Predictive reliability for AC photovoltaic modules based on electro-thermal phenomena

AC photovoltaic modules promise many performance enhancements but expose their embedded power electronics to rigorous environmental stressors. The ability to identify reliability consequences in the early design phase is critical to rapid market adoption. This paper develops the capability to generate long-term reliability predictions based on electric and thermal performance over multiple time scales. The model is validated with measurements from centralized inverters as well as power electronics that are integrated into the module. The resulting model will enable manufacturers to optimize the design and physical layout of next-generation AC modules for microgrid systems prior to mass production. Preliminary results indicate direct electrical performance impact due to heat transfer between the module and embedded power electronics. Non-unity power factor experiments show that an operation factor of 0.85 can cause a 60% increase in DC voltage ripple, while raising voltage to approximately 90% of the open circuit value.