Yarn Quality Assessment and Monitoring Using ML in Textile Industry

The monitoring and alert systems in most textile industries are manual or partially automated. Temperature and humidity are manually captured and recorded in textile units, but there is no use with this raw data without immediate analysis and necessary actions. Some of the issues in textile automation are lack of automated monitoring systems, poor data analysis after data collection, lack of real-time alerts and inability to integrate with existing systems. Live sensor-based data collection and data analysis are essential to take corrective or preventive measures in textile units. The proposed system collects the data from the sensors and stores it in a database. The raw material for each production run is determined by the technological value of cotton called the fibre quality index. During the daily production run, temperature, humidity and pollution level inside spinning units are measured and transferred by the sensor and monitoring unit. At the end of every production run, the imperfection level and count strength product are measured and updated by the quality analysts for every fibre quality index. Once the data was collected, it should be analysed to identify any patterns or correlations between the different variables. After analysing the data, a supervised machine learning (ML) algorithm was implemented to select high-quality yarn.