cases, without any wind generation.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 4
Table 1: Winter Peak Load - No Wind Generation
Generators Capacity MW Spin MW Spin MW Spin
Healy Clean Coal 61.9 61.9 0.0 61.9 0.0 61.9 0.0
Healy 1 28.5 28.5 0.0 28.5 0.0 28.5 0.0
EvaCreek Wind 24.0 0.0 0.0 0.0 0.0 0.0 0.0
Chena5 24.0 24.0 0.0 24.0 0.0 24.0 0.0
North Pole 1 64.0 0.0 0.0 0.0 0.0 0.0 0.0
North Pole 2 64.0 54.9 9.1 40.7 23.3 0.0 0.0
North Pole CC1 53.0 53.0 0.0 53.0 0.0 39.7 13.3
North Pole CC2 12.0 12.0 0.0 12.0 0.0 9.9 2.1
DeltaWind 25.6 0.0 0.0 0.0 0.0 0.0 0.0
Univ. AK Gen 22.6 7.6 0.0 7.6 0.0 7.6 0.0
Fort WainwrightGen 20.0 6.0 0.0 6.0 0.0 6.0 0.0
Eielson 25.0 15.0 0.0 15.0 0.0 15.0 0.0
Bradley 15.2 0.0 0.0 14.0 4.6 14.7 4.6
BESS 25.0 0.0 25.0 0.0 25.0 0.0 25.0
Purchase N/A 0.0 0.0 0.1 0.0 59.8 0.0
Total Generation 262.9 34.1 262.6 52.9 267.1 45.0
Required Generation 262.9 262.6 266.9
Total Load 257.4 257.4 257.4
Losses 5.5 5.2 9.5
Additional Military Load (MW) 28.6 28.6 28.6
Auxiliary Loads (MW) 17.1 17.1 17.1
GVEA Net Load (MW) 211.7 211.7 211.7
WinterPeakNo Wind
Islanded Nominal Import Maximum import
Note that the BESS spin has been unadjusted from the base value of 25 MW.
2.5 Generation Configuration – Summer Peak
Starting from the 2020 base cases, EPS has created the proposed summer peak load cases.
The GVEA generation was changed in order to get the three Anchorage-Fairbanks tie flow
conditions. Table 2 shows the load, losses, import, unit commitment, and unit dispatch for the
three cases, without any wind generation.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 5
Table 2: Summer Peak Load - No Wind Generation
Generators Capacity MW Spin MW Spin MW Spin
Healy Clean Coal 61.9 61.9 0.0 61.9 0.0 58.0 3.9
Healy 1 26.4 26.0 0.4 26.0 0.4 26.0 0.4
EvaCreek Wind 24.0 0.0 0.0 0.0 0.0 0.0 0.0
Chena5 24.0 24.0 0.0 24.0 0.0 24.0 0.0
North Pole 1 39.5 0.0 0.0 0.0 0.0 0.0 0.0
NorthPole 2 40.0 30.3 9.7 17.5 22.5 0.0 0.0
NorthPole CC1 33.0 33.0 0.0 33.0 0.0 0.0 0.0
NorthPole CC2 7.0 7.0 0.0 7.0 0.0 0.0 0.0
DeltaWind 25.6 0.0 0.0 0.0 0.0 0.0 0.0
Univ. AK Gen 22.6 9.2 0.0 9.2 0.0 9.2 0.0
Fort WainwrightGen 20.0 12.0 0.0 12.0 0.0 12.0 0.0
Eielson 25.0 15.0 0.0 15.0 0.0 15.0 0.0
Pump9Gen(FWP,FGP) 8.0 0.0 0.0 0.0 0.0 0.0 0.0
Bradley 15.2 0.0 0.0 14.0 4.6 14.0 4.6
BESS 0.0 0.0 25.0 0.0 25.0 0.0 25.0
Purchase N/A 0.0 0.0 0.0 0.0 59.8 0.0
Total Generation 218.4 35.1 219.6 52.5 218.0 33.9
Auxiliary Loads 15.94 15.94 15.94
Required Generation 218.44 219.54 220.74
Total Load 215.34 215.34 215.34
Losses 3.1 4.2 5.4
GVEA NetLoad (MW) 163.2 163.2 163.2
Additional Military Load (MW) 36.2 36.2 36.2
Islanded Nominal Import Maximum import
SummerPeak No Wind
2.6 Generation Configuration – Summer Valley
Starting from the 2020 base cases, EPS has created the proposed summer valley load cases.
The GVEA generation was changed in order to get the three Anchorage-Fairbanks tie flow
conditions discussed previously. Table 3 shows the load, losses, import, unit commitment, and
unit dispatch for the three cases, without any wind generation.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 6
Table 3: Summer Valley Load - No Wind Generation
Generators Capacity MW Spin MW Spin MW Spin
Healy Clean Coal 61.9 58.8 3.1 57.9 4.0 38.0 23.9
Healy 1 28.5 26.5 2.0 26.5 2.0 18.0 10.5
EvaCreek Wind 24.0 0.0 0.0 0.0 0.0 0.0 0.0
Chena 5 24.0 24.0 0.0 24.0 0.0 24.0 0.0
North Pole 1 47.0 0.0 0.0 0.0 0.0 0.0 0.0
North Pole 2 48.0 0.0 0.0 0.0 0.0 0.0 0.0
North Pole CC1 43.0 6.5 36.5 0.0 0.0 0.0 0.0
North Pole CC2 10.0 0.0 0.0 0.0 0.0 0.0 0.0
DeltaWind 25.6 0.0 0.0 0.0 0.0 0.0 0.0
Univ. AK Gen 22.6 0.0 0.0 0.0 0.0 0.0 0.0
Fort WainwrightGen 20.0 0.0 0.0 0.0 0.0 0.0 0.0
Eielson 25.0 0.0 0.0 0.0 0.0 0.0 0.0
Bradley 15.2 0.0 0.0 8.0 4.6 8.0 4.6
BESS 0.0 0.0 25.0 0.0 25.0 0.0 25.0
Purchase N/A 0.0 0.0 0.0 31.0 0.0
Totals 115.8 66.6 116.4 35.6 119.0 64.0
Required Generation 115.8 116.4 119
Total Load 113.7 113.7 113.7
Losses 2.1 2.7 5.3
Additional Military Load (MW) 0 0 0
Auxiliary Loads 20.4 20.4 20.4
GVEA Net Load (MW) 93.3 93.3 93.3
SummerValley No Wind
Islanded Nominal Import Maximum import
2.7 Wind Generation Cases
2.7.1 Winter Peak Generation Cases
The winter peak wind generation cases with wind generation are shown below in table 4. The
addition of the wind reduced the amount of generation needed from the North Pole units as
compared with the base cases without wind. The maximum import condition with the maximum
amount of wind resulted with the North Pole Combined Cycle unit at its minimum while the other
units remain near their maximum capabilities.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 7
Table 4: Winter Peak Wind Generation Cases
Generators Capacity MW Spin MW Spin MW Spin MW Spin MW Spin MW Spin
Healy Clean Coal 61.9 61.8 0.1 61.9 0.0 57.2 4.7 61.9 0.0 61.9 0.0 59.1 2.8
Healy 1 28.5 28.5 0.0 28.5 0.0 28.5 0.0 28.5 0.0 28.5 0.0 28.5 0.0
EvaCreek Wind 24.0 24.0 0.0 24.0 0.0 24.0 0.0
Chena 5 24.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0
North Pole 1 64.0
North Pole 2 64.0 28.8 35.2 15.3 48.7 10.0 54.0
North Pole CC1 53.0 53.0 0.0 53.0 0.0 24.0 29.0 48.5 4.5 45.8 7.2 6.0 47.0
North Pole CC2 12.0 12.0 0.0 12.0 0.0 6.0 6.0 12.0 0.0 11.5 0.6
Delta Wind 25.6 25.6 0.0 25.6 0.0 25.6 0.0 25.6 0.0 25.6 0.0 25.6 0.0
Univ. AK Gen 22.6 7.6 0.0 7.6 0.0 7.6 0.0 7.6 0.0 7.6 0.0 7.6 0.0
Fort Wainwright Gen 20.0 6.0 0.0 6.0 0.0 6.0 0.0 6.0 0.0 6.0 0.0 6.0 0.0
Eielson 25.0 15.0 0.0 15.0 0.0 15.0 0.0 15.0 0.0 15.0 0.0 15.0 0.0
Bradley 15.2 14.0 4.6 14.7 4.6 14.0 4.6 14.7 4.6
BESS 25.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0
Purchase N/A 0.0 0.0 59.6 0.0 0.1 0.0 59.5 0.0
Total Generation 262.3 60.3 262.9 78.3 268.2 69.3 263.1 83.5 263.8 37.4 270.0 79.4
Required Generation 262.3 262.8 268.2
257.4 257.4 257.4
5.4 10.8
263 263.7 270
5.6 6.3 12.6
257.4 257.4
Losses 4.9
Total Load 257.4
17.1 17.1 17.1 17.1 17.1Auxiliary Loads (MW) 17.1
28.6 28.6 28.6 28.6 28.6Additional Military Load (MW) 28.6
211.7 211.7 211.7 211.7 211.7GVEA NetLoad (MW) 211.7
WinterPeak No Wind WinterPeak Delta Wind WinterPeak Delta and Eva Wind
Islanded Nominal Imp Max Imp Islanded Nominal Imp Max Imp
2.7.2 Summer Peak Wind Generation Cases
The summer peak wind generation cases are shown in table 5 below. While creating these
cases, the condition with the maximum import, Delta Junction wind, and Eva Creek wind
resulted in more generation capability than the load would allow. To resolve this issue, either the
wind or import could be reduced. In order to simulate a worst case stability condition EPS kept
the import at its limit while curtailing the wind generation. While this may not be a viable
operational strategy, it will provide a worst case scenario for a stability simulation.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 8
Table 5: Summer Peak Wind Generation Cases
Generators Capacity MW Spin MW Spin MW Spin MW Spin MW Spin MW Spin
Healy Clean Coal 61.9 61.9 0.0 61.9 0.0 41.0 20.9 61.9 0.0 61.9 0.0 38.0 23.9
Healy 1 26.4 26.0 0.4 26.0 0.4 22.0 4.4 26.0 0.4 26.0 0.4 18.0 8.4
EvaCreek Wind 24.0 24.0 0.0 24.0 0.0 16.5 0.0
Chena5 24.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0
NorthPole 1 39.5
NorthPole 2 40.0 15.3 24.7
NorthPole CC1 33.0 24.0 9.0 25.7 7.3 22.2 10.8 8.9 24.1
NorthPole CC2 7.0 6.0 1.0 6.2 0.8
DeltaWind 25.6 25.6 0.0 25.6 0.0 25.6 0.0 25.6 0.0 25.6 0.0 16.2 0.0
Univ. AK Gen 22.6 9.2 0.0 9.2 0.0 9.2 0.0 9.2 0.0 9.2 0.0 9.2 0.0
FortWainwrightGen 20.0 12.0 0.0 12.0 0.0 12.0 0.0 12.0 0.0 12.0 0.0 12.0 0.0
Eielson 25.0 15.0 0.0 15.0 0.0 15.0 0.0 15.0 0.0 15.0 0.0 15.0 0.0
Bradley 15.2 14.0 4.6 14.0 4.6 14.0 4.6 14.0 4.6
BESS 0.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0
Purchase N/A 60.5 0.0 60.4 0.0
Total Generation 219.0 60.1 219.6 38.1 223.3 54.9 219.9 36.2 220.6 54.1 223.3 61.9
SummerPeakDeltaWind SummerPeakDeltaandEvaWind
Max Imp Islanded Nominal Imp Max Imp
GVEA NetLoad(MW) 163.2 163.2
Islanded Nominal Imp
163.2 163.2 163.2 163.2
Additional Military Load (MW) 36.2 36.2 36.2 36.2 36.2 36.2
Auxiliary Loads 15.94 15.94 15.94 15.94 15.94 15.94
Total Load 215.34 215.34
Losses 3.7
219.54
215.34 215.34 215.34 215.34
4.2
Required Generation 219.04 223.24 219.84 220.54 223.24
7.9 4.5 5.2 7.9
2.7.3 Summer Valley Wind Generation Cases
The summer valley cases are largely similar to each other due to the low levels of load and
excess generation capability. It is likely that these cases produce more operational issues than
stability issues. It can be seen that the summer valley nominal and maximum import cases are
very similar due to the low load levels. When both wind projects are online, all of the cases need
curtailment of the wind. In these cases, the minimum generation levels of the coal units may be
the limiting condition for accepting wind energy.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 9
Table 6: Summer Valley Wind Generation Cases
Generators Capacity MW Spin MW Spin MW Spin MW Spin MW Spin MW Spin
Healy Clean Coal 61.9 39.4 22.5 40.0 21.9 38.0 23.9 38.0 23.9 38.0 23.9 38.0 23.9
Healy 1 28.5 20.0 8.5 18.0 10.5 18.0 10.5 18.0 10.5 18.0 10.5 18.0 10.5
EvaCreek Wind 24.0 17.5 6.5 13.5 10.5 14.0 10.0
Chena5 24.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0 24.0 0.0
North Pole 1 47.0
North Pole 2 48.0
North Pole CC1 43.0 6.5 36.5
North Pole CC2 10.0
Delta Wind 25.6 25.6 0.0 25.6 0.0 25.6 0.0 18.0 7.6 14.1 11.5 14.6 11.0
Univ. AK Gen 22.6
FortWainwright Ge 20.0
Eielson 25.0
Bradley 15.2 0.0 0.0 8.0 4.6 8.0 4.6 0.0 0.0 8.0 4.6 8.0 4.6
BESS 0.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0 0.0 25.0
Purchase N/A 2.0 0.0 0.1 0.0 1.1 0.0
Totals 115.5 92.5 115.6 62.0 115.6 64.0 115.5 73.5 115.5 86.0 115.5 85.0
SummerValley DeltaWind SummerValley Deltaand EvaWind
Islanded Nominal Imp Max Imp Islanded Nominal Imp Max Imp
GVEA NetLoad (MW) 93.3 93.3 93.3 93.3 93.3 93.3
Additional Military Load (MW)0 0 0 0 0 0
Auxiliary Loads 20.4 20.4 20.4 20.4 20.4 20.4
Total Load 113.7 113.7 113.7
Losses 1.8
115.5 115.5
113.7 113.7 113.7
1.8 1.8
Required Generation 115.5 115.4 115.4 115.4
1.7 1.7 1.7
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 10
3 Power Flow Contingency Analysis
EPS performed a contingency analysis for the 27 power flow cases presented above. EPS
outaged each 138 and 69kV transmission line in the GVEA system. Line flow violations seen in
the winter peak cases are shown below in table 7.
Table 7: Winter Peak Contingency Results – Line Overloads
Savecase From Bus To Bus kV From Bus To Bus kV Rating(MVA) % Overload
wp_open North Pole Carney 138 Highway Park Dawson 69 29 159.9
wp_open North Pole Carney 138 Dawson Eielson 69 29 139.3
wp_open North Pole Carney 138 Eielson Johnson Tap 69 29 132.4
wp_brad NorthPole Carney 138 Highway Park Dawson 69 29 160.3
wp_brad NorthPole Carney 138 Dawson Eielson 69 29 139.8
wp_brad NorthPole Carney 138 Eielson Johnson Tap 69 29 132.9
wp_75 NorthPole Carney 138 Highway Park Dawson 69 29 166.3
wp_75 NorthPole Carney 138 Dawson Eielson 69 29 140.7
wp_75 NorthPole Carney 138 Eielson Johnson Tap 69 29 134.4
wp_75 Wilson FTWW 138 Gold Hill Aurora 69 43 157.6
wp_75 Wilson FTWW 138 Aurora Zehnder 69 43 138.7
wp_75 Wilson FTWW 138 Gold Hill ChenaTap 69 43 137.8
wp_75 Wilson FTWW 138 University Tap International 69 43 121.7
wp_75 Wilson FTWW 138 ChenaTap University Tap 69 43 121.7
d_wp_open Wilson FTWW 138 Gold Hill Aurora 69 43 153
d_wp_open Wilson FTWW 138 Aurora Zehnder 69 43 134
d_wp_open Wilson FTWW 138 Gold Hill ChenaTap 69 43 133.7
d_wp_open Wilson FTWW 138 University Tap International 69 43 117.6
d_wp_open Wilson FTWW 138 ChenaTap University Tap 69 43 117.6
de_wp_open Wilson FTWW 138 Gold Hill Aurora 69 43 103.6
de_wp_brad Wilson FTWW 138 Gold Hill Aurora 69 43 120.4
de_wp_brad Wilson FTWW 138 Gold Hill ChenaTap 69 43 106.3
de_wp_brad Wilson FTWW 138 Aurora Zehnder 69 43 100.9
de_wp_75 Wilson FTWW 138 Gold Hill Aurora 69 43 183.4
de_wp_75 Wilson FTWW 138 Aurora Zehnder 69 43 165.1
de_wp_75 Wilson FTWW 138 Gold Hill ChenaTap 69 43 159.8
de_wp_75 Wilson FTWW 138 ChenaTap University Tap 69 43 144
de_wp_75 Wilson FTWW 138 UniversityTap International 69 43 144
de_wp_75 Wilson FTWW 138 International PegerRoad 69 43 121.4
de_wp_75 Wilson FTWW 138 PegerRoad SSS Tap 69 43 104.4
de_wp_75 Wilson FTWW 138 SSS Tap S. Fairbanks 69 43 104.4
Contingency Outage Overload
Two contingencies result in line overloads. The contingencies are the 138kV lines from Wilson
to Fort Wainwright and North Pole to Carney. The North Pole to Carney contingency results in
violations on the 69kV line from Highway Park to Eielson substations. When the Delta Junction
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 11
wind farm is providing energy, the loading along the 69kV circuit is reduced under this
contingency preventing flow violations. In order to mitigate these violations, the line sections
between Highway Park and Johnson Tap could be re-conductored to increase ampacity from
the 4/0 cable that is shown in the GVEA impedance diagram.
