Development and experimental validation of a comprehensive thermoelectric dynamic model of photovoltaic modules. (1st March 2017)

• Record Type: Journal Article
• Authors: Migliorini, Luca; Molinaroli, Luca; Simonetti, Riccardo; Manzolini, Giampaolo
• Is Part Of: Solar energy. Volume 144(2017)
• Page Start: 489
• Page End: 501
• Publication Date: 2017-03-01
• Subjects: PV forecast power production -- PV equivalent electrical circuit -- Dynamic thermo-electric models -- Five layers thermal model; Solar energy -- Periodicals; Solar engines -- Periodicals; 621.47
• DOI: 10.1016/j.solener.2017.01.045
• Languages: English
• ISSNs: 0038-092X

Abstract

Highlights: An extensive dynamic thermo-electric model for PV forecasting is developed. The model is validated through experimental results on two different modules. Dynamic model can reduce the error of the forecasting around 3 W. The dynamic model reduces the error of conventional approaches by 50%. When using weather forecast, the importance of a dynamic model drops. Abstract: In order to increase PV penetration and guarantee the grid stability, it is fundamental to accurately predict the power produced from PV plants. This work discusses a comprehensive thermo-electric model accounting for PV module dynamic behavior. Specifically, the thermal part considers five different sections featuring the five layers of the PV module. The electrical behavior of the PV module is described by the five parameters electric equivalent circuit where the operating temperature is calculated within the thermal model. The accuracy of the model is evaluated against the power produced by two monocrystalline silicon modules at the SolarTech Lab of Politecnico di Milano using both actual weather information and the weather forecast. Results highlight that dynamic models adequately characterize the behavior of the modules when actual weather measurements are available; the resulting error indexes WMAE and RMSE for the dynamic model are reduced with respect to steady-state and conventional approaches based on NOCT by 50%. Indeed, the WMAE is approximately 1%, which is comparable to the uncertainty related to measuring instrumentations, whereas the RMSE is included between 3 W and 4 W. When using the long term weather forecast, no advantages in the development of dynamic models can be noted.