Low-Voltage Microgrid Planning Strategies for an Isolated Village — A Case Study in Cambodia

Abstract

The purpose of this paper is to propose three different solutions to handle the issue of excess electricity to the MV grid from LV microgrid with PV. These scenarios are compared based on their TOTEX (total costs) calculated for 30 years. Firstly, the Mixed Integer Linear Programming (MILP) approach is used to allocate power to each household load across all three phases (phase A, B, C) of the system. Secondly, the network is built using the Shortest Path (SP) algorithm to establish radial topology by defining all line connections. Thirdly, a Genetic Algorithm (GA) is used to determine the size and location of the photovoltaic (PV) and decentralized battery energy storage (DeBES). After having simulated the LV microgrid for 30 years, the power at the slack bus is analyzed to identify the occurrence of excess electricity when the PVgenerated power exceeds the load consumptions. To handle this issue, three different scenarios are proposed: selling the reversed energy back to the MV grid, MV grid with centralized battery energy storage (CeBES), and LV generator with CeBES. For the case study, a non-electrified Cambodian village with 73 households is selected. Based on the simulation results, in terms of the smallest TOTEX, scenario 3 (LV generator with CeBES) is the most interesting solution compared to the other two scenarios.