A development of real-time load forecasting system based on chaos theory

Abstract

Accurate performance of short-term load forecasting is indispensable in order to avoid undesirable disturbances in power system operations. To this end, the authors have developed forecasting methods of this type based on conventional chaos theory. However, this approach is unable to provide accurate forecasts in cases of anomalous loads where the loads consecutively exceed the historical maximum or are lower than the minimum. This paper presents an improved forecasting method based on chaos theory and a real-time forecasting system. In particular, the potential of the local fuzzy reconstruction method is fully exploited in order to achieve as accurate a forecast as possible. The effectiveness of the proposed method and the real-time load forecasting system is verified by the actual load at the Chugoku electric power Co.