Design and Simulation of Neuro-Fuzzy Based MPPT Controller for PV Power System

This paper introduces a full simulation and design of a photovoltaic power generation system to supply Al-Mahmoudia hospital in Baghdad/Iraq. The design steps include; daily average load estimation, annual solar irradiance and temperature monitoring and logging. PV panel selection and datasheet study to design the PV array and decide the number and connection method for these panels. (DC-to-DC) boost converter is designed. Then design and simulate a new type of (MPPT) controller based on mixed technique of artificial neural networks (ANN) with fuzzy Logic Controller (FLC) or (neuro-fuzzy) technique. The power generated from photovoltaic (PV) panels rely on the sun radiation and the ambient temperature. For this reason, the use of (MPPT) controller is a necessity to get as much as possible of the photovoltaic power from the system despite of variation in climatic conditions. The most famous method used for (MPPT) control is the Perturb and Observe (P&O) algorithm. Despite of the simplicity of this method, it is suffering from many problems including the oscillation around the (MPP), delay in system response and the maximum overshoot especially when the perturbation step size is large. The proposed (neuro-fuzzy) controller has been designed to get rid of the drawbacks of the classical (MPPT) methods including the P&O method. MATLAB/Simulink software has been used to design a photovoltaic module and array to supply the load of the hospital with 459kW. In addition the (DC-to-DC) boost converter and (MPPT) neuro-fuzzy controller are modeled. The performance has been compared between the proposed controller with the classical P&O and fuzzy logic controllers. The simulation results illustrated an enhanced performance in terms of oscillation, the response time, power harvesting and highest power overshoot.