FFT-BASED VIBRATION SIGNAL ANALYSIS FOR FAULT DETECTION IN WEAVING MACHINES
Keywords:
FFT analysis, vibration signal processing, bearing fault frequencies, BPFO, envelope analysis, weaving machine, fault detection, spectral analysis, MATLAB, condition monitoring.Abstract
This thesis investigates the application of Fast Fourier Transform (FFT)-based vibration signal analysis for fault detection in weaving machines. The mathematical foundation of spectral analysis is presented alongside its practical implementation in the weaving machine diagnostic context. Bearing defect frequencies (BPFO, BPFI, BSF, FTF) are analytically derived from machine geometry and verified through MATLAB simulation. Three signal processing approaches — time-domain threshold monitoring, 512-point FFT spectral analysis, and envelope analysis — are compared for five weaving machine fault types. Results demonstrate that FFT spectral analysis combined with envelope demodulation achieves 96.1% bearing fault detection accuracy — 8.4% higher than simple RMS threshold monitoring — and provides fault localization to specific bearing components, enabling targeted component replacement rather than full bearing assembly overhaul.
References
