Power System Load Forecasting Based on Fuzzy Clustering and Gray Target Theory

Abstract

Accuracy of power system load forecasting is affected directly by existing considerable uncertainties, which are accompanied with some correlation among load factors. And this relationship between variables can be eliminated with fuzzy clustering based on historical load in a large number of observations. Carrying out the load factors on the reduction tactics, the impact factors have been classified. Then the target theory is adopted under the standard model of predictors. Then the weight coefficients between the factors category and the predictor are determined in the ways of calculating contributing degrees to various components indicators. For verification, an actual data is provided from a power grid and sorted into four types. Contributing degrees for predicting of the 4 factors are found. It is shown by result analysis that the combined method takes advantages of accuracy and efficiency in prediction.