Crack fault identification in rotor shaft with artificial neural network

Based on the truth of the change of the mode shapes of cracked structure and ANN´s strong capability on nonlinear approximation, a novel method by combining modal analysis of cracked rotor system and artificial neural network (ANN) and is proposed for fast and precise identification of crack fault in rotor shaft, including single crack case and multi-crack case. To obtain the specific mode shapes of cracked rotor system, fracture mechanics theory and the energy principle of Paris are introduced into the modeling of finite element (FE) model of the cracked rotor system. Thus, a set of different mode shapes of a rotor system with one or several localized on-edge non-propagating open cracks in different positions and depths will be produced to be fed into the pre-designed ANN model with back-propagation learning algorithm. Then the validation of the method is verified by several selected crack cases. The results show that the trained ANN models have good performance to identify the crack location and depth, single crack and dual cracks, with higher accuracy and efficiency. The idea can be used in fast identification of crack fault in rotating machinery.