Materion is the top global producer of high-performance advanced engineered materials for semiconductor, aerospace, automotive, defense, consumer electronics, and medical appliance industries. Since the early days of NASA, Materion has had a front seat in space exploration. Their high-end specialty materials outfit heat shields, reflectors, and filters and are on various Mars rovers, spacecrafts, and space telescopes.

For years, the company faced a persistent problem of low yield due to multiple defect types, which caused them to scrap hundreds of thousands of pounds of material at various processing steps. Materion’s quality and process engineers had no way to test for root cause analysis (RCA) to identify the underlying issues. Materion also suffered from serious data challenges: lack of data, multiple data-capture systems creating fragmentation even across a single plant, history available for only a subset of processes, sensors that had never even been hooked up, inconsistent defect labeling, and non-existent product-process genealogy.

Materion chose Quality Flow because quality is essential to their customers. Quality Flow was first put to work at one plant to perform RCA on one defect: edge lamination. Because the Noodle.ai software is designed to handle disparate data sources, it efficiently mapped the required data (sensor, defect, production process, and product genealogy) to data models.

After data ingestion and model training, Quality Flow analyzed more than 1 million parameter combinations, using a time-series granular process and sensor data for hundreds of parameters to identify the top six combinations, or rule groups, that were the leading causes for the lamination defect.

By the time Materion deployed Quality Flow at three of its plants, the Noodle.ai software had:

• Generated millions of dollars in ROI.
• Reduced defects by nearly half.
• Achieved significant improvement in overall yield.
• Reduced scrap by hundreds of thousands of pounds.
• Enabled further savings by allowing Materion to switch to lower-cost raw materials.

For the first time, Materion engineers could navigate reams of data from years of production to diagnose the likely cause of defects and take preventive actions in production. The AI-powered Quality Flow allowed them to:

• Find complex patterns hidden in vast amounts of data.
• Validate defect diagnosis with explanatory analysis.
• Schedule AI-recommended operating rules into production.
• Monitor alerts and mitigate defects.
• Collaborate with colleagues across the company.