Quantitative Electroluminescence Imaging Analysis for Performance Estimation of PID-Influenced PV Modules

Karl G. Bedrich, Wei Luo, Mauro Pravettoni, Daming Chen, Yifeng Chen, Zigang Wang, Pierre J. Verlinden, Peter Hacke, Zhiqiang Feng, Jing Chai, Yan Wang, Armin G. Aberle, Yong Sheng Khoo

Research output: Contribution to journalArticlepeer-review

44 Scopus Citations


An empirical method for estimating relative power losses caused by potential-induced degradation (PID) for p-type solar cells and modules using quantitative electroluminescence (EL) analysis (QELA) is presented. First, EL images are corrected for camera- and perspective distortion. The relative power loss map is then calculated from the logarithmic ratio of two EL images, taken either before and after PID degradation or at different applied currents. Only the cell average of the resulting power loss map is evaluated. The highest power loss across each string is averaged to obtain the overall power loss. Consequently, for modules with three strings, three cells are averaged. The resulting power loss depends on the current applied. The conversion to equivalent irradiance allows for comparison of measured and estimated device performance. The analysis of roughly 2000 EL images and related current-voltage ( I-V) curves indicates a good agreement between flash-test-measured and performance estimated using QELA. A relative root mean square error of 1-3% can be achieved.

Original languageAmerican English
Article number8410677
Pages (from-to)1281-1288
Number of pages8
JournalIEEE Journal of Photovoltaics
Issue number5
StatePublished - 2018

Bibliographical note

Publisher Copyright:
© 2011-2012 IEEE.

NREL Publication Number

  • NREL/JA-5K00-72198


  • Electroluminescence (EL)
  • metrology
  • performance evaluation
  • photovoltaic (PV) modules
  • potential-induced degradation (PID)
  • solar cells


Dive into the research topics of 'Quantitative Electroluminescence Imaging Analysis for Performance Estimation of PID-Influenced PV Modules'. Together they form a unique fingerprint.

Cite this