PV Field Reliability Status--Analysis of 100,000 Solar Systems

Dirk Jordan, Bill Marion, Chris Deline, Teresa Barnes, Mark Bolinger

Research output: Contribution to journalArticlepeer-review

55 Scopus Citations


In this study, we analyzed annual production data from 100 000 photovoltaic systems as well as comments relating to their performance and maintenance. Our analysis revealed that 80% to 90% of all systems performed within 10% of the predicted production or better. Also, 56% of the systems were still performing above P50 or the median at 5 years. However, a small but significant tail of about 7000 systems perform below P90 expectations. In general, residential systems have a lower rate of failure than utility or commercial systems. Despite higher rates of component failures, utility systems lose less power than residential or commercial systems. This outcome is likely due to closer monitoring and better operations and maintenance practices. Inverters are still the components that reportedly fail most often (4%-6%), but other failures such as unspecified repair and meters cause more production loss. Reported module failures are relatively rare (0.2%) and are within the range of historical values. Installation quality affects performance and safety as indicated by data showing connector, wiring, breaker, and fuse failures due to undersizing, electrical design, and improper connection. Lastly, early detection of degradation and proactive response resulted in less impact on production than reactive, unplanned repairs.

Original languageAmerican English
Pages (from-to)739-754
Number of pages16
JournalProgress in Photovoltaics: Research and Applications
Issue number8
StatePublished - 2020

Bibliographical note

Publisher Copyright:
© 2020 John Wiley & Sons, Ltd.

NREL Publication Number

  • NREL/JA-5K00-75023


  • field failure
  • field performance
  • photovoltaic
  • reliability


Dive into the research topics of 'PV Field Reliability Status--Analysis of 100,000 Solar Systems'. Together they form a unique fingerprint.

Cite this