Clustering load distribution substation based on similarity of load curves using statistic-fuzzy methods

Having accurate information of load is one of the key points of the proper operation and investment in distribution networks. Identification and classification of substation loads and consumer services is the first step in: estimation and reconstruction of load distribution substations, proper operation, load forecasting, comprehensive plans and other studies. So far, different methods based on neural networks and also fuzzy methods have been presented for this aim. In this study, a statistic-fuzzy method is used for clustering substations loads. In this method, recorded load curves in statistical history are used for generation membership function of fuzzy relationship matrix. After inner multiplication of fuzzy matrix, nearness degree coefficient of load curves which it shows the similarity of curves is calculated. At the end, this method is used in a real geographical area (Bandarabbas) and the results are presented.