The second contingency of the Wilson-Fort Wainwright 138kV line results in several 69kV line
overloads for the lines leaving Gold Hill substation. The line overloads seen from this
contingency are quite large. The cases with the Eva Creek wind farm online are more severe
than the other cases. When the Eva Creek generation was added to the case, the North Pole
generation was reduced. As such more generation is coming out of the Healy area resulting in
greater flow on the lines going north into Fairbanks. Additionally, the worst cases are the
maximum import cases that also stress the lines going north into Fairbanks. These violations
could be mitigated by reconductoring the 69kV lines or unit dispatches that reduce the line flow
north into Fairbanks. Due to the apparent severity of the Wilson-Fort Wainwright 138kV
contingency, this event was added to the dynamic stability analysis.
The summer peak line flow violations are included in table 8 below.
Table 8: Summer Peak Contingency Analysis Results – Line Overloads
From Bus To Bus kV FromBus To Bus kV Rating(MVA) % Overload
sp_open North Pole Carney 138 Highway Park Dawson 69 29 111.9
sp_brad North Pole Carney 138 Highway Park Dawson 69 29 112.0
sp_75 North Pole Carney 138 Highway Park Dawson 69 29 114.8
sp_75 North Pole Carney 138 Dawson Eielson 69 29 102.8
sp_75 Wilson FTWW 138 Gold Hill Aurora 69 43 160.0
sp_75 Wilson FTWW 138 Aurora Zehnder 69 43 145.4
sp_75 Wilson FTWW 138 Gold Hill ChenaTap 69 43 138.1
sp_75 Wilson FTWW 138 University Tap International 69 43 129.1
sp_75 Wilson FTWW 138 ChenaTap University Tap 69 43 129.1
sp_75 Wilson FTWW 138 International PegerRoad 69 43 112.8
d_sp_75 Wilson FTWW 138 Gold Hill Aurora 69 43 135.3
d_sp_75 Wilson FTWW 138 Aurora Zehnder 69 43 120.6
d_sp_75 Wilson FTWW 138 Gold Hill Chena Tap 69 43 116.6
d_sp_75 Wilson FTWW 138 University Tap International 69 43 107.5
d_sp_75 Wilson FTWW 138 ChenaTap University Tap 69 43 107.5
de_sp_open Wilson FTWW 138 Gold Hill Aurora 69 43 107.2
de_sp_brad Wilson FTWW 138 Gold Hill Aurora 69 43 123.7
de_sp_brad Wilson FTWW 138 Aurora Zehnder 69 43 109.1
de_sp_brad Wilson FTWW 138 Gold Hill ChenaTap 69 43 106.2
de_sp_75 Wilson FTWW 138 Gold Hill Aurora 69 43 146.8
de_sp_75 Wilson FTWW 138 Aurora Zehnder 69 43 132.2
de_sp_75 Wilson FTWW 138 Gold Hill ChenaTap 69 43 126.6
de_sp_75 Wilson FTWW 138 ChenaTap University Tap 69 43 117.5
de_sp_75 Wilson FTWW 138 University Tap International 69 43 117.5
de_sp_75 Wilson FTWW 138 International PegerRoad 69 43 101.2
Savecase Contingency Outage Overload
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 12
The summer peak contingencies result in very similar results. Again, the problem contingencies
are the Wilson-Fort Wainwright and North Pole-Carney 138kV transmission lines. Due to the
lower load levels, the line overloads are not as severe as seen for the winter peak cases.
The summer valley cases have no overloads for any transmission outages in the GVEA area.
Only the winter peak cases have contingencies that result in voltages below 0.90 pu. The
results of the voltage violations are shown below in table 9.
Table 9: Winter Peak Contingency Analysis - Voltage Violations
Case FromBus To Bus kV Bus kV Voltage (PU)
d_wp_75 Jarvis DeltaWind 138 CANTWELL 138 0.88
de_wp_75 Healy Healy Wind 138 CANTWELL 138 0.91
de_wp_75 Healy Healy Wind 138 NENANA 138 0.89
de_wp_75 Jarvis DeltaWind 138 CANTWELL 138 0.89
Outage Violation
The loss of generation in the heavy import cases result in a large flow from the Anchorage.
These results are misleading as the BESS would auto-schedule for a loss of the Healy-Healy
Wind 138kV line. This action mitigates the low voltages seen at Cantwell and Nenana. Also, by
losing the Jarvis – Delta Wind 138kV line, the wind generation itself will be lost. The North Pole
and Healy governors would respond to the loss in generation reducing the import from
Anchorage. This response would mitigate the low voltages seen at Cantwell for this
contingency. No remediation actions are required.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 13
4 Disturbances
The list of disturbances run with the power flow cases discussed in section 2 is shown below.
The transient stability results were analyzed for possible system stability problems. In these
planning cases, the largest unit contingency has been reduced from an AMLP Plant 2 Unit 7 or
a CEA Beluga 7 outage to a CEA Southcentral Power Plant (SPP) outage. This largest unit
outage is smaller in total MW’s lost than exists in today’s system, and therefore decreases the
spin requirements of the entire grid. The SPP outage for the Winter Peak season is 67.5 MW.
Table 5 lists the disturbances used for this study. Additionally, for the islanded cases, Healy
Clean Coal is the largest unit in the GVEA area at 61.9 MW of capacity.
The line fault and trip scenarios were selected as the anticipated worst case fault scenarios as
seen from the proposed Delta Junction wind farm. The Douglas – Healy fault will be simulated
with the fault being located at both ends of the line as seen in disturbances D1 and D2. The
distribution faults will be evaluated to determine the voltage impact at the point of
interconnection for the Delta Junction wind farm.
Table 10: Contingency List
Contingency Voltage
(kV) Fault Location
Clearing Time (Cycles)
Name Line/Generator Local Remote
D1 Douglas Healy 138.0 Douglas 5 30
D2 Douglas Healy 138.0 Healy 5 30
D3 Jarvis Fort Greely 138.0 Jarvis 5 5
D4 Carney Highway Park 69.0 Carney 5 5
D5 Carney North Pole 138.0 Carney 5 5
D6 Missile Defense Feeder 12.5 12.5 kV Feeder 5 N/A
D7 Pump 9 Feeder 13.8 13.8 kV Feeder 5 N/A
D8 Wilson Fort Wainwright 138 Wilson 5 5
G1 Healy Clean Coal unit trip 13.8 N/A
G2 SPP G2 unit trip 13.8 N/A
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 14
5 Dynamic Stability Results
5.1 Transmission Fault and Trip Scenarios
All of the line fault scenarios exhibited acceptable system performance. All cases were stable.
All frequency and voltage swings are within acceptable limits and all cases exhibit a well-
damped response and return to nominal voltage and frequency. The worst case line fault and
trip scenarios were the d1 and d2 contingencies. In cases with heavy import, the loss of the line
causes GVEA to shed load, however, the GVEA system remains stable after the load shed. A
sample plot showing the d2 event is shown below in figure 1. The top left plots show the several
bus frequencies in the GVEA area and frequency trace in Anchorage. The top right plot shows
selected bus voltages in the GVEA area. The bottom left plot shows the relevant wind farm
information including the MW, MVAR, and voltage at the wind farm collector buses. The bottom
right shows line MW flows as well as BESS output.
Based on the line fault and trip transient stability results, the addition of Delta Junction wind farm
or the combination of Delta Junction and Eva Creek wind farms have no negative impact on the
system stability.
Figure 1: d_sp_brad_d2 Sample Plot
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 15
The fault scenario results can be found in Appendix A.
5.1.1 Delta Junction Wind Farm LVRT/ Protection Coordination
The Delta Junction Wind farm connects to the Jarvis Creek substation. Distribution feeders that
can impact the 138 kV voltage at the wind farm’s POI are the Jarvis Creek substation 24.9 kV
Jarvis Creek feeders, the Missile Defense substation feeder, Doyon’s 24.9 kV Fort Greely
feeders and the 13.8 Pump 9 substation feeders. An example of a fault and clear of the Pump 9
distribution feeder is shown in figure 2. It can be seen that the Jarvis Creek 138kV voltage is
reduced to ~0.45pu. The GE wind turbines have a low voltage ride-through (LVRT)
characteristic that will trigger for voltages below 0.75 pu. It is clear that the distribution faults in
the Jarvis Creek substation area can cause the voltage at Jarvis Creek to dip into the GE LVRT
characteristic. Typically the distribution relays will clear faults on the system much faster than
the GE Low Voltage event characteristic, however the initial fault clearing is not necessarily the
area of concern with regards to LVRT. Based on EPS experience, the LVRT logic is not
properly modeled in the PSSE models provided by GE. A quick discussion is warranted.
Figure 2: Summer Peak Maximum Import with Delta Junction Wind - Pump 9 Feeder Fault
The GE LVRT scheme uses a seal-in timer that begins timing at the first low voltage event. This
timer will continue to count even after voltage returns to a healthy operation point (>0.75 pu)
until such a time that the low voltage time is released. Since distribution feeders typically utilize
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 16
auto-reclosing in their protection, the GE LVRT characteristic can be triggered for normally-
cleared, nearby distribution faults with reclosing on the second or third reclose event. The
following event would result in a trip if the seal-in timer is set at 2 seconds.
An example of this undesired wind trip is shown in figure 3. A distribution fault occurs at time =
0”. The voltage at the point of interconnection is reduced below the low voltage pickup of the GE
LVRT to 0.4 pu. The relay picks up and clears the fault in 0.1 seconds and the POI voltage
recovers to its nominal voltage. The distribution relay then automatically recloses at 2 seconds.
The fault still exists and the relay clears the fault again at 2.1 seconds. Even though the fault is
active for a total time of only 0.2 seconds, the GE LVRT will activate and trip the wind farm for
this scenario.
Figure 3: GE LVRT Undesired Tripping Event
EPS recommends that GVEA and Doyon review their reclosing schemes in nearby distribution
feeders to evaluate the impact of the reclosing on the low voltage ride through of the Delta
Junction Wind Farm. The GE LVRT characteristics are programmable and offer a great deal of
flexibility with regards to the tripping characteristics. We believe that between a combination of
reclosing intervals and programming changes to the LVRT characteristics, inadvertent tripping
can be avoided for faults on local distribution circuits.
5.2 Generator Trip Scenarios
EPS ran a series of seven unit trip scenarios for each of the power flow cases. In running the
simulations for the trip of Healy clean coal unit there were a few scenarios that resulted in out of
step conditions along the tie to Anchorage when GVEA is at its maximum import condition. As
such, EPS needed to perform additional investigation into the unit trip events. EPS added the
following events:
G3 – Trip of Healy Clean Coal, and a transfer trip of the Healy-Cantwell line
G4 – Trip of Chena 5 unit
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 17
G5 – Trip of Healy 1 unit
G6 – Trip of North Pole Combined Cycle unit
G7 – Trip of Healy Clean Coal, with auto-scheduling enabled in BESS
G8 – Trip of North Pole Combined Cycle, with auto-scheduling enabled in BESS
The G1 contingency resulted in out of step conditions for the following cases:
Winter peak with maximum import
Summer peak with maximum import
Winter peak with maximum import and Delta Junction Wind farm.
Winter peak with maximum import and both wind farms online.
The plot of the winter peak with maximum import case under the G1 contingency is shown
below in figure 4.
Figure 4: Winter Peak Maximum Import - Healy Clean Coal Trip
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 18
Figure 4 shows the GVEA system going out of step. The red trace in the top left shows the
Healy 1 frequency separating from the other GVEA system frequencies. Additionally, the
voltage collapse seen in the top right plot suggests the system has gone out of step.
The G3 contingency was simulated as a mitigation step for the loss of the HCCP unit. This
scenario assumed a transfer trip of the Healy line to Cantwell six cycles after the unit tripped.
This simulation was to see if a transfer trip prevents the out of step condition seen in the G1
simulations highlighted in figure 2. The G3 contingency does resolve the out of step conditions
seen in the G1 contingencies. The G3 disturbance applied to the winter peak maximum import
case is shown below in figure 5.
Figure 5: Winter Peak Maximum Import - HCCP Trip + Transfer Trip Healy-Cantwell
Figure 5 illustrates that the transfer trip of the Healy-Cantwell tie in response to the trip of Healy
Clean Coal can prevent the out of step conditions. All G3 contingency cases resulted in stable
simulations.
In addition to the transfer trip of the Healy-Cantwell line, EPS simulated the effects of the auto-
scheduling function of the BESS. The BESS auto-scheduling is a communication aided BESS
mode trip. GVEA provided the following list of seven events which will trigger the auto-
scheduling component of the BESS operation.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 19
1. Healy Unit 1 138kV unit breaker trip
2. North Pole Units 1&2 138kV breaker trip
3. North Pole Units 1&2 13.8kV unit breaker trip
4. North Pole Combined Cycle units 1&2 13.8kV breaker trip
5. Healy to Douglas 138kV trip at healy breaker B17
6. Healy to Wilson 138kV breaker trip at Healy B21
7. Wilson to Healy 138kV trip at Wilson ring-bus breakers B4 & B5
Each of these events communicates with the BESS and pre-loads a unique MW value for each
event. When any of the seven events listed below occurs, a high speed communication system
calls for the BESS to rapidly ramp to the pre-loaded output level. This BESS enable signal is
triggered at the same time the relay decides to trip the breakers, as such, the BESS is assumed
to begin ramping at three cycles after the breaker opens to account for some communication
delay. This rapid BESS response prevents the system from going out of step in the simulations.
The winter peak maximum import condition with the BESS auto-scheduling enabled.
Figure 6: Winter Peak Maximum Import - HCCP trip + BESS Auto-Scheduling
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 20
When the BESS provides its maximum output three cycles after the initial HCCP trip, the system
remains in synchronism. With the auto-scheduling feature, all generator trip simulations resulted
in stable simulations such as this one.
The additional unit trips of Chena 5, Healy 1, and North Pole Combined Cycle were expected to
be less severe than the trip of HCCP. These were chosen to define a unit trip limit for out of step
conditions without the BESS auto-scheduling.
The results for the G4 and G5 disturbances result in stable, well-damped systems suggesting
that the limit is larger than the 28.5MW seen on the Healy 1 unit. However, the loss of North
Pole Combined Cycle 1 for the summer valley case with maximum import results in an out of
step condition. In order to mitigate this condition, EPS modeled the auto-scheduling for the
NPCC1 unit trip. The results for the out of step condition as well as the mitigation by auto-
scheduling the BESS are shown below in figures 7 and 8.
Figure 7: Winter Peak Maximum Import - NPCC1 Trip
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 21
Figure 8: Winter Peak Maximum Import - NPCC1 Trip BESS Auto-Scheduled
Figure 8 shows the impact of the BESS’s fast response capability to a large loss of generation.
By providing MW within 1/10th of a second, the out of step condition is avoided. Assuming
GVEA has the capability to start generation before the BESS energy is expended, the loss of
the NPCC1 unit is survivable under all contingency scenarios studied. Table 11 shows the
simulation results observed in the Delta Junction Wind Farm dynamic simulations.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 22
Outage G1 G2 G3 G4 G5 G6 G7 G8
PowerFlow
Case HCCP SCPP 3
HCCP +
TTHly
Cantwell Chena5 Healy 1 NPCC1 HCCP +BESS NPCC1+BESS
wp_open Stable Stable Stable Stable Stable Stable Stable Stable
wp_brad Stable Stable Stable Stable Stable Stable Stable Stable
wp_75 Out of Step Stable Stable Stable Stable Outof Step Stable Stable
sp_open Stable Stable Stable Stable Stable Stable Stable Stable
sp_brad Stable Stable Stable Stable Stable Stable Stable Stable
sp_75 Outof Step Stable Stable Stable Stable Stable Stable Stable
sv_open Stable Stable Stable Stable Stable Stable Stable Stable
sv_brad Stable Stable Stable Stable Stable N/A Stable N/A
sv_75 Stable Stable Stable Stable Stable N/A Stable N/A
d_wp_open Stable Stable Stable Stable Stable Stable Stable Stable
d_wp_brad Stable Stable Stable Stable Stable Stable Stable Stable
d_wp_75 Out of Step Stable Stable Stable Stable Stable Stable Stable
d_sp_open Stable Stable Stable Stable Stable Stable Stable Stable
d_sp_brad Stable Stable Stable Stable Stable Stable Stable Stable
d_sp_75 Stable Stable Stable Stable Stable N/A Stable N/A
d_sv_open Stable Stable Stable Stable Stable Stable Stable Stable
d_sv_brad Stable Stable Stable Stable Stable N/A Stable N/A
d_sv_75 Stable Stable Stable Stable Stable N/A Stable N/A
de_wp_open Stable Stable Stable Stable Stable Stable Stable Stable
de_wp_brad Stable Stable Stable Stable Stable Stable Stable Stable
de_wp_75 Outof Step Stable Stable Stable Stable Stable Stable Stable
de_sp_open Stable Stable Stable Stable Stable Stable Stable Stable
de_sp_brad Stable Stable Stable Stable Stable Stable Stable Stable
de_sp_75 Stable Stable Stable Stable Stable N/A Stable N/A
de_sv_open Stable Stable Stable Stable Stable N/A Stable N/A
de_sv_brad Stable Stable Stable Stable Stable N/A Stable N/A
de_sv_75 Stable Stable Stable Stable Stable N/A Stable N/A
Table 11: Dynamic Stability Results Table
This table highlights two surprising results. First is that the G6 contingency results in an
unstable system without wind during a winter peak maximum import case, whereas the cases
with wind result in a stable system. The second surprising result is that the G1 contingencies do
not result in an out of step condition for the summer peak case with wind, but can run into out of
step with no wind.
