Home / Case studies / Industry reference
Industry referenceEL imaging · 56 MW

Night-time EL testing leads to a warranty claim on 5,000+ modules

Clean Energy Associates (CEA) · 56.49 MW plant, England

Not a Retron project. This example is from Clean Energy Associates (CEA). It shows the kind of work involved.
~180,000modules in the plant
~90%of modules imaged
~5%with cold soldering, plant-wide
5,000+modules replaced under warranty

The situation

An operations and maintenance provider saw power losses on a large plant and engaged CEA in 2018 to find the cause.

Background

At this scale a loss of a few percent is a large amount of money every year, but the plant was too big to test module by module with a lab flash tester. The question was which part of the plant, and which defect, to chase.

What was tested, and how that method works

On-site night-time electroluminescence (EL) imaging of about 90% of the plant's modules, out of roughly 180,000, over three months.

EL imaging powers a module in the dark so that healthy cells glow evenly. Cracked or poorly soldered areas stay dark. Doing it at night, on site, avoids shipping modules to a lab and lets very large numbers be imaged.

Our service: Electroluminescence (EL) imaging →

What was found

  • Cold soldering in about 5% of all modules, along with microcracks.
  • At one location of nearly 19 MW, about 80% of modules were affected by cold soldering or microcracks.

How it unfolded

2018CEA is engaged after the owner's O&M provider sees losses.
Over 3 monthsAbout 90% of modules are imaged at night.
After analysisDefects are traced back to production.
OutcomeThe manufacturer agrees to a warranty claim.

Outcome

The defects were traced back to production, so the manufacturer agreed to a warranty claim and replaced more than 5,000 modules, worth almost EUR 500,000 at market value according to CEA.

Lessons for a plant owner

  • Field EL on a large share of modules can supply the evidence a manufacturer claim needs.
  • Defects can cluster in one block, so sampling across blocks matters.
  • Dark images alone do not give a power loss. They need to be paired with IV readings.

Questions to ask about your own plant

  • Do we have EL images from delivery or commissioning to compare against?
  • Which blocks have the weakest strings in our own data?
  • What does our warranty require as proof of a manufacturing defect?

How we would approach a similar site

We would start with a representative sample of modules, confirm the pattern with IV curve tracing, and document findings so they can support a claim.

Published by the testing company itself. The plant is in England, not India. Source: Clean Energy Associates blog. Summarised in our own words.