Gray-Regression Variable Weight Combination Model for Load Forecasting

Abstract

A gray model and regression model based middle and long term load forecasting method using variable weight combination model is proposed. In view of the shortcomings of grey prediction model is not very suitable for middle and long term load forecasting, the equivalent dimensions additional data processing technology is adopted to build the equivalent dimensions additional grey model to improve the model. At the same time, there are some characteristic within mid-long term load forecasting such as the long study time span, the complex factors with large uncertainty which have great influence on load forecasting, and the possible original error occurring in basic data of forecasting, the time-varying weight combinational prediction method is adopted to overcome the shortcomings of the fixed weight, it is more practical. The example results show that this model is applicable in the long-term load forecasting, and it has a high forecasting accuracy.