Reliability evaluation of DC power optimizers for photovoltaic systems: Accelerated testing at high temperatures with fixed and cyclic power stresses

This paper presents a long-term and large-scale time-to-failure reliability study to determine the acceleration factor and field-use lifetime for dc optimizers. Forty optimizers were tested at two high static temperatures (85 degrees C and 95 degrees C) and two input power profiles (fixed power and cyclic power) over a period of 6400 h. For degradation analysis, continuous electrical and thermal data were collected. On average, the dc optimizer units degraded by about 1%, while no hard failures occurred over 6400 h of high-temperature stress testing. Using the Arrhenius life-stress model along with average field measured temperature, it is predicted that the dc optimizers should be able to survive for 39-73 years in field operating conditions with a reliability of 79%, having a lower one-sided confidence bound of 90%, demonstrating similar or even exceeding the lifetime of the associated PV modules. Nevertheless, with the determined activation energy of 0.33 eV, derived from the observed degradation percentage in cyclic power profiles, the projected lifespan for dc optimizers is significantly curtailed. Under field conditions, our estimations suggest a service life ranging between 9 and 12 years. Furthermore, the cyclic power profile at higher stress temperature demonstrated higher levels of degradation compared to the fixed power profile. This implies the cyclic thermal mechanisms at higher stress temperature (which are not captured in the Arrhenius equation) account for a higher contribution to degradation. This has been confirmed through the use of Weibull distribution model where the analysis indicates that the estimated lifetime for dc optimizer can significantly decrease based on a higher stress temperature along with the power cycling of the units.