In production, on-line monitoring, off-line diagnosis and field maintenance should be considered as part of all-round quality control measures.

In this paper, the application of neural network is presented in the control system fault diagnosis. The method of on-line and off-line fault diagnosis are given.

The diagnosis system is able to collect fault data on-line and analyse the fault off-line, which achieves classification of turbine shafting vibration fault and detect early slight fault.