This is a two-part thesis assessing the long-term reliability of photovoltaic modules.

Part 1: Manufacturing dependent reliability - Adapting FMECA for quality control in PV module manufacturing

This part is aimed at introducing a statistical tool in quality assessments in PV module manufacturing. Developed jointly by ASU-PRL and Clean Energy Associates, this work adapts the Failure Mode Effect and Criticality Analysis (FMECA, IEC 60812) to quantify the impact of failure modes observed at the time of manufacturing. The method was developed through analysis of nearly 9000 modules at the pre-shipment evaluation stage in module manufacturing facilities across south east Asia. Numerous projects were analyzed to generate RPN (Risk Priority Number) scores for projects. In this manner, it was possibly to quantitatively assess the risk being carried the project at the time of shipment of modules. The objective of this work was to develop a benchmarking system that would allow for accurate quantitative estimations of risk mitigation and project bankability.

Part 2: Climate dependent reliability - Activation energy determination for climate specific degradation modes

This work attempts to model the parameter (Isc or Rs) degradation rate of modules as a function of the climatic parameters (i.e. temperature, relative humidity and ultraviolet radiation) at the site. The objective of this work was to look beyond the power degradation rate and model based on the performance parameter directly affected by the degradation mode under investigation (encapsulant browning or IMS degradation of solder bonds). Different physical models were tested and validated through comparing the activation energy obtained for each degradation mode. It was concluded that, for the degradation of the solder bonds within the module, the Pecks equation (function of temperature and relative humidity) modelled with Rs increase was the best fit; the activation energy ranging from 0.4 – 0.7 eV based on the climate type. For encapsulant browning, the Modified Arrhenius equation (function of temperature and UV) seemed to be the best fit presently, yielding an activation energy of 0.3 eV. The work was concluded by suggesting possible modifications to the models based on degradation pathways unaccounted for in the present work.