[go: up one dir, main page]
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
5.3
2.5
Journals
Applied Sciences
Special Issues
Intelligent Fault Diagnosis and Health Detection of Machinery
applsci-logo
Submit to Applied Sciences Review for Applied Sciences Propose a Special Issue

Journal Menu

► Journal Menu

Journal Browser

► Journal Browser
share announcement

Intelligent Fault Diagnosis and Health Detection of Machinery

A special issue of Applied Sciences (ISSN 2076-3417). This special issue belongs to the section "Computing and Artificial Intelligence".

Deadline for manuscript submissions: closed (20 September 2023) | Viewed by 28056

Special Issue Editors

Dr. Xingxing Jiang

E-Mail Website
Guest Editor
School of Rail Transportation, Soochow University, Suzhou 215131, China
Interests: machinery intelligent fault diagnosis; health monitoring of rotating machines; adaptive signal decomposition
Special Issues, Collections and Topics in MDPI journals
Dr. Xiaojian Yi

E-Mail Website
Guest Editor
School of Mechatronical Engineering, Beijing Institute of Technology, Beijing 100811, China
Interests: system reliability; system design of prognostic and health management; RAMS engineering
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear Colleagues,

Modern machinery is usually characterized by a complex electromechanical or mechanical-electro-liquid system. As these systems become increasingly complex, higher standards of reliability and safety are required. To ensure the reliable operation of machines, it has always been an issue of significance to comprehensively and accurately diagnose the latent faults of the machinery. In recent years, a multitude of techniques for intelligent fault diagnosis and health detection of machinery have been developed and described in the literature. This Special Issue welcomes any original and high quality papers dealing with but are not limited to:

(1) Early weak fault detection method of machines;

(2) Advanced signal processing techniques for feature extraction;

(3) Deep learning–based intelligent fault diagnosis of machines;

(4) Fault detection of machines under varying speed conditions;

(5) Health condition monitoring of electromechanical and mechanical-electro-liquid systems;

(6) Reliability analysis and evaluation of electromechanical and mechanical-electro-liquid systems.

Dr. Xingxing Jiang
Dr. Xiaojian Yi
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Applied Sciences is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Published Papers (17 papers)

Download All Papers
Order results
Result details
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:

