An Advanced Approach to the Bending Fatigue in Spur Gears

The testing and use is described of an advanced signal processing technique, called the continuous wavelet transform, for the detection and location of defects in gear teeth caused by bending fatigue. Defects of this kind were simulated on a real spur gear which was then tested on a back-to-back- gear rig, the resulting vibration signals being detected by accelerometers located around the bearing housing closest to the faulty gear. Accelerometer and timing signals were recorded, and the vibration data was averaged over three pinion rotations. the faulty gear. Accelerometer and timing signals were recorded, and the vibration data was averaged over three pinion rotations. The continuous wavelet transform (in which the Morlet wavelet was used as analysing wavelet) was applied to each set of experimental data as the basis for fault detection. It was concluded that the wavelet transform readily detects surface damage of very low severity.