Statistical analyses of electricity use during the hottest and coolest days of summer for groups of residences with and without air-conditioning

Residential air-conditioning demand is a significant contributor to electric utility peaks which typically occur on very hot summer afternoons. Shaving or shifting air-conditioning use to off-peak periods in the day could reduce utility problems associated with high peak demand. As a prerequisite to searching for means of peak reduction, it is necessary to gain a better physical understanding and to acquire quantitative estimates of how whole-house electricity consumption, both daily and hourly, varies for individual as well as groups of residences during the hottest days. The present study specifically deals with these issues by first creating, from a load-experiment data base gathered by a New York-New Jersey utility, four subsets of data on hourly electricity use during a number of the hottest and coolest days in summer corresponding to those residences with and without air-conditioning. Differences in daily total consumption between these groups are investigated. We present and discuss results of regressing diurnal electricity use aggregated group-wise with lagged values of ambient dry-bulb temperature. We also define appropriate stochastic variables to account for day-to-day and resident-to-resident behavior within each group, present the probability distributions of these random variables, and discuss the physical parameters which influence them.