Research

25 pages, 24728 KiB  
Article
Application of Knowledge Graph Technology with Integrated Feature Data in Spacecraft Anomaly Detection
by Xiaojian Yi, Peizheng Huang and Shangjie Che
Appl. Sci. 2023, 13(19), 10905; https://doi.org/10.3390/app131910905 - 30 Sep 2023
Viewed by 1795
Abstract
Given the complexity of spacecraft system structures and functions, existing data-driven methods for anomaly detection face issues of insufficient interpretability and excessive dependence on historical data. To address these challenging problems, this paper proposes a method for applying knowledge graph technology with integrated [...] Read more.
Given the complexity of spacecraft system structures and functions, existing data-driven methods for anomaly detection face issues of insufficient interpretability and excessive dependence on historical data. To address these challenging problems, this paper proposes a method for applying knowledge graph technology with integrated feature data in spacecraft anomaly detection. First, the ontology concepts of the spacecraft equipment knowledge graph are designed according to expert knowledge, and then feature data are extracted from the historical operation data of the spacecraft in various states to build a rich spacecraft equipment knowledge graph. Next, spacecraft anomaly event knowledge graphs are constructed based on various types of anomaly features. During spacecraft operation, telemetry data are matched with the feature data in the knowledge graph, enabling anomaly device location and anomaly cause judgment. Experimental results show that this method, which utilizes spacecraft anomaly prior knowledge for anomaly detection and causes interpretation, has high practicality and efficiency. This research demonstrates the promising application prospects of knowledge graph technology in the field of spacecraft anomaly detection. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Knowledge graph-based spacecraft anomaly detection algorithm architecture.</p> Full article ">Figure 2
<p>Knowledge graph-driven spacecraft anomaly detection and processing system framework.</p> Full article ">Figure 3
<p>Technical route of the spacecraft anomaly detection method based on a knowledge graph.</p> Full article ">Figure 4
<p>Entity recognition model.</p> Full article ">Figure 5
<p>Entity annotation example.</p> Full article ">Figure 6
<p>Relationship extraction model.</p> Full article ">Figure 7
<p>Visualization of the data processing workflow.</p> Full article ">Figure 8
<p>Ontology design of a spacecraft equipment knowledge graph.</p> Full article ">Figure 9
<p>Ontology design of a spacecraft abnormal event graph.</p> Full article ">Figure 10
<p>Ontology design of an event chain subgraph.</p> Full article ">Figure 11
<p>Unstructured corpus example (sensitive information removed).</p> Full article ">Figure 12
<p>The interface of doccano for anomaly report text annotation (partial).</p> Full article ">Figure 13
<p>Outlier removal visualization example.</p> Full article ">Figure 14
<p>Missing value imputation visualization example (regression imputation).</p> Full article ">Figure 15
<p>CMG system parameter correlation heatmap.</p> Full article ">Figure 16
<p>Visualization of feature data for each group.</p> Full article ">Figure 17
<p>Knowledge graph visualization.</p> Full article ">Figure 18
<p>Clustering result visualization example.</p> Full article ">Figure 19
<p>The abrupt changes of telemetry parameters (green part).</p> Full article ">Figure 20
<p>Spacecraft CMG system anomaly handling experimental case.</p> Full article ">
21 pages, 8940 KiB  
Article
A Study of Fault Signal Noise Reduction Based on Improved CEEMDAN-SVD
by Sixia Zhao, Lisha Ma, Liyou Xu, Mengnan Liu and Xiaoliang Chen
Appl. Sci. 2023, 13(19), 10713; https://doi.org/10.3390/app131910713 - 26 Sep 2023
Cited by 2 | Viewed by 986
Abstract
In light of the challenges posed by the complex structural characteristics and significant coupling of vibration signals in rotating machinery, this study proposes an adaptive noise reduction method called Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN). Additionally, an enhanced threshold screening [...] Read more.
In light of the challenges posed by the complex structural characteristics and significant coupling of vibration signals in rotating machinery, this study proposes an adaptive noise reduction method called Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN). Additionally, an enhanced threshold screening Singular Value Decomposition (SVD) algorithm is introduced to address the issues pertaining to noise identification and feature extraction in the context of vibration signals from rotating machinery, which are subjected to complex noise interference. The effectiveness of the proposed approach is substantiated through the evaluation of key metrics, such as the signal-to-noise ratio (SNR), as well as the utilization of advanced signal analysis techniques, including Short Time Fourier Transform (STFT) and Continuous Wavelet Transform (CWT). The experimental results validate the finding that the combination of the improved CEEMDAN and the enhanced threshold screening SVD algorithm effectively reduces noise interference in vibration signals from rotating machinery. This integrated denoising approach successfully preserves the informative characteristics of the vibration signals, thereby laying a foundation for the subsequent fault diagnosis of rotating machinery. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Noise Reduction Technology Roadmap.</p> Full article ">Figure 2
