Short-term load forecasting with artificial neural network and fuzzy logic

Short-term load forecasting, which forecasts the electric load of the future day or week, is basis not only for power generation and operation, but also for power market contract. Short-term load forecasting has become an essential part of modern control centers. In the past three decades, many forecasting models have been presented. This paper proposes a new model, which divides the electric load into two parts: the load scaled curve and the day maximal load and minimal load. The load scaled curve is forecasted using five artificial neuron networks. The day maximal load and minimal load are forecasted using fuzzy logic. In this model, weather condition, seasonal index, Test shows the proposed method in this paper can improve the forecasting accuracy.