Classification of a cracked-rotor system during start-up using Deep learning based on convolutional neural networks

This article addresses an improvement of a classification procedure on cracked rotors through Deep learning based convolutional neural networks (CNNs). At first, rotor-bearing system is modeled by the finite element method (FEM), then throughout its start-up, related time-domain responses are calculated numerically. In following, as pre-processing stage, continuous wavelet transform (CWT) and Short-time Fourier (STFT) applied three various health conditions, i.e. without crack, shallow-cracked, relatively deep-cracked shafts. The plots CWT’s coefficients STFT’s in these classes used input dataset CNNs introduced output. AlexNet with 25 layers employed network. results testing phase demonstrated that not only this expanded has reasonable capacity healthy rotors, but it also can classify different crack depths negligible error.