<p>Simulation Signal Time-Domain Diagram.</p> Full article ">Figure 3
<p>IMF time-domain graphs of each order after decomposition. (<b>a</b>) EMD decomposition; (<b>b</b>) EEMD decomposition; (<b>c</b>) CEEMD decomposition; (<b>d</b>) CEEMDAN decomposition; (<b>e</b>) CEEMDAN decomposition.</p> Full article ">Figure 4
<p>A comparison of magnitude spectrum between Improved CEEMDAN and CEEMDAN decompositions.</p> Full article ">Figure 5
<p>A comparison of power spectra between Improved CEEMDAN and CEEMDAN decompositions.</p> Full article ">Figure 6
<p>A time–frequency diagram of the original signal; (<b>a</b>) Noiseless STFT; (<b>b</b>) Noiseless CWT; (<b>c</b>) Noise-containing STFT; (<b>d</b>) Noise-containing CWT.</p> Full article ">Figure 6 Cont.
<p>A time–frequency diagram of the original signal; (<b>a</b>) Noiseless STFT; (<b>b</b>) Noiseless CWT; (<b>c</b>) Noise-containing STFT; (<b>d</b>) Noise-containing CWT.</p> Full article ">Figure 7
<p>The time–frequency plot of the noise-containing raw signal after direct SVD thresholding noise reduction; (<b>a</b>) STFT; (<b>b</b>) CWT.</p> Full article ">Figure 8
<p>The time–frequency plot of the noise-containing signal after SVD thresholding noise reduction reconstruction with improved CEEMDAN decomposition; (<b>a</b>) STFT; (<b>b</b>) CWT.</p> Full article ">Figure 9
<p>Signal Denoising Effect Comparison.</p> Full article ">Figure 10
<p>Experimental combine harvester and problem injection locations.</p> Full article ">Figure 11
<p>The measurement point locations of the experimental combine harvester.</p> Full article ">Figure 12
<p>A time-domain diagram of the test signal.</p> Full article ">Figure 13
<p>The decomposition of the IMF of each order using improved CEEMDAN decomposition.</p> Full article ">Figure 14
<p>A time–frequency diagram of the original signal. (<b>a</b>) Original signal STFT; (<b>b</b>) Original signal CWT.</p> Full article ">Figure 15
<p>The time–frequency plot after direct SVD threshold denoising; (<b>a</b>) STFT; (<b>b</b>) CWT.</p> Full article ">Figure 16
<p>The time–frequency plot after improved CEEMDAN decomposition with SVD thresholding noise reduction reconstruction; (<b>a</b>) STFT; (<b>b</b>) CWT.</p> Full article ">
13 pages, 2730 KiB  
Article
Detection of Broken Bars in Induction Motors Using Histogram Analysis of Current Signals
by Veronica Hernandez-Ramirez, Dora-Luz Almanza-Ojeda, Juan-Jose Cardenas-Cornejo, Jose-Luis Contreras-Hernandez and Mario-Alberto Ibarra-Manzano
Appl. Sci. 2023, 13(14), 8344; https://doi.org/10.3390/app13148344 - 19 Jul 2023
Cited by 4 | Viewed by 1554
Abstract
The lifetime of induction motors can be significantly extended by installing diagnostic systems for monitoring their operating conditions. In particular, detecting broken bar failures in motors is important for avoiding the risk of short circuits or other accidents with serious consequences. In the [...] Read more.
The lifetime of induction motors can be significantly extended by installing diagnostic systems for monitoring their operating conditions. In particular, detecting broken bar failures in motors is important for avoiding the risk of short circuits or other accidents with serious consequences. In the literature, many approaches have been proposed for motor fault detection; however, additional generalized methods based on local and statistical analysis could provide a low-complexity and feasible solution in this field of research. The proposed work presents a methodology for detecting one or two broken rotor bars using the sums and differences histograms (SDH) and machine learning classifiers in this context. From the SDH computed in one phase of the motor’s current, nine texture features are calculated for different displacements. Then, all features are used to train two classifiers and to find the best displacements for faults and health identification in the induction motors. A final experimental evaluation considering the best displacements shows an accuracy of 98.16% for the homogeneity feature and a few signal samples used in a decision tree classifier. Additionally, a polynomial regression curve validates the use of 50 samples to obtain an accuracy of 88.15%, whereas the highest performance is achieved for 250 samples. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Proposed methodology for HLT, 1BRB, and 2BRB fault classification.</p> Full article ">Figure 2
<p>Data acquisition module.</p> Full article ">Figure 3
<p>Single -phase current signals of an induction motor for (<b>a</b>) HLT, (<b>b</b>) 1BRB, (<b>c</b>) 2BRB.</p> Full article ">Figure 4
<p>Squirrel cage rotor images: (<b>a</b>) healthy motor, (<b>b</b>) 1BRB, (<b>c</b>) 2BRB.</p> Full article ">Figure 5
