Traditional degradation or reliability analysis of photovoltaic (PV) modules has historically consisted of some combination of accelerated stress and field testing, including field deployment and monitoring of modules over long time periods, and analyzing commercial warranty returns. This has been effective in identifying failure mechanisms and developing stress tests that accelerate those failures. For example, BP Solar assessed the long term reliability of modules deployed outdoor and modules returned from the field in 2003; and presented the types of failures observed. Out of about 2 million modules, the total number of returns over nine year period was only 0.13%. An analysis on these returns resulted that 86% of the field failures were due to corrosion and cell or interconnect break. These failures were eliminated through extended thermal cycling and damp heat tests. Considering that these failures are observed even on modules that have successfully gone through conventional qualification tests, it is possible that known failure modes and mechanisms are not well understood. Moreover, when a defect is not easily identifiable, the existing accelerated tests might no longer be sufficient. Thus, a detailed study of all known failure modes existed in field test is essential. In this paper, we combine the physics of failure analysis with an empirical study of the field inspection data of PV modules deployed in Arizona to develop a FMECA model. This technique examines the failure rates of individual components of fielded modules, along with their severities and detectabilities, to determine the overall effect of a defect on the module's quality and reliability.