Edited by Cori Watkins, contributing editor, Plastics Decorating

Krevera has introduced its AI-powered vision and autonomy system for automated defect detection and reduction in plastic injection molding, with a focus on in-mold labeling applications. The vision system detects even the most challenging defects, such as contamination, label rotation, label folds, blisters and more – providing detailed insights into the type, size and location of each defect. This data is integrated with injection molding machines and connected robotics to perform causal analysis. Leveraging this information, the AI system enables root cause analysis, process drift detection and preventative maintenance, which results in an up to 30% reduction in waste.

Many current vision systems require monthly and sometimes daily recalibration. Factors such as changes in lighting, the presence of dust or dirt, camera movement and shifts in product orientation all can cause legacy vision systems to misidentify products as defective. These systems are highly sensitive to environmental variations, and even minor inconsistencies can disrupt their ability to accurately distinguish between acceptable and defective items.

Krevera’s proprietary vision system requires zero recalibration, even with changes such as lighting, product orientation, camera placement or dust/dirt on the lens. It continually works with no management or coding experience. Vision systems on the market today typically are limited to making simple, binary decisions about each product to determine if the product being produced is good or bad. This AI-powered vision system can categorize defects into one of 30 types, determine the size of the defect and pinpoint its location, which are important for increasing production efficiency.

Addressing Production Challenges and Optimizing Processes
The in-depth outcomes data that the vision system provides can be connected with the data from the injection molding machine to create a closed loop, allowing the system to elevate process windows, provide predictive maintenance and boost machine uptime.

The vision system enables companies to reduce labor by providing 100% inspection, which decreases the need for floor inspection and allows for automated packaging. Krevera’s inspection system also offers features for machine management and mold degradation watch. This allows companies to proactively fix potential concerns before the issues result in weeks of downtime.

Production challenges the system address includes the following:

• Process drift and hidden root causes. Drift is flagged within a handful of cycles.
• Labor shortages. Repetitive part inspection and in‑process tuning are handed off to the system.
• Inconsistent lighting/fixture set-ups. The vision stack is robust to camera placement, dust and ambient variation.
• Slow changeovers. Rather than having to set up the vision system to detect defects every time the consumer has a new product or generate and manually label data for the vision system to function, the AI agent offers rapid set-up with no data collection and no coding requirement.

Krevera’s solution also is built to address quality concerns for in-mold labeling applications. For manufacturers that are producing plastic containers with labels fused into the mold, the Krevera system can detect if a label is rotated, folded, misaligned or blistered, all in real time. With this data, the system can determine what is causing the label defects and alert the production team. This means the container branding is consistent, rejects are low and customers stay happy.

Technology and Sustainability
The solution leverages a relatively small number of third-party cameras, after which the software and AI system present data on every defective product in production and evaluate metrics, such as the defect rate. The overall equipment effectiveness (OEE) indicator automatically updates with automated trigger alerts for any defects that are detected. The vision system also provides an optimized dynamic process window for each plastic injection molding machine. The system is equipped with built-in (native) software drivers that use the OPC UA protocol for communication. This allows the machine to connect directly and seamlessly to other devices, systems or software platforms and covers > 90 % of global injection molding presses.

Krevera is collaborating with leading OEMs and automation cell companies to build a strong partnership ecosystem focused on sustainability. By leveraging the company’s advanced vision system, manufacturers can significantly reduce scrap, resulting in the conservation of tons of plastic and the saving of thousands of kilowatt-hours of energy per plant each year. This commitment not only drives operational efficiency but also delivers substantial environmental impact.

Plant managers have reported that the Krevera solution is easy to set up and provides “the clearest path to lights-out manufacturing,” with no churn to date. The numbers back up this high praise with an estimated return on investment of 130,000 to 400,000 per injection molding machine. n

For more information, visit www.krevera.com.