Factory AI Quality Inspection
A leading global manufacturer of lead-acid batteries faced high defect rates during production, primarily due to assembly flaws, welding defects, and plating issues.The traditional manual quality control system lacked real-time monitoring, making it difficult to detect defects promptly, which negatively impacted maintenance and equipment management efficiency.
To address these challenges, Edgegenesis partnered with the company to co-develop a customized Edge AI-driven IoT quality inspection platform. This system captures and processes production defects in real time, significantly enhancing the accuracy and responsiveness of quality monitoring. As a result, defect rates have been reduced, equipment management has been optimized, and both production efficiency and product quality have improved—ultimately lowering operating costs and strengthening market competitiveness.
Solution Highlights:
- Improved Yield Rate: The system promptly identifies defective products and issues early warnings. It automatically controls the production line to manage anomalies, effectively increasing the yield rate and ensuring stable product quality.
- Accelerated Deployment Efficiency: With support for rapid remote deployment and device configuration, the solution reduces manual operations and equipment downtime, boosting overall production efficiency.
- Enhanced Recognition Accuracy: Continuous updates to detection models enable quick adaptation to new products, improving both recognition speed and accuracy.
- Scalability for Future Expansion: The system’s high flexibility allows for expansion across additional production lines and scenarios, supporting long-term development.
Challenges
- Device Compatibility: Variations in hardware and software platforms across production stages complicate system integration and data sharing.
- Real-Time Data Processing: The system must handle large volumes of real-time data to detect and evaluate product defects accurately, demanding high responsiveness.
- Operator Adaptation: Transitioning from manual to intelligent systems requires operators to quickly adapt and become proficient, reducing the learning curve.
- System Integration: The solution must integrate seamlessly with existing workflows without disrupting the current production line's stability.
Edgenesis Solution
Automated Deployment
Utilizing SUSE EIB / Elemental for plug-and-play edge gateways, simplifying deployment and reducing configuration time.
Cloud-Edge Management
Leveraging k3s and Rancher to enable rapid application iteration and smooth device integration, boosting stability and security.
IoT Interoperability
Using Shifu to abstract devices as API endpoints, enabling unified access and fast development, enhancing production flexibility.
Edge AI Quality Inspection
Vision-based edge AI enhances the precision of defect detection and improves yield rates.
Result
Following system implementation, the company achieved over 99% detection accuracy, enabling high-precision, real-time monitoring of battery components. This allowed for accurate identification of defects in battery terminals and the conveyor system. As a result, the yield rate improved by 6%, delivering annual cost savings of approximately 2.2 million USD through reduced rework and waste. Overall, the company saw a marked improvement in production efficiency and strengthened its market competitiveness.
