Intertek CEA's 100% EL Review enables third-party inspection of every module in a shipment using electroluminescence imaging
The EL-AI software can analyze up to 20,000 EL images per hour and identify up to 20 categories of module defects
The inspection results include module-level defect identification, validated by quality experts, and are delivered as a traceable report for future reference
Quality assurance and technical advisory company Intertek CEA offers a 100% electroluminescence (EL) imaging review to assess every module that leaves a manufacturing facility. It uses the company’s proprietary AI-based software called EL-AI.
Currently, the quality assurance processes follow internationally recognized standards, such as ISO 2859 for sampling, to inspect and qualify a batch/lot of products. While inspecting the entire production lot before shipping is desirable, it is not common practice. Module manufacturers incorporate EL imaging into assembly lines to detect defects and rework, sort, or reject modules. Although some manufacturers have AI-based defect detection integrated into EL imaging, several others do not, and manual inspection isn’t 100% foolproof.
Intertek CEA's EL-AI enables a third-party review of EL image of every single module that would be shipped to a customer. Based on this assessment, modules can be rejected or replaced before reaching the customer. Additionally, it creates a historical record of every module before it ships and can be used to verify future inspections after shipping, installation, and even after weather events the module may experience during its lifetime.
The AI-supported software can inspect up to 20,000 EL images per hour. It can be configured to identify up to 20 defect categories including several types of cracks, scratches, soldering defects, disconnected cells and cell parts, and cell mismatch. For the tool to accurately analyze and identify the images, the EL image should have a minimum resolution of 10 million pixels and at least 1 MB in size. The modules are rated according to the severity of the defect identified by the software, which in turn is validated by the company’s quality expert. After the final analysis and review, results of the inspected lot are provided in an Excel file that includes a list of identified defective module serial numbers along with the defect description.
