Predicting damp heat degradation in heterojunction PV modules using machine learning

Due to the ongoing advancements in the efficiency of solar cells, photovoltaic-generated electricity is now the most affordable energy source globally. Nevertheless, to unlock the full potential of photovoltaic systems, their reliability needs to be improved. The capability to accurately predict the performance of photovoltaic modules over their years of operation using fast and cheap methods can be a game changer. In this study, the performance of photovoltaic modules is monitored during 1,500 hours of damp heat testing. Classical machine learning models were then developed to predict their performance at the end of the test, using ONLY 10% of the measurements. This research represents a crucial step toward predicting the long-term performance of photovoltaic modules in the field, a capability that will revolutionize the photovoltaic industry.