An Electricity Price Zone Forecasting Method with Fuzzy Inference Net

This paper proposes a Fuzzy Inference Net (FIN) method for electricity price zone forecasting. Under smart grid environment, it is important for players to maximize profits and minimize risks though power markets while introducing renewable energy into grids. The time series of electricity price becomes more complicated due to the nonlinearity and uncertainties. To capture the behavior of the time series appropriately, more sophisticated methods are required to overcome them as a prediction tool. In this paper, a new method is proposed for price zone forecasting. The proposed method makes use of FIN that evaluates the association probability of unknown data to predetermined clusters with fuzzy inference and self-organization. The selection of input variables is determined by the variable importance of the CART algorithm of data mining. The association probability is used determine which zone the one-step ahead electricity price belong to. The proposed method is tested for real data in comparison with the conventional artificial neural network.