Combining Engineering and Statistical Approaches to Estimate End-Use Load Shapes Volume 2:Methodology and Results

摘要

Previous methods for estimating load curves for individual appliances include (1) engineering methods that incorporate data on weather and building characteristics along with known principles of thermodynamics but do not reflect customers' behav¬ioral responses or differences in these responses due to differences in income and other characteristics, and (2) statistical methods that reflect behavioral responses but do not incorporate a priori engineering information about appliances. In this report we develop and demonstrate methods for estimating end-use load curves that combine both the engineering and statistical approaches. These methods are applied on data from three areas (Los Angeles, Wisconsin, and Southern California Edison's service territory) and found to be capable of accurately capturing the effect of the major conditioning factors, namely, weather, building characteristics, and customer characteristics on appliance-specific loads.