Investigation of temperature coefficients of PV modules through field measured data

Varying broadband irradiance and temperature are generally known as the major factors influencing the performance of PV modules, but studies have also shown the substantial impact of spectral variations. In this work, a simple and efficient method to calculate the temperature coefficient using long term data is demonstrated. Temperature coefficients of PV modules are estimated from long term performance data following IEC 60891 standard with additional spectral correction, and are compared against the datasheet values. Significant improvement of correlation coefficient from -0.89 to -0.97 is observed during the regression for maximum power temperature coefficient of two poly-crystalline modules, after spectral correction by spectral factor (SF). Also, the standard deviation of yearly estimated values of these coefficients reduced from 5-7 % to 1-2 %. In another setup involving spectral measurements and various PV technologies, the annual mean of 1.62 eV for average photon energy in 350-1700 nm range, suggests a general blue shift of the spectrum. Higher averages than reference values of useful fraction (UF) for c-Si, CIGS and HIT technologies also validate the blue shift of spectrum. Results show SF produces maximum power temperature coefficients closer to the datasheet values compared to UF, suggesting better applicability of SF as an index for spectral correction. The coefficient values were found closer to STC values and the results from Mann and Kendall test, employed to detect any underlying monotonic trend in the development of temperature coefficients over eight years, showed no increasing or decreasing trend and hence no degradation of temperature coefficients for the long-term exposed PV modules.

Automated summary

Broadband irradiance and temperature are significant factors affecting the performance of PV modules, but spectral variations also play a substantial role. A method for calculating temperature coefficients using long term data with spectral correction has been demonstrated in this study. The results show that spectral correction improves the accuracy of temperature coefficients and closer aligns them with datasheet values, with no degradation observed in long-term exposed PV modules.