How IoT and AI Eliminated Costly Downtime in High-Speed Can Production
The Tear-off Cans Probable Cause Analysis highlights how one of the world’s largest beverage can manufacturers addressed a recurring challenge of torn-off and short cans. Leveraging IoT-enabled production data, the company, supported by advanced data engineering, analysis, and machine learning, sought to identify and maintain safe operating zones to ensure high-quality cans.
With production speeds averaging 300 cans per minute per machine, any downtime due to defective cans had a significant impact. Each bad can caused the machine to halt, and the average recovery time of 13 minutes translated into 3,900 missed cans per incident. Reducing these stoppages became a critical objective to improve efficiency and output.
Through detailed investigation, data patterns revealed that combinations of key controls such as Air Pressure, Speed, and Die Tool Deviation influenced can quality. Frequent item set mining was applied to pinpoint these patterns, and interactive PowerBI™ dashboards were developed to provide stakeholders with actionable insights. Plans were also made to incorporate additional data on coil materials and tooling for deeper analysis.
Challenges
Machine stoppages due to torn-off or short cans resulting in significant production losses.
Complex interplay of machine settings and operational parameters affecting can quality.
Large volumes of IoT production data requiring thorough analysis to uncover actionable patterns.
Solutions
-
Data-Driven Root Cause Analysis
Applied frequent item set mining to identify combinations of operational parameters affecting quality. -
Safe/Unsafe Operating Zones
Defined optimal control settings to consistently produce good cans and prevent defects. -
Visual Insights for Stakeholders
Developed PowerBI™ dashboards to track performance, monitor parameters, and guide operators in real time.
Impact and Transformative Outcomes
-
Improved Efficiency
Reduction in machine downtime, increasing total production capacity. -
Proactive Quality Control
Operators can make informed adjustments before defects occur. -
Data-Backed Decision Making
Real-time visualization and analytics drive continuous process improvement.
Conclusion:
By embracing a data-first approach, the beverage can manufacturer transformed its production process from reactive to proactive. Advanced analytics and machine learning provided a clear path to minimize defects, improve efficiency, and safeguard product quality. With optimized operating zones and actionable insights, the company positioned itself to achieve consistent high-volume, high-quality output — ensuring competitive advantage in a demanding market.