The reason the G6 contingency is causing out of step for the winter peak without wind case is
that it is tripping a full 41MW. The cases with wind displace the North Pole Combined Cycle
generation. The dispatch level of NPCC is the primary reason for the improvement, not the
addition of wind resources. The G1 differences are similar in that by increasing the wind output,
the Healy generation will be reduced resulting in a smaller outage.
The loss of HCCP is actually the largest contingency seen by GVEA in these cases. The loss of
approximately 61MW of generation is larger than the BESS can provide. In islanded cases, the
size difference can result in load shedding when spinning reserve is insufficient to cover the 15-
20MW of generation difference. GVEA may need to use its SILOS (Shed in lieu of spin) scheme
to balance generation and load since the BESS’s 15 minute design is for 25MW not 46MW.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 23
The results of the dynamic simulations study are that the cases with wind do not negatively
impact the dynamic stability of the GVEA system.
6 Wind Model/Analysis
The fundamental step in determining the required amount of regulation reserve lies with the
analysis of the wind farm’s power fluctuation. All the power changes from the wind farm must
be accounted for and balanced by the Control Area. The response of the Control Area to a
change in power is complex and must be examined in different time periods. There are several
time frames that must be considered when doing the review and are listed as follows in
ascending order:
Primary Control – The Primary Control responds to the change in frequency and occurs
in the first few seconds. The frequency response from the utility system is provided by
the governor controls on the generators. Certain types of loads also respond to changes
in system frequency. However, while the governor control attempts to mitigate changes
to frequency, it does not attempt to restore system frequency.
Secondary Control - The Secondary Control maintains frequency in the minute to minute
basis. In the GVEA system, this is achieved through the AGC system which controls
GVEA generation and tie-line schedules.
Tertiary Control – These are normally manual controls taken by operators that are in the
time frame from 10 minutes to an hour. This includes starting units that are off-line,
changing hydro schedules or changing interchanges with neighboring utilities. Once
tertiary control is undertaken, it generally comes under the control of the utility’s AGC
system.
The Primary Control is the initial response to the change in frequency, but it does not return
frequency back to normal. Secondary Control (GVEA AGC) will correct the frequency error to
the extent possible with the resources under AGC control. So long as the unloaded capacity
under control of AGC meets or exceeds the amount of energy produced by the wind farm,
GVEA will be able to return the system frequency to 60 Hz. The variability of the AEP wind farm
will increase the amount of short term frequency error on the system from historical levels due
to the increased variability of the wind resource. However, this variability is not expected to
exceed the capability of the GVEA thermal generation units expected to provide regulation. The
main focus of this section is to characterize Delta Junction Wind Farm expected power output,
its variability and to estimate the cumulative effect of adding Eva Creek Wind Farm in
conjunction with the AEP project.
6.1 Wind Data
EPS’s converted AEP’s MET tower data into a net power output time series to be used for
power fluctuations. For this analysis, the MET data was used to get estimated 10-min average
and 1 minute average time series to be used for ramp rate analysis. This analysis is not meant
to evaluate the available long term energy from the wind farm. EPS also received wind data
from GVEA for their proposed Eva Creek wind farm that had been converted to power data by
GVEA’s sub-contractor.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 24
6.1.1 Delta Junction Wind Farm
The Delta Junction Wind Farm is located approximately 4.4 miles east of the city of Delta
Junction, Alaska. EPS contracted with Wind Logics to do the work to convert the MET tower
data received from AEP into 10 minute time series. A chronological listing of the data received
can be found in Appendix A that has Windlogic’s Report, “Delta Junction, Alaska Power Ramp
Analysis”. Wind Logic’s scope was to analyze the onsite meteorological data collected and to
calculate the net energy production of the wind farm based on the proposed 16 tower locations,
hub height, and wind turbine type. A second task was to look at wide scale ramps associated
with large scale weather patterns.
Wind Logic reviewed the MET data from 9/27/2007 through 5/13/2010 and found there were two
extended periods in 2009 where there was no data. The data review also confirmed that no
sensor errors which were prevalent in sensors of this time period were observed in the data.
AEP’s Directwind 54/900 KW turbine was utilized to verify wind shear estimates from the MET
tower. Wing Logic recommended that the time series data collected subsequent to January
2010 should not be used for further analysis due to concerns of turbine waking affecting the
MET data. As a result of the data gaps in 2009 and Wind Logics recommendation, EPS
selected the most recent contiguous year of data for further analysis.
Wind Logic’s assessed wide area ramps associated with large scale weather patterns. They
defined a ramp event of at least 30% of the capacity as well as the ramp event to be sustained
for a period of no less than three hours. This is just a subset of what needs to be examined but
the results are still useful. Based on Windlogic’s definition, 350 ramp events, equally distributed
between up and down ramps were identified. Winter months had more ramp events than the
summer months. The signature of the ramp events varied by season, with the summer events
being shorter in duration than the winter months. The summer months are effected more by the
heating and cooling cycle. However, the winter months have extreme storms caused more by
pressure systems moving through the state.
EPS contracted with Clarity Analytical to convert the net 10 minute average power output data
received from Wind Logics to a net 1 minute average power output data that is representative of
the wind farm. Clarity Analytical also used historical 1 minute generation data from an operating
wind farm to perform this task. A chronicle of the work done can be found in Appendix A that
has Clarity Analytical’s Memorandum, “Preparation of simulated time series of 1-minute power
output of the proposed Delta Junction Wind Farm”.
6.1.2 Eva Creek Wind Farm
GVEA is evaluating a project to construct a 24 MW Eva Creek Wind Farm near Healy. The 10-
minute averages, completed by V-Bar, LLC and provided by GVEA had the simulated net
capacity factor in percent. The simulation used GE-1.5xle/e (82.5-m rotor) with a hub height of
80m but there was no notation of wind farm size. The data range was from 6/24/2008 to
4/20/2010. GVEA stated by e-mail that the net power output to the grid used for our analysis
should be 24 MW.
The time intervals of the EVA Creek and AEP wind farms are not synchronized, however EPS
assumed the data from the wind farms was reasonably synchronized when performing the “net”
impacts analysis.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 25
6.2 Wind Analysis
EPS examined the expected ramp rate of each of the data sets looking forward from the starting
point. For example, one set of tables and graphs show the number of occurrences of a
particular MW change, both positive and negative, for different time periods from the starting
point.The time periods are 10, 20, 30, 40, 50 and 60 minutes ahead. This is a rolling
calculation and made for every point that data is allowed. 60 minutes would be the difference
between the 1st and the 6th value. Next, the distribution of the samples was taken and then
split up into ranges and the tables and graphs were made that showed the entire range.
Another set of tables and graphs were made that had a smaller range and showed only a
portion of the points. This was done to show more detail in the center part of the distribution
range.
Figure 9: Graph and Chart of Distribution of MW Changes in 2 MW bins.
Another set of graphs were made to look at only the negative ramps. Negative ramps are
normally the most disruptive to the system since nothing can be done to supply short fall other
than the control area’s own generation or the implementation of a BESS. This series of scatter
charts show the negative ramp or negative difference in net power output looking ahead some
time period from a starting load. The time periods are 10, 20, 30, 40, 50 and 60 minutes ahead.
This is a rolling calculation and made for every point that data is allowed. These graphs are
very useful to determine the required regulating reserve based on risk of shortfall in the
reserves. GVEA has indicated that the control area will be dispatched without wind in order to
establish the preferred unit commitment. Once the unit commitment is established, wind will be
added to the system to back of generation production, but units will not be de-committed due to
the non-firm nature of an isolated wind farm in an islanded network. This procedure ensures
that Secondary control, or AGC control will have the ability to provide regulation for all wind
variability in the downward direction from the average production level.
Power Range 10 minutes 20 minutes 30 minutes 40 minutes 50 minutes 60 minutes
lower mw
range
[>=]
upper mw
range [<]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
-27 -25 0 0 0 0 0 1
-25 -23 0 2 4 9 8 13
-23 -21 0 2 10 16 22 23
-21 -19 1 9 16 20 37 33
-19 -17 6 12 28 44 43 70
-17 -15 6 26 46 76 89 116
-15 -13 21 65 95 105 169 199
-13 -11 41 114 188 245 268 293
-11 -9 103 217 315 417 493 532
-9 -7 245 487 629 729 841 887
-7 -5 697 1071 1234 1399 1502 1589
-5 -3 1799 2379 2667 2754 2810 2966
-3 -1 5658 5940 5867 5899 5862 5765
-1 1 35100 31851 30292 29077 28217 27559
1 3 5886 5880 5874 5881 5862 5749
3 5 1972 2510 2698 2830 2964 3021
5 7 702 1082 1310 1477 1495 1617
7 9 260 537 637 745 799 872
9 11 118 246 353 383 490 533
11 13 47 120 190 241 291 324
13 15 18 67 101 151 164 213
15 17 10 38 59 74 109 141
17 19 1 18 34 49 69 76
19 21 1 13 24 36 40 40
21 23 1 4 14 18 26 32
23 25 1 3 5 13 15 20
25 27 0 0 2 3 5 5
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 26
Figure 10: Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes Ahead
Similar scatter charts were made for the 1 minute data. The same methodology was used
except that the time periods are 1, 2, 3, 4, 5 and 6 minutes ahead. These graphs were only
done for Delta Junction Wind Farm as opposed to the net impact of the DJWF plus the Eva
Creek facility. The minute based graphs plus the 10 minute ahead chart can be used to judge
the secondary control requirements. The secondary controls maintain frequency in the minute
to minute basis. Normally this would be via automatic generation control.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 27
Figure 11: Scatter of Negative Ramp Distribution from Starting Power Level 1 Minutes Ahead
Lastly, another graph is made to look at the volatility of the data as well as the distribution. The
intent of this graph is to show the distribution of the hourly minimum, maximum and average of
the 10 minute averages when all the data samples are sorted by the Average from largest to
smallest. The hours really indicate the number of samples.
All the charts are included in Appendix B and only selected charts are in the main body of the
report.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 28
Figure 12: Delta Junction distribution of Hourly 10 minute Min-Max-Avg
6.2.1 Delta Junction Wind Farm
Examining the 10 min ahead scatter graph for DJWF indicates that the predominant maximum
negative ramp is equal to the starting power level. When the starting power level reaches 8,000
KW, the knee, the predominant maximum negative ramp becomes less than the starting power
level and ultimately reaches a level of -11,000 KW. This is of course excluding spurious points
and is plotted in Figure 13. The knee is identified by the circle in the chart which is typical for all
of them.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 29
Figure 13: Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes Ahead with
Negative Ramp Curve
Examining the 20 min ahead scatter graph, the point where the knee starts occurs at a higher
starting power level than the 10 minute ahead graph and ends ultimately at a lower power level
of -15000 KW. This can be seen in figure 14.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 30
Figure 14: Scatter of Negative Ramp Distribution from Starting Power Level 20 Minutes Ahead with
Negative Ramp Curve
Examining the 60 min ahead scatter graph, the point where the knee starts occurs at a very high
starting power level far beyond where the 20 minute ahead graph knee is. The 60 minute
ahead knee is at 20,000 KW starting power level and ends ultimately at a lower power level of -
21,500 KW. This can be seen in Figure 15.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 31
Figure 15: Scatter of Negative Ramp Distribution from Starting Power Level 30 Minutes Ahead with
Negative Ramp Curve
In systems which de-commit units based on available wind energy, the amount of required on-
line regulation is generally equivalent to either the 10 or 20-minute regulation amounts,
depending upon the starting time of the off-line units. Based on the indications from GVEA
concerning unit commitment and dispatch conditions, we believe the appropriate regulation level
for the GVEA system is the level defined by the 10-minute curve above. It is however
recognized that the GVEA system must be able to operate for all instances up to and including
the 60-minute regulation requirements.
6.2.2 Eva Creek Wind Farm
The simulation for Eva creek indicates more volatility in 10-minute average energy data than the
AEP wind farm although the size and type of wind turbine generator are similar. This can be
seen by comparing the min-max-average charts between the two wind farms. Figure 16 shows
both DJWF and Eva Creek WF next to each other.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 32
Figure 16: Delta Junction and Eva Creek Distribution of Hourly 10 minute Min-Max-Avg
There are two notable differences between the graphs. When compared to the DJWF graph,
the Eva Creek WF chart has a greater volatility for a greater range of the output level, even
though the wind turbines for the two farms are very similar. For example, beginning at about 20
MW, the Eva Creek wind farm has instances where the output of the wind farm varies between
full output and 0 MW. For the DJWF, the average output of the wind farm is at only 10 MW
when the variance reaches 0 MW output. The other interesting thing to note is that Eva Creek
has 770 hours where the output never drops below full output, compared to 270 hrs for DJWF.
The Eva Creek full output duration seems to be abnormally high given the load factor of the
plant. However, this graph was based on the wind analysis provided by GVEA.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 33
6.2.3 Delta Junction and Eva Creek Wind Farm Combined
The 10 min ahead scatter graph for the combined net impact to the GVEA system indicates that
the predominant maximum negative ramp is equal to the starting power level. When the starting
power level reaches 5,000 KW, the knee, the predominant maximum negative ramp becomes
less than the starting power level and ultimately reaches a level of -14,000 KW. This excludes
spurious points and is plotted in Figure 17. This is a decrease of 3,000 KW when compared to
the volatility for DJWF by itself.
Figure 17: Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes Ahead with
Negative Ramp Curve
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 34
Examining the 20 min ahead scatter graph, the point where the knee starts occurs at a higher
starting power level than the 10 minute ahead graph and ends ultimately at a lower power level
of -19,000 KW. This can be seen in figure 18. This is a decrease of 4,000 KW when
compared to DJWF by itself.
Figure 18: Scatter of Negative Ramp Distribution from Starting Power Level 20 Minutes Ahead with
Negative Ramp Curve
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 35
Examining the 60 min ahead scatter graph, the point where the knee starts occurs at a very high
starting power level far beyond where the 20 minute ahead graph knee is. The 60 minute
ahead knee is at 24,000 KW starting power level and ends ultimately at a lower power level of -
28,000 KW. This can be seen in figure 20. This is a decrease in the magnitude of the power
drops of 9,000 KW when compared to just the DJWF by itself. The combination of DJWF and
Eva Creek energy results in a significantly less volatile energy input to the GVEA system than if
the volatility of both projects were simply added together. Since the exact time stamp of each
project is unknown, the exact amount of diversity cannot be accurately determined, however,
logically, diversity is expected and does appear present given the assumptions in this study.
Figure 19: Scatter of Negative Ramp Distribution from Starting Power Level 60 Minutes Ahead with
Negative Ramp Curve
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 36
7 Regulation
As defined in the context of this study, regulation is the amount of generation available within
the GVEA control area to counteract the variability of wind energy and provide constant
frequency regulation and scheduled interchanges with neighboring utilities. Regulation is
generally divided into two categories, upward and downward regulation. Upward regulation is
the ability of the control area’s generation to increase its power output to make up for a
decrease in wind farm output power. Downward regulation is the ability of a control area’s
generation to decrease its power output in order to accept more wind power on its system.
The DJWF has the ability to curtail or limit its upward generation volatility to match the capability
of the GVEA system, therefore downward regulation in a control area can always be met by
limiting the upward ramps of the wind farm and is not discussed in this report.
A wind farm has no ability to limit or constrain its reduction in power resulting from a reduction in
wind. The utility control area must have the capability to increase its power output on its own
generation in order to make up for the reduction in power from the wind farm in order to
maintain system frequency and interchange schedules.
The amount of regulation required by the control area is a function of the acceptable risk to the
control area, normal regulation reserves currently available, unit commitment philosophy and
starting time of emergency generation units.
In islanded systems with single wind farms or wind farms with limited diversity, generation units
are usually committed based on the amount of wind that can be considered “firm” to the utility.
Many times this results in a unit commitment without consideration of wind energy due to the
difficulty in providing accurate forecasts of wind energy for isolated wind plants.
In this operating condition, the utility will provide 100% of the regulation requirements for the full
output of the wind energy, since wind energy will reduce the outputs of utility generation, but no
units will be removed from the system. While the utility does not capture the full benefit of wind
energy since it does not reduce the number of operating units, it does capture the incremental
cost of energy production by reducing the energy produced by the operating units.
As utilities gather more experience with wind characteristics, or they are able to provide unit
commitments on less than 24-hour notice, more benefits of the wind energy can be realized as
units are de-committed based on the amount of wind energy available to the system.
In islanded systems, the ability of the control area to provide energy changes in its generation
output through AGC is predicated on there being sufficient capacity under AGC control for the
required energy changes. Therefore, although energy changes may only be analyzed in 1-
minute increments, capacity requirements necessary to make the energy changes must be
analyzed in the time increment required to make capacity decisions in the control area. In the
case of GVEA, should units be de-committed based on expected energy from wind farms, the
GVEA system must have sufficient capacity available to enable a change in energy schedule by
its AGC system for the period of time it takes for additional capacity to be brought on-line.
