Differential evolution algorithm featuring novel mutation combined with Newton-Raphson method for enhanced photovoltaic parameter extraction

Abstract

Accurate parameter extraction in photovoltaic (PV) cells and modules is crucial for optimizing performance, modeling, and predicting behavior under varying environmental conditions. In this context, we propose a novel hybrid algorithm, Mean Differential Evolution with Newton-Raphson (MDE-NR), which combines the strengths of Mean Differential Evolution (MDE) and the Newton-Raphson (NR) method to enhance the precision of parameter extraction. MDE, recognized for its ability to balance exploration and exploitation, employs an innovative mean-based mutation strategy that reduces the risk of premature convergence. However, while MDE effectively performs a global search, achieving the lowest possible error often requires further refinement. This is where the NR method comes into play, offering fast local convergence by using the optimal parameters generated by MDE as initial guesses. The combination of these two methods in MDE-NR significantly reduces the Root Mean Square Error (RMSE) in the final estimation. The effectiveness of the MDE-NR algorithm is validated through comprehensive comparisons with well-known metaheuristic algorithms across Single Diode Model (SDM), Double Diode Model (DDM), and Photovoltaic Module Model (PMM), achieving minimal RMSE values with standard deviations as low as 10E-19 to 10E-21 over 30 runs, far superior to those of 10 other metaheuristic algorithms. The algorithm demonstrates rapid convergence and outperforms its counterparts in computational efficiency. Moreover, MDE-NR effectively handles varying environmental conditions, such as constant irradiation with variable temperature and vice versa, achieving highly accurate results across different PV technologies. This hybrid approach establishes MDE-NR as a highly effective and reliable tool for the precise extraction of PV parameters, providing significant improvements in both accuracy and computational efficiency.