Mark B. Lively
Utility Economic Engineer
Creating a Market for Wind Variances Instead of Imbalance Penalties
Mark B. Lively
Utility Economic Engineer
Electric systems must balance supply and demand instantaneously. This system requirement has led to some utilities imposing penalties for generator imbalances. A dynamic market for imbalances could forestall imbalance penalties and reflect the declining economies of scale associated with electricity reliability issues.
Electric systems are operated in a Procrustean manner, generally changing generator output to match consumer load exactly, but sometimes controlling consumer load to match available generation exactly. This Procrustean system concept has led some utilities to tariff designs that punish generators for every failure to match the delivery schedule.
FERC seemed to tout the penalty approach of Bonneville Power Administration in Preventing Undue Discrimination and Preference in Transmission Services, FERC Docket No. RM05-25-000 and RM05-17-000. After a small dead band, the Bonneville tariff imposed 25% penalties for additional imbalances. Surpluses are purchased at 75% of the nominal price. Shortages are charged 125% of the nominal price. These were not the 100% haircuts that Procrustes imposed on his guests, but the Bonneville penalties are still hurtful.
The physical operational requirement for matching load and generation is a system need. The tariff penalty approach imposes the system requirement onto individual generators. Yes, if
then the system would be in balance. But these individual balances almost never happen. Instead, the various imbalances mostly net out against each other, creating a small net system imbalance.
SYSTEM IMBALANCE AS PRICE DRIVER
It is the imbalance for the entire system that should be of concern for driving the tariff. The imbalance on the entire system should determine when there is a dead band. The imbalance on the entire system should determine when the price of the imbalances is 75% or 125% of the nominal price. The use of system imbalances to drive the cash-out of imbalances changes the tariff from Bonneville’s penalty structure to a market structure. I call such a pricing mechanism WOLF, for Wide Open Load Following.
Figure 1 presents the simulated imbalance for a 17 MW wind system, based on actual generation data by a wind system in Kodiak, Alaska, on 2010 October 1. The imbalances were calculated as the difference between the actual generation every two seconds versus the average generation each hour. This approach for calculating imbalances is the equivalent of perfect hourly forecasting and scheduling. The horizontal axis indicates how much the imbalance would have distorted system frequency each second, assuming an effect of 100 MW/Hertz. The vertical axis is an accumulation of the imbalances associated with each level of imbalance-driven frequency distortion.
The left side of Figure 1 represents when the system is short of power, at least when the balance is considered only from the perspective of the wind generator and its effect on system frequency. During such times, economic theory suggests higher prices for any associated imbalances, such as 125% of the nominal price, as is used in the BPA balancing tariff. All of the imbalances on the left side of Figure 1 are negative.
In contrast, the right side of Figure 1 represents when the system has a surplus of power, at least when the balance is considered only from the perspective of the wind generator and its effect on system frequency. During such times, economic theory suggests lower prices for any associated imbalances, such as 75% of the nominal price, as is used in the BPA balancing tariff. All of the imbalances on the right side of Figure 1 are positive.
In some respects, Figure 1 organizes the data in a manner that is consistent with the Bonneville imbalance tariff. Any positive imbalance will end up on the right of Figure 1 and be paid a low price. Any negative imbalance will end up on the left of Figure 1 and be charged a high price. The very center of Figure 1 can be considered to be a dead zone, where all unscheduled energy is paid the full value.
The simulated imbalances in Figure 1 were 53,296.21 MW-seconds, plus and minus. The net of these imbalances is zero, as is shown in Table 1. The penalty associated with the BPA balancing tariff is -25,312.04 MW-seconds, about 47% of the plus/minus imbalance. The penalty is energy that the simulated system delivered to the system at no compensation.
Figure 2 contains the same imbalance data as Figure 1 but organized based on a simulated joint imbalance of the Kodiak data with ACE data for Anchorage, Alaska, for 2011 October 23/24. The imbalance distributions are different between Figure 1 and Figure 2 but reflect the same total imbalance 53,296.21 MW-seconds, plus and minus, just spread across the horizontal axis in a different manner.
Table 2 evaluates the information in Figure 2 following the same format used in Table 1. The difference between Tables 2 and 1 represents the difference between a market for imbalances versus a penalty structure. Under a market structure in Figure 2 and Table 2, imbalances can be rewarded. Under the penalty structure of Figure 1 and Table 1, imbalances are always penalized, except in the rare case of the imbalance falling in the dead zone. In Table 2, the wind system would incur a cost of 21,895.40 MW-seconds, which is 41% of the total imbalance 53,296.21 MW-seconds, plus and minus. That cost is much less than the cost developed in Table 1 for Figure 1. This cost difference is the result of some of the surplus imbalances being rewarded by facing a multiplier greater than 1.0. Similarly, some of the shortage imbalances are rewarded by facing a multiplier that is less than 1.0. Also, a greater portion of the energy falls into the dead zone and faces no penalty or reward.