<p>Representation of the sliding window used to calculate the SDH in the signal <math display="inline"><semantics><mrow><mi>S</mi><mo>(</mo><mi>n</mi><mo>)</mo></mrow></semantics></math> using <math display="inline"><semantics><mo>Δ</mo></semantics></math> displacements. The displacements to the right compute <math display="inline"><semantics><msub><mi>h</mi><mi>s</mi></msub></semantics></math> and to the left, <math display="inline"><semantics><msub><mi>h</mi><mi>d</mi></msub></semantics></math>.</p> Full article ">Figure 6
<p>Representation of displacements in S. (<b>a</b>) For <math display="inline"><semantics><msub><mi>h</mi><mi>s</mi></msub></semantics></math>, its right neighbor value is added, (<b>b</b>) <math display="inline"><semantics><msub><mi>h</mi><mi>d</mi></msub></semantics></math> is subtracted from its left neighbor value.</p> Full article ">Figure 7
<p>Hyperparameter optimization performance graphs. (<b>a</b>) Objective function model. (<b>b</b>) Min objective vs. function evaluations.</p> Full article ">Figure 8
<p>Classification results for the homogeneity texture (<math display="inline"><semantics><msub><mi>C</mi><mn>7</mn></msub></semantics></math>) of <a href="#applsci-13-08344-t003" class="html-table">Table 3</a>. (<b>a</b>) Confusion matrix for <math display="inline"><semantics><msub><mi>C</mi><mn>7</mn></msub></semantics></math>. (<b>b</b>) Polynomial regression curve for decision tree.</p> Full article ">Figure 9
<p>Confusion matrices obtained for homogeneity feature. (<b>a</b>) K-nearest neighbor. (<b>b</b>) Bagged decision tree.</p> Full article ">
19 pages, 5933 KiB  
Article
A Novel Rolling Bearing Fault Diagnosis Method Based on MFO-Optimized VMD and DE-OSELM
by Yonghua Jiang, Zhuoqi Shi, Chao Tang, Jianan Wei, Cui Xu, Jianfeng Sun, Linjie Zheng and Mingchao Hu
Appl. Sci. 2023, 13(13), 7500; https://doi.org/10.3390/app13137500 - 25 Jun 2023
Cited by 1 | Viewed by 1124
Abstract
Rolling bearings are critical in maintaining smooth operation of rotating machinery and considerably influence its reliability. The signals collected from rolling bearings in field conditions are often subjected to noise, creating a challenge to extract weaker fault features. This paper proposes a rolling [...] Read more.
Rolling bearings are critical in maintaining smooth operation of rotating machinery and considerably influence its reliability. The signals collected from rolling bearings in field conditions are often subjected to noise, creating a challenge to extract weaker fault features. This paper proposes a rolling bearing fault diagnosis method that addresses the above-mentioned problem through the moth-flame optimization algorithm optimized variational mode decomposition (MFO-optimized VMD) and an ensemble differential evolution online sequential extreme learning machine (DE-OSELM). By using the dynamic adaptive weight factor and genetic algorithm cross operator, the optimization accuracy and global optimization ability of the moth-flame optimization (MFO) are improved, and the two basic parameters of VMD decomposition level and quadratic penalty factor are adaptive selected. Since the vibration characteristics of the signal cannot be fully interpreted by a single index, The effective weighted correlation sparsity index (EWCS) is utilized to extract the relevant intrinsic mode functions (IMF) of VMD decomposition and extract their energies as features. In order to improve the classification accuracy, The energy feature set is subsequently inputted into DE-OSELM for training and classification purposes, and the proposed method is assessed via a sample set with four different health states of actual rolling bearings. Our proposed method results are compared with other diagnosis methods, proving its feasibility to diagnose rolling bearing faults with higher classification accuracy. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>The flowchart of the improved MFO.</p> Full article ">Figure 2
<p>The iterative optimization convergence curves of MFO, GWO, and PSO. (<b>a</b>) f1, (<b>b</b>) f2, (<b>c</b>) f3, and (<b>d</b>) f4.</p> Full article ">Figure 2 Cont.
<p>The iterative optimization convergence curves of MFO, GWO, and PSO. (<b>a</b>) f1, (<b>b</b>) f2, (<b>c</b>) f3, and (<b>d</b>) f4.</p> Full article ">Figure 3
<p>(<b>a</b>) Time–domain of the simulation signal. (<b>b</b>) Frequency–domain diagram of the simulation signal.</p> Full article ">Figure 4
<p>The improved MFO convergence curve of the simulation signal.</p> Full article ">Figure 5
<p>The effective component envelope spectrum.</p> Full article ">Figure 6
<p>(<b>a</b>) The envelope spectrum by method-1. (<b>b</b>) The envelope spectrum by method-2.</p> Full article ">Figure 7
<p>The fault diagnosis process based on the MFO-optimized VMD and DE-OSELM.</p> Full article ">Figure 8
<p>The rolling bearing fault simulation test bench.</p> Full article ">Figure 9
<p>The improved MFO convergence curve of the bearing signal.</p> Full article ">Figure 10
<p>(<b>a</b>) original signal; (<b>b</b>) reconstructed signal.</p> Full article ">Figure 11
<p>Comparison of the DE-OSLEM and OSELM classification results.</p> Full article ">Figure 12
<p>Drivetrain dynamics simulator.</p> Full article ">Figure 13
<p>(<b>a</b>) original signal; (<b>b</b>) reconstructed signal.</p> Full article ">Figure 14
<p>A comparison of the DE-OSLEM and OSELM classification results.</p> Full article ">
