Fuzzy systems application to electric short-term load forecasting. II. Computational results

For pt.I see ibid., p.125-30 (2003). In the first part of this paper, different fuzzy load models are developed for short-term load forecasting. The first model is a harmonic model with fuzzy coefficients. It is a function only of the hour in question, and it is suitable for summer and winter season. The second model is a hybrid model. It is a function of both the hour in question and the temperature at that hour and the previous hours. The coefficients of this model are fuzzy. Both models are implemented to predict the load of a large utility company. The obtained results show that when using such fuzzy models, the variations in the load power can be accounted for, since the estimated load power is within upper and lower limits.