Research On Spatial Load Forecast Of Distribution Network Based On Geographic Information System

Spatial load forecast is the basis for automatically programming distribution network. Only after determining the intending load needed in all the power supply area of the distribution network, can it be possible to program the position and capacitence of the transformer substation, the types and paths of the feeders, the installation and use of switches, and other decision-variables. Therefore, it is necessary to forecast not only the quantity of the intending load but also the position where the load increases, namely, the spatial distribution of load.Spatial load forecast covers bulk spatial information. GIS (geographic information system) provides a very wonderful platform for collecing data, processing data, displaying the forecasting result for spatial load forecast. Thus, using the strong capability of extracintg and analyzing spatial data of GIS, this paper collects, extracts and analyzes original data according to the requirements of spatial load forecast. Meanwhile, the fuzzy logical system, as well as the actual mode of patial load forecast of distribution network, is used to set up a base of fuzzy knowledge. Fuzzy logical techniques are applied to carry out fuzzification, inference and defuzzification of the information collected by GIS, and thus the grade of each power supply area can be obtained. Given the fact that the fuzzy sets and the fuzzy rules of fuzzy inference are subject to continuous adjustment in practial use, genetic algorithms are introduced to train the fuzzy rules, which can work very well. Finally, the detailed steps are put forward for the spatial load forecast of distribution network based on GIS: (1) divide the power supply area; into power supply area with the same size by using GIS and grade each power supply area by using fuzzy inference; (2) make the decision on how to use the ground according to the genetic algorithm; (3) calculate the forecasting values for these power supply area.The application in practical engineering shows that the steps brought forth for spatial load forecast of distribution network in this paper is effective and especially suitable for the case of limited resources.