Data Clustering for Optimal Photovoltaic-Distributed Generation Placement in an Active Distribution Network

Abstract

The growing interest in extensive distribution generating unit (DG) integration into the grid creates new challenges for power system planners and operators. The non-optimal placement of renewable energy sources-based (RES) DG units in distribution networks impacts grids’ reliability and stability. This paper proposes a K-means clustering technique for the optimal placement of multiple photovoltaic distributed generation (PVDG) units using voltage stability index (VSI), the voltage risk index (VRI), and the hosting capacity under different PVDG local voltage control techniques. The proposed technique was tested on the IEEE 33-bus radial distribution network. After the optimal placement of the PVDG units, the VSI and VRI for the base network were improved 32.76% and 90.1%, respectively by the best simulation case. The voltage profile was improved by 3.6% and the total active power loss was reduced by 67.7%. The results show that the proposed technique is effective to allocate multiple PVDG units for distribution network planning and expansion as well for increasing the PVDG penetration while minimizing the voltage stability and risk indices.