Citation: You, K.; Qiu, G.; Gu, Y.
Rolling Bearing Fault Diagnosis
Using Hybrid Neural Network with
Principal Component Analysis.
Sensors 2022, 22, 8906. https://
doi.org/10.3390/s22228906
Academic Editors: Yangquan Chen,
Subhas Mukhopadhyay,
Nunzio Cennamo, M. Jamal Deen,
Junseop Lee, Simone Morais and
Yolanda Vidal
Received: 8 October 2022
Accepted: 15 November 2022
Published: 17 November 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 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
Rolling Bearing Fault Diagnosis Using Hybrid Neural Network
with Principal Component Analysis
Keshun You, Guangqi Qiu and Yingkui Gu *
School of Mechanical and Electrical Engineering, Jiangxi University of Science and Technology,
Ganzhou 341000, China
* Correspondence: guyingkui@163.com
Abstract:
With the rapid development of fault prognostics and health management (PHM) technology,
more and more deep learning algorithms have been applied to the intelligent fault diagnosis of rolling
bearings, and although all of them can achieve over 90% diagnostic accuracy, the generality and
robustness of the models cannot be truly verified under complex extreme variable loading conditions.
In this study, an end-to-end rolling bearing fault diagnosis model of a hybrid deep neural network
with principal component analysis is proposed. Firstly, in order to reduce the complexity of deep
learning computation, data pre-processing is performed by principal component analysis (PCA) with
feature dimensionality reduction. The preprocessed data is imported into the hybrid deep learning
model. The first layer of the model uses a CNN algorithm for denoising and simple feature extraction,
the second layer makes use of bi-directional long and short memory (BiLSTM) for greater in-depth
extraction of the data with time series features, and the last layer uses an attention mechanism for
optimal weight assignment, which can further improve the diagnostic precision. The test accuracy
of this model is fully comparable to existing deep learning fault diagnosis models, especially under
low load; the test accuracy is 100% at constant load and nearly 90% for variable load, and the test
accuracy is 72.8% at extreme variable load (2.205 N
·
m/s–0.735 N
·
m/s and 0.735 N
·
m/s–2.205 N
·
m/s),
which are the worst possible load conditions. The experimental results fully prove that the model has
reliable robustness and generality.
Keywords:
PHM; intelligent fault diagnosis; complex extreme variable loading; hybrid deep neural
network; robustness and generality
1. Introduction
Due to complex working conditions and frequently changing loads in actual produc-
tion, a large number of mechanical system failures are caused by faults in bearings [
1
]. The
mechanism of bearing damage is very complex; the machine operating environment [
2
],
frequent fluctuations in load [
3
–
5
], and improper installation, etc., can all cause different
types of bearing faults, mainly including abrasion failure, fatigue failure, corrosion failure,
and cavitation failure [
6
]. It is very difficult and unrealistic to analyze and diagnose faults
by only studying the mechanism [
7
], but some studies have modeled bearing dynamics
in terms of the radial internal clearance of rolling bearings as a way of analyzing bearing
failure and life [
8
,
9
], which provide good references. Therefore, we can combine mecha-
nism analyses to research a better intelligent fault diagnosis method. Rolling bearings, as
important rotating parts in machinery and equipment, are also one of the important sources
of faults in machinery and equipment [
10
]. Rolling bearings are one of the most common
and widely used kinds of bearing; therefore, the fault diagnosis method of rolling bearings
has been one of the key technologies in the development of machinery fault diagnosis [
11
].
Fault prognostic and health management (PHM) systems need to have a complete,
practical, intelligent, reliable, and systematic solution for rolling bearing health man-
agement [
12
,
13
], which includes raw data pre-processing, feature value selection and
Sensors 2022, 22, 8906. https://doi.org/10.3390/s22228906 https://www.mdpi.com/journal/sensors
