Enhancement of Maximum Power Extraction from a PV-Based Residential DC Microgrid under Partial Shading Conditions using Grey Wolf Algorithm

Abstract

In recent years, the whole world is focusing on the utilization of solar energy because of the depletion of fossil fuels and its impact. However, PV system with multiple modules is exposed to non-uniform irradiance due to long building, trees, cloud, flying birds and etc. As the result, when certain section or parts PV modules receive different levels of sunlight, Partial Shading Condition (PSC) formed in the system. This leads to overall power loss and multiple peaks in Power-Voltage (P-V) characteristics curve. Multiple peaks in the P-V curve resulted when unshaded and shaded modules operate at different voltage during shadowing, which gives difficulty to extract maximum power from the PV system. In this condition, conventional maximum power point tracking (MPPT) algorithms like Perturb and Observe (P&O), may fail to track the global peak of P-V curve. To extract the maximum power generated by a PV device in this complex condition, improved MPPT approach needs to be applied. Thus, this thesis proposed Grey Wolf Optimization (GWO) algorithm to extract maximum available power under PSC, for a residential DC load micro grid. The GWO algorithm is utilized to find the best duty cycle for DC-DC boost converter in the PV system to adjust the maximum operating point among local and global peaks of PV curve. It inspired from the social hierarchy and hunting habits of grey wolves. In this study, centralized (single) MPPT controller configuration has been used to process power of the PV panels. Mathematical modeling of minimum load with 105W up to 4300W peak load for residential DC microgrid, solar energy analysis depending on meteorological data, boost converter design and formulation of GWO and P&O techniques has been done. Performance of P&O algorithm and proposed algorithm is compared under four Shading Patterns by using Matlab/Simulink 2018a. The findings confirmed that the proposed system operated with less output power oscillation, faster tracking time, and higher efficiency than traditional P&O algorithm.