A Simple Yet Fully Adaptive PSO Algorithm for Global Peak Tracking of Photovoltaic Array Under Partial Shading Conditions
Saba Javed, Kashif Ishaque, Shoaib Ahmed Siddique, Zainal Salam
Abstract
This article proposes a simple yet fully adaptive particle swarm optimization (PSO) algorithm to find the global peak (GP) of a photovoltaic array under partial shading condition. It exploits a constrained optimization approach having a very simple PSO structure with two adaptive parameters. The first one that controls the magnitude of velocity is adaptively varied according to the absolute distance of each particle with respect to best particle's position. The other parameter controls the search space, has been varied based on a penalty condition that decides the participation of particle in the next iteration. Regardless of a wide range of population size, the proposed scheme precisely locates the GP without jeopardizing the tracking speed—consumes a maximum of 16 perturbations. The algorithm is implemented on a Cuk converter and compared to two well-known tracking methods. It is also validated through run length distribution (RLD) test. The obtained RLD results reveal that proposed PSO outperforms other two methods, in terms of convergence speed and success rate. It always obtains 100% rate by utilizing fewer perturbations of voltage. When tested for a whole day environmental profile, the average efficiency of proposed algorithm is found to be 99.65%.