National Grid Plc · LSE:NG.
National Grid — Technical Presentation (2011)
34 slides from National Grid's technical presentation, 2011
Slide 1: Quality of Wind BMU PNs pp11_47
Quality of Wind BMU PNs pp11_47 Place your chosen image here. The four corners must just cover the arrow tips. For covers, the three pictures should be the same size and in a straight line. by Erol Chartan
Slide 2: Introduction
Introduction This presentation will discuss the idea of good industry practice for PN submissions. It will look at definitive / comparative benchmarks and how to present the results. It will include; Examples of Common PNs vs Metering Comparing all Wind BMUs Possible Industry Benchmarks Wind BMU PNs
Slide 3: Examples of Common PNs vs Metering
Examples of Common PNs vs Metering Place your chosen image here. The four corners must just cover the arrow tips. For covers, the three pictures should be the same size and in a straight line. by Erol Chartan
Slide 4: Examples of Common PNs vs Metering
Examples of Common PNs vs Metering 4
Slide 5: Examples of Common PNs vs Metering
Examples of Common PNs vs Metering 5
Slide 6: Examples of Common PNs vs Metering
Examples of Common PNs vs Metering This is a common example of a wind BMU whose PN data increases in resolution and accuracy as the horizons decrease This approach to PN submissions is ideal There is little difference between the errors of each PN This particular wind BMU ranked 19/30 for FPN accura
Slide 7: Examples of Common PNs vs Metering
Examples of Common PNs vs Metering 7
Slide 8: Examples of Common PNs vs Metering
Examples of Common PNs vs Metering 8
Slide 9: Examples of Common PNs vs Metering
Examples of Common PNs vs Metering This is also a common example. The day and 8 hour ahead PNs are approximately constant around a “best guess” for average load factor This is a simpler approach to PN submissions The errors show that they have underestimated their average load factor for May 2011 A
Slide 10: Comparing all Wind BMUs
Comparing all Wind BMUs Place your chosen image here. The four corners must just cover the arrow tips. For covers, the three pictures should be the same size and in a straight line. by Erol Chartan
Slide 11: Comparing all Wind BMUs
Comparing all Wind BMUs It would be interesting to see how the accuracy of PN submissions from wind BMUs compare with each other Some factors that can effect the accuracy of PN submissions need to be considered when deciding an industry benchmark using analysis over one month Wind predictability in
Slide 12: Average Absolute Percentage Errors for Wind BMU PNs
Comparing all Wind BMUs Average Absolute Percentage Errors for Wind BMU PNs in Half Hourly Resolution over May 2011 Ordered in Accuracy of FPN 0.4 0.35 0.3 0.25 Percentage 0.2 0.15 0.1 0.05 0 a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad x BMU Day Ahead PN 8 Hour Ahead PN Final PN
Slide 13: Comparing all Wind BMUs
Comparing all Wind BMUs The majority of the wind BMUs have a similar percentage error; around 15%. This is very close to Windy Millar’s performance which was our wind power forecasting tool in May 2011. Total MW error varies with capacity. The “x” category at the end is the average error for each ho
Slide 14: Average Absolute Errors for Wind BMU PNs
Comparing all Wind BMUs Average Absolute Errors for Wind BMU PNs in Half Hourly Resolution over May 2011 100 90 80 70 60 MW 50 40 30 20 10 0 a b d e g h n o p q u -1 aa f t ac c k s v x y z i j l w r ab m W IN D E BMU Day Ahead PN 8 Hour Ahead PN Final PN 14
Slide 15: Comparing all Wind BMUs
Comparing all Wind BMUs Still ordered in FPN percentage error accuracy Percentage error was similar throughout and the larger wind farms are producing more error Previously conventional plant that produced an absolute error in their FPN of more than 100 MW would justify a significant event report. W
Slide 16: All half hourly Errors for Wind BMU PNs
Comparing all Wind BMUs All half hourly Errors for Wind BMU PNs summed up (Positive means Over Estimating PN) 40000 20000 0 a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad x MW half-hours -20000 -40000 -60000 -80000 -100000 -120000 BMU Day Ahead PN 8 Hour Ahead PN Final PN 16
Slide 17: Standard Deviation of the Absolute Errors
Comparing all Wind BMUs Standard Deviation of the Absolute Errors for Wind BMU PNs in Half Hourly Resolution over May 2011 0.6 0.5 Standard Deviation 0.4 0.3 0.2 0.1 0 a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad x BMU Day Ahead PN 8 Hour Ahead PN Final PN 17
Slide 18: Comparing all Wind BMUs
