An interactive distribution load forecasting methodology for minicomputer use based upon a Markov-type process

An efficient comprehensive methodology is described that allows interactive forecasting of distribution-substation load growth and load sharing, with application to the determination of an economically feasible timing schedule for adding substation capacity. A BASIC-language implementation illustrating some of its capabilities is discussed. Load growth is expressed by a Gompertz model using parameters evaluated from historical data; however, other load-growth models can be easily implemented. Load transfer, which can be set to begin at any percentage of the substation capacity, is accomplished stochastically in the algorithm by means of a quasi-Markov process. Default load transfer threshold, load transfer flags, and other parameters are under direct control of the user and can be set to conform to individual base-case and historical data. The implementation of this algorithm for minicomputer use as an interactive planning and instructional tool is discussed, and an example is presented. >