Analysis of Electric Utility Load Forecasting
RCED-83-170, Jun 22, 1983
In response to a congressional request, GAO provided information on electric utilities' load forecasting methodologies, the state-of-the-art in forecasting, key variables that are considered, and how energy conservation and cogeneration are addressed. It also identified weaknesses in current load forecasting procedures.
GAO found that utilities use a variety of forecasting methodologies. Many small utilities rely on trend analysis which predicts future power demand by assuming that the factors that influenced power consumption will remain stable. Large utilities use more complex methods such as econometric and end-use forecasting. Econometric forecasting also assumes that the variables affecting electricity consumption will remain constant, while end-use forecasting breaks electricity consumption into user sectors and needs an extensive inventory of data. Data collection for end-use forecasting is expensive and time consuming; however, it readily reflects changes in the factors that influence consumption. A sum-of-the-utilities forecast is sometimes used to give a regional or national perspective on utility demand. According to utility officials, the forecasting state-of-the-art involves several techniques to mitigate and minimize the uncertainty inherent in forecasting. By using a range of forecasts, the effects of changes in key assumptions and variables can be readily observed. GAO found that all forecasting methodologies consider the population of the service area, the income of its customers, the price of electricity, economic growth, and conservation measures. For the most part, utilities do not factor conservation into their forecasts.