Assuming GVEA can start and load an off-line unit within 10 minutes results in the requirement
for GVEA to utilize the 20-minute scatter graphs to determine the required regulation. Changes
which occur of thirty minutes would be handled by a combination of the 20-minute reserve
capacity and additional unit(s) which would be started to compensate for the ramp event. In
practice, this is difficult to administer since the duration of a ramp event can only be judged in
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 37
retrospect, that is, when a ramp event starts it may be 10-15 minutes before a decision is made
that the event justifies a unit start.
GVEA has indicated that due to the 24-hour scheduling requirements for its hydro energy and
inter-utility scheduling requirements, it will be committing its units without consideration of wind
and using the wind energy to reduce its incremental cost of energy production. As such, the
amount of regulation capacity required by the wind farm is somewhat immaterial as by virtue of
unit commitment procedures, GVEA will always have sufficient regulation available. GVEA will
however, not realize the 1:1 relationship between increased wind energy and the average cost
of displaced thermal energy by virtue of the same procedures.
7.1 Ramp Rate Analysis
The intent of this section is to look at the capacity requirement for secondary controls (AGC
controlled generation). The analysis primarily uses the scatter charts showing the negative
ramp or drop in power from a starting 10-minute average energy power level. The complete set
of the graphs can be found in Appendix C for one minute increments of time looking ahead to
determine the ability of GVEA’s thermal generation to maintain frequency control with the
maximum expected ramp rates. The predominant characterization of the DJWF is that the
expected negative change in power output will be greatest in the middle of the wind turbine
generator power curve. This can be seen in the Figure 21 for the 1-minute series which shows
the drop in power looking one minute ahead with the predominate curve shown as the solid line.
Figure 20: Scatter of Negative Distribution from Starting Power Level 1 Minute Ahead
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 38
In general it shows that -2.1 MW/min can be expected in the middle of the power curve on a
regular basis, with decreased ramp rates toward the higher or lower starting power levels.
Occasional ramp rates approaching -4.0 MW/minute can be expected. Depending upon the
time of these ramp rates with respect to the GVEA system loading and the number of units
available to respond to this ramp, the GVEA generation should be able to handle these ramp
rates. However, during times of extreme volatility, it can be expected that the wind farm will be
curtailed by GVEA in order to provide a smoothing impact of the wind farm’s output.
Although it is expected that diversity does exist between DJWF and Eva Creek, there will be
times when the diversity is additive and as such, the expected ramp rate can be expected to
double from the average of -2.1 to -4.2 MW/min with extreme rates approaching -8.0
MW/minute. It is assumed that during extreme volatility, the wind farms will be curtailed to
mitigate excessive ramp rates.
7.2 Excess Energy – Curtailment
There may be instances where curtailment of the wind farm may be necessary for reasons other
than excessive ramp rates. For example, if the amount of energy provided by wind farms
exceeds the downward regulating capability of the balancing area’s generation, including a
downward regulating margin, the wind farm energy must be curtailed or thermal units de-
committed from the system. This condition is often referred to as an “excess energy” condition.
EPS did an analysis to see how much curtailment could be expected due to excess energy
based on historical loads and estimated energy projections for the Delta Junction and the Eva
Creek Wind Farms. Using the 10-minute wind farm 10-minute power averages discussed in
Section 6 Wind Model/Analysis, EPS estimated the expected number of hours and energy
curtailments for wind energy. When both DJWF and Eve Creek are on-line, the curtailment is
not estimated per wind farm, but by total wind energy production of the two plants.
The examined data includes 7,740 hours of date from 6/24/2008 through 5/17/2009. A full
year’s worth of data (8,760 hours) could not be used due to missing information between the
two respective sets of data. Load data was received from GVEA for the system net load as well
as the Fort Knox load to allow a sensitivity analysis to be completed both with and without the
Fort Knox load. Fort Knox is a very large load on the system that maintains a fairly constant
load of approximately 30 MW. The 15 minute time increment load information was converted to
10 minute intervals using linear interpolation so that the wind energy data and GVEA load data
could be used in the analysis. The system load was assumed to be net generation to grid,
including all losses and loads. The minimum generation capability was obtained in an e-mail
from GVEA that stated the minimum load following capability (LFCMIN) of the units shown in
Table 7-1. The minimum downward regulating reserve of 3 MW was also obtained from GVEA
to account for loss of load during minimum conditions. Table 12 indicates that the minimum
generation level is 88 MW.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 39
Table 12: Summer Valley Load-Nominal Import
The minimum generation level sets the threshold of allowable wind energy on the system during
off-peak conditions. That is, the total system load minus wind energy must meet or exceed 88
Mw or a curtailment conditions will exist. For purposes of this analysis, the wind farm
generation is treated as a negative load. The wind farm’s output is subtracted directly from the
net system load and the minimum generation level is used as the curtailment point for every
historical 10 minute interval. The negative difference (adjusted load is below the threshold) is
accumulated over the analysis time period and converted to MWH to provide an approximation
of energy curtailment.
The cases that EPS examined were:
Case 1 – GVEA net load with Delta Junction Wind Farm
Case 2 – GVEA net load with Delta Junction Wind Farm and Eva Creek Wind
Farm
Case 3 – GVEA net load with Fort Knox load removed and with Delta Junction
Wind Farm
Case 4 - GVEA net load with Fort Knox load removed and with Delta Junction
Wind Farm and Eva Creek Wind Farm
To account for changes in downward regulation requirements, EPS completed evaluated the
sensitivity by evaluating the regulating reserve at 3, 6 and 9 MW. The results are shown in
Table 13.
Table 13
Case 1 had no curtailment required for normal conditions with the stated downward regulation
reserve of 3 MW. Case 2 had a minimum expected curtailment of 388 MWH. This represents
59 hours were some curtailment was in effect. When the Fort Knox load is removed from the
GVEA system, there is more curtailment with Case 3 (just Delta Junction Wind Farm) and
MinimumGeneration Scenario
Bradley Scheduled 8 MW
RegulatingReserve Down 3 MW
LFCMIN
Healy Clean Coal 35 MW
Healy 1 18 MW
Chena5 24 MW
Minimumallow generation 88 MW
RegulatingReserve Down
3MW 6MW 9MW
MWH MWH MWH
Case 1 0 0 3
Case 2 388 590 878
Case 3 1110 1814 2859
Case 4 9901 12874 16481
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 40
significant increased curtailment with Case 4 (Delta Junction Wind Farm plus Eva Creek Wind
Farm). Case 3’s curtailment represents 189 hours. Case 4 represents 5, 091 hours of
curtailment and the curtailed energy represents approximately 8.1% of the combined energy
output of the two wind farms. Curtailment of either the Delta Junction Wind Farm or the Eva
Creek wind farm or both due to excess non-firm generation does not appear to involve a
significant amount of energy.
7.3 Voltage Regulation/Coordination
The proposed GE turbines have excellent voltage control characteristics and are capable of
regulating the voltage at the POI or providing constant power factor control depending on the
control mode selected for the turbine. The wind farm will be interconnected with a relatively
large SVC on a weak system. The turbine controls must ensure that control instability or
hunting does not occur between the static voltage regulator of the WTG and the SVC at Jarvis
Creek station.
During conditions when the wind farm becomes islanded on the GVEA system without
synchronous generation, there is a possibility of severe overvoltages or abnormal frequency
operation until the turbine controls can detect and shut down the machines. The possibility can
only be examined through the use of EMTP (electro-magnetic transient programs) and detailed
models of the GVEA and wind turbine systems. There are multiple ground sources on the 138
kV system in the area of the proposed wind farm and additional mitigating measures may not be
required, however, an EMTP study should be completed during the design stage of the project
to ensure insulation coordination with the GVEA system in the area.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 41
8 Wind Ramp Dynamic Simulations
EPS performed a series of wind ramp simulations to determine GVEA’s capability to maintain
system stability for a sustained wind ramp event. Based on 1-minute wind ramp analysis, the
worst case 1-minute ramp was approximately 4 MW/min for the Delta Junction Wind farm. In
order to show a worst case ramp condition, EPS assumed that this 4 MW/min ramp would
continue until the wind plant reaches 0MW generation.
The permissible ramp rate values used in the AGC model seemed suspect. The turbines at
North Pole have been updated with ramp rates of 5%/min. The Healy coal units are modeled
with ramp rates of 3%/min but only respond in emergency situations when the frequency dips
below 59.90 Hz. In the end, these ramp rates are of little importance for these studies so long as
the regulating generators ramp rates are larger than the wind ramp rates.
The AGC model was not designed for use in simulating regulation, but rather it was designed to
measure the system response during severe under frequency events. In addition, the AGC
model for the southern utilities represents the obsolete AGC system no longer in service at
AML&P and Chugach. Conclusions based solely on the results of this section of the report
should be drawn with care. The actual sustained ramp rates used in the AGC model have been
approximated throughout the interconnected system. Additionally, the unit pulse rate, pulse
timing, emergency mode logic, and varying control modes have not been verified.
The intent of these simulations was not to confirm the absolute regulation capability of GVEA,
but rather to evaluate the stability of the GVEA system for sustained wind ramp events and its
ability to maintain interchange schedules with neighboring utilities. Four power flow cases were
selected for the wind ramp stability simulations. The cases were selected to satisfy the
following:
The power flow cases must have sufficient spin to cover the loss of all the wind
generation, ie no unit starts or BESS required for regulating capacity.
An islanded and an interconnected case from both winter and summer should be
selected.
The winter peak and summer valley cases in nominal import and islanded conditions were
selected. Table 14 below shows the regulation amounts with respect to the amount of wind
generation online.
Table 11: Available Regulation for Wind Ramp Events
Power Flow Case Available Capacity BESS Spin Regulation DWF Eva Total Wind Reg Difference
d_wp_open 60.3 25 35.3 25.6 0 25.6 9.7
d_wp_brad 78.3 25 53.3 25.6 0 25.6 27.7
d_sv_open 92.5 25 67.5 25.6 0 25.6 41.9
d_sv_brad 62 25 37 25.6 0 25.6 11.4
Regulation = Capacity BESS Spin Generation
With a maximum wind generation level of 25.6 MW, and a wind ramp rate of 4 MW/min, the
wind can reach 0 MW in 6.4 minutes (384 seconds). EPS used a custom load model to simulate
the loss of the wind in a linear ramp. The wind ramp simulations were run to 400 seconds to
capture the loss of all wind generation. The system frequencies, GVEA generation, wind
generation, and important GVEA line flows were recorded for plotting.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 42
Two sets of plots are shown below in figures 21 and 22. Figure 21 shows the d_wp_open case.
This case only has one unit (North Pole 2) on regulation with room to move. With the wind
ramping down at 4 MW/min, North Pole 2 does not have a sufficient ramp rate (3MW/min) to
keep up with the wind. As such, the GVEA frequency declines to 59.8 Hz. The top left plot
shows the system frequencies on a scale from 59.1 to 60.1Hz. The top right set of traces shows
the GVEA unit generation with the orange trace being the North Pole unit 2 generation. The
bottom left set of traces shows the wind ramp from 25.6 MW to 0 MW. The bottom right trace
shows the important line flows in the GVEA system with the red trace being the tie import.
Figure 21: Wind Ramp in Islanded Condition with One Unit Regulating
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 43
Figure 22: Wind Ramp in Connected System with Coal Providing Regulation
Figure 22 shows a simulation in which GVEA has generation capacity, and is connected to the
rest of the system. It can be seen that the HCCP unit (red trace in top right) does not ramp at
the beginning due to the large AGC deadband for this unit. However, at approximately half way
through the simulation, HCCP begins to ramp due to AGC pulses. Due to the response of
generation in the rest of the grid, the GVEA tie import (red trace bottom right) increases as
GVEA cannot keep up with the wind ramp.
The wind ramp dynamic simulation results have been included in Appendix C. Based on the
results of these simulations, EPS concludes that there are no stability concerns caused by the
ramping of wind generation. However, to avoid leaning on the other utilities in the Railbelt, the
regulation should be spread across multiple units such that the cumulative ramp rate capability
of the units is larger than the maximum expected wind ramp rate.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 44
9 Interconnection
The Delta Junction Wind Farm is located approximately 4.4 miles east of the city of Delta
Junction, Alaska. Present there are two operational turbines at the site that are at the site but
are not part of this interconnection. This interconnection is for 16 GE 1.6xle wind turbines for a
total gross capacity of 26.6 MW.
9.1 Description
The Delta Junction Wind Farm (DJWF) substation is a 138 KV station served radially from the
GVEA Jarvis Creek switching station via an overhead transmission line. The DJWF substation is
the interconnection point for the proposed wind farm project. The proposed DJWF Substation
Single Line Diagram (SLD), DJWF-EL-0010 is shown in Appendix D. The point of demarcation
at DJWF Substation is the lockable gang operated 138 KV disconnect switch (SW D1). DJWF
Substation consists of a 138 KV circuit breaker to a grounded wye-delta, 138 – 34.5 KV, step-
down transformer (T1) to a 34.5 KV three breaker switchgear. Between the 138 KV breaker
(52-D) and SW D1 are a set of PTs and CTs to be used for the revenue metering. The revenue
metering is to be per GVEA standards. There will be an additional 138 KV PT between the
circuit breaker 52-D and transformer T1. This PT will be used for synch-check purposes across
circuit breaker 52-D. The 34.5 KV switchgear will have three circuit breakers 52-A, 52-B and
52-C. The 34.5 KV switchgear consists of three 1200 amp breakers and rated 500 MVA. The
breakers are 52-A, 52-B and 52-C that connect to the step-up transformer T1, collector circuit 1,
and collector circuit 2 respectively. The switchgear will also have 3 PTs as well as a three
phase station service transformer.
The proposed DJWF Collector Circuits SLD, DJWF-EL-0011 is shown in Appendix D. The
collection circuits 1 and 2 will have eight wind turbine generators each. Each collection circuit
will have the eight turbines and a grounding transformer in a switchable normally open loop
system with a radial tap to the collector circuit breaker. The step up transformers will be
grounded wye-delta, 575Y/332-34500 V, 1750 KVA OA transformers as well as the grounding
transformer.
9.2 Protection Requirements
It is recommended that the relaying used for the 138 KV line protection between DJWF
Substation and Jarvis Switching Station be SEL-311L relays for both primary and backup at
both terminals. These relays should be configured to provide complete redundancy if one relay
needs to be taken out of service for some reason. The communication path should be
independent if possible to prevent a common point of failure.
The primary protection for step-up transformer T1 will be differential protection 87T1 (SEL-387)
that covers from the line side CT of breaker 52-D to bus side CT of 52-A. Also within the zone
of protection are lightning arrestors (108KV MCOV), one un-fused potential transformers (80.5
KV, 700/1200:1) that is used for synch-check. The primary protection will trip a lock out relay
86T1 which will trip and block close circuit breakers 52-D and 52-A. The backup protection
351T (SEL-351) will use directional ground and phase overcurrent protection utilizing the line
side CTs of breaker 52-D and the 138KV PT source. 351T will also provide 27, 59 and 81 O/U
protection that must be coordinated with GVEA. 351T will also provide the synch-check function
for circuit breaker CB 52-D. It is recommended that closing of circuit breaker 52-D only be
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 45
allowed for DB-DL and DB and LL. The backup protection will trip a lock out relay 86T2 which
will trip and block close circuit breakers 52-D and 52-A.
Circuit breaker 52-D will have breaker failure protection (SEL-501) and will trip lock out 86BF.
86BF will trip circuit breaker 52-D as well as initiate a direct transfer trip to Jarvis Switching
Station via the 87L protection. It is recommend that and immediate re-trip function be used from
the SEL-501 to prevent inadvertent operation of the scheme.
The primary protection of the 34.5 KV switchgear will be differential protection 87B (SEL-587Z)
that covers from the transformer side CT of 52-A and the feeder circuit side of 52-B and 52-C.
The differential protection 87B trips lockout relay 86B1 which trips and blocks close of circuit
breakers 52-A, 52-B and 52-C. The backup protection 351-A (SEL-351) will use directional
ground and phase overcurrent protection utilizing the transformer side CTs of breaker 52-A and
the 34.5 KV bus PT source. The backup protection 351-A also provides backup protection for
collector circuits 1 and 2. Relay elements 27 and 59 will be utilized for alarming functions. The
backup protection 351-A trips lockout relay 86B2 which trips and blocks close of circuit breakers
52-A, 52-B and 52-C.
The primary protection of the collection circuits 1 and 2 will be 151-B (SEL-351) and 151-C
(SEL-351). The 151-B and 151-C will use directional ground and phase overcurrent protection
utilizing the bus side CTs of their respective breakers and the 34.5 KV bus PT source.
There is a grounding transformer on each of the collection circuits to maintain a zero sequence
source with the collection circuit is tripped for a ground fault. Each grounding transformer
neutral will have a CT input into the SEL-351 relay for ground detection. The bus also uses the
PTs for a back up zero sequence source for ground detection. There are some possible
modification to this scheme and can change during the detail design.
Each of the wind turbine generator step-up transformer will have their own high side fuse
protection and low voltage protection. The exact configuration may change during the design
portion.
The GVEA system between Jarvis Creek and Carney Substation will require additional
protection upgrades. Due to the limited short circuit current available from the wind turbines,
line current differential relaying will be required on each line section between the Jarvis Creek
station and Carney Substation.
In addition to the line current differential relaying and its associated communication
requirements, a communication system and interface to GVEA’s existing BESS control system
will be required at the DJWF and each possible point of islanding between Jarvis Creek and
Carney substations.
