A short-term load forecasting expert system

This paper describes a new practical knowledge-based expert system (called LoFy) for short-term load forecasting equipped with graphical user interfaces. This system uses AI and other computing techniques. Various visual objects like calendar, chart, grid and dialog box have been included to increase the facility of interaction. Also, various forecasting models like trending, multiple regression, artificial neural networks, a fuzzy rule-based model and the relative coefficient model have been included to increase the forecasting accuracy. The simulation based on historical sample data shows that the forecasting accuracy is improved when compared to the results from the conventional methods. Through the fuzzy rule-based approach, the forecasting accuracy has improved remarkably.