In-Situ Characterization of Potential-Induced Degradation in Crystalline Silicon Photovoltaic Modules Through Dark I–V Measurements

Abstract

A temperature correction methodology for in-situ darkI-V(DIV) characterization of conventional p-type crystalline silicon photovoltaic (PV) modules undergoing potential-induced degradation (PID) is proposed. We observe that the DIV-derived module power temperature coefficient (γ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> ) varies as a function of the extent of PID. To investigate the relationship between γ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> and DIV-derived module power (P <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> (T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> ), measured in situ and at the stress temperature) two parameters are defined: change in the DIV-derived module temperature coefficient (Δγ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> ) and DIV-derived module power degradation at the PID stress temperature (ΔP <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> (T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> )). It is determined that there is a linear relationship between Δγ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> and ΔP <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> (T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> ). Based on this finding, we can easily determine the module γdark at various stages of PID by monitoring P <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> (T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> ) in situ. We then further develop a mathematical model to translate P <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dark</sub> (T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> ) to that at 25 °C (Pdark (25 °C)), which is correlated with the module power measured at the standard testing conditions (PST C). Our experiments demonstrate that, for various degrees of PID, the temperature correction methodology offers a relative accuracy of ±3% for predicting PST C. Furthermore, it reduces the root-mean-square error (RMSE) by around 70%, compared with the PSTC estimation without the temperature correction.