9.3 Wind Turbine Options
The following options are recommended to be purchased with the DJWF turbines.
Intertial Control
GE’s Windinertia option provides simulated inertia from the wind turbines to the power grid
during off-frequency performance. This option is required for Railbelt wind turbines to prevent
the degradation of frequency due to the loss of inertia in the Railbelt in the event that wind
turbines displace higher inertia gas turbines or hydro units on the system.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 46
WindControl System
GE’s Windcontrol system controls several features of the wind farm as a whole, including the
ability to provide power frequency droop control, real and reactive power control and real power
ramp rate control. Frequency droop control may or may not be beneficial depending upon the
operating parameters of the plant and its commercial terms. However, real power control will be
required to implement full or partial curtailments and ramp rate control will be required in order
to transition from a curtailment period to a non-curtailment period. Ramp rate control may also
be required to minimize regulation costs, however the requirement for ramp rate control during
normal operation is embedded in the commercial issues associated with wind farm regulation.
WindRideThru
WindRideThru provides the GE turbines with both low and high voltage ride-through ability.
This requirement will not be the typical GE LVRT requirement as this application is not intended
for a FERC regulated transmission system. This option will be required in order for the turbines
to continue operation through Railbelt grid or GVEA subtransmission or distribution
disturbances.
Voltage control
The FIWf will be operating in a voltage control mode as opposed to a power factor or constant
MVAr mode. The FIWF will be required to maintain a constant voltage at its point of
interconnection without respect to WTG output. This requirement may force the installation of
WindFree as a WTG option.
9.4 Interface
The DJWF will have equipment necessary to initiate, control the level of, and remove
curtailments. The implementation of the Curtailment Control Interface will allow GVEA’s System
Operator to initiate the curtailment, vary the level of curtailment, and remove the curtailment
remotely. The Curtailment Control Interface shall include a demarcation cabinet for analog and
digital signals and other devices necessary to connect to the GVEA RTU. The Curtailment
control interface should provide provisions for an analog and a status feedback indicating the
level of curtailment as well as the curtailment still being in effect. When a curtailment lower
control signal is received by DJWF’s control interface, the corresponding action shall be initiated
without delay. DJWF shall be capable of and allow Company to curtail output at a rate of up to
2 MW/min.
9.5 Voltage Regulation
The DJWF shall be capable of voltage regulation at the interconnection. DJWF’s equipment
shall deliver reactive capacity at the point of the interconnection from 95% leading to 90%
lagging while delivery full power to GVEA.
9.6 Curtailment Procedures
During periods of excess energy or other times when the wind farm requires curtailment, the
wind farm must have a single point controller, capable of establishing an operating point
selectable by the utility system or its designee. When curtailment is instituted, the facility should
provide a ramped response to the point of curtailment. When the curtailment is released, the
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 47
controller should provide a ramped response from the curtailed condition to the full generating
capability of the plant.
The ramp rate control should be selectable to match system capability at the time of the
curtailment.
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 51
Appendix D – EPS Wind Analysis
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 1
Appendix zz – Wind Charts and Tables
Scatter Graphs – 10 Minute Distribution of Negative Ramps
Delta Junction Wind Farm ………………………………………………………………..zz-2
Eva Creek Wind Farm ………………………………………………………………….…zz-5
Delta Junction and Eva Creek Wind Farms Combined ……………………………….zz-9
10 Minute Distribution of Negative Ramps MW changes Series of Graphs
Delta Junction Wind Farm ……………………………………………………………….zz-12
Eva Creek Wind Farm …………………………………………………………………….zz-13
Delta Junction and Eva Creek Wind Farms Combined ……………………………….zz-14
10 Minute Hourly Min-Max-Avg Distribution Graphs
Delta Junction Wind Farm ……………………………………………………………….zz-16
Eva Creek Wind Farm …………………………………………………………………….zz-17
Delta Junction and Eva Creek Wind Farms Combined ……………………………….zz-18
1 Minute Delta Junction Wind Farm Graphs
Scatter Charts Distribution of Negative Ramps………………………………………..zz-16
Distribution of MW changes Series of Graphs…………………………………………zz-19
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 2
Scatter Graphs – 10 Minute Distribution of Negative Ramps
Ramp rates from wind farms can occur in both the upward and downward direction
corresponding to increases and decreases, respectively, in wind speed. Increases in wind
energy are generally ignored when defining regulation requirements for modern wind
installations due to the control features available with new turbine systems. Modern control
systems have the capability of limiting increases in wind farm energy by controlling the blades of
the wind turbine to limits which are acceptable to the utility system. Therefore, should the utility
experience increases in wind energy beyond the capability of its generation; it can control the
wand farm’s ramp rate to an amount allowable by the utility system. This type of control option
is recommended for the DJWF and as such, upward ramp rates are not considered a problem in
the GVEA system.
Negative ramps are normally the most disruptive to the system since they cannot be controlled
by the wind farm itself, but must depend on either utility generation, battery storage or other
methods to make up for the loss of wind energy. In the GVEA system, the supply short fall
created by the wind farm’s negative ramp is expected to be compensated for by GVEA’s
thermal generation. The amount of generation required to make up for expected negative ramp
rates is dependent upon the time frame under consideration. For systems with assured
capacity, the negative ramp rates can be estimated using minute-minute changes in wind farm
output, generally using less than 10 minutes as the interval to define the required energy
changes.
However, in isolated power systems such as GVEA’s, regulating capacity is limited to the
number of units operating in the system. This means that regulation is limited to the capacity
defined by the unit commitment schedule of the utility. If the unit commitment schedule uses
capacity requirements from 10-minute scatter charts, it means that the utility is only planning on
having on-line capacity required to meet the maximum change in wind farm energy that occurs
over only a 10-minute period. We would not recommend capacity planning periods of less than
20 minutes in islanded systems due to delay times for starting units, the uncertainty of ramp
conditions and the delay between the initiation of a ramp event and the decision to start
additional units.
The series of scatter charts included in this appendix show the negative ramp or negative
difference in net power output looking ahead some time period from a starting load. The time
periods are 10, 20, 30, 40, 50 and 60 minutes ahead.
GVEA has indicated that thermal units will not be de-committed based on available wind energy
from a single wind farm. Unit commitment decisions will be made based on no wind energy
input. Following the unit commitment decision, the wind energy will be allowed to reduce the
thermal units to their minimum values, assuring GVEA that they have on-line capacity to
account for the total loss of wind energy in the GVEA system.
Delta Junction Wind Farm
The data used for Delta Junction Wind farm is from Wind Logics (WL). The WL data consists of
approximately two and one half years of data from 9/27/2007 through 5/13/2010 that developed
from the MET tower data from the site. The most recent contiguous year was used which is
from 5/17/2008 through 5/17/2009. There are a total of 6 graphs.
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 3
Figure 1 Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes Ahead
Figure 2 Scatter of Negative Ramp Distribution from Starting Power Level 20 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 4
Figure 3 Scatter of Negative Ramp Distribution from Starting Power Level 30 Minutes Ahead
Figure 4 Scatter of Negative Ramp Distribution from Starting Power Level 40 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 5
Figure 5 Scatter of Negative Ramp Distribution from Starting Power Level 50 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 6
Figure 6 Scatter of Negative Ramp Distribution from Starting Power Level 60 Minutes Ahead
Eva Creek Wind Farm
The data used in the calculation is from GVEA and the estimation made by V-Bar, LLC. The
GVEA data is from 6/24/2008 to 4/20/2010. At contiguous year was used which is from
6/24/2008 through 6/23/2009 which is close to the data range for Delta Junction Wind Farm.
Figure 7 Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 7
Figure 8 Scatter of Negative Ramp Distribution from Starting Power Level 20 Minutes Ahead
Figure 9 Scatter of Negative Ramp Distribution from Starting Power Level 30 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 8
Figure 10 Scatter of Negative Ramp Distribution from Starting Power Level 40 Minutes Ahead
Figure 11 Scatter of Negative Ramp Distribution from Starting Power Level 50 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 9
Figure 12 Scatter of Negative Ramp Distribution from Starting Power Level 60 Minutes Ahead
Delta Junction and Eva Creek Wind Farms Combined
The following graphs are of Delta Junction WF and Eva Creek WF combined. The data used in
the calculation is from the same sources listed in the individual graphs. Because the time dates
for the contiguous years are slightly different, the data show is slightly less than a year. The
data is from 6/24/2008 through 5/17/2009.
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 10
Figure 13 Scatter of Negative Ramp Distribution from Starting Power Level 10 Minutes Ahead
Figure 14 Scatter of Negative Ramp Distribution from Starting Power Level 20 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 11
Figure 15 Scatter of Negative Ramp Distribution from Starting Power Level 30 Minutes Ahead
Figure 16 Scatter of Negative Ramp Distribution from Starting Power Level 40 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 12
Figure 17 Scatter of Negative Ramp Distribution from Starting Power Level 50 Minutes Ahead
Figure 18 Scatter of Negative Ramp Distribution from Starting Power Level 60 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 13
10 Minute Distribution of Negative Ramps MW changes Series of Graphs
The intent of this series of graphs and charts are to show the number of occurrences of a
particular MW change, both positive and negative, for different time periods from the starting
point.The time periods are 10, 20, 30, 40, 50 and 60 minutes ahead. This is a rolling
calculation and made for every point that data is allowed. 60 minutes would be the difference
between the 1st and the 6th value. Next, the distribution of the samples was taken and then
split up into MW ranges or bins and the table and graph are shown. The first graph and table
has a larger MW range or bin that covers the total count. The second graph and table has a
smaller MW range or bin that shows more details of the distribution centered on zero. Each wind
farm sections are covered separately.
Delta Junction Wind Farm
Delta Junction Wind Farm full count graph and table use 2 MW bins for the full count and is
shown in figure 19. Figure 19 is quite useful for the larger deviations but is congested in the +/-
5 MW range. The detailed partial count graph and table use 0.5 MW bins shown in Figure 20
and covers +/- 5.25 MW.
An example of the graph and table shown in Figure 19, using the 10 minute ahead, there was
one occurrence that happen in the range of -19.0 MW to -21.0 MW. The number of occurrences
between -15.0 MW to -21.0 MW would be the sum of the individual ranges (6 + 6 + 1) or 13.
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 14
Figure 19 Graph and Chart of Distribution of MW Changes in 2 MW bins.
Figure 20 Graph and Chart of Distribution of MW Changes in 0.5 MW bins.
Power Range 10 minutes 20 minutes 30 minutes 40 minutes 50 minutes 60 minutes
lower mw
range
[>=]
upper mw
range [<]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
-27 -25 0 0 0 0 0 1
-25 -23 0 2 4 9 8 13
-23 -21 0 2 10 16 22 23
-21 -19 1 9 16 20 37 33
-19 -17 6 12 28 44 43 70
-17 -15 6 26 46 76 89 116
-15 -13 21 65 95 105 169 199
-13 -11 41 114 188 245 268 293
-11 -9 103 217 315 417 493 532
-9 -7 245 487 629 729 841 887
-7 -5 697 1071 1234 1399 1502 1589
-5 -3 1799 2379 2667 2754 2810 2966
-3 -1 5658 5940 5867 5899 5862 5765
-1 1 35100 31851 30292 29077 28217 27559
1 3 5886 5880 5874 5881 5862 5749
3 5 1972 2510 2698 2830 2964 3021
5 7 702 1082 1310 1477 1495 1617
7 9 260 537 637 745 799 872
9 11 118 246 353 383 490 533
11 13 47 120 190 241 291 324
13 15 18 67 101 151 164 213
15 17 10 38 59 74 109 141
17 19 1 18 34 49 69 76
19 21 1 13 24 36 40 40
21 23 1 4 14 18 26 32
23 25 1 3 5 13 15 20
25 27 0 0 2 3 5 5
Power Range 10 minutes 20 minutes 30 minutes 40 minutes 50 minutes 60 minutes
lower mw
range
[>=]
upper mw
range [<]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
-5.25 -4.75 270 404 457 480 515 520
-4.75 -4.25 370 460 588 548 565 675
-4.25 -3.75 404 605 643 684 680 733
-3.75 -3.25 574 710 826 834 876 869
-3.25 -2.75 771 920 911 1035 988 993
-2.75 -2.25 1001 1220 1193 1198 1164 1131
-2.25 -1.75 1373 1336 1398 1463 1424 1431
-1.75 -1.25 1937 1913 1803 1754 1800 1736
-1.25 -0.75 2771 2785 2780 2712 2660 2680
-0.75 -0.25 4889 4196 3967 3678 3666 3516
-0.25 0.25 22282 20422 19281 18658 18053 17472
0.25 0.75 4681 4035 3825 3593 3442 3428
0.75 1.25 2710 2673 2636 2538 2466 2526
1.25 1.75 1825 1742 1711 1720 1769 1662
1.75 2.25 1334 1330 1352 1352 1396 1360
2.25 2.75 1005 1079 1103 1134 1063 1130
2.75 3.25 800 881 929 969 1011 952
3.25 3.75 599 768 827 829 887 859
3.75 4.25 471 600 574 661 712 745
4.25 4.75 344 473 573 597 600 623
4.75 5.25 264 372 456 471 483 524
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 15
Eva Creek Wind Farm
The data used in the calculation is from GVEA and the estimation made by V-Bar, LLC. The
GVEA data is from 6/24/2008 to 4/20/2010. At contiguous year was used which is from
6/24/2008 through 6/23/2009 which is approximately the same as Delta Junction Wind Farm.
Eva Creek Wind Farm full count graph and table use 2 MW bins for the full count and is shown
in figure 19. Figure 19 is quite useful for the larger deviations but is congested in the +/- 5 MW
range. The detailed partial count graph and table use 0.5 MW bins shown in Figure 20 and
covers +/- 5.25 MW.
Figure 21 Graph and Chart of Distribution of MW Changes in 2 MW bins.
Figure 22 Graph and Chart of Distribution of MW Changes in 0.5 MW bins.
Power Range 10 minutes 20 minutes 30 minutes 40 minutes 50 minutes 60 minutes
lower mw
range
[>=]
upper mw
range [<]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
-27 -25 0 0 0 0 0 0
-25 -23 14 25 36 48 64 74
-23 -21 1 7 14 38 40 54
-21 -19 11 13 32 42 57 71
-19 -17 5 26 54 68 90 112
-17 -15 14 40 76 113 143 175
-15 -13 17 84 126 159 206 256
-13 -11 64 151 226 279 334 376
-11 -9 107 254 350 457 519 585
-9 -7 239 486 620 735 828 834
-7 -5 574 900 1158 1282 1356 1455
-5 -3 1486 2027 2196 2395 2459 2616
-3 -1 5039 5355 5437 5319 5431 5208
-1 1 37545 33923 31994 30651 29576 28969
1 3 4920 5272 5270 5330 5325 5144
3 5 1422 1978 2242 2390 2436 2519
5 7 585 907 1157 1278 1411 1467
7 9 276 523 618 771 815 940
9 11 110 236 390 433 508 586
11 13 54 138 205 283 349 352
13 15 25 76 147 195 221 286
15 17 11 54 82 110 145 183
17 19 10 31 48 72 101 118
19 21 11 19 24 34 47 67
21 23 4 7 18 19 26 34
23 25 15 26 37 55 68 73
25 27 0 0 0 0 0 0
Power Range 10 minutes 20 minutes 30 minutes 40 minutes 50 minutes 60 minutes
lower mw
range
[>=]
upper mw
range [<]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
-5.25 -4.75 200 342 388 410 461 459
-4.75 -4.25 263 390 456 499 535 569
-4.25 -3.75 361 456 534 578 564 660
-3.75 -3.25 432 596 622 698 695 732
-3.25 -2.75 666 765 781 815 865 834
-2.75 -2.25 826 939 1048 1039 1101 1070
-2.25 -1.75 1125 1296 1309 1287 1247 1276
-1.75 -1.25 1634 1683 1657 1618 1659 1544
-1.25 -0.75 2467 2322 2151 2083 2110 1950
-0.75 -0.25 4513 4039 3828 3750 3445 3493
-0.25 0.25 26023 23424 22209 21284 20555 20042
0.25 0.75 4346 4006 3625 3443 3385 3292
0.75 1.25 2346 2141 2167 2000 1938 1962
1.25 1.75 1646 1758 1581 1593 1582 1588
1.75 2.25 1101 1236 1333 1320 1450 1263
2.25 2.75 810 907 995 1073 1002 1017
2.75 3.25 587 742 778 785 773 808
3.25 3.75 432 588 700 704 696 744
3.75 4.25 369 488 547 622 574 602
4.25 4.75 243 402 419 463 557 538
4.75 5.25 225 327 374 425 490 459
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 16
Delta Junction and Eva Creek Wind Farms Combined
The following graphs and tables are for Delta Junction WF and Eva Creek WF combined. The
data used in the calculation is from the same sources listed in the individual table and graphs.
Because the time dates for the contiguous years are slightly different, the data show is slightly
less than a year. The data is from 6/24/2008 through 5/17/2009.
The combined wind farm’s full count graph and table use 4 MW bins and is shown in figure 23.
Figure 23 is quite useful for the larger deviations but is congested in the +/- 5 MW range. The
detailed partial count graph and table use 0.5 MW bins shown in Figure 24 and covers +/- 5.25
MW.
Figure 23 Graph and Chart of Distribution of MW Changes in 4 MW bins.
Figure 24 Graph and Chart of Distribution of MW Changes in 0.5 MW bins.
