Use-Case: Photovoltaic Manufacturer

How FireVisor helps to save 600,000 USD for a 1 GW plant

Solar is the future with bright market prospects but there are problems to solve to stay competitive.

Challenges In The Solar Industry

High cost of poor quality

Poor quality products are devastating to profits and brand value as buyers expect warranties for more than 20 years.

Data full of false signals

Up to 70% of the defect data could be incorrect due to misclassification of defects. Data has to be cleaned first before any analysis is possible.

Inexplorable Data

Machines like EL imaging provide valuable data for yield improvement but process engineers have no tools to analyze image data.

The power of data science for every process engineer

FireVisor’s AI-powered descriptive analytics solution

Seemless Integration

Integrates easily to existing MES and inspection systems such as EL Imaging

Accurate Data cleaning

Cleans defect data from false positive with 98.5% accuracy

Tailored to solar

Gives powerful data exploration tools with visual data analytics capability tailored for solar cell and module line.

Data Cleaning

  • Confidently remove falsely classified defects with 98.5% accuracy and enable process engineering to gain real insights from true datasets.
Misalignment
Rollermarks
Grains
Marks

Defect Localization
Heatmap

  • Overlay of hundred thousands of defects allows process engineers to analyze defect patterns in one glance.
  • The software is specifically written for the solar industry and allows engineers to do analysis on module and cell level.

Segmentation

  • In this setting a single EL machine gets modules from 4 different stringer machines. Our platform is able to untangle the data knowing which defects are contributed by which machine.

Clustering

  • Analysis gets hard when multiple defects are happening at the same time. We help engineers to identify identify distinct defect causes with our proprietary machine learning algorithms. Now engineers are able to tackle the problems one by one.

NegTracing

  • Trace back where defects come from through NegativeTracing.
  • As engineers no longer have to spend time to find where defects come from, they can immediately focus on solving the issues.

PosTracing

  • Our systems identifies machines that perform exceptionally well. Engineers use those machines as reference models and use similar configurations to optimize other machines.

Results

Automated root cause Analysis

Real-time defect analysis solution that singles out root cause among hundreds of variables.

50% of Engineering Analysis time

Process Engineer no longer spend hours collecting, cleaning and organizing data. Instead they can focus on line improvements.

Savings of 600k USD for a 1GW Plant

Get In touch