Using Historical Thickness and Sensor Data to Predict Future Quality Control Performance (2026-EU-30MP-2763)

Maintaining process control is essential in manufacturing to ensure product quality, consistency, and compliance with critical parameters. Traditional quality control methods often require halting production to run test samples through systems or machines, resulting in significant downtime and reduced output. In many cases, production cannot resume until the sample passes all required specifications, with testing phases lasting several hours. While lean manufacturing techniques offer some relief, their impact is limited when further efficiency gains are needed.

This study explores a data-driven approach to reduce testing time and improve production continuity. Using JMP, we identified key process variables and developed a predictive model to estimate test parameter values prior to actual measurement. Analysis of a large data set yielded a high R squared value, validating the model’s accuracy and reliability. This enabled “risk” production to proceed ahead of sample confirmation, significantly minimizing output loss and improving operational efficiency.

The results demonstrate that predictive analytics can complement traditional quality control methods, offering a scalable solution for reducing downtime in manufacturing environments. This approach highlights the potential of leveraging historical data to enhance decision making and to streamline production processes without compromising quality standards.