Power Range 10 minutes 20 minutes 30 minutes 40 minutes 50 minutes 60 minutes
lower mw
range
[>=]
upper mw
range [<]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
-46 -42 0 1 2 1 1 0
-42 -38 0 0 1 1 3 5
-38 -34 0 1 0 4 3 5
-34 -30 0 1 1 6 9 11
-30 -26 1 2 9 12 24 37
-26 -22 13 32 55 87 116 133
-22 -18 17 56 122 159 206 269
-18 -14 72 184 287 394 465 545
-14 -10 238 512 755 915 1063 1184
-10 -6 994 1802 2135 2458 2662 2795
-6 -2 5801 6841 7217 7342 7334 7401
-2 2 32800 28392 26107 24443 23475 22443
2 6 5927 6782 6985 7263 7175 7253
6 10 1047 1800 2275 2482 2727 2889
10 14 215 514 783 960 1084 1218
14 18 53 193 294 418 504 595
18 22 22 65 114 153 204 250
22 26 19 36 63 99 126 130
26 30 3 4 8 15 24 33
30 34 0 3 6 2 7 13
34 38 0 0 1 4 6 8
38 42 0 0 0 1 0 0
Power Range 10 minutes 20 minutes 30 minutes 40 minutes 50 minutes 60 minutes
lower mw
range
[>=]
upper mw
range [<]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
Delta between
10 min avg
[counts]
-5.25 -4.75 379 560 630 598 666 700
-4.75 -4.25 527 615 747 744 759 812
-4.25 -3.75 638 752 852 923 879 928
-3.75 -3.25 715 949 961 981 1000 975
-3.25 -2.75 1030 1107 1111 1173 1131 1126
-2.75 -2.25 1305 1397 1396 1379 1364 1285
-2.25 -1.75 1674 1623 1591 1566 1490 1485
-1.75 -1.25 2231 2079 2050 1834 1863 1769
-1.25 -0.75 3051 2752 2510 2425 2244 2256
-0.75 -0.25 4681 3816 3441 3126 3073 2893
-0.25 0.25 11538 9674 8726 8224 7760 7326
0.25 0.75 4496 3739 3338 3077 2937 2782
0.75 1.25 2862 2652 2465 2333 2181 2144
1.25 1.75 2166 2025 1901 1796 1798 1751
1.75 2.25 1680 1559 1594 1551 1455 1420
2.25 2.75 1206 1281 1299 1327 1247 1266
2.75 3.25 1031 1090 1122 1154 1101 1011
3.25 3.75 807 919 990 968 953 988
3.75 4.25 676 809 765 872 902 878
4.25 4.75 487 624 679 713 787 800
4.75 5.25 394 583 578 600 673 714
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 17
10 Minute Hourly Min-Max-Avg Distribution Graphs
The intent of this graph is to show the distribution of the hourly minimum, maximum and
average of the 10 minute averages when all the data samples are sorted by the Average from
largest to smallest. The hours on the graph really indicate the number of samples. The graphs
show the volatility of the wind farm across the spectrum of average power output.
Delta Junction Wind Farm
The data used for Delta Junction Wind farm is from Wind Logics (WL). The WL data consists of
approximately two and one half years of data from 9/27/2007 through 5/13/2010 that developed
from the MET tower data from the site. Previously it was mentioned that the most recent
contiguous year was used which is from 5/17/2008 through 5/17/2009.
Figure 25 Delta Junction distribution of Hourly 10 minute Min-Max-Avg
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 18
Eva Creek Wind Farm
The data used in the calculation is from GVEA and the estimation made by V-Bar, LLC. The
GVEA data is from 6/24/2008 to 4/20/2010. At contiguous year was used which is from
6/24/2008 through 6/23/2009 which is close to the data range for Delta Junction Wind Farm. It
is interesting to see that there are almost a 771 hours where there was no minimum less than
full output. Another point to pull away from the graph is volatility.
Figure 26 Eva Creek distribution of Hourly 10 minute Min-Max-Avg
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 19
Delta Junction and Eva Creek Wind Farms Combined
The following graphs are of Delta Junction WF and Eva Creek WF combined. The data used in
the calculation is from the same sources listed in the individual graphs. Because the time dates
for the contiguous years are slightly different, the data show is slightly less than a year. The
data is from 6/24/2008 through 5/17/2009
Figure 26 Combined WF Distribution of Hourly 10 minute Min-Max-Avg
1 Minute Delta Junction Wind Farm Graphs
The data used for Delta Junction Wind farm is from Clarity Analytical (CA). The WL data
consists of approximately two and one half years of data from 9/27/2007 through 5/13/2010 that
developed from the MET tower data from the site. Previously it was mentioned that the most
recent contiguous year was used which is from 5/17/2008 through 5/17/2009. There are a total
of 6 graphs.
Negative ramps are normally the most disruptive to the system since nothing can be done to
supply short fall other than the control area’s own generation or the implementation of a BESS.
This series of scatter charts show the negative ramp or negative difference in net power output
looking ahead some time period from a starting load. The time periods are 10, 20, 30, 40, 50
and 60 minutes ahead. This is a rolling calculation and made for every point that data is
allowed. There are a total of 6 graphs. The reason for the numerous time periods ahead shows
the amount of reserve required depending on the response time.
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 20
Scatter Charts Distribution of Negative Ramps
Figure 27 Scatter of Negative Ramp Distribution from Starting Power Level 1 Minutes Ahead
Figure 28 Scatter of Negative Ramp Distribution from Starting Power Level 1 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 21
Figure 29 Scatter of Negative Ramp Distribution from Starting Power Level 3 Minutes Ahead
Figure 30 Scatter of Negative Ramp Distribution from Starting Power Level 4 Minutes Ahead
Alaska Environmental Power & GVEA
Delta Junction Wind Plant Impact Study
September 7, 2010 Page 22
Figure 31 Scatter of Negative Ramp Distribution from Starting Power Level 5 Minutes Ahead
Figure 32 Scatter of Negative Ramp Distribution from Starting Power Level 6 Minutes Ahead
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 52
Appendix E – Clarity Analytical Wind Analysis
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 53
Appendix F – Wind Logic - Wind Analysis
Electrical Power Systems Inc.
Delta Junction, Alaska
Power Ramp Analysis
November 24, 2010
Confidential Business Information
Client Name Electrical Power Systems Inc.
Contact Michael Bradley
Document ID 1110-835.001-01
Issue B
Version Final
Date November 24, 2010
Prepared by: Ryan Butterfield
Technical Consultant
WindLogics
1021 Bandana Boulevard East, Suite 111
Saint Paul, Minnesota 55108 USA
651.556.4200
support@windlogics.com
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page i
Final 1110-835.001-01
DISCLAIMER
The following document was prepared by WindLogics Inc. for the use of Electrical Power
Systems Inc., 3305 Arctic Boulevard, Suite 201, Anchorage, Alaska, its successors and/or
assigns. The report is expressly and exclusively for the sole use and benefit of Electrical Power
Systems Inc. and is not for the use or benefit of, nor may be relied upon by, any other person or
entity without the advanced written consent of WindLogics Inc. This document is a trade secret
and its confidentiality is strictly maintained.
Disclaimer: WindLogics Inc. has prepared this report based on available third party historical
weather information and use of our predictive software and analysis methods. We cannot
guarantee the accuracy of historical weather data. Historical weather information also does not
necessarily allow accurate prediction of future weather patterns. WE ARE THEREFORE
PROVIDING THIS REPORT TO YOU WITHOUT ANY EXPRESS OR IMPLIED WARRANTY OF
FITNESS FOR A PARTICULAR PURPOSE OR MERCHANTABILITY. Our sole responsibility is the
preparation and delivery of this report. By accepting this report, you agree that our liability in any
situation is limited to the amount paid for our report. In no event will we be liable for any special or
consequential damages arising from use, or misuse, of our report or information in it.
REVISION HISTORY
Issue Issue Date Revision Summary
A 11/05/2010 Original Issue
B 11/24/2010 Revisions to Meteorological
Discussion
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page ii
Final 1110-835.001-01
TABLE OF CONTENTS
EXECUTIVE SUMMARY ......................................................................................................................... iii
1. INTRODUCTION .................................................................................................................................. 1
1.1 Overview of Objective and Scope ................................................................................................... 1
1.2 Site Location and Description ......................................................................................................... 1
1.3 Customer-Supplied Data................................................................................................................. 2
2. THEORETICAL WIND FARM PRODUCTION CALCULATION METHODOLOGY ............................. 3
2.1 Meteorological Tower Data Processing Methods ........................................................................... 3
2.1.1 Met Tower Sensor Configuration ............................................................................................................. 4
2.1.2 Met Tower Data Processing Overview .................................................................................................... 5
2.1.3 Data Quality and Data Recovery Results ................................................................................................ 6
2.2 Production Data Processing Methods ............................................................................................. 7
2.2.1 Power Production Data Inventory ............................................................................................................ 7
2.2.2 Nacelle-Mounted Anemometer Filtering .................................................................................................. 8
2.3 Hub-Height Wind Speed Extrapolation ........................................................................................... 8
2.4 Full-Farm Weighting........................................................................................................................ 9
2.5 Energy Production Calculation Methods ....................................................................................... 10
2.6 Project-Specific Loss Analysis Methods ....................................................................................... 11
2.6.1 Wake Losses ......................................................................................................................................... 11
2.6.2 Electrical Efficiency Losses ................................................................................................................... 12
2.6.3 Wind Sector Management Control Losses ............................................................................................ 12
3. POWER RAMP CHARACTERIZATION METHODOLOGY ............................................................... 12
3.1 Ramp Characterization ................................................................................................................. 12
3.2 Ramp Analysis Methods ............................................................................................................... 13
4. RESULTS ........................................................................................................................................... 14
4.1 Meteorological Overview............................................................................................................... 14
4.2 Ramp Occurrences ....................................................................................................................... 15
4.2.1 Monthly Power Ramp Distribution ......................................................................................................... 15
4.2.2 Diurnal Pattern ....................................................................................................................................... 16
4.3 Summary and Discussion of Data Set Limitations ........................................................................ 17
5. CONCLUSIONS ................................................................................................................................. 18
6. REFERENCES ................................................................................................................................... 20
EXHIBIT A – SEASONAL RAMP OCCURANCES ................................................................................ 21
ATTACHMENTS
APPENDICES
Met Tower Data QA/QC Processing System Appendix A
Global Weather Archive Appendix B
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page iii
Final 1110-835.001-01
EXECUTIVE SUMMARY
The purpose of this report is to analyze historical wind data at the Delta Junction,
Alaska site in order to provide context around wind power ramps for a wind farm
planned for development. WindLogics understands that the output data will be
used by Electrical Power Systems Inc. (EPS) for power flow modeling, transient
stability analysis and other distributional-power grid modeling methods to assess
utility operator control strategies.
The objective of this effort is to provide time series data deliverables and analysis
of power ramp events the Delta Junction wind project. The scope of work is to
analyze onsite meteorological data collection (met) towers and available power
production data from an operational wind turbine to calculate a time series of full-
farm power production data for the sixteen planned GE 1.6XLE wind turbines.
The power production data was then analyzed by WindLogics to provide context
regarding the positive and negative power ramp events. A power ramp definition
and power ramp rate characterization were provided along with a meteorological
pattern analysis to determine the drivers behind the temporal fluctuations of
energy production.
The general approach of the ramp characterization and data analysis is to focus
on the fundamentals to derive a theoretical energy production data set, which
are: 1) meteorological data collection (met) tower data quality, 2) hub-height
extrapolation (wind shear) calculations, and 4) the calculation of full-farm net
energy from wind speed and gross energy values. Of particular concern is how
these power production simulations relate to the integration of wind energy on
the electrical grid network and utility control room.
The results of the analysis show a generally equal trend of positive and negative
power ramp events by season and an overall drastic increase in power ramps
during the transitional and winter seasons. It was also discussed in detailed how
the seasonal weather patterns are intrinsic to the daily ramp occurrences.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 1
Final 1110-835.001-01
1. INTRODUCTION
The purpose of this report is to analyze historical wind data at the Delta Junction,
Alaska site in order to provide context around wind power ramps from a
theoretical wind farm. WindLogics understands that the output data will be used
by Electrical Power Systems Inc. (EPS) for power flow modeling, transient
stability analysis and other distributional-power grid modeling methods to assess
utility operator control strategies for the Alaskan utility operator.
1.1 Overview of Objective and Scope
The objective of this effort is to provide time series data deliverables and analysis
of power ramp events at the Delta Junction wind project. The scope of work is to
analyze onsite meteorological data collection (met) towers and available power
production data from an operational wind turbine to calculate a time series of full-
farm power production data for the sixteen planned GE 1.6XLE wind turbines.
The power production data will be analyzed to provide context regarding the
positive and negative power ramp events. A power ramp definition and power
ramp rate characterization is provided along with a meteorological pattern
analysis to determine the drivers behind large fluctuations of energy production.
The general approach of the ramp characterization and data analysis is to focus
on the fundamentals to derive a theoretical energy production data set, which
are: 1) meteorological data collection (met) tower data quality, 2) hub-height
extrapolation (wind shear) calculations, and 4) the calculation of full-farm net
energy from wind speed and gross energy values. Of particular concern is how
these power production simulations relate to the integration of wind energy on
the electrical grid network and utility control room.
1.2 Site Location and Description
The Delta Junction Wind Farm is located approximately 7 kilometers (km) east of
the city of Delta Junction, Alaska. The current makeup of the wind farm is two
operational wind turbine generators, a EWT Directwind 54/900kW and a
NorthWind 100B, with another sixteen GE 1.6XLE (82m RD) wind turbines
currently being planned for construction, for a total of 26.6 megawatts (MW) of
nameplate capacity.
Figure 1, below, shows a site map that includes the operational turbine locations
(white hatch dots) and the planned turbine locations (hollow white dots).
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 2
Final 1110-835.001-01
Figure 1: Map of Met Tower and Turbine Locations
1.3 Customer-Supplied Data
EPS supplied WindLogics with the following information and data files to be
considered in the analysis:
Manufacturer-supplied technical specifications for the GE 1.6XLE and
EWT Directwind 54/900 kW wind turbine generators. The documentation
contained power curves, thrust coefficients and acoustic parameters.
Digital photographs of the EWT Directwind 54/900 kW turbine
construction including photographs of surrounding land taken from the
nacelle.
Turbine-location coordinates for the planned GE 1.6XLE wind turbines.
Ten-minute average Supervisery Control and Data Acquisition (SCADA)
data for an EWT Directwind 54/900 kW wind turbine generator. The data
files were transmitted electronically to WindLogics and contained data for
wind speed (nacelle-mounted anemometer), wind direction (nacelle-
mounted wind vane), active power, reactive power and turbine-specified
event codes.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 3
Final 1110-835.001-01
The ten-minute average, met tower data files were transmitted
electronically to WindLogics for one (1) onsite meteorological data
collection (met) tower.
The above information was used by WindLogics to characterize the power
performance and analyze power ramp events of the Delta Junction wind farm.
Specific descriptions on data handling methodologies will be discussed in
Section 2.
2. THEORETICAL WIND FARM PRODUCTION CALCULATION METHODOLOGY
This section describes the methods used by WindLogics to create a theoretical
data set of full-farm power production data at the Delta Junction site.
The methodology for creating time series data for full-farm production data can
be summarized in the bulleted points below. Full description of the data handling
and energy production methodologies are listed, below in Sections 2.1 to 2.6.
The methodologies can be summarized as:
On site met tower data were quality checked and poor-quality data filtered
from the data set,
On site met tower data were then extrapolated to hub-height on a
timestep-by-timestep basis using the power law equation,
Extrapolated hub-height wind speed data was verified and compared to
nacelle-mounted anemometer data collected at 75 meters above ground
level (m AGL) from a nearby operational EWT Directwind 54/900 kW wind
turbine,
Extrapolated hub-height wind speed data was adjusted, on a timestep-by-
timestep basis to represent a turbine-location average using power law
dynamics, turbine layout information and turbine-specific elevations,
The turbine-location adjusted wind speed data was then converted to
power production on a timestep-by-timestep basis using manufacturer-
supplied power curve information for the GE 1.6XLE wind turbine,
turbulence intensity and site-representative air density,
Power production data was then scaled to represent gross energy output
for the full-farm (16 turbines), and
Instantaneous energy production losses were estimated and applied, on a
timestep-by-timestep basis.
The output of these methodologies were time series data of theoretical wind farm
production for the entire data collection period of the Delta Junction met tower
and will be the basis for the analysis of power ramp events.
2.1 Meteorological Tower Data Processing Methods
EPS supplied WindLogics meteorological data collection (met) tower data for one
onsite tower. Table 1, below lists the coordinate locations and data collection
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 4
Final 1110-835.001-01
period for the met tower. This section describes the methods used to quality
check the met tower data collected at the Delta Junction site.
Latitude Longitude Elevation
(WGS 84) (WGS 84) (meters AMSL)
Delta Junction Met N 64.01375 W 145.59675 403 9/27/2007 - 5/13/2010
Tower Data Collection Period
Table 1: On-Site Met Tower Location
All data collected at the Delta Junction met tower location following the
installation of EWT Directwind 54/900 kW wind turbine was reviewed for possible
data contamination due to turbine waking. The affects of turbine waking were not
seen in the tower data signature but data collected subsequent to January 2010
should not be considered for further analysis.
2.1.1 Met Tower Sensor Configuration
EPS supplied WindLogics with met tower data files in ASCII text format for the
Delta Junction met tower location. The file header and metadata contained in
these files were used to determine the configuration of the met tower and were
entered into the WindLogics Data Management database as-is.
