A fault diagnostics system for helicopter main gearbox(MGB) was constructed using discrete wavelet transform(DWT) and neural network

and the DWT was used to extract feature vectors and neural network was used for fault distinguish and classification. Firstly

the theory of DWT

Parseval theorem and GRNN were introduced

and then

the flowchart of fault diagnostics system for helicopter MGB was proposed

finally

the system was verified by the vibration data which were collected during helicopter flying

and both generalized regression neural network(GRNN) and back-propagation neural network(BPNN) were used and compared on the experimental platform. Experimental results indicate that the proposed diagnostics system can distinguish the MGB fault better

which would provide a new technical reference for the further development of helicopter MGB fault diagnostics system.