Detection of crack and unbalancing in a rotor system using artificial neural network

Abstract

Crack and unbalance are two important effects experienced by a rotor system during its motion. These are a common source of high vibration and undesirable functioning in rotating machinery. The methods of detection of rotor faults have improved from time being such as single fault identification at any given instance, but generally more than one fault is existent in a rotor system simultaneously which has not been discussed so far. In this work, a method is being devised to critically identify unbalance and crack in rotor system using artificial neural networks (ANN) by statistical features. Moreover, a confusion matrix is also obtained by statistical features. Then, by the help confusion matrix, the class of crack and unbalance was decided. Validation and testing of the neural network have been done with simulation data.