ew 18: e3

Research Article

On the effects of hyper-parameters adjustments to the PSO-GMPPT algorithm for a photovoltaic system under partial shading conditions

Download52 downloads
  • @ARTICLE{10.4108/eai.13-7-2018.160981,
        author={Andr\^{e}  Luiz Marques  Leopoldino and Cleiton  Magalh\"{a}es  Freitas and Lu\^{\i}s  Fernando Corr\"{e}a  Monteiro},
        title={On the effects of hyper-parameters adjustments to the PSO-GMPPT algorithm for a photovoltaic system under partial shading conditions},
        journal={EAI Endorsed Transactions on Energy Web: Online First},
        volume={},
        number={},
        publisher={EAI},
        journal_a={EW},
        year={2019},
        month={10},
        keywords={Photovoltaic Energy Generation, Maximum Power Point Tracking, Particle Swarm Optimization},
        doi={10.4108/eai.13-7-2018.160981}
    }
    
  • André Luiz Marques Leopoldino
    Cleiton Magalhães Freitas
    Luís Fernando Corrêa Monteiro
    Year: 2019
    On the effects of hyper-parameters adjustments to the PSO-GMPPT algorithm for a photovoltaic system under partial shading conditions
    EW
    EAI
    DOI: 10.4108/eai.13-7-2018.160981
André Luiz Marques Leopoldino1, Cleiton Magalhães Freitas1,*, Luís Fernando Corrêa Monteiro1
  • 1: Rio de Janeiro State University, Campus Maracanã, Zip Code 20559-900, Rio de Janeiro-RJ, Brazil
*Contact email: cleiton.freitas@uerj.br

Abstract

This paper exploits the performance of the particle swarm optimization (PSO) algorithm for a photovoltaic system under partial shading condition (PSC). Essentially our main contribution consists on analyzing the hyper-parameters adjustment of the PSO algorithm to determine the minimum particle numbers, such that the assertiveness to identify the Global Maximum Power Point (GMPP) be higher than 99%. The database was obtained throughout 5760 simulations based on different test cases. From these test cases, the PSC was applied in 2880 simulations. In the previous work, it was shown the best results based on 5 particles. In this update version, it is also shown the best results for 3, 7 and 9 particles, together with a comparison among them. Furthermore, this paper also presents the simulation results to evaluate the performance of the developed algorithm under transient- and steady-state conditions.