Evaluation of different methods for solar cells/modules parameters extraction

摘要

Reliable modelling techniques are required to optimise the performance of photovoltaic (PV) systems. This paper considers the implicit form of the single diode model for solar cell to introduce and compare four new approaches for extracting the five characteristics parameters of PV cells/modules. Measured I-V curves for three different PV samples (mono-crystalline cell, polycrystalline module and amorphous module) under different irradiance conditions and temperatures were used to implement and validate each method. Numerical and optimisation methods were applied to solve different formulations of the single diode equivalent model. All methods are mathematically described, and the primary considerations when writing the MATLAB codes are highlighted. Comparison factors such as the mean absolute percentage error (MAPE), the algorithm convergence time (ACT), the coefficient of determination (R-2) and the error at the maximum power point (delta), are used to evaluate each method. All proposed methods were verified, and the parameters were extracted with excellent performance. The method based on the least square fitting and solved by Newton-Raphson method provides the best results for the mono-crystalline sample compared to the other three methods while using particle swarm optimisation as a solving algorithm provides better results for the other two sample modules. The result of this work represents a guideline for extracting the five parameters of the single diode model from the measured solar cells/modules I-V curves. Although the SDM model does not describe the physical structure of the amorphous solar module, since this consists of a tandem p-i-n diode structure, the developed methods still deliver results with errors comparable to that of the single junction PV devices.