Short term load forecasting using fuzzy model of the orthogonal least square method

An algorithm for short term load forecasting based on a fuzzy model of the orthogonal least square method (OLS) is presented. The generation of fuzzy rules, and the selection of the model inputs and membership functions are all obtained from historical data. Therefore, the bottle-neck difficulty in knowledge acquaintance by questionnaire and experience is overcome. In the mathematic model, identification of the premise part and consequent part is separately accomplished. In the practical examples, the accuracy of this method is higher than that of the RBF method by 0.3%-0.5%. With this method, the computing time is only one second for every model.