Comparing all Wind BMUs Generally the wind farms aren’t over or under forecasting Small bars don’t necessarily mean more accurate as errors can cancel each other out, depends on the SD of error. E.g. a low SD of error with a largely positive / negative sum of error would show a constant over / under
Slide 19: Possible Industry Benchmarks
Possible Industry Benchmarks Place your chosen image here. The four corners must just cover the arrow tips. For covers, the three pictures should be the same size and in a straight line. by Erol Chartan
Slide 20: Possible Industry Benchmarks
Possible Industry Benchmarks Throughout this chapter there are 4 possibilities for Industry Benchmarks that will be explained and discussed Absolute – A specific percentage Relative – A percentage relative to the best performing wind farm Average Relative – A percentage relative to the average perfo
Slide 21: Possible Industry Benchmarks
Possible Industry Benchmarks Noting the variability of wind speed predictability observe two example two-tier methods, 1) An absolute Industry Benchmark: Percentage Error > 15% Î Notified and warned Percentage Error > 20% Î Further action taken 2) A relative Industry benchmark: Percentage Error > pe
Slide 22: Possible Industry Benchmarks
Possible Industry Benchmarks The performance of wind BMU PNs for May 2011 for the 2 Industry Benchmarks proposed earlier is as below: 1) Absolute Industry Benchmark: 14/30 had performed acceptably 10/30 would have been notified and warned 6/30 would have had further action taken 2) Relative benchmar
Slide 23: Possible Industry Benchmarks
Possible Industry Benchmarks Relative Benchmark Advantages Adapts to different levels of predictability due to weather, forecast horizon etc There is constant motivation for the Wind Farms to improve Relative Benchmark Disadvantages If there was at least one wind farm with nearly no wind all month t
Slide 24: Possible Industry Benchmarks
Possible Industry Benchmarks Average Relative Benchmark Possibility A possibility is to have the Industry Benchmark as the average error of the 5 most accurate wind BMUs out of the 10 wind BMUs with highest load factor for each time period. This would: Remove disadvantage of windy sites struggling t
Slide 25: Wind BMU PNs Vs National Grid Forecasting
Wind BMU PNs Vs National Grid Forecasting Place your chosen image here. The four corners must just cover the arrow tips. For covers, the three pictures should be the same size and in a straight line. by Erol Chartan
Slide 26: Wind BMU PNs Vs National Grid Forecasting
Wind BMU PNs Vs National Grid Forecasting Our own forecasting tool is essential as well as having PNs. Below is the comparison of day ahead percentage error from our wind power forecasts with similar wind BMU wind power forecasts. WindyMillar’s average percentage error was 16.4 % Windy Millar’s weig
Slide 27: Wind BMU PNs Vs National Grid Forecasting
Wind BMU PNs Vs National Grid Forecasting Overall, WPFS does forecast better than WM and this period was unusual. The difference in forecasting error month on month varies a lot. Basing an Industry Benchmark that is not variable on less than one year is perhaps unfair. It is worth noting that The ba
Slide 28: Proposed Industry Benchmark
Proposed Industry Benchmark Place your chosen image here. The four corners must just cover the arrow tips. For covers, the three pictures should be the same size and in a straight line. by Erol Chartan
Slide 29: Proposed Industry Benchmark
Proposed Industry Benchmark Using the absolute method; Final PN Percentage Error > 8% Î Notified and warned Percentage Error > 12% Î Further action taken 8 hour ahead PN Percentage Error > 10% Î Notified and warned Percentage Error > 14% Î Further action taken Day ahead PN Percentage Error > 12% Î N
Slide 30: Presenting PN Performance
Presenting PN Performance Place your chosen image here. The four corners must just cover the arrow tips. For covers, the three pictures should be the same size and in a straight line. by Erol Chartan
Slide 31: Presenting PN Performance
Presenting PN Performance Present 3 tables, one for each PN horizon Sort the Wind BMUs in order of percentage error Highlight those who fall in different bands of percentage error based on what we set as an Industry Benchmark Green = Acceptable PN Orange = Notified and warned Red = Further action ta
Slide 32: Presenting PN Performance
Presenting PN Performance 32
Slide 33: Presenting PN Performance
Presenting PN Performance 33
Slide 34: Presenting PN Performance
Presenting PN Performance 34