一种基于种群分割的多变异差分进化算法用于光伏模型参数提取

Abstract Accurate modeling of photovoltaic (PV) cells is essential for the simulation, evaluation, control, and optimization of PV systems. Among PV models, the diode circuit model and the Sandia Array Performance Model are more extensively used. However, these models exhibit nonlinear, multi-modal, and multi-variable features, making it challenging to accurately obtain model parameters. To address this issue, this study proposed a population segmentation-based multi-mutation differential evolution algorithm (PSMDE) for parameters extraction. In this algorithm, the proposed segmentation method dynamically divided the population into three subpopulations, each updated with different mutation strategies to effectively tackle complex problems. Additionally, a parameter adaptation strategy was introduced to balance the algorithm’s exploration and exploitation capabilities, and a restart strategy was implemented to enhance the algorithm’s ability to escape local optima. To evaluate the performance of PSMDE, the algorithm was used to extract parameters for various types of PV models and compared against 8 other state-of-the-art algorithms. The results demonstrated that PSMDE outperforms other algorithms in terms of accuracy, stability, and convergence speed for the parameter extraction of PV models. Furthermore, the static and dynamic experimental tests were conducted to verify the reliability of this parameter extraction method for calculating the module’s output power under various temperature and irradiance conditions, as well as real weather conditions. The experimental and simulated results showed good agreement. Specifically, the daily mean relative errors between the simulated and dynamic experiment results ranged from 3.74% to 5.91% during the measurement period.