A HYBRID NEURAL NETWORK MODEL FOR DAILY PEAK LOAD FORECASTING USING A NOVEL CLUSTERING APPROACH

This paper proposes a hybrid neural network model for daily electrical peak load forecasting (PLF). Since similar data patterns usually can be seen in peak loads, classification of data would enhance the accuracy of the forecasts. Most classification attempts in literature have been intuitive with no justifications. In this paper, we propose a novel approach for clustering data by using a self organizing map. The Davies-Bouldin validity index was introduced to determine the best clusters. A feedforward neural network (FFNN) for each cluster has been developed to forecast the PLF. Four training algorithms have been used in order to train the proposed FFNNs. The application of the principal component analysis (PCA) reduced the dimensions of the network’s inputs and led to simpler architecture. To evaluate the effectiveness of the proposed hybrid model (PHM), forecasting has been performed by developing a FFNN which uses the unclustered data. The results proved the superiority and effectiveness of the PHM. Linear regression (LR) models have also been developed for PLF, and the results indicated that the PHM produces significantly better forecasts than those of LR models. The results also indicate that the proposed clustering approach significantly improves the forecasting results on regression analysis too.

Load forecasting, Artificial neural networks, Clustering, Self-organizing map