Statistical Quality Control in Ceramic Manufacturing

The Challenge

Ceramic tile manufacturers face significant variation in their testing procedures that can mask the actual process variations they’re trying to identify. Florida Tile’s quality control staff discovered that their testing variation was as great as or greater than the process variation they were attempting to measure. Traditional destructive break strength testing also provides weaker correlation with process variables, making it difficult to optimize manufacturing parameters.

The Solution

Florida Tile implemented GrindoSonic non-destructive testing to correlate process parameters. By tapping tiles with a small mallet and sensing vibrations with a piezoelectric probe, the testing configuration achieved near-zero standard deviation in readings across operators and conditions. The torsional vibration mode provided the strongest correlation with tile properties.

The testing configuration proved highly repeatable and reproducible, with minimal variation from probe placement, operator technique, or tapping strength. Combining NDT data with statistical correlation software enabled analysis of 20+ variables simultaneously, identifying which processing parameters most strongly influenced final product quality.

Results

Testing 586 production tiles demonstrated strong correlations between GrindoSonic readings and physical properties:

• Fired weight showed strongest correlation (T-value: -14.78)
• Thickness correlation (T-value: 4.11) and size correlation (T-value: -2.92)
• GrindoSonic model explained 78.9% of variation (RSQ = .789) vs only 63.8% for break strength
• In process parameter experiments, GrindoSonic explained 87.9% of variation while break strength explained only 47.8%

An Experimental Design Optimizer selected 8 efficient experiments from 36 possible combinations, identifying CBF (batch moisture) quantity as a key process variable. The method provides a starting point for understanding processing variable interactions while enabling continuous quality improvement.