At the time of his passing, Mark was preparing the following article for publication in USAEE Dialogue. Mrs. Tracy Lively, his widow, very kindly sent this near-to-completion "work in progress" to the Editor so that Mark's work could be shared with his colleagues, as he always did.
Too Much of a Good Thing: Wind in AlbertaMark B. Lively
“Renewable Electric Power—Too Much of a Good Thing: Looking At ERCOT” discusses the impact of wind on the wholesale prices in Texas. There was so much wind generation in the West Texas generation pocket that 25% of the settlement periods in April 2009 had negative prices in West Texas. The constraints to the rest of ERCOT were such that the rest of ERCOT experienced negative prices for the equivalent of only 3 to 4 hours, or less than 1% of the month.
Alberta, Canada, is experiencing a surge in the amount of wind generation. Some analysts have shown that wind generation is being paid an average rate that is less than the average rate paid for all generation. The Alberta Wind Energy Report by EDC Associates showed in November 2015 an average pool price of $21.17/MWH versus an average received price for wind of $17.08/MWH, a discount of 19.3% or $4.09/MWH.
In West Texas too much wind dramatically suppressed the prices paid for all generation in that region. Similarly, wind can be shown to suppress the prices for all generation in Alberta. In particular too much wind can be shown to overly suppress the prices for wind generation in Alberta. Further, wind is generally less available during times when the AESO settlement price spikes.
The Alberta Electric System Operator (AESO) provides historic hourly load levels and pricing information on its web site. AESO also has posted responses to information requests that provide 10 minute wind and load levels for recent quarters. There are minor differences in the hourly load levels and adjustments must be made for daylight savings time changes.
Data from these AESO sources are in general agreement with the Alberta Wind Energy Report, as shown in Table 1. The differences are generally attributed to timing difference, that is, that wind generation is thought to be stronger when pool prices are lower.
Figure 1 is a plot of AESO hour prices versus the hourly load for the fourth quarter of 2015. There is a very slight slope to the data, as is shown by the regression equation. The price increases $0.0052/MWH for every MW that load increases. Thus, a 1000 MW increase in load will increase the price by $5.2/MWH. Conversely, a 1000 MW decrease in load will decrease the price by $5.2/MWH. These are first approximations.
The effect of wind is close to being the same as a negative change in load. Thus, a 1000 MW increase in wind should reduce the settlement price by about $5.2/MWH. During the fourth quarter of 2015, the average wind generation was 566 MW, which suppressed the simple average of all prices by $2.90/MWH. This price suppression would be applicable to all prices, not just to wind prices, and would have no effect on average to the difference between wind prices and all prices. Under this analysis, both prices during the fourth quarter of 2015 shown in the right column of Table 1 have been lowered by $2.90/MWH, while the difference between those two prices are unchanged.
During the fourth quarter of 2015, hourly wind production varied from 0 MW to 1,347 MW. The concurrent price suppression thus varied from $0.00/MWH to $6.99/MWH. The greatest price suppression would have occurred when wind generation was the greatest. The simple average of these hourly price suppressions would be the above calculated $2.90/MWH.
The weighted average of these price suppressions, using the wind generation as the weighting factors, was $4.35/MWH. Thus, too much wind (or wind variability) caused the average wind price to be suppressed by $4.35/MWH. This $4.35/MWH price suppression for wind should be compared to the suppression of $2.90/MWH for all prices. Wind variability, independent of other timing issues, was thus responsible for $1.45/MWH of the $3.82/MWH difference shown in the right column of Table 1.
Figure 1 has eight outliers, where the price soared in each case to greater than $100/MWH. Table 2 shows the impact of wind performance during these eight hours. The Pool Price is as reported by AESO, $909.47/MWH during hour 10 of 2015 November 30. The Price Variance is relative to the average price of $21.46/MWH for all hours. This price of $21.46/MWH was shown above in Table 1. Subtracting the average price produces a Price Variance of $888.01/MWH during the specified hour.
Wind Generation is the energy reported by AESO during the relevant hour, 107 MWH during the specified hour. The Wind Variance, shown in the fourth column, is relative to the average wind generation during the quarter. The average wind generation during the quarter was 566 MW. The subtraction produces a Wind Variance of 452 MWH during the specified hour. The Lost Revenue is the extension of the two differences, that is, the simple multiplication of the Price Variance times the Wind Variance.
The Lost Revenue is spread across the wind generation during the quarter to produce the Average Impact in the right column. The Average Impact is the change in monthly average wind price during the quarter based on the lost revenue spread over the 566 MW during all hours in the quarter. Thus, these eight hours explain $1.47/MWH of the of the $3.82/MWH difference shown in the right column of Table 1.
According to the above analysis, the variability in the level of wind production plus the low production during extreme price events together explain much of the price variance identified in the Alberta Wind Energy Report. This suggests that the two explanations are at least partially duplicative.
The same analysis was done for the third quarter. The slope was $16.27/MWH for 1000 MW change in load, substantially than for the fourth quarter. The higher slope is due to the greater number of hours when the price soared above $100/MWH, 21 in the third quarter versus the 8 in the fourth quarter. That these soaring prices occurred during high load hours affects the slope.
A slope of $16.27/MWH and a average wind production level of 363 MW resulted in an average price effect of $5.91/MWH. The weighted average price effect was $9.98/MWH. Thus, the variability of wind production itself created $4.07/MWH of the difference of $6.52/MWH identified in Table 1.
The low production levels during the 21 hours when prices soared above $100/MWH contributed $3.24/MWH to the price difference. That the two price effects combined equal more than actual difference in prices. This suggests, as stated above, that the two price effects are duplicative, double counting some of the effects.
Wind generation is allowing AESO to back out some fossil fired generation. This reduction in fossil fired generation has resulted in a general lowering of the average settlement price in Alberta. The reduction in the average settlement price has been accentuated in regards to the average settlement price of wind, generally because of the variability of wind. Wind variability has resulted in significant price suppression when wind is at full power, causing a decline in the weighted average price paid to wind. Wind variability has also led to lower amounts of wind generation during periods when prices have spiked.
Dialogue, United States Association for Energy Economics, 2009 August,
 $0.0052/MWH/MW x 566 MW = $2.90/MWH
 $0.0052/MWH/MW x 1,347 MW = $6.99/MWH