The Delta Junction met tower was equipped with wind speed sensors at 30 and
50 m AGL. Full met tower commissioning information, as interpreted from the file
header, is listed in Table 2.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 5
Final 1110-835.001-01
Table 2: On-Site Met Tower Configuration
2.1.2 Met Tower Data Processing Overview
The ten-minute average, met tower data files, in ASCII text format, were
transmitted electronically to WindLogics. The data were quality checked and
stored in the WindLogics Data Management database. See Appendix A for more
information regarding WindLogics Met Tower Data QA/QC Processing System.
It is important to note that NRG sensor defaults for scale and offsets were used
in the wind speed data processing. A 90 degree directional offset was applied to
the wind direction data based on met tower documentation and configuration
notes provided to WindLogics.
Following the data filtering, WindLogics examined tower shadow plots for the
redundant sensors at 50m and 30m AGL, and some shadowing is seen, which is
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 6
Final 1110-835.001-01
to be expected. The effect of tower shadowing or other mast effects have on the
data and extrapolation to hub height was minimized by using the maximum value
between the redundant sensors at each measurement height.
During the analysis process, WindLogics also screened the data for the NRG
#40 vibratory mode issue (Clark, May 2009) (commonly referred to as sensor
dragging). This issue was not seen in the data screening for the Delta Junction
met tower. An additional check can be done using NRGs web interface by
inputing sensor-specific serial numbers to ensure the manufacturing of the
anemometer did not occur within a defined period by NRG. Serial numbers were
not provided in the metadata, as such, WindLogics did not perform this
verification.
2.1.3 Data Quality and Data Recovery Results
Table 3 and Figure 2, below show the amount of data recovered from the data
files and how many timestamps were flagged during the data quality screening
process.
Channel Name/Height Variable Type
Total
Timestamps
Missed
Timestamps
Percent
Received
Flagged
Timestamps
Total
Missing/Flagged
Timestamps
Percent Not
Missing/Flagged
Ch. 01 NRG Anem. 50m Wind Speed 138,070 14,572 89.45% 7,691 22,263 83.88%
Ch. 02 NRG Anem. 50m Wind Speed 138,070 14,572 89.45% 7,716 22,288 83.86%
Ch. 03 NRG Anem. 30m Wind Speed 138,070 14,572 89.45% 8,108 22,680 83.57%
Ch. 04 NRG Anem. 30m Wind Speed 138,070 14,572 89.45% 10,229 24,801 82.04%
Ch. 05 NRG Anem. 10m Wind Speed 138,070 14,572 89.45% 12,139 26,711 80.65%
Ch. 06 NRG Anem. 10m Wind Speed 138,070 14,572 89.45% 11,879 26,451 80.84%
Ch. 07 NRG Vane 50m Wind Direction 138,070 14,572 89.45% 8,462 23,034 83.32%
Ch. 08 NRG Vane 50m Wind Direction 138,070 14,572 89.45% 7,153 21,725 84.27%
Ch. 09 NRG Vane 30m Wind Direction 138,070 14,572 89.45% 7,300 21,872 84.16%
Ch. 10 NRG Vane 30m Wind Direction 138,070 14,572 89.45% 9,499 24,071 82.57%
Ch. 11 NRG Temp. 0m Temperature 138,070 14,572 89.45% 108 14,680 89.37%
Ch. 12 NRG Humid. 0m Humidity 138,070 14,572 89.45% 106 14,678 89.37%
Ch. 13 NRG Press. 0m Pressure 138,070 14,572 89.45% 69,743 84,315 38.93%
Ch. 14 NRG Volt. 0m Voltage 138,070 14,572 89.45% 0 14,572 89.45%
Total: 1,932,980 204,008 89.45% 160,133 364,141 81.16%
Delta Junction Met - 09/2007 to 5/13/2010
Table 3: Met Tower Data Recovery Summary
.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 7
Final 1110-835.001-01
Figure 2: Tower Data Coverage and Data Recovery
Note that in the figure above, the white areas are time periods not part of the
data record, but that the light-yellow areas are part of the data record (missing
data).
Several gaps in the data record exist, but did not occur for an extended period of
time, with exception of long data outages in the summer and fall of 2009. The
reason for missing data in those months was not documented in the supplied
information and is unknown to WindLogics at this time. The minor gaps that were
seen with the sensors were periods of calm wind (NRG #40 minimum reading).
2.2 Production Data Processing Methods
This section describes the process used by WindLogics to quality check and
analyze the power production data from the EWT Directwind 54/900 kW wind
turbine.
2.2.1 Power Production Data Inventory
EPS supplied WindLogics with turbine production data that were extracted from
the SCADA software. The ten-minute average data files were transmitted
electronically to WindLogics and contained data for wind speed (nacelle-mounted
anemometer), wind direction (nacelle-mounted wind vane), active power, reactive
power and turbine-specified event codes, which were entered into the
WindLogics Data Management database as-is.
For this analysis WindLogics only considered wind speed and wind direction from
the nacelle-mounted instruments. All power production data from the EWT
Directwind 54/900 kW turbine were deemed to be out of the scope of this
analysis and, as such, were disregarded.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 8
Final 1110-835.001-01
It is also important to note that no met towers were installed following
construction to allow for wake-free wind speed/direction readings. Consequently,
all data collected at the Delta Junction met tower location following installation of
EWT Directwind 54/900 kW was reviewed for possible data contamination due to
turbine waking. Time series data collected subsequent to January 2010 should
not be considered for further analysis.
2.2.2 Nacelle-Mounted Anemometer Filtering
To quality check the nacelle-mounted instrumentation for the EWT turbine, the
production data files were processed using the ExcelTM and WindographerTM
software programs. Automated and manual data filtering algorithms were not
used for this data.
2.3 Hub-Height Wind Speed Extrapolation
Estimating hub-height wind speeds from measurements at lower elevations
required the use of a wind shear exponent, commonly referred to as (alpha),
which is a value showing the relationship between wind speeds at an upper and
a lower height. This value is then used in the power law equation to calculate
estimated hub-height wind speeds.
Met tower wind speed data collected at the Delta Junction met tower location
was extrapolated to the hub height of 80 m AGL using the wind shear coefficient
calculated from data collected at the 30 and 50 m AGL sensor levels. The
calculated alpha value was applied via the power law equation to the uppermost
sensor velocity on a timestep-by-timestep basis.
WindLogics used the data collected from the nacelle-mounted anemometry on
the EWT Directwind 54/900 kW turbine to verify wind shear estimates from the
Delta Junction met tower. To do this, a comparison of monthly and seasonal
alpha values were reviewed. Table 3 below summarizes the calculated alpha
values for the met tower and operational turbine.
Upper Anemometer Lower Anemometer
Upper Anem. Wind
Speed Average
(m/s)
Lower Anem. Wind
Speed Average
(m/s)
Shear Exponent
(alpha)
Estimated 80m
AGL Wind Speed
(m/s)
50m(1) & 50m(2) Maximum 30m(1) & 30m(2) Maximum 5.57 4.68 0.297 6.67
EWT - 75m Nacelle-Mounted EWT - 10m 6.84 6.84 0.009 ~6.84
Table 4: Calculated Wind Shear
As anticipated, the nacelle-mounted anemometer wind speed was similar to the
extrapolated hub-height wind speed at the met tower location. For the same data
collection period (Jan 2010 to May 2010), the met tower extrapolated value was
6.96 m/s while nacelle-mounted anemometer had an average wind speed of 6.84
m/s. The cursory comparison added confidence in the extrapolated hub-height
wind speeds at the met tower location.
WindLogics did not incorporate any other adjustments to the wind speed profile
based on the data comparison for two reasons. These are:
1. Nacelle-mounted anemometer is waked from turbine rotor blades which
degrade the wind velocity, and
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 9
Final 1110-835.001-01
2. Lower wind speed instrumentation, identified at 10 m AGL within the data
file, shows an identical average value, which lessened the confidence in
the validity of the data.
The close geographic proximity, similar elevations and exposure of the met tower
and EWT turbine also suggested that the wind speeds would be similar.
2.4 Full-Farm Weighting
WindLogics performed a qualitative analysis to determine the representativeness
of the met tower location to that of the anticipated turbine locations. The majority
of the planned turbine locations are located to the north of the met tower location
in lower elevations and with less exposure to the east-southeast (dominant) wind
flow, as shown above in Figure 1.
To calculate the effects of elevation differences on overall (full-farm) average
wind speed, absent any wind flow modeling or detailed information of the spatial
variability at the site, WindLogics estimated the average wind speed at the
turbine locations using known properties of the change in wind speed with height
calculated using the power law equation and met tower data.
To do this, the calculated wind shear exponent (alpha) at the met tower location
was used in the power law equation to estimate an effective hub-height wind
speed representative of all turbine locations. Calculation of a turbine-specific
average wind speed was determined by:
met
met
trb
mettrb Z
ZWSWS (1)
Where:
WS = average wind speed in m/s at the turbine locations (trb) or met
tower location (met)
Z = elevation in meters above sea level at the turbine locations (trb) or
met tower location (met)
= wind shear exponent
The power law method used to estimate turbine-location specific wind speed is
an oversimplified approximation of the wind flow over terrain and forested areas.
A detailed numerical weather or wind flow model is typically used to assess the
spatial variability of a wind energy development site to estimate expected energy
yields but the simplistic assumptions used here were deemed appropriate
because of the relatively small wind farm size and overall intent of the analysis. A
detailed wind flow model accounting for atmospheric physics properties (Navier-
Stokes equations), stability, surface roughness and high-resolution elevation/land
use would likely result in different magnitudes at each turbine location but the
aggregate wind farm production, specifically instances of power ramps, will not
be materially different.
WindLogics estimated that the overall average wind speed at the turbine
locations was approximately 2% lower than the measured wind speeds at the
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 10
Final 1110-835.001-01
met tower location. The hub-height wind speed data was adjusted on a timestep-
by-timestep basis to account for this. The calculated turbine-by-turbine wind
speed and full-farm average wind speed is listed in Table 5.
Turbine
Number
Latitude
(WGS 84)
Longitude
(WGS 84)
Elevation
(m)
PowerLaw Wind
Speed (m/s)
Met 64.0138 -145.5968 403.00 6.67
1 64.0132 -145.5890 390.45 6.61
2 64.0136 -145.5959 404.47 6.68
3 64.0154 -145.5921 388.32 6.60
4 64.0160 -145.5981 407.52 6.69
5 64.0175 -145.5927 391.67 6.61
6 64.0185 -145.5984 395.02 6.63
7 64.0201 -145.5932 373.68 6.52
8 64.0215 -145.5985 387.40 6.59
9 64.0229 -145.5927 370.64 6.51
10 64.0240 -145.5984 377.65 6.54
11 64.0256 -145.5959 372.77 6.52
12 64.0259 -145.5903 364.54 6.47
13 64.0240 -145.5861 363.32 6.47
14 64.0213 -145.5862 366.37 6.48
15 64.0186 -145.5864 371.55 6.51
16 64.0159 -145.5859 373.68 6.52
381.19 6.56Turbine Average
Table 5: Calculated Turbine Average Wind Speed
It is important to note that this qualitative assessment is based on elevation
influences to the wind velocities and does not account for seasonal vegetation
porosity, roughness length and is prior to any degradation of wind speeds from
inter-turbine waking.
2.5 Energy Production Calculation Methods
Hub-height wind speed and air density values were applied to the manufacturer
wind turbine power curve to calculate power values, which were then converted
to gross energy production and gross capacity factor.
The hub-height wind speed and air density were applied to the power curve, on a
timestep-by-timestep basis, to calculate a set of power values in kilowatts (kW).
For the GE 1.6XLE, with multiple power curves, one for each air density, the air
density was used as a lookup value to select the appropriate manufacturers
power curve, which was then used along with the raw wind speed, to calculate
the corresponding power value. It is important to note that the wind speed to
power lookup values were interpolated from the manufacturers power curve wind
speed bins of 0.5 m/s to a resolution within the thousandths (0.001 m/s).
The result of this process was a set of gross energy production and gross
capacity factor values that took into account both the hub-height wind speed and
air density values at the site, on a timestep-by-timestep basis, for the entire data
collection period. This data was then weighted to represent the output of 16 wind
turbines by simply scaling by a factor of sixteen.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 11
Final 1110-835.001-01
2.6 Project-Specific Loss Analysis Methods
This section describes the process used by WindLogics to estimate the amount
of wind plant energy losses that will likely affect the energy production profile at
the Delta Junction site and thus the rate of power ramps.
Again, the intent of this analysis is to create a theoretical wind farm production
time series that represents energy production transmitted to the electrical
distribution grid. Thus, accrued energy losses for turbine availability, seasonal
icing, blade degradation and power curve performance were not considered, nor
were losses to energy applied within the data time series. However,
instantaneous energy losses due to turbine waking, electrical line losses and
high wind hysteresis1 were considered and described in more detail below.
2.6.1 Wake Losses
To account for energy losses due to inter-turbine wake effects, WindLogics used
professional judgment to estimate the percentage of wake loss. The turbine rotor
diameter (82 m) and turbine spacing were assessed. It was determined that the
overall wake effect, or loss of energy production, was approximately 8%.
Wake losses were applied on a timestep-by-timestep basis using a dynamic
wake profile (see Figure 1) that is a function of ambient wind speeds. The
resulting net energy was applied to create a load duration curve with a granularity
of one tenth of a m/s.
0.01
0.04
0.09
0.14
0.19
0.24
0.29
0.34
Wind Speed (m/s)
WakeProfilebased on 8.0% wakeloss
Figure 3: Wake Effect versus Wind Speed
1 High wind hysteresis was calculated during the wind speed to power (or power curve calculation) process and not
discounted as an operational energy loss.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 12
Final 1110-835.001-01
2.6.2 Electrical Efficiency Losses
WindLogics was not supplied with estimates of electrical line losses. Based on
professional judgment and operational experience, electrical line losses were
estimated to be 2%. This value was applied on a timestep-by-timestep basis to
represent the energy loss from each of the turbines locations to that of the
interconnection point.
2.6.3 Wind Sector Management Control Losses
Unless proper control system precautions are taken, winds from non-prevailing
directions could result in a measurable loss of energy with the current turbine
placements. WindLogics did not account for energy losses due to turbine
shutdown from a manufacturer-defined wind direction sectors or any other
turbine control losses. Based on the provided turbine layout however, these
losses may affect energy production of the Delta Junction wind farm. It is
recommended that the project developer consults the turbine manufacturer for
further details on available wind sector management programs.
3. POWER RAMP CHARACTERIZATION METHODOLOGY
Increased utility regulatory demands have placed added restrictions on wind
plant performance, specifically power ramps. Recent wind turbine generator
control system requirements include low voltage ride-through capability, voltage
control, power-output control and ramp rate control (UWIG May 2006).
For areas with limited wind energy penetration, like Alaska, the sensitivity to
variable power sources can be exaggerated. Moreover, system stability studies
are highly focused on the correlation between system load profiles and patterns
of wind plant output (daily, monthly, and seasonal) as well as small-scale power
ramps with the objective to plan for other dispatchable generation sources to
meet system demand.
The impact of additional wind penetration typically varies depending on many
factors. Nonetheless anticipating large energy variations and steep power ramps
over a short period of time can be very challenging, especially for balancing
areas with long distances to load centers, as is the case for this region of Alaska.
Using theoretical wind plant output to simulate the variable wind resource is
essential for planning wind integration. Upon commercial operation status and
live power flow to the distribution grid it is often critical to have a wind power
forecast in place to anticipate wind plant output and plan for power ramp events
to ensure continued reliability and stability of the power system.
This section describes the methods used by WindLogics to characterize a ramp
event and methods used to analyze the full-farm power production data at the
Delta Junction site. The intent is to provide a data time series that can be used
for future system-stability studies and detailed power-flow modeling.
3.1 Ramp Characterization
A power ramp is often described as a steep increase (or decrease) in power
output. The exact definition of what constitutes a significant ramp often varies
from plant operator to utility off taker. Each utility will have its own thresholds
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 13
Final 1110-835.001-01
which will be based on system load, total wind generation and the dispatchable
generation available on their system. From the utility perspective, it is also the
aggregate wind generation that is most important (barring significant
transmission constraints) for defining wind energy ramps For example, the
passage of a frontal system will cause a system-wide wind energy ramp, but that
ramp will be spread out over a longer period of time than at any individual wind
plant.
For this analysis a ramp is characterized by two criteria. These can be
summarized as follows:
1. Total ramp size must be at least 30% or more of rated capacity of the
wind farm. In the case of Delta Junction the rated capacity is 25.6 MW
equating to a ramp magnitude of 7.7 MW.
2. To generalize a ramp characteristic, a ramp event is also defined to be
sustained for a period of no less than three hours. This is done to isolate
power ramps that were caused by large-scale weather patterns and not
isolated spikes in power due to independent (yet weather related) wind
gusts, which could be the result of various micro-scale meteorological
induces. Stated differently, the scope of this analysis was to review large-
scale weather patterns influencing wide-spread power ramps and not
necessarily an in depth analysis of distinct instances of short-lived and
diverse power ramps.
It is WindLogics opinion that the above criteria best characterizes a weather-
related increase or decrease to wind plant output. Other control strategies or
external wind plant operations (imposed curtailments, system-wide faults, etc.)
can result in data signatures that mimic that of a power ramp. These cases
should be included in any power flow or system stability simulations but are not
described in this analysis as power ramps nor are they represented in the
derived data set. Moreover, power ramps and wind plant output may show much
more variability (or volatility) to a utility control room operator than captured within
this analysis. See Section 4.3 for additional discussion on temporal resolution,
data smoothing and other assumptions that may limit the use of the derived data
set.
