Identification of shallow cracks in rotating systems by utilizing convolutional neural networks and persistence spectrum under constant speed condition

Nima Rezazadeh,Mohammad-Reza Ashory,Shila Fallahy

doi:10.21595/jmeacs.2021.22221

Nima Rezazadeh, Mohammad-Reza Ashory + Show 1 more

Open Access

https://doi.org/10.21595/jmeacs.2021.22221

Copy DOI

Abstract

The positive benefits of early faults detection in rotating systems have led scientists to develop automated methods. Although unbalancing is the most prevalent defect in rotor systems, this fault normally is accompanied by other defects such as crack. In this article, an effective self-acting procedure is addressed in identifying shallow cracks in rotor systems throughout the steady-state operation. To classify rotor systems suffering cracks with three various depths, firstly, healthy and cracked systems are modeled by employing the finite element method (FEM). In the following, systems' vibration signals are calculated in different situations numerically; for pre-processing stage, the persistence spectrum is implemented. Finally, by using a supervised convolutional neural network (CNN), rotor systems are classified by regarding the crack depths. The result of the testing step revealed that this hybrid method has rational capacity in distinguishing shallow cracks in steady-state operation where many other methods are somehow powerless.

Highlights

During the last century has been increased the use of rotating machines noticeably and faults detection methods in these instruments have been witnessed significant improvements
For three cracks’ depths equal to a = 0.12r, a = 0.2r, and a = 0.32r and without crack, operational conditions, i.e. angular velocity, and the innate parameters, i.e. disc’s mass, eccentricity, bearings’ stiffness, damping ratio, shaft’s density, length, diameter, and elasticity modulus have changed randomly, vibration signals are captured during 16 seconds of the steady-state operation
The persistence spectrum of these signals is stored. 85 % of the data were randomly selected for the training phase, and the rest of the images were allocated for the testing step

Summary

Introduction

During the last century has been increased the use of rotating machines noticeably and faults detection methods in these instruments have been witnessed significant improvements. Various types of defects can occur in rotor systems such as unbalance, shaft cracks, misalignment, fluid-induced instability, bearing failure, rub/looseness, blade cracks, and shaft bow, some of them such as shaft cracks are more prevalent, their late detection can bring about catastrophic failures. Since many of these faults in their advanced stages have similar symptoms in the methods that researchers have introduced so far, the priority belongs to the procedures that can detect defects in their very early stages where the negligible abnormality can be seen properly. Parallel with remarkable advancements in the graphical units of computers, scientists tend to deal

Methods

Results

Conclusion