A Dual-Feedback Adaptive Clone Selection Algorithm With Golden Sine Search for Parameter Identification of Photovoltaic Models

Converting solar energy to electricity efficiently requires an accurate model with well-estimated parameters in a photovoltaic (PV) system. However, identifying parameters for the PV model remains challenging attributed to the nonlinearity and multimodality inherent in diode circuit models. In this paper, we propose a dual-feedback adaptive clone selection algorithm with a golden sine search. The clonal selection algorithm (CSA), known for its inherent diversity maintenance ability, offers an advantage in handling multimodal optimization problems. The critical factor is the cloning operator, particularly the determination of clone size, crucial for balancing exploration and exploitation. To address the conflict between clone size and convergence speed, we introduce a dual feedback adaptive cloning strategy. This strategy considers both the fitness difference between the hypermutated population and the original population and the harmonic mean distance of the population. Additionally, to improve accuracy, we introduce an elite learning strategy based on the golden sine mechanism. The proposed algorithm’s effectiveness is assessed through comprehensive experiments on PV cells and three PV modules, considering both single diode and double diode models, respectively. The experimental results demonstrate that the accuracy of parameter identification achieved by the proposed algorithm exceeds 99.9% for all three types of PV modules. Moreover, we validate the algorithm’s performance using manufacturer’s datasheets with varying irradiations and temperatures. All experiments consistently indicate the stability and effectiveness of the proposed method in accurately identifying the parameters of PV models.