A Fuzzy Inference Approach to Short-term Load Forecasting in Power Systems

Abstract

This paper proposes an optimal fuzzy inference approach to short-term load forecasting. In recent years, artificial neural net based approaches to short-term load forecasting seem to be promising for universal approximation of any nonlinear functions. However, they still have drawbacks that they do not clarify the relationship between input and output variables due to a black-box description, and that they do not necessarily provide satisfactory predicted values due to overfitting and insufficient data normalization. The proposed method constructs an optimal structure of the simplified fuzzy inference that minimizes model errors and the number of the membership functions to grasp nonlinear behavior of power system short-term loads. The optimal model is identified by simulated annealing and the steepest descent method. The advantage of the model allows us to understand the relationship between input and output variables and provide better predicted values. The effectiveness of the proposed approach is demonstrated in examples.