Load forecasting based on intelligence information processing

Abstract

In electricity market, it is widely accepted that short-term load forecast is a key problem of market operation. In this paper, a novel model for load forecasting based on intelligence information processing is presented. Here, we make full use of the excellent property reconstruction ability of independent component analysis, which is a new intelligence information processing technology for separating signals and making them independent mutually, and presents STLF model based independent property reconstruction. The load properties of different kinds are restructured to enhance its representation ability and simplifying STLF modeling by ANN. After neural network is trained by new properties with lower dimension, STLF model is built. Finally, the real load data of spot market in New England is applied to demonstrate the validity of the proposed approach