Static photovoltaic models’ parameter extraction using reinforcement learning strategy adapted local gradient Nelder-Mead Runge Kutta method

Chen, Zhiqing; Kuang, Fangjun; Yu, Sudan; Cai, Zhennao; Chen, Huiling

doi:10.1007/s10489-023-04761-8

Cited by 5 publications

(1 citation statement)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There has been much research in the solar PV parameter identification field. Commonly used are traditional methods such as pattern search optimization (AlRashidi et al 2011), formula resolution (Chen et al 2023), and numerical analysis (Farivar and Asaei 2011). However, these conventional methods are often subject to various limitations and may not adequately meet the complex and changing needs of practical applications.…”

Section: Introductionmentioning

confidence: 99%

Mutated opposition-based learning and Nelder-Mead simplex enhanced grey wolf optimizer for photovoltaic model parameter estimation

Wu,

Chen,

Cai

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Accurate assessment of solar cell conversion efficiency relies heavily on the precise determination of unobserved parameters within photovoltaic (PV) cell models. To address this need, we present a novel approach named MOLNMGWO in this paper. This method aims to streamline the process of identifying latent parameters in PV models. MOLNMGWO ingeniously combines the Nelder-Mead simplex method with a mutated opposition-based learning mechanism derived from the gray wolf optimizer (GWO).Gaussian-Cauchy mutation in mutated opposition-based learning mechanism provides the ability to jump out of the local optimum. The opposition-based learning ensures that the local exploitation ability of the algorithm guarantees that it always obtains a better solution. In addition, the introduced Nelder-Mead simplex method also enhances its capability for local search and convergence precision. Evaluating the efficacy of MOLNMGWO involved a comprehensive comparison with various cutting-edge algorithms using the CEC2014 function set. The findings underscore the remarkable superiority of MOLNMGWO over these diverse algorithms. Subsequently, the application of MOLNMGWO to discern enigmatic PV parameters in the context of the single-diode model (SDM), double-diode model (DDM), and three-diode model (TDM) was pursued under consistent temperature and illumination circumstances.The MOLNMGWO can identify these parameters accurately and reliably. Finally, MOLNMGWO was used to identify the three commercial PV module parameters under diverse temperature and light conditions, and satisfactory results were obtained. All the above three experimental results show that MOLNMGWO has excellent competitiveness and that MOLNMGWO can accurately and efficiently identify photovoltaic parameters even in harsh environments. Overall, the MOLNMGWO is a novel, reliable and effective method to overcome the problem of the estimation of parameters for solar cell PV models.

show abstract

Section: Introductionmentioning

confidence: 99%