Estimating residential air conditioner loads using consumer survey information

In planning demand-side strategies, electric utilities are vitally interested in the effects of weather-sensitive end uses on system load. Residential air conditioners, in particular, can be major contributors to peak loads. However, quantitative information weather-related end-use loads difficult and expensive to obtain. The paper uses billing records, available weather data and survey information to offer an inexpensive alternative to household monitoring. The average amount of energy that is used by air conditioners during some time period is estimated by multivariate regressions explaining consumers demands as functions of the energy-using equipment that they own, prevailing weather conditions, their demographic and socioeconomic characteristics, prices of energy and the characteristics of their dwellings. The correlations between air conditioner load and weather can then be measured. The approach can be used as a weather normalization technique for the summer months to adjust energy consumption to account for weather variations