It is also anticipated that the power ramp rate definition will be sufficient to
identify large frequency deviations in weather-related wind plant output. With that
being said, the occurrence of events is directly correlated to the power ramp
definition. If EPS has a need for a modified power ramp rate based on
expectations of both frequency and temporal rate of change, WindLogics will be
accommodating in updating the analysis.
3.2 Ramp Analysis Methods
The power production time series data was flagged for each time step that met
the above criteria. The identified power ramp events were then analyzed in
conjunction with WindLogics weather archives using various data analysis
software applications.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 14
Final 1110-835.001-01
WindLogics weather archives were used to review the large-scale weather
events that corresponded to the same time period as the power ramp events.
The weather archive consists of National Weather Service issued forecast model
data, of which WindLogics has been actively archiving for over a decade. The
extensive weather data archive is obtained by WindLogics in the standard file
format for gridded meteorological data (GRIdded Binary or grib format). This data
is then converted to a proprietary data format to be accessed across different
functioning groups throughout the company. Proprietary scripts to parse through
and analyze the surface and low level wind speed/direction data were used in
order to corroborate the results throughout this report. Although the horizontal
grid resolution ranging from 20 km to approximately 80 km2 is very coarse,
especially for the complex terrain of the region, this data was used to review
large-scale frontal patterns, pressure systems and overall wind flow patterns. The
weather archives were not used to validate the overall wind velocity magnitudes
or ramp rates since the smaller-scale features (site-specific terrain and land
characteristics) are typically not resolved. The small-scale features can play a
significant role in the magnitude of the wind resource. See Appendix B for a
detailed description of the weather archives.
WindLogics also used the ExcelTM and WindographerTM software programs to
analyze the power ramp events and compile the results within this report.
4. RESULTS
This section describes the results of the ramp analysis at the Delta Junction site.
4.1 Meteorological Overview
During the winter and transitional seasons, the wind regime of this region is
highly influenced by transient and developing synoptic-scale weather systems.
These cool/cold season weather systems track closely with the position of the
upper level jet stream. The position of the jet stream, which is near the Aleutian
Islands and the Gulf of Alaska during these seasons, results in the frequent
passage of vigorous low-pressure systems south of the Delta Junction site.
These low pressure systems often result in power ramps and sustained energy
production for multiple days. As the low weakens or moves out of the area, the
winds subside creating a sharp drop in production or negative power ramp.
During the summer months, the wind regime is influenced by a large seasonally-
persistent high pressure system in the eastern Pacific. As a result, wind speeds
over much of the region moderate. The Alaskan interior often experiences
afternoon thunderstorms in the early summer months (May and June) that drive
short-lived ramps in energy production. These often subsist within a few hours.
Mountainous terrain plays a very important role in modifying the larger scale
synoptic wind patterns that influence the Delta Junction site. The interaction of
steep terrain, prevailing winds, and forested valleys results in a variety of wind
conditions that include windy mountain tops and sheltered valleys. Specifically
the Delta Junction site is located in the plains between two significant terrain
2 Horizontal spatial resolution varies between the archived data sources.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 15
Final 1110-835.001-01
features to the north and south, respectively. Wind flow parallel to the mountains
within the Tanana River Valley drives the east-southeastern direction of winds.
4.2 Ramp Occurrences
Using the power ramp definition, WindLogics identified a total of 350 ramp events
during the met tower data collection period. Table 6, below identifies the total
positive and negative power ramp occurrences described by season and time of
day. A more detailed discussion of the seasonal and diurnal occurrences of ramp
events is located in Sections 4.2.1 and 4.2.2, respectively.
Ramp Occurrences
Hour Summer Winter Total Summer Winter Total
1-6 6 28 34 16 41 57
7-12 8 37 45 8 41 49
13-18 13 44 57 4 20 24
19-0 16 29 45 10 29 39
Total Ramps 43 138 181 38 131 169
Positive Negative
Table 6: Overall Ramp Occurrences
Not surprising is the difference in frequency of power ramps between winter and
summer. Please note that winter (and transitional) months spanned from October
through April, whereas summer months are May through September.
The signature of power ramp events also varies by season. In the summer, a
ramp up/down is short-lived, ~12-hour, representing the heating/cooling of the
surface. Whereas winter time ramp up/down tend to last one to three days (even
longer in extreme winter months of January and December) based on the
position and duration of low pressure systems in the Bay of Alaska.
4.2.1 Monthly Power Ramp Distribution
The monthly distribution of power ramp occurrences is generally correlated to the
monthly wind resource, with the transitional seasons having greater instances of
large ramp rates. Figure 4 shows the monthly power ramp occurrences.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 16
Final 1110-835.001-01
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
0
5
10
15
20
25
30
Jan. Feb. Mar Apr May Jun Jul Aug Sep Oct Nov Dec
MonthlyWind Speed and PowerRampOccurrences
Negative Ramps
PositiveRamps
Monthly Average Wind Speed
Figure 4: Monthly Pattern of Power Ramp Events and Monthly Average Wind Speed
It is suspected that the months of January and December show a significant
decrease in ramp events compared to surrounding months because of the
persistence of low pressure systems positioned in the Bay of Alaska during
extreme winter months which elongates periods of increased energy production.
4.2.2 Diurnal Pattern
The power ramp occurrences were summarized by hour to analyze the daily
pattern. This is depicted in Figure 5, below. Seasonal diurnal graphs are included
in Exhibit A.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 17
Final 1110-835.001-01
4.0
5.0
6.0
7.0
8.0
9.0
10.0
0
2
4
6
8
10
12
14
16
18
20
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23HourAKST (9 GMT)
Annual Daily RampOccurrences
PositiveRamps
Negative Ramps
WindSpeed
Poly.(Positive Ramps)
Poly.(Negative Ramps)
Figure 5: Diurnal Pattern of Power Ramp Events and Wind Speed
The highest occurrences of downramps occurs as the wind resource statistically
is increasing, and the largest number of upramps occurs when the wind
resource is statistically decreasing. What this likely suggests is that the diurnal
cycle isnt a large factor in creating wind energy ramps at least not in the
annual average. This is not totally surprising given that the sun is up all day in
summer, and near or below the horizon in winter.
On a seasonal basis the ramp occurrences tend to track well with the diurnal
pattern (Shown in Exhibit A). The depiction of diurnal wind patterns overlain with
ramp occurrences is a helpful visual of when to plan for positive or negative
energy fluctuations in the wind plant operations strategy.
4.3 Summary and Discussion of Data Set Limitations
A data set of theoretical wind farm output was compiled to analyze instances of
power ramp events and the impact of those events to the utility balancing area.
The data set is intended to support the planning of wind energy integration, but
certain data-compilation assumptions need to be conveyed in order for proper
utilization and data interpretation. Some points of concern are summarized
below:
Depending on the user requirements, ramp rates are often described as
MW change per minute. The onsite data available to perform this analysis
was at ten-minute increments.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 18
Final 1110-835.001-01
Average 10-minute wind speed and direction time series data measured
at a nearby met tower was used to derive the data. The use of 10-minute
data is widely accepted for this type of analysis but does present an issue
of data smoothing. Actual power output from an operational wind facility
will show much more variability of the minute-by-minute power
production.
The identification of power ramp occurrences within this report is based
on the simplistic assumption that all power ramps are based on weather-
related events. Large fluctuations to power output can (and do) occur in
any operational wind facility and may be greatly under-represented within
the derived data set.
The wind farm output is based on calculations from wind distributions
measured at a specific-point location that may not be representative of all
turbine locations and assumes that a weather system will affect all turbine
outputs simultaneously.
5. CONCLUSIONS
The objective of this effort was to provide time series data deliverables and
analysis of power ramp events the Delta Junction wind project. Specifically, the
scope of work was to analyze onsite meteorological data collection (met) towers
and available power production data from an operational wind turbine to calculate
a time series of full-farm power production data for the sixteen planned GE
1.6XLE wind turbines.
The power production data was then analyzed by WindLogics to provide context
regarding the positive and negative power ramp events. A power ramp definition
and power ramp rate characterization were provided along with a meteorological
pattern analysis to determine the drivers behind the temporal fluctuations of
energy production.
The general approach of the ramp characterization and data analysis was to
focus on the fundamentals to derive a theoretical energy production data set,
which are: 1) meteorological data collection (met) tower data quality, 2) hub-
height extrapolation (wind shear) calculations, and 4) the calculation of full-farm
net energy from wind speed and gross energy values. Of particular concern was
how these power production simulations relate to the integration of wind energy
on the electrical grid network and utility control room.
The results of the analysis show an increase in power ramps during the
transitional and winter seasons. It was also discussed in detailed how the
seasonal weather patterns are intrinsic to the daily ramp occurrences.
It is recommended that EPS continue to work with WindLogics to define the ramp
rate and ramp characterization that may be more accommodating to EPS and the
Alaskan utilities. Specific follow on work that is considered relevant to assist EPS
in utility stability analysis include, but not limited to: a more granular review of
ramp events, creation of additional data sources to include 1-minute timeseries
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 19
Final 1110-835.001-01
distributions, and simulating the effects of wind integration and wind variability at
the Delta Junction site.
Implementation of wind plant forecasting for both power market operation and
utility control room operations is also a recommended long-term step to
accommodate increasing amounts of wind penetration.
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 20
Final 1110-835.001-01
6. REFERENCES
AWS Truewind, LLC. February 10 2010. Overview of the State-of-the Art and Current Issues in Wind Ramp
Forecasting. Presentation at UWIG Forecasting Workshop, Albuquerque, New Mexico.
Clark, Steven H. Investigation of the NRG #40 Anemometer Slowdown, May 2009, AWEA Windpower 2009
Enernex Corporation. May 2006. Xcel Wind Integration Study for Public Service Company of Colorado - Final
Report. Knoxville, TN.
http://www.nrel.gov/wind/systemsintegration/pdfs/colorado_public_service_windintegstudy.pdf.
National Renewable Energy Laboratory. September 2004. Wind Power Plant Behaviors: Analysis of Long-Term
Wind Power Data. Golden, CO. http://www.nrel.gov/docs/fy04osti/36551.pdf.
Mitchell Jr., Murray. Strong Surface Winds at Big Delta, Alaska – An Example of Orographic Influence on Local
Weather. Monthly Weather Review, January 1956. http://docs.lib.noaa.gov/rescue/mwr/084/mwr-084-01-0015.pdf
Utility Wind Integration Group (UWIG). Utility Wind Integration State of the Art. Golden, CO. May 2006,
http://www.uwig.org/UWIGWindIntegration052006.pdf.
Wang, Jianhui, et al. Impact of Wind Power Forecasting on Unit Commitment and Dispatch. Argonne National
Laboratory. http://www.dis.anl.gov/pubs/65610.pdf.
WindLogics, Iberdrola Renewables. May 26 2010. An Introduction to Wind Ramps: Atmospheric Causes and
Forecasting. Presentation at WindPower 2010, Dallas, Texas.
Zheng, Haiyang. Kusiak, Andrew. Prediction of Wind Farm Power Ramp Rates: A Data-Mining Approach. Journal
of Solar Energy Engineering, August 2003, Volume 131. http://www.icaen.uiowa.edu/~ankusiak/Journal-
papers/Wind_Ramp_Rates_2009.pdf
Delta Junction, Alaska C O N F I D E N T I A L Power Ramp Analysis - Page 21
Final 1110-835.001-01
EXHIBIT A – SEASONAL RAMP OCCURANCES
4.0
5.0
6.0
7.0
8.0
9.0
10.0
0
2
4
6
8
10
12
14
16
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23HourAKST (9 GMT)
Winter Daily RampOccurrences
PositiveRamps
Negative Ramps
WindSpeedDiurnal
Poly.(Positive Ramps)
Poly.(Negative Ramps)
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23HourAKST (9 GMT)
Summer Daily RampOccurrences
Positive Ramps
Negative Ramps
WindSpeedDiurnal
Poly.(Positive Ramps)
Poly.(Negative Ramps)
C O N F I D E N T I A L Appendix A
WindLogics Tower Data QA/QC Processing System
Introduction
The intent of this document is to provide more information on the WindLogics Tower Data Quality
Assurance / Quality Checking (QA/QC) processing system. The QA/QC processing system is used to
screen data flows from a variety of sources, including (but not limited to) the following:
Prospecting meteorological data collection (met) towers
Permanent met towers
Turbine-mounted data collection equipment
SODAR/LIDAR units (yet to be fully implemented)
The QA/QC processing system is made up of the following three phases:
Ingest: The first phase of QA/QC processing is to ingest data in various formats. The purpose of
this phase is to convert data from the many and various data formats from different equipment
vendors to a common format for internal use so that subsequent processing methods are uniform
and consistent.
Evaluate: The second phase of QA/QC processing is to evaluate the quality of the data by adding
QA/QC “flags” (using both automatic and manual methods). It is important to note that at no point
during the data evaluation phase are data actually removed from the original data set – in fact,
information is added (in the form of the QA/QC flags) during this process.
Filter: The third and final phase of QA/QC processing is to filter the data for further processing,
based on the QA/QC flags. Only at this point are data removed from the data flow, and only from
the output data set, not the original. It is important to note that the needs of various end users have
to be kept in mind at this stage, as different end users will likely have different criteria for filtering,
depending on what their intended use for the data is.
Note: The default filter behaviors listed in the table below can be changed as needed during the
output data file creation process.
QA/QC Flags
The following table lists the types of QA/QC flags and the default filtering behaviors used to generate
QA/QC’d data files.
Variable Check Value Default Filter
Behavior
Wind Speed Max. 10-minute Avg. >= 50 m/s REJECT
Wind Speed Min. 10-minute Avg. <= 0.4 m/s REJECT
Wind Speed Max. 10-minute Std. Dev. > 4 m/s REJECT
Wind Speed Min. 10-minute Std. Dev. < 0.1 m/s REJECT
Wind Speed Max. 10-minute Delta > 30 m/s REJECT
Wind Speed Min. 1-hour Delta < 0.1 m/s REJECT
Wind Speed Previous value missing --- ALLOW
Wind Speed Negative Shear --- ALLOW
C O N F I D E N T I A L Appendix A
Variable Check Value Default Filter
Behavior
Wind Direction Max. 10-minute Avg. > 360 degrees REJECT
Wind Direction Min. 10-minute Avg. < 0 degrees REJECT
Wind Direction Max. 10-minute Std. Dev. > 75 degrees REJECT
Wind Direction Min. 10-minute Std. Dev. < 0.1 degrees REJECT
Wind Direction Min. 1-hour Delta < 0.1 degrees REJECT
Wind Direction Previous value missing --- ALLOW
Speed and
Direction Tower Shadow <= 10 degrees of
tower orientation ALLOW
Temperature Max. 10-minute Avg. > 50 degrees C REJECT
Temperature Min. 10-minute Avg. < -40 degrees C REJECT
Temperature Max. 10-minute Std. Dev. > 5 degrees C REJECT
Temperature Min. 10-minute Std. Dev. < 0 degrees C REJECT
Temperature Max. 1-hour Delta > 30 degrees C REJECT
Temperature Min. 3-hour Delta < 0.1 degrees C REJECT
Temperature Previous value missing --- ALLOW
Pressure Max. 10-minute Avg. > 1100 mb REJECT
Pressure Min. 10-minute Avg. < 700 mb REJECT
Pressure Max. 10-minute Std. Dev. > 5 mb REJECT
Pressure Min. 10-minute Std. Dev. < 0 mb REJECT
Pressure Max. 1-hour Delta > 40 mb REJECT
Pressure Min. 12-hour Delta < 0.1 mb REJECT
Pressure Previous value missing --- ALLOW
Where the units are:
m/s = meters per second
degrees = radial degrees (0-360)
degrees C = degrees Celsius
mb = millibars
C O N F I D E N T I A L Appendix B
The WindLogics Global Weather Archive
WindLogics has compiled a multi-year database of global weather through archival of analysis data
from Global Forecast System (GFS). The GFS is a global numerical weather prediction model run by
the National Centers for Environmental Prediction (NCEP). This GFS analysis data represents the
National Weather Service’s best available, real-time analysis of the state of the atmosphere every six
hours over the globe. Observational data from numerous platforms (see below) are assimilated onto a
three-dimensional grid to create these datasets. Sophisticated techniques are used to assimilate the
observational data while maintaining dynamic consistency. The analysis data are used to initialization
runs of the GFS model. WindLogics archives a version of the analysis data and first 3 hour forecast
that has 1 degree (111 km at equator, finer spacing in longitude away from equator)) horizontal grid
spacing and 26 vertical levels.
Data Input to GFS at NCEP
Aircraft weather data through the
ACARS, ASDAR and MDCARS
systems
Rawinsondes and dropwindsondes
VAD winds from WSR-88D NEXRAD
radars
SSM/I wind speeds
SSM/I precipitation estimates
TRMM precipitation estimates
MODIS IR and water vapor channel
winds
GMS, METEOSAT, GOES cloud drift
IR and visible winds
GOES water vapor channel cloud top winds
Quickscat wind speed and direction
NOAA and AQUA polar orbiter radiances
GOES-12 radiances
NOAA 405 MHz wind profilers
Boundary-layer (915 MHz) wind profilers
Surface / METAR station data
Buoy data
Ship reports
Final Draft
Alaska Environmental Power & GVEA
Delta Junction Wind Farm Impact Study
December 13, 2010 Page 54
Appendix G – EPS Proposed One-Line For AEP Wind Farm