Enhanced PSO Update Mechanism for Efficient Single-Diode Modeling of Photovoltaic Generators: Experimental Performance Evaluation

Abstract

Accurate parameter identification of photovoltaic (PV) models is crucial for performance evaluation and realworld system optimization. In this work, we focus on the singlediode model (SDM) and propose an enhanced variant of Particle Swarm Optimization incorporating a Trend-Aware Mechanism (TAM-PSO) to overcome the premature convergence commonly observed with conventional approaches when processing real data. The method is validated using experimentally measured I-V characteristics of the ET-M53630 module, obtained under five real operating conditions (C1-C5) with varying irradiance and temperature. A first analysis presents the parameters extracted by TAM-PSO along with their corresponding RMSE values, highlighting the stability and consistency of the method across all test conditions. A second comparative evaluation between TAM-PSO, standard PSO, and the JAYA algorithm demonstrates that TAMPSO consistently achieves lower error metrics including MAE, MAPE, IAE_avg, and RMSE of 2.69E-3 indicating superior accuracy and robustness in SDM parameter estimation. Finally, the reconstructed I-V curves confirm the capability of TAM-PSO to closely reproduce the real measurements under all scenarios. These results demonstrate the effectiveness of the proposed trendaware update mechanism for reliable and efficient modeling of photovoltaic generators under real operating conditions.