Development of a photovoltaic MPPT control based on neural network

The Maximum Power Point Tracking (MPPT) is an important factor to increase the efficiency of the solar photovoltaic (PV) system. This paper presents a solar PV system containing a solar PV array, a DC/DC boost converter and a load. Different MPPT algorithms have been established with their features. The conventional algorithms (Perturb and Observe, Incremental Conductance and Open Circuit Voltage) show a lot of drawbacks. The major issue is the tracking of the Maximum Power Point (MPP) when environmental conditions change faster. So, a MPPT technique based on Neural Network (NN) was developed and which can enhance the efficiency and gathers the advantages of a lot of techniques. A multi layer neural network with back-propagation algorithm is used in order to have a small Mean Squared Error (MSE). The inputs of NN are irradiance, temperature and the output is the duty cycle that controls the boost converter. Finally, it is discussed the results and made comparison in terms of performance of the different algorithms, covering the overshoot, time response, oscillation and stability.