Early Failure Detection of Rapier Weaving Loom Using Sensors and Machine Learning Approach

Abstract

The textile industry is the oldest and widespread across the world.  This has high demand due to its impact on daily life. Textile production is not a simple task and hence textile industry uses many machines to meet daily production demands. Few of the common machineries used in textile industries are spinning machines, weaving looms, knitting machines, dyeing machines, and finishing machines. For weaving purpose, the terry and similar textile industries rely on high-speed rapier weaving looms. These machines are designed for longer operational time. The Terry and similar textile industries are frequently experienced breakdown in functioning of rapier looms due to involvement of complex electronic circuitry, sensors, and their failures. In a survey it is observed that, more faults related to electronics is the main cause of weaving loom stoppage as compared to faults related to mechanical reasons. The local working conditions in the industry are responsible for premature failure in electronic circuitry and sensor parts. Hence, the industry now experiencing higher faults in electronic circuits, sensors and power supply as compared to mechanical related faults. The main objective of this paper is to design an IoT framework for condition monitoring of rapier loom to generate data representing its current working condition. A classifier based on supervised learning is designed to distinguish between safe and unsafe working condition of rapier loom.