An Improved Crack Breathing Model and Its Application in Crack Identification for Rotors

The crack breathing model and identification method for rotors using nonlinearity induced by cracks are studied in this work. Firstly, the finite element is utilized to a rotor–bearing system with response-dependent obtain numerical data identification. During modelling, an improved proposed, focused on unreasonable assumption about closure line original position (CCLP) model. Compared model, can reflect nonlinear behavior of better. Secondly, based established super-harmonic features at 1/3 1/2 critical rotating speeds under different locations depths extracted Additionally, from two measurement points used as inputs into artificial neural network Levenberg–Marquardt back-propagation algorithm, corresponding positions outputs. robustness tested examining results levels noise. show that proposed efficient simultaneous depth operating rotors.