Page 107 - IJOCTA-15-4

An International Journal of Optimization and Control: Theories & Applications
ISSN: 2146-0957 eISSN: 2146-5703
Vol.15, No.4, pp.649-669 (2025)
https://doi.org/10.36922/IJOCTA025170086

RESEARCH ARTICLE

Rolling bearing fault diagnosis method based on GJO–VMD,
multiscale fuzzy entropy, and GSABO–BP neural network

1*
2
1
1†
2†
Jingsong Zhang , Xiaolong Zhou , Soo Siang Yang , Min Keng Tan , Yanzhen Wang , Bin
2
1
Zheng , Jing Zhe , and Haoyu Li 2
1
Faculty of Engineering, Universiti Malaysia Sabah, Kota Kinabalu, Sabah, Malaysia
2
College of Mechanical Engineering, Beihua University, Jilin, Jilin, China
zhangjingsong851@gmail.com, xlzhou1987@163.com, ssyang@ums.edu.my, tanminkeng@ums.edu.my,
wyz200012051@163.com, binz74609@gmail.com, jingzhe20041004@163.com, Lihaoyu030722@163.com
ARTICLE INFO ABSTRACT
Article History:
Received: April 21, 2025 Accurate fault diagnosis of rolling bearings is hindered by the weak nature of
1st revised: May 22, 2025 early fault signals and the limited availability of labeled data, especially under
2 revised: June 9, 2025 small-sample conditions. To overcome these challenges, this paper proposes
Accepted: June 9, 2025 a novel method combining golden jackal optimization (GJO) with improved
Published Online: August 11, 2025 variational mode decomposition (VMD), enhanced feature extraction, and op-
Keywords: timized classification. First, GJO is used to optimally determine the key de-
Rolling bearing composition parameters of VMD, thereby improving the accuracy of vibration
Fault diagnosis signal decomposition. A comprehensive discrimination factor algorithm then
Variational mode decomposition selects fault-sensitive intrinsic mode functions, and the signal is reconstructed
Golden jackal optimization to enhance fault characteristics. Multiscale fuzzy entropy is calculated from
Multiscale fuzzy entropy the reconstructed signals at multiple scales to build distinct state feature vec-
Golden sine subtraction-average-based tors. These vectors are fed into a back-propagation neural network optimized
optimizer–back-propagation neural via the golden sine subtraction-average-based optimizer for precise fault clas-
network sification. The method’s effectiveness is verified through simulation and ex-
perimental data. Compared with conventional approaches, it shows superior
performance in extracting weak fault features and maintaining high diagnostic
accuracy under small-sample scenarios. This integrated framework presents a
robust solution for rolling bearing fault diagnosis.

1. Introduction rolling bearing fault diagnosis is not only theoret-
ically significant but also highly valuable in prac-
Rolling bearings used in today’s machinery tical engineering applications. Recurring faults in
are essential components, especially for rotat- rolling bearings include fatigue spalling and lo-
ing machines such as motors, automobiles, and calized damage on inner/outer rolling elements.
aircraft. 1,2 As key load-bearing elements, the op- The mechanical system generates abnormal vi-
erational status of rolling bearings has a direct im- bration signals when a bearing failure occurs,
pact on the performance and service life of the ma- carrying critical information that reflects fault
5
chinery. Failure in these bearings can lead to a de- characteristics. The extraction of fault features
cline in equipment performance, increased energy from complex vibration signals and their accurate
consumption, and even severe mechanical fail- fault diagnosis and prognostics have become a hot
ures or safety incidents. 3,4 Therefore, research on research topic. The vibration signal analyses can
†
These authors contributed equally to this work.
*Corresponding Author
649