The economics of commercial demand response for spinning reserve

Abstract

Demand response (DR) for spinning reserve may be appropriate for customers whose operational constraints preclude participation in energy and capacity DR programs. We investigate the private business case of an aggregator providing spinning reserve in California across customer end uses and business segments. Revenues are calculated using end use level hourly load profiles. With average annual revenue of $\sim$$35/kW, steady end uses (e.g., lighting) are more than twice as profitable as seasonal end uses (e.g., cooling) because spinning reserve is needed year-round. Business segments with longer operating hours, such as groceries or lodging, have more revenue potential. Total costs for participation would need to be under $250/kW for many end uses and business segments to have payback periods less than 5 years, which is plausible given equipment cost data from California’s Automated Demand Response programs. Avoided carbon emission damages from using DR instead of fossil fuel generation for spinning reserve could justify incentives for DR resources.

Appendices

Appendix A: DR availability proportional to load

Figure 7 displays DR clearing MW in each hour of the day across 4 seasons for spinning reserve. The Pearson correlation coefficient between the median DR MW cleared in a given hour across all days of 2012–2013 and the median load for that hour of the day was 0.92. DR clearing amounts in the summer appears quite low—this may be due to other more lucrative DR opportunities (such as capacity) during those times.

Fig. 7

Demand response MW cleared in spinning reserve market for each hour of the day in PJM during the period 2012–2013. The pattern of cleared demand response mimics the typical overall load pattern seen in each season. The heavy horizontal line in the middle of each box marks the median. The range of the box represents the interquartile range. The whiskers extend to the extremes of the distribution

Appendix B: Reasoning for removal of projects from cost information in SCE

Figure 8 displays the incentive information from PG&E and SCE.

Fig. 8

a Incentives Provided by PG&E for AutoDR. b Incentives Provided by SCE for AutoDR

The authors conducted an investigation into the AutoDR program costs and found that nearly all of the projects which had incentives of $300/kW in the SCE territory were likely from one contractor that received money from the American Recovery and Reinvestment Act (ARRA) grant funds. We surmise that the use of ARRA funds may have led to different recruitment practices and cost reporting. Thus, we do not believe that the incentive information reported for these projects is representative of the rest of the project population. The list below provides details on why the authors believe that these projects were from one contractor.

• An AutoDR program report stated that “the U.S Department of Energy’s $11.4 million American Recovery and Reinvestment Act grant influenced a larger load shed and enablement cost in the SCE territory” [19].

• ARRA records show a total AutoDR project cost of $22.8M in SCE [42] attributable to one company. The 50% cost sharing required by ARRA leads to a grant of $11.4 million.

• There are 348 facilities in the project incentive database from SCE that had project incentives of $300/kW. These projects have a total load response of 67MW. The total rebate amount given to these participants was just over $20M, which closely matches the ARRA project cost report.

We believe that most, if not all of the projects with incentive values at $300/kW were not representative of the true costs to install, program, and commission this equipment. This is especially apparent when you compare the incentive distribution from SCE with that of PG&E. There may be other projects in the database with incentive costs of less than $300/kW that were implemented by this DR contractor. However, we have no way of differentiating those projects.