Short-term load forecasting via fuzzy neural network with varied learning rates

Due to the lack of natural resources, the majority of energy in many countries must depend on import, and the corresponding cost is expensive and affected by international market fluctuation and control. In recent years, an intelligent microgrid system composed of renewable energy sources is becoming one of the interesting research topics. The forecasting of short-term loads enables the intelligent micro-grid system to manipulate an optimized loading and unloading control by measuring the electrical supply each hour for achieving the best economical and power efficiency. Therefore, this study investigates a fuzzy neural network (FNN) with varied learning rates for the short-term load forecasting (STLF), and compares its better forecasting performance with a conventional neural network (NN) by numerical simulations of a real case in Taiwan campus.