Statistically adjusted engineering (SAE) models of end-use load curves

Abstract

We develop and demonstrate models that combine engineering and statistical approaches to estimating customer-specific end-use load curves. Simulated end-use loads from engineering methods enter as explanatory variables in statistical models, and estimated parameters adjust the engineering loads on the basis of customers' observed loads. The resulting end-use loads, called statistically adjusted engineering (SAE) loads, depend on a variety of conditioning variables, including weather and the size and type of the customer's dwelling (which enter the engineering simulations) and the income and other characteristics of the household (which enter the statistical adjustment). Using data from a Los Angeles sample of households, several SAE models are estimated that differ in the flexibility that they allow in the adjustment of the engineering loads.