Heterogeneous differential evolution algorithm for parameter estimation of solar photovoltaic models

The behavior of a photovoltaic (PV) system can be derived from its current–voltage characteristics, depending on its unknown circuit model parameters. It is significant to accurately and efficiently extract the parameters of the PV model because of the nonlinear, multivariable, and multimodal characteristics. Many meta-heuristic algorithms have been proposed, in which differential evolution (DE) is known for simple structure, ease of use, and fast convergence. However, the performance of DEs still has room for further improvement because exploration and exploitation are highly coupled during the iterative process for most DE variants. To solve this problem, in this paper, a novel heterogeneous differential evolution algorithm (HDE) is presented to extract the parameters of PV. By proposing a concise heterogeneous mechanism, introducing two new mutation strategies, and establishing an implicit information exchange mechanism, exploration and exploitation processes of the HDE are simultaneously enhanced without one process crippling the other. The efficiency of the HDE has been demonstrated by considering six different PV modules models including the single diode model, the double diode model, the triple diode model, and three different PV modules. The statistical results indicate that the HDE is a representative DE variant in terms of robustness, reliability, solution accuracy, and execution efficiency compared with other 14 state-of-the-art algorithms for the estimation of PV model parameters.

Перевод пока недоступен