Towards real time automated early gear failure detection

The ability to stop a gear fatigue test before catastrophic failure has many advantages. However, today, widely accepted approach is not available. This case study applies vibration-based condition monitoring methodology detect early failures. studied takes part in an all-wheel-drive drivetrain system. Vibration signals from four run-to-failure tests at two constant torque-speed combinations were used as input time-synchronous averaging and autoregression model generation. applied shows promising results for detection, the process feasible implementation automated environment. Real time analysis also possible since generates healthy state TSA signal during testing stages. detection varies with operating conditions, low sensitivity high-speed low-torque conditions.