Citation: Zhang, D.; Zhang, L.;
Zhang, N.; Yang, S.; Zhang, Y. Early
Fault Diagnosis of Rolling Bearing
Based on Threshold Acquisition
U-Net. Machines 2023, 11, 119.
https://doi.org/10.3390/
machines11010119
Academic Editor: Ahmed Abu-Siada
Received: 12 December 2022
Revised: 8 January 2023
Accepted: 12 January 2023
Published: 15 January 2023
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Early Fault Diagnosis of Rolling Bearing Based on Threshold
Acquisition U-Net
Dongsheng Zhang
1,2,3,
*, Laiquan Zhang
1
, Naikang Zhang
1
, Shuo Yang
1
and Yuhao Zhang
1
1
School of Construction Machinery, Shandong JiaoTong University, Jinan 250357, China
2
Laboratory of Traffic Construction Equipment and Intelligent Control Engineering of Shandong, Shandong
JiaoTong University, Jinan 250357, China
3
Laboratory of Traffic Construction and Safety Technology Engineering of Shandong, Shandong JiaoTong
University, Jinan 250357, China
* Correspondence: zhangdongsheng@sdjtu.edu.cn
Abstract:
Considering the problem that the early fault signal of rolling bearing is easily interfered
with by background information, such as noise, and it is difficult to extract fault features, a method of
rolling bearing early fault diagnosis based on the threshold acquisition U-Net (TA-UNet) is proposed.
First, to improve the feature extraction ability of U-Net, the channel spatial threshold acquisition
network (CS-TAN) and the dilated convolution module (DCM) based on different dilated rate
combinations are introduced into the U-Net to construct the TA-UNet. Among them, the CS-TAN
can adaptively learn the threshold, reduce the interference of noise in the signal, and the DCM can
improve the multi-scale feature extraction ability of the network. Then, the TA-UNet is used for
early fault diagnosis, and the method is divided into two steps: The model training phase and the
vibration signal fault feature extraction phase. In the first step, additive gaussian white noise is added
to the vibration signal to obtain the noise-added vibration signal, and the TA-UNet is trained to
learn how to denoise the noise-added vibration signal. In the second step, the trained TA-UNet is
used to extract the fault features of vibration signals and diagnose the early fault types of rolling
bearing. The two-step method solves the problem that U-Net, as a supervised neural network, needs
corresponding labeled data to be trained, as it realizes the fault diagnosis of unlabeled data. The
feature extraction capability of the TA-UNet is evaluated by denoising the simulated signal of rolling
bearing. The effectiveness of the proposed diagnostic method is demonstrated by the early fault
diagnosis of open-source datasets.
Keywords: U-Net; threshold denoising; early fault diagnosis
1. Introduction
Vibration monitoring is one of the most effective tools for the fault diagnosis of rolling
bearings [
1
–
3
]. When a local fault occurs, the bearing can generate different characteristic
frequencies depending on which surface the fault affects, and people diagnose the fault by
analyzing the vibration signal [
4
,
5
]. Wear is the inevitable progressive material loss of a
part and is one of the most common forms of failure of parts (e.g., bearing, gear) [
6
]. For
example, foreign matter entering the bearing causes excessive grinding damage, pockmarks,
and abrasions or grooves on the rolling elements and raceway, etc. [
7
], and the failure
gradually becomes serious over time, eventually leading to more serious consequences.
However, the fault feature is very weak in the initial stage of the bearing wear [
8
], and
with the interference of the surrounding environment (collectively referred to as noise in
this paper), the vibration signal becomes complex, resulting in a decrease in the accuracy
of diagnosis [
9
]. Therefore, it is imperative to effectively extract the fault characteristic of
rolling bearing vibration signals under noise background.
In recent years, scholars have proposed many signal processing methods and applied
them to the vibration signal fault feature extraction of rolling bearing. Yu et al. [
10
] used
Machines 2023, 11, 119. https://doi.org/10.3390/machines11010119 https://www.mdpi.com/journal/machines
