Energy Reports (Nov 2021)
Supply demand optimization algorithm for parameter extraction of various solar cell models
Abstract
This article illustrates a new application of the Supply-Demand-Based Optimization (SDO) algorithm to accurately extract the electrical parameters of different PV models. The key goal is to minimize the total error between the experimental data and the proposed approach by optimizing the electrical parameter of three different models. The SDO simulates the stability and instability modes of the supply–demand mechanism in market economy, where the quantity and price converge and diverge from the equilibrium point, respectively. Three different marketable PV modules were provided, and the findings correlate with the experimental data and other well-known optimization techniques denoting the superiority of SDO. These modules are STM6_40_36, STP6_120_36, and Photowatt-PWP 201 are adopted. The validation is introduced, and the findings are correlated with experimental data and other recently well-known optimization techniques which are grey wolf optimization (GWO), Crow search Optimizer (CSO), Bernstein–Levy Search Differential Evolution Algorithm (BSDE), and Manta Ray Foraging Optimizer (MRFO), Backtracking Search Algorithm (BSA). In the article a deep analysis has been carried out for the best parameter extraction in a PV module. An online parameter extraction is extended based on the SDO algorithm via SDM under different sunshine irradiation of 200, 400, 600, 800 and 1000 W/m2 and temperature of 25, 50 and 75 °C. The standard deviations of the fitness values, over 30 runs, for three models for the above-mentioned modules are less than 1 × 10−18, 10−17 and 10−6, respectively which denotes the superiority of the SDO. Also, a unique set of the extracted parameters is provided with significant validations of the proposed SDO using measured I–V curves for different irradiance and temperature values. Also, the obtained results were compared to different algorithms with a detailed statistic characterizations for convergence, achieving good fitting correlations. Therefore, the SDO are extremely consistent, competitive one among other algorithms.
