Fault identification in rotating machinery using artificial neural networks

In this paper, an artificial neural network system is designed and employed for fault prediction of rotating machinery systems. Multi-layer feedforward networks, constituted of non-linear neurons, have been employed. A normalization scheme is implemented on the input and output vectors. The performance of the expert structure is optimized to encounter input data with different intensities and non-regular data. More than 40 rotating machinery faults are introduced into the algorithm. To train the network, the data in the vibration identification chart consisting of vibration signals of common rotating machinery faults are used. Computer software is developed to detect machinery faults by using the above techniques and is validated for fault detection of different machinery systems. It is found that the designed network is capable of identifing unknown faults in rotating machinery. The effectiveness of the proposed neural network algorithm is displayed by several tests.