AUTOMATIC DIAGNOSTIC SYSTEM FOR FAULT DETECTION IN WEAVING MACHINES
Keywords:
weaving machine, automatic diagnostics, fault detection, vibration analysis, FFT spectral analysis, neural network, PLC, SCADA, condition monitoring, predictive maintenanceAbstract
This thesis examines the development of an automatic diagnostic system for fault detection in weaving machines. Five primary fault types are identified and characterized: weft thread breakage, warp thread breakage, roller bearing wear, tensioning mechanism failure, and main shaft vibration. A three-tier diagnostic system is proposed based on a multi-sensor data acquisition layer, Siemens S7-1500 PLC control level with 512-point FFT signal processing, and WinCC SCADA supervisory level with a multilayer perceptron (MLP) neural network classifier. The system achieves 94.7% average fault classification accuracy with a 0.09 s reaction time, reduces unplanned downtime by 78.3%, and decreases defective fabric output by 4.2 times compared to conventional manual monitoring.
References
