Maximum Power Point Tracking of PV System under Uniform Irradiance and Partial Shading Conditions using Machine Learning Algorithm Trained by Sailfish Optimizer

摘要

Solar energy is a viable solution to the damage caused by the conventional power sources to the environment. Temperature and irradiance levels have a high impact on the power generation of photovoltaic modules, but due to non-uniform irradiance levels, PV modules generate non-linear P-V curves. Maximum power point tracking control is introduced to harvest maximum power from PV modules. In this paper, a general regression neural network trained with sailfish optimizer (GRNN-SFO), a hybrid MPPT technique is presented. Highly effective global optimization of sailfish optimizer combined with precise estimation capability of the general regression neural network makes GRNN-SFO highly effective for MPPT control. Comparison is made with GRNN-PSO and GRNN-P&O to check the performance of the proposed technique. Two cases are presented in order to validate the superior performance of GRNN-SFO. The comparison shows that GRNN-SFO tracks the global maxima with greater than 99.9% efficiency and 12 ms faster tracking time under fast varying irradiance and partial shading condition. The analysis of statistical data has also been exhibited in order to examine the robustness and responsiveness of the proposed technique.