Application of gamma process model to estimate the lifetime of photovoltaic modules

摘要

The analysis of long term data for degradation of PV modules suffers from volatility and uncertainty due to intrinsic and extrinsic factors. The low rate of degradation causes analysis complexity and ambiguity. In this study, methods used to estimate the PV module lifetimes were reviewed in terms of degradation of power output with time. Under the assumption that degradation is continuous, gradual, and monotonic, the gamma process model can explain the sampling and temporal uncertainties of lifetime data. Examples are provided to demonstrate the use of gamma process model for long term and accelerated lifetime test (ALT) data. Three types of lifetime estimation method were compared for long term operation data. Although they all gave similar estimated lifetime, the gamma process model gave the most applicable results to determine warranty life. The gamma process model can also express the condition variation at inspection and the lifetime variation at failure level as probability distributions. A method to determine warranty life is proposed using an age based replacement policy. For ALT data, we estimated the lifetime from degradation data using the Arrhenius equation' for standard environmental conditions and applied the gamma process model to obtain time varying probability distributions for condition and lifetime. Service life was estimated as the median, while warranty life was estimated as the minimum rate of increase of optimal replacement time. (C) 2017 Elsevier Ltd. All rights reserved.