Mutual infonnation and non-fixed ANNs for daily peak load forecasting

In this paper, a new method for the daily peak load forecasting which uses mutual information (MI) and non-fixed artificial neural networks (ANNs) is presented. Although ANNs based predictors are more widely used for short-term load forecasting in recent years, there still exist some difficulties in choosing the proper input variables and selecting an appropriate architecture of the networks. Since there are lots of factors that may affect the load series, and the influencing factors are varied in different seasons, a varied structure ANN model including four ANN modules is proposed. The mutual information theory is first briefly introduced and employed to perform input selection and determine the initial weights of ANNs. Then each ANN module is trained using historical daily load and weather data selected to perform the final forecast. To demonstrate the effectiveness of the approach, daily peak load forecasting was performed on the Hang Zhou Electric Power Company in China, and the testing results show that the proposed model is feasible and promising for load forecasting.