Visual Parameter Space Analysis for Optimizing the Quality of Industrial Nonwovens

Area of Science:

• Materials Science and Engineering
• Industrial Manufacturing
• Data Science and Analytics

Background:

• Technical textiles, especially nonwovens used in medical masks, are critical but challenging to optimize.
• Manufacturing process parameters for nonwovens significantly impact product quality.
• Optimization of these parameters is difficult in live industrial settings.

Purpose of the Study:

• To present a visual analytics framework for interactive parameter space exploration and optimization in nonwoven production.
• To integrate strategies for optimizing industrial nonwoven manufacturing processes.
• To enable real-time interaction and rapid quality computations.

Main Methods:

• Survey of analysis strategies for optimizing industrial nonwoven production processes.
• Augmentation of a digital twin with a machine learning surrogate model for fast quality predictions.
• Integration of sensitivity analysis mechanisms for consistent product quality.

Main Results:

• Demonstrated interactive exploration of parameter spaces for nonwoven manufacturing.
• Identified optimal parameter sets and investigated input-output relationships.
• Conducted sensitivity analysis to determine robust quality settings.

Conclusions:

• The visual analytics framework effectively supports parameter optimization in nonwoven production.
• The integration of machine learning and digital twins enables real-time quality control.
• Sensitivity analysis ensures product quality consistency under varying parameters.