Machine Learning for Yarn Quality Prediction
topic
Machine learning models trained on historical spinning process data including fibre properties, machine settings, ambient conditions, and measured yarn quality outcomes predict yarn quality parameters from upstream process data before physical yarn testing, enabling early intervention when predicted quality is outside specification and building the process-property relationship knowledge embedded in mill historical data.
Role
Enables proactive yarn quality management by predicting quality outcomes from process input data before conventional laboratory testing could detect problems, shifting quality management from reactive defect detection to predictive process control and building explainable process-quality models that transfer spinning expertise from retiring expert operators into data-driven decision support systems.