Development and validation of coupled thermal-electric transient model of a photovoltaic system

Abstract

Installed capacity of renewable energy systems keeps growing worldwide in response to climate and socio-economic changes. Among these, photovoltaic power can increase the solar energy contribution in the global mix of energy sources in the near future. Crystalline (mono or poly-crystalline) silicone is the material used in the most widespread photovoltaic modules. Their efficiency depends on several factors and thus modeling the conversion of solar radiation into electricity is crucial and can have various applications, from the study of the different parameters that affect this conversion to the design, maintenance and power output forecasting of photovoltaic systems. To that end, various models have been proposed in the literature, many resorting to empirical data and based on specific modules/systems. In this work, a coupled thermal-electric model was developed for crystalline silicon modules and validated with experimental data from four photovoltaic technologies. The model was developed to be used without resorting to system measurements but only using data provided by the manufacturers. To avoid biases towards a specific technology or location empirical data is not used in its development. The thermal model is based on the energy conservation principle and the heat transfer processes that occur in illuminated photovoltaic modules. The electrical model used is the single diode - five parameters equivalent electrical circuit. The proposed model is transient thus providing the temporal variation of the temperature of the module and electric power output simultaneously at an imposed time step, which is an advantage for modeling at high temporal resolutions with applications in inverter operation and electric grid stability. Moreover, the model outputs can be easily averaged or integrated in order to obtain mean values of system operation. The validation of the model was done with 10 minute data of global tilted irradiance, air temperature, wind speed and direction, temperature of the modules, power output of arrays of four different crystaline silicone technologies with peak power ranging from 1904.8 to 2000 W located in Koriyama, Japan (37.4495; 140.3144). The data used resulted in 24543 data points for each photovoltaic system without shadowing of the photovoltaic cells caused by other rows or snow deposition. The results show a slight overestimation of photovoltaic temperature (up to 2.52°C in mean bias error) and power output (up to 6.44% in relative mean bias error) for all systems. Although in terms of generated power, when comparing the developed model with the single diode - five parameters equivalent electric circuit model but using the measured temperature of the modules, the proposed model showed better estimations for all systems but one.