A new study has evaluated the accuracy of mathematical models used in the PV industry and research to evaluate the performance of solar modules under varying environmental conditions such as irradiance and temperature.
The findings are described in the paper Evaluation of solar module equivalent models under real operating conditions—A review, published in the Journal of Renewable and Sustainable Energy.
The researchers analyzed the basic models for evaluating the I-V profile of modules – a graphical representation of the relationship between the voltage applied across an electrical device and the current flowing through it.
The conventional modeling tools are a basic, simplified three-parameter system which does not incorporate series and shunt resistance for I-V profiles; an extended five-parameter model which considers the effect of series and shunt resistance; and a seven-parameter version which also includes the effect of temperature and irradiance variation on solar cells.
“While all these models reasonably predict IV profiles of solar modules at small variations from standard testing conditions (STCs), their performance in modeling the module performance at low irradiances and high temperatures is far from ideal,” the researchers said.
The accuracy of the models was measured for crystalline silicon technology (c-Si) solar products and for thin-film devices based on amorphous silicon (a-Si); copper, indium, gallium and selenium (CIGS); and cadmium telluride (CdTe).
“The major contribution of this work lies in the detailed assessment of the models where we quantify the accuracy of all three models over a wide range of varying irradiance profiles for both TF [thin-film] and c-Si modules,” the researchers wrote.
Crystalline solar panels made by Mitsubishi, Sharp, Suntech, Adani Solar, Amerisolar, Canadian Solar, Gintech, Tamesol and Solar Power Mart were used, as well as thin-film products from First Solar, Q-Cells and Solar Frontier, among others.
The scientists found the three-parameter model generally over-predicted by a significant margin the maximum power point (MPP) – and therefore the power output – of c-Si and thin-film panels.
The five-parameter model, which is said to require higher computational complexity and, often, numerical solutions, was considered a better baseline for more accurate results as it was better able to predict the I-V profile for thin-film modules in multiple irradiance conditions. In low-irradiance conditions, however, the five-parameter model still did not accurately model results, according to the paper.
The researchers said the seven-parameter model was the most accurate for c-Si modules. “Slightly improved performance of the seven-parameter model is due to the fact that it includes the low-irradiance performance factor; it brings the power over-prediction down to 11.5%.” the report noted. “It still overestimates the power produced but is generally a better fit for modeling performance of c-Si solar cell technology.”