PV Degradation Curves: Non-Linearities and Failure Modes

Research output: Contribution to journalArticlepeer-review

106 Scopus Citations


Photovoltaic (PV) reliability and durability have seen increased interest in recent years. Historically, and as a preliminarily reasonable approximation, linear degradation rates have been used to quantify long-term module and system performance. The underlying assumption of linearity can be violated at the beginning of the life, as has been well documented, especially for thin-film technology. Additionally, non-linearities in the wear-out phase can have significant economic impact and appear to be linked to different failure modes. In addition, associating specific degradation and failure modes with specific time series behavior will aid in duplicating these degradation modes in accelerated tests and, eventually, in service life prediction. In this paper, we discuss different degradation modes and how some of these may cause approximately linear degradation within the measurement uncertainty (e.g., modules that were mainly affected by encapsulant discoloration) while other degradation modes lead to distinctly non-linear degradation (e.g., hot spots caused by cracked cells or solder bond failures and corrosion). The various behaviors are summarized with the goal of aiding in predictions of what may be seen in other systems. Published 2016. This article is a U.S. Government work and is in the public domain in the USA.

Original languageAmerican English
Pages (from-to)583-591
Number of pages9
JournalProgress in Photovoltaics: Research and Applications
Issue number7
StatePublished - 2017

Bibliographical note

Publisher Copyright:
Published 2016. This article is a U.S. Government work and is in the public domain in the USA.

NREL Publication Number

  • NREL/JA-5J00-65860


  • durability
  • failure mode
  • non-linearity
  • PV degradation
  • reliability


Dive into the research topics of 'PV Degradation Curves: Non-Linearities and Failure Modes'. Together they form a unique fingerprint.

Cite this