LBNL Report Number
The primary goals of this scoping study were to
- summarize existing methods for estimating demand response,
- evaluate these methods' abilities to accurately estimate residential demand response for the purpose of program evaluation,
- recommend a preferred approach, and
- outline any remaining knowledge gaps. This study was motivated by the CPUC directive (D.05-11-009) of developing measurement and evaluation protocols for demand response.
Our evaluation considers both day-matching and regression techniques, outlining the following alternative methods:
- prior-day averaging,
- weather-matching techniques,
- regression-based load profile comparison, and
- econometric demand analysis.
Based on a review of these methods for evaluating demand response, we find that customer-specific regression analysis is likely to give accurate, transparent and intuitive results. Depending on program requirements, this method can be modified to estimate hourly demand response before, during and after events, providing hourly kW response results and load profiles.
Beyond basic demand response estimation, several issues need to be addressed before a practical method for residential demand response program evaluation can be determined. Among them are the ability to evaluate multiple events on consecutive days, an understanding of how advance notification affects demand response, and incorporation of considerations affecting the extrapolation of results from a voluntary pilot to a large-scale program.