Clustering-based optimal sizing and siting of photovoltaic power plant in distribution network

This article focuses on sitting and sizing of PV units aiming at minimizing the total power loss in distribution network. Due to fluctuation in output power of photovoltaic and variable nature of the loads in electricity network, for evaluating power loss and the bus voltages, the grid should be simulated for all days of the year and all hours of the day. However, this simulation is time consuming and costly. Consequently the clustering method is employed to find some days as representatives and simulate the grid for these days rather than all days of the year. If the representatives indicate the nature of whole data very well, the error of simulating the grid in this case is negligible comparing with simulating it for all days. In order to assess the performance of the clustering methods and finding the optimal number of clusters, three clustering validity indices are used. The 33 buses test system and irradiation profile of Tehran in a year are considered. Various types of loads are assigned to loads in selected network. The network is simulated for all days and its result is compared with clustering-based method.