Application of random search method for maximum power point tracking in partially shaded photovoltaic systems

The power-voltage (P-V) curve of a photovoltaic (PV) power generation system under partially shaded conditions (PSCs) is largely non-linear and multimodal, and hence, global optimisation techniques are required for maximum power point tracking. A traditional optimisation algorithm is proposed here, namely random search method (RSM) for tracking the global maximum power point in a solar power system under PSC. The RSM is based on the use of random numbers in finding the global optima and is a gradient independent method. The major advantage of RSM is its very simple computational steps, which requires very less memory. The performance of RSM in tracking the peak power is studied for a variety of shading patterns and the tracking performance is compared with two-stage perturb and observe (P&O) and population-based particle swarm optimisation (PSO) methods. The simulation results strongly suggest that the RSM is far superior to two-stage P& O method and better than PSO method. Experimental results obtained from a 120-watt prototype PV system validate the effectiveness of the proposed scheme.