Comparative analysis of different computational intelligence techniques for maximum power point tracking of PV systems

Abstract

The performance of a photovoltaic (PV) module can be improved by employing maximum power point tracking (MPPT) controllers. MPPT controllers are algorithms that are included in PV battery charge controllers or inverters to extract the maximum available power from PV modules for any given temperature and irradiance. Several studies report that the use of PV modules without MPPT controllers results in power losses, which ultimately results in the need to install more solar panels for the same power requirement. Numerous techniques of varying complexities have been proposed in the literature to solve the MPPT objective function. This paper presents a comparative analysis of three computational intelligence (CI) based MPPT techniques namely, the fuzzy logic (FL) based controller, artificial neural networks (ANN) based controller, adaptive neuro-fuzzy inference system (ANFIS) based controller and one conventional technique, the perturbation and observation (P&O) controller. These MPPT controllers are designed, simulated and analysed in the MATLAB/Simulink environment. The performance of the studied MPPT techniques is evaluated under steady-state weather conditions, rapidly changing weather conditions and varying load conditions. CI-based MPPT controllers are found to be more efficient than the P&O controller. Moreover, the ANFIS-based MPPT controller shows an outstanding MPPT performance for all the scenarios studied.