Invenergy Beech Ridge Wind Farm Update West Virginia Wind Working Group – September 18, 2007.
Wind Working Group
10
Wind Working Group June 13
description
Wind Working Group. June 13. Wind Working Group. Seasonality and direction of wind issues Study of wind direction at Elkhorn indicated that splitting wind into two basic directions improved the forecast - PowerPoint PPT Presentation
Transcript of Wind Working Group
Wind Working GroupJune 13
Wind Working Group
• Seasonality and direction of wind issues– Study of wind direction at Elkhorn indicated that
splitting wind into two basic directions improved the forecast
– Light load versus heavy load MAE since April are 57/54 MW…respectively. We have wayyyy to much error at night.
• We are patching holes in the dike and not fixing the problem
• A more comprehensive approach is needed
Wind Working Group
• Seasonality…
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 230
20
40
60
80
100
120
140
160
180 Power Versus Time of Day
June AprilTime of day
Pow
er (M
W)
What is this?
Daily max-imum power shifts to later in the day
Average Power
decreases from Spring to Summer
Spring phenomena?
Wind Working Group
• Wind Direction…
0 50 100 150 200 250 300 350 4000
5
10
15
20
25
30
35
40
45
50
Wind Power versus Model Forecast Direction
Cassia
Direction (To)
Pow
er (M
W)
Wind Working Group• Direction…
0-19 20-39 40-59 60-79 80-99 100-119
120-139
140-159
160-179
180-199
200-219
220-239
240-259
260-279
280-299
300-319
320-339
340-360
0
2
4
6
8
10
12
14
16
18
20
Average Power versus Wind Direction
CassiaDirection (To)
Aver
age
Pow
er (M
W)
Wind Working Group• Direction…
0-19 20-39 40-59 60-79 80-99 100-119
120-139
140-159
160-179
180-199
200-219
220-239
240-259
260-279
280-299
300-319
320-339
340-360
0
5
10
15
20
25
30
35
40
45
50
Average Power versus Wind DirectionCassiaAverage Frequency/10
Direction (To)
Aver
age
Pow
er (M
W)
Wind Working Group
• Average last four runs of model (or less if unavailable)– Provides the best forecast over
single deterministic model run– 60 days updated every month
• Cutting the pie too small results in loss of statistical significance
• Addresses seasonality– Difference between model
runs can be used to develop confidence interval
Average last 4 model runs
Carry forward average model
forecast along with max and min wind
speeds
Two coefficient databases need to be developed; 1st is a function of wind direction/time of day; 2nd for hourly
averaged forecast
Wind Working Group• Divide data by three directions (NW,SE, All
others)– Customize per site
• Golden Triangle may need four directions– Use model forecast wind (WRF-NAM)
• Divide data into eight 3 hour time blocks• Hourly data would be best but there is not
enough data (yet) to be statistically significant• Addresses time of day/seasonality
• Split by model forecast 80m and 10m winds
• Convert wind speed to power and develop coefficients
Directional Component
80m and 10m wind
3 Hour Time blocks (23-00-01)
Use Power curves to convert wind speed to power; actuals and data passed to Excel/Matlab to develop
regression coefficients as a function of time of day and wind direction
Wind Working Group• Second step is developing coefficients
for the first and second hour for the hourly averaged forecast– We can either develop the
coefficients as a function of time of day or present method of hour one and two
• Same procedure would be needed as other models are added
• Coefficients database will be updated every 30 days
• Whenever a forecast is generated, the coefficients database would be used with forecast data to obtain the actual forecast
• Same procedure would be accomplished for hour-ahead forecasts with the addition of “persistence” as a forecast
F0,F1,10m,80m variables passed to Excel to develop regression
coefficients for hourly averaged forecast
This establishes the coefficients database.
Time of day coefficients
F0 and F1 coefficients
Wind Working GroupWRF-NAM
ChronologicalBreakout into3hr blocks
DirectionalBreakout
Wind SpeedBreakout
00z(Hr 6-72+)
Hrs 1-3
Hourly Average
06z(Hr 6-72+)
12z(Hr 6-72+)
18z(Hr 6-72+)
Hrs 4-6
Hrs 1-3NW
Hrs 1-3SE
Hrs 1-3Other
Hrs 1-3NW10m
Hrs 1-3NW80m
Hrs 1-3SE
10m
Hrs 1-3SE
80m
Hrs 1-3Other10m
Hrs 1-3Other80m
Hrs 4-6NW10m
Hrs 4-6NW80m
Hrs 4-6SE
10m
Hrs 4-6SE
80m
Hrs 1-3Other10m
Hrs 1-3Other80m
Hrs 1-3NW
Hrs 1-3SE
Hrs 1-3Other
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
CoeffSet
