Citation: Zhu, F.; Liu, C.; Yang, J.;
Wang, S. An Improved MobileNet
Network with Wavelet Energy and
Global Average Pooling for Rotating
Machinery Fault Diagnosis. Sensors
2022, 22, 4427. https://doi.org/
10.3390/s22124427
Academic Editors: Kim Phuc Tran,
Athanasios Rakitzis and Khanh T.
P. Nguyen
Received: 8 May 2022
Accepted: 9 June 2022
Published: 11 June 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
An Improved MobileNet Network with Wavelet Energy and
Global Average Pooling for Rotating Machinery
Fault Diagnosis
Fu Zhu
1,2
, Chang Liu
1,2,
* , Jianwei Yang
1,2
and Sen Wang
1,2
1
Key Laboratory of Advanced Equipment Intelligent Manufacturing Technology of Yunnan Province,
Kunming University of Science & Technology, Kunming 650500, China; 20202203144@stu.kust.edu.cn (F.Z.);
yjw326101@163.com (J.Y.); wangsen0401@kust.edu.cn (S.W.)
2
Faculty of Mechanical & Electrical Engineering, Kunming University of Science & Technology,
Kunming 650500, China
* Correspondence: liuchang3385@gmail.com; Tel.: +86-159-2523-5670
Abstract:
In recent years, neural networks have shown good performance in terms of accuracy and
efficiency. However, along with the continuous improvement in diagnostic accuracy, the number of
parameters in the network is increasing and the models can often only be run in servers with high
computing power. Embedded devices are widely used in on-site monitoring and fault diagnosis.
However, due to the limitation of hardware resources, it is difficult to effectively deploy complex
models trained by deep learning, which limits the application of deep learning methods in engineering
practice. To address this problem, this article carries out research on network lightweight and
performance optimization based on the MobileNet network. The network structure is modified
to make it directly suitable for one-dimensional signal processing. The wavelet convolution is
introduced into the convolution structure to enhance the feature extraction ability and robustness
of the model. The excessive number of network parameters is a challenge for the deployment of
networks and also for the running performance problems. This article analyzes the influence of the
full connection layer size on the total network. A network parameter reduction method is proposed
based on GAP to reduce the network parameters. Experiments on gears and bearings show that
the proposed method can achieve more than 97% classification accuracy under the strong noise
interference of −6 dB, showing good anti-noise performance. In terms of performance, the network
proposed in this article has only one-tenth of the number of parameters and one-third of the running
time of standard networks. The method proposed in this article provides a good reference for the
deployment of deep learning intelligent diagnosis methods in embedded node systems.
Keywords: fault diagnosis; lightweight network; pooling method; wavelet convolution
1. Introduction
In mechanical equipment, key components often fail due to high speed and heavy
load, variable working conditions and other harsh working conditions, resulting in the
failure of the entire transmission system and even major safety accidents. Real-time
condition monitoring [
1
] and fault diagnosis of key components of machinery equipment
are important to ensure the safe and smooth operation of mechanical systems.
Traditional fault diagnosis techniques are usually signal processing methods based on
manual feature extraction, such as Fourier transform [
2
], wavelet transform [
3
], empirical
mode decomposition [
4
] and other methods to extract features to achieve fault diagnosis.
Zhang et al. [
5
] used the kurtosis index and correlation coefficient to construct measurement
indexes. They use the maximum weighted kurtosis index to analyze sensitive patterns
and finally achieve fault diagnosis by extracting fault features. Han et al. [
6
] combined
the Teager energy operator and the signal processing method of complementary ensemble
Sensors 2022, 22, 4427. https://doi.org/10.3390/s22124427 https://www.mdpi.com/journal/sensors
