Boosted backtracking search optimization with information exchange for photovoltaic system evaluation

• Published in: Energy science & engineering 2023-01, Vol.11 (1), p.267-298
• Main Authors: Weng, Xuemeng, Liu, Yun, Heidari, Ali Asghar, Cai, Zhennao, Lin, Haiping, Chen, Huiling, Liang, Guoxi, Alsufyani, Abdulmajeed, Bourouis, Sami
• Format: Article
• Language: English
• Subjects: Algorithms; Design; Diodes; Efficiency; Electrical measurement; Energy sources; Environment models; Environmental conditions; Evaluation; Feature selection; Genetic algorithms; Machine learning; metaheuristics; Neural networks; Optimization; Optimization algorithms; Parameter estimation; parameter extraction; Parameter identification; Parameters; Photovoltaic cells; photovoltaic models; Photovoltaics; Power consumption; R&D; Research & development; Simulation; solar cell; Solar energy; swarm intelligence; Traveling salesman problem
• ISSN: 2050-0505; 2050-0505
• DOI: 10.1002/ese3.1329

The determination of photovoltaic (PV) parameters is of great importance for the reliability of solar system operation, continuity of the load power consumption, and control management of the energy source. Therefore, this study proposes an advanced backtracking search optimization algorithm (BSA) equipped with teaching and learning‐based optimization (TLBO), named TLBOBSA, to accurately simulate the PV model. During the evaluation of the proposed algorithm, the concept of teaching from TLBO is introduced into the BSA to guide optimal individuals, thus improving the convergence rate of the algorithm. The learning behavior among individuals in the student phase of TLBO facilitates interindividual learning and provides beneficial information for its evolution, which is introduced into the BSA to ensure the diversity of the population. The comprehensive test results of different PV module models in different environmental conditions show that the proposed algorithm is more advantageous for parameter extraction than other existing algorithms. This can be seen in the simulation experiments of two commercial PV models, where the simulated current is consistent with the measured current at each measured voltage. This demonstrates that the proposed TLBOBSA is an accurate and reliable tool for evaluating unknown parameters of PV models. (1) An improved backtracking search algorithm with teaching and learning‐based optimization (TLBOBSA) is proposed to extract parameters of the photovoltaic system. (2) The performance of TLBOBSA is compared with some well‐known competitive algorithms. (3) TLBOBSA was evaluated under different irradiance levels and temperature levels. (4) TLBOBSA has improved the convergence speed and obtained optimal accuracy among all competitive algorithms.