Electrical loads can be forecasted for distribution planning. [Time-series and multivariate spatial analyses]

Accurate forecasts of electrical demand in small geographic areas are needed if a utility is to develop comprehensive plans for expanding its transmission and distribution systems. The limiting item in using optimization methods is the value to be used for future small-area demands. This paper describes two techniques for developing small-area demand forecasts that show considerable improvement over existing techniques. The first technique, time-series trend analysis, describes small-area demand as a time function based solely on historical data. Forecasts are obtained from projections of the historical trend. The second technique, multivariate spatial analysis, describes small-area demands as a function of both historical data and correlative variables. Forecasts are derived from estimates of future values for the appropriate correlative variables.