DIS-ECONOMIES OF SCALE
Figures 1 and 2 represent the simulated imbalances for a system that is 17 MW, the approximate size of the wind farm being proposed for the Anchorage area. The simulated generation was 4.0 times the generation of the Kodiak wind system. Figures 3 and 4 use the actual Kodiak wind data, effectively one fourth of the energy imbalances used in Figures 1 and 2.
In Figure 3, much of the wind imbalances now fall within the dead zone at 0.0 Hertz frequency error, which actually the range of -0.005 Hertz to +0.005 Hertz. The plus/minus imbalance is now 13,324.05 MW-seconds. The penalty is 4,340.91 MW-seconds, as shown in Table 3, or 33% of the plus/minus imbalance. This fraction of 33% is much less than the 47% penalty developed in Table 1.
In Figure 4, much of the wind imbalances now fall within the dead zone at 0.0 Hertz frequency error, as in Figure 3. The plus/minus imbalance is again 13,324.05 MW-seconds. The price discount is 2,249.91 MW-seconds, or 17% of the plus/minus imbalance. The reduction in the price discount is due to more relative energy imbalance appearing in the dead zone, but also to more energy imbalance having a sign (plus or minus) that is opposite to the sign of the concurrent ACE.
The difference between the first two figures and tables versus the last two figures and tables can be considered to be dis-economies of scale, and are summarized in Table 5. These dis-economies of scale are the result of shifting the valuation of imbalances between a penalty based driver versus a market based driver. As discussed in the difference between Figures 1 and 2, Figure 1 is purely a penalty based driver with the size of the generator imbalance setting the price for the imbalance. In contrast, Figure 2 is a market based driver with system frequency error setting the price for the imbalance. But the system imbalance includes the generator imbalance. As the generator imbalance gets larger, the system imbalance increasingly is driven by the generator imbalance. Since a generator imbalance price driver results in a penalty based system, increasing the size of the generator moves the imbalance pricing toward a penalty based system reducing the payments to the generator for any imbalance that it incurs.
CONTINUOUS PRICING CURVES
The BPA method for penalizing imbalances includes three discreet prices, a discount of 25% for excess surpluses, a surcharge of 25% for excess shortages, and a neutral zone with zero discount or surcharge. The penalty analyses presented in Tables 1 and 3 reflect these discreet surcharges and discounts. Similarly the pricing analyses presented in Tables 2 and 4 reflected the same discreet surcharges. Generally I have suggested the prices applicable to imbalances should be continuous, such that a frequency error or 0.001 Hertz would face a different price than a frequency error of 0.0011 Hertz. Such a continuous pricing curve could be applicable to the examples presented above, but with a gradation of 0.01 Hertz, which is the gradation presented in the various figures in this paper.
One approach to a continuous pricing function would have the percentage discount be ten times the frequency error. Thus a positive 0.01 Hertz frequency error would result in a price only 90% of the nominal price, or a hair cut of 10%. Conversely, a negative 0.05 Hertz frequency error would result in a price equal to 150% of the nominal price. This would be a bonus of 50% for surpluses and a penalty of 50% for shortages. Applying such a continuous pricing function for imbalances would result in the following reductions in the amount of energy for which the standard price is applicable.
The penalty system for imbalances can be converted to a market system by changing the independent variable driving the pricing mechanism. Using the size of the generator imbalance to drive the pricing mechanism creates a penalty mechanism. Using the size of ACE to drive the pricing mechanism creates a competitive pricing mechanism.
A comparison of Figure 2 and 4 raises an issue as to the size of the wind farm, especially in regard to the size of the rest of the grid. Though larger generators provide great cost savings both in regard to installed cost and operating costs, the cost savings can be overcome by the cost of imbalances when those imbalances are priced using ACE as an independent variable in the pricing mechanism.
The discreet penalties and prices included in Tables 1 through 4 can be replaced with continuous prices, prices that vary directly with the frequency error under consideration as the independent variable in the pricing formula.
 See "Tie Riding Freeloaders--The True Impediment to Transmission Access," Public Utilities Fortnightly, 1989 December 21, available at http://livelyutility.com/