A novel hybrid approach combining Differentiated Creative Search with adaptive refinement for photovoltaic parameter extraction

Abstract

Accurate parameter extraction from Photovoltaic (PV) models using current-voltage (I-V) data is essential for optimizing and simulating photovoltaic systems. Despite the existence of various techniques, many face challenges in achieving a balance between precision, robustness, computational efficiency, and execution time. In this paper, we present a novel hybrid algorithm, Differentiated Creative Search combined with Newton-Raphson (DCS-NR), designed to improve the accuracy and efficiency of PV parameter extraction. DCS employs a dual-strategy mechanism that balances exploration and exploitation through divergent and convergent thinking, ensuring a comprehensive search for solutions. The Newton-Raphson method further refines the parameters optimized by DCS, minimizing the discrepancy between estimated and measured currents, and consequently improving power estimation. The proposed approach is evaluated on three distinct models: Single Diode Model (SDM), Double Diode Model (DDM), and PV Module Model (PMM). Among the different models tested, DCS-NR consistently delivers superior accuracy. For example, it achieves an RMSE of 7.75392 × E−04 for the RTC France SDM and 1.77454 × E−04 for the PVM 752 cell, outperforming ten state-of-the-art metaheuristic algorithms. Moreover, DCS-NR demonstrates remarkable computational efficiency, requiring only 0.830 s on average for the RTC France SDM, which is considerably faster than algorithms such as Flying Foxes Optimization (251.5 s). Furthermore, it proves highly effective in real-world conditions, under varying irradiance and constant temperature, as well as vice versa. The method consistently converges within approximately 100 iterations, showcasing rapid optimization capabilities. These findings highlight the potential of DCS-NR as a powerful and versatile tool for photovoltaic parameter extraction, capable of addressing diverse and challenging scenarios.