The opportunity gap in economic load response

On March 25, 2013, PJM Interconnection released an analysis of its implementation of FERC Order 745, titled ‘2012 Economic Demand Response Performance Report,’ Order 745 directed independent system operators (ISOs) to create a mechanism for paying a demand response resource the locational marginal price (LMP) at its location for curtailed load whenever the hourly price is above a net benefits threshold.

In only five months after PJM implemented the order in April 2012 the volume of energy that was curtailed for this program was about equal to the volume that had been curtailed over the previous three and a half years. While this is certainly an achievement, there are still many opportunities to create increased flexibility in PJM’s demand side, and these have largely not been addressed in discussions to date.

Resource location — Not driven by price signals

In examining participation in economic load response since its post-745 rejuvenation, we see that more than 42 percent of participation has occurred in the Dominion zone that stretches from Northern Virginia to the southern border of PJM in North Carolina.

A ranking of PJM zones by average peak-time LMP in the real time market, where most economic participation has happened, put Dominion at about average. Zones in more congested areas — like Northern New Jersey, Washington D.C. and Baltimore — had average prices that were $3/MWh to $5/MWh more than Dominion.

While a ranking like this changes significantly depending on which hours are defined as “peak hours” and what statistic the ranking is based on, Dominion’s prices rarely deviate far from average. Since there are several zones just as large with price distributions similar to Dominion’s, the zone’s large geographic footprint is also not enough to explain the high participation in the zone.

The reason for this relatively high participation in Dominion likely stems from the fact that 97 percent of these energy market participants are larger than 5 MW. In fact, 86 percent are larger than 10 MW.

As PJM notes, most participation is coming from “a small number of very large customers.” In other words, a majority of the uptick we’ve seen in economic load response since last April has been from a few large industrial customers in remote, uncongested areas of the PJM grid.

Hours of participation

It also appears that most participation in this program is coming from facilities that are responding less to the market and more to the convenience of curtailing, resulting in hours of resource participation with only slight correlation to peak prices.

Participants are not following the market and optimizing revenue by responding to its signals through short-term behavioral changes. They are instead shedding load when convenience and elevated prices — even if only moderately elevated — intersect.

For example, the average amount earned per MWh curtailed from April 2012 to October 2012 across all of PJM was $65.30. This price is half a standard deviation higher than the weighted average price across all zones during the same months during hours when an industrial facility might be in full operation, and even higher than the median.

Nonetheless, it shows that participants are not taking advantage of the long tail at the high end of the price distribution. This skewed price distribution means participants have the opportunity to make revenue far higher than what they are currently making.

In fact, the average of the top 1500 hours during the relevant time period in Dominion is still significantly higher than the average price participants made per hour there. Participants in Dominion curtailed for far fewer than 1500 hours, so we cannot attribute an extremely large volume of participation hours to bringing down their average.

A larger role for economic participation?

We should now step back and look at participation in economic load response from a broader perspective. Economic participation is only a small part of demand response in PJM. The report notes that enrollment in economic load response (energy market) is only 3 percent of the enrollment in emergency load response, a mechanism that allows end-users to bid into the capacity market. Most demand response resources are enrolled in the emergency load response program.

When reviewing the information in Joule Assets’ database, one will find that capacity clearing prices for the upcoming delivery year range from $10.12/kW/year in low demand areas like Dominion to $90.21/kW/year in congested areas like PEPCO, the zone that contains Washington, D.C. Emergency load response events typically last 4 to 5 hours, but may last up to 6 hours.

And unlike economic load response, where participants voluntarily curtail during hours when they would like to be paid the LMP, in emergency load response participants are subject to mandatory curtailment requests from PJM in exchange for consistent capacity payments. The number of events called per year ranges. During 2010 in Pepco territory, there were 5 calls; but in other zones events are never called. ComEd, the zone that includes Chicago, has never had an event. Yet participants in ComEd have made an average of $30.30/kW/year just for being available to PJM operators in the rare case they are needed.

PJM recently elicited a strong response from several of the more active demand response providers after it changed the rules for curtailment service providers (CSPs) in Manual 18, which describes the business rules for PJM’s capacity market. CSPs will essentially have to fix the location of their loads before bidding into the base residual auction for the 2016/2017 delivery year, which takes place next month.

PJM instituted this because, as they noted in a statement, the current demand response requirements “do not provide enough information for PJM to adequately understand the quantity and location of demand response to include in reliability planning.” This will be problematic for CSPs, as they typically bid a fixed amount of MW into the base residual auction — three years before the delivery year — and either fill up their quota or buy the capacity back in the annual Incremental Auctions as the delivery year approaches.

PJM’s desire for more locational information about its resources is undoubtedly caused, at least partly, by the 97 percent of demand side resources that are responding to zonal capacity prices as opposed to granular energy prices. The disadvantage to a demand response market dominated by zonal capacity pricing without a robust LMP-based energy market has begun making itself prominent, which is an impetus for broader participation in the energy market by demand resources.

Increased enrollment may suppress capacity prices, forcing demand response providers to look for other ways to participate in the market. EnerNOC, for example, recently announced it will be taking a smaller position in ISO-New England’s forward capacity market, where capacity prices have been consistently low. It will also be increasing its position in the ISO’s energy markets, by enrolling more of its customers in price responsive demand.

Conclusion

There is currently a significant opportunity gap for economic load response since participation in the program is coming primarily from a few large participants responding mostly as a result of convenience to their own operations. We should focus on enabling more diverse participation through the use of new methods and technologies.

This may be use of advanced building management systems, facility optimization platforms, two-way communication devices, or real time price feeds. It may also include practices such as auto DR and optimized market bidding strategies. Furthermore, we need a variety of end-users at specific LMPs that engage in strategies for participation in a larger quantity of more lucrative hours in order to fully capitalize on economic load response.

PJM’s implementation of FERC 745 has created greater participation in economic load response. With market barriers removed, we can now focus on tapping into more value on the demand side. By relieving congestion in ways that other types of participants are incapable of, demand side resources have the capability of reaping high revenues at specific times and specific LMPs. With optimization of the participation profile, economic load response has the potential to create value that no other market mechanism can.

Previous articleMassachusetts solar mixes business sense with public engagement
Next articleStudy: 1,300 demand response programs underway worldwide

No posts to display