Condition monitoring for helicopter main gearbox based on wavelet packet transform and wavelet neural network

This paper describes a conditon monitoring systems for helicopter main gearbox using wavelet packet transform (WPT) and wavelet neural network (WNN). According to the fault characteristics of main gearbox, a fault diagnosis method that combining WPT and WNN with threshold is proposed. First the noise is removed from vibration signals, then the denoising signals are decomposed by WPT, extract standard deviation coefficients of each level as the input of WNN, the learning rates and momentum factors are used to adjust the network, the method of batch training is applied and it can diagnose fault quickly, which can monitor the condition of main gearbox. Theoretical and practical application shows that this method is effective and feasible, its diagnostic speed is rapid and result is accuracy, which provides a new technical reference for the development of helicopter fault diagnostic systems.