19 pages, 36702 KiB  
Article
Unsupervised Method Based on Adversarial Domain Adaptation for Bearing Fault Diagnosis
by Yao Li, Rui Yang and Hongshu Wang
Appl. Sci. 2023, 13(12), 7157; https://doi.org/10.3390/app13127157 - 15 Jun 2023
Viewed by 1275
Abstract
This paper contributes to improving a bottleneck residual block-based feature extractor as a set of layers for transforming raw data into features for classification. This structure is utilized to avoid the issues of the deep learning network, such as overfitting problems and low [...] Read more.
This paper contributes to improving a bottleneck residual block-based feature extractor as a set of layers for transforming raw data into features for classification. This structure is utilized to avoid the issues of the deep learning network, such as overfitting problems and low computational efficiency caused by redundant computation, high dimensionality, and gradient vanishing. With this structure, a domain adversarial neural network (DANN), a domain adversarial unsupervised model, and a maximum classifier discrepancy (MCD), a domain adaptation model, have been applied to conduct a binary classification of fault diagnosis data. In addition, a pseudo-label is applied to MCD for comparison with the original one. In comparison, several popular models are selected for transferability estimation and analysis. The experimental results have shown that DANN and MCD with this improved feature extractor have achieved high classification accuracy, with 96.84% and 100%, respectively. Meanwhile, after using the pseudo-label semi-supervised learning, the average classification accuracy of the MCD model increased by 15%, increasing to 94.19%. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Structure of DANN.</p> Full article ">Figure 2
<p>The general process of MCD.</p> Full article ">Figure 3
<p>Process of pseudo-label semi-supervised learning.</p> Full article ">Figure 4
<p>The brief structure of the revised feature extractor.</p> Full article ">Figure 5
<p>Target domain data.</p> Full article ">Figure 6
<p>Different stages of target data.</p> Full article ">Figure 7
<p>The process of data preprocessing.</p> Full article ">Figure 8
<p>Accuracy and loss plot of DANN.</p> Full article ">Figure 9
<p>Accuracies and discrepancy loss plot of MCD and MCD with the pseudo label.</p> Full article ">
21 pages, 14593 KiB  
Article
Research on Fault Diagnosis Algorithm of Ship Electric Propulsion Motor
by Fengxin Ma, Liang Qi, Shuxia Ye, Yuting Chen, Han Xiao and Shankai Li
Appl. Sci. 2023, 13(6), 4064; https://doi.org/10.3390/app13064064 - 22 Mar 2023
Cited by 1 | Viewed by 1316
Abstract
The permanent magnet synchronous motor (PMSM) has been used in electric propulsion and other fields. However, it is prone to the stator winding inter-turn short-circuit, and if no effective measures are taken, the ship’s power system will be paralyzed. To realize intelligent diagnosis [...] Read more.
The permanent magnet synchronous motor (PMSM) has been used in electric propulsion and other fields. However, it is prone to the stator winding inter-turn short-circuit, and if no effective measures are taken, the ship’s power system will be paralyzed. To realize intelligent diagnosis of inter-turn short circuits, this paper proposes an intelligent fault diagnosis method based on improved variational mode decomposition (VMD), multi-scale principal component analysis (PCA) feature extraction, and improved Bi-LSTM. Firstly, the stator current simulation dataset is obtained by using the mathematic model of the inter-turn short-circuit of PMSM, and the parameters of VMD are optimized by the grey wolf algorithm. Then, the data is coarse-grained to obtain multi-scale features, and the main features are selected as the sample data for fault classification by PCA. Subsequently, the Bi-LSTM neural network is used for training and analyzing the data of the sample set and the test set. Finally, the learning rate and the number of hidden-layer nodes of the Bi-LSTM are optimized by the whale algorithm to increase the diagnosis accuracy. Experimental results show that the accuracy of the proposed method for inter-turn short-circuited fault diagnosis is as high as 100%, which confirms the effectiveness of the method. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>The diagnosis process of the proposed method.</p> Full article ">Figure 2
<p>PMSM equivalent model of a phase inter-turn short-circuited fault.</p> Full article ">Figure 3
<p>The structure of the LSTM model.</p> Full article ">Figure 4
<p>The structure of the Bi-LSTM network.</p> Full article ">Figure 5
<p>The time-frequency characteristic method at a scale of 3.</p> Full article ">Figure 6
<p>The flowchart of using the whale algorithm to optimize the hyperparameters of the Bi-LSTM diagnostic model.</p> Full article ">Figure 7
<p>The simulation diagram of the permanent magnet synchronous motor a phase stator winding inter-turn short-circuit.</p> Full article ">Figure 8
<p>(<b>a</b>) The waveform of the stator current signal. (<b>b</b>) The stator current. (<b>c</b>) The simulation of the stator current with a short-circuit turn ratio of 0.2.</p> Full article ">Figure 8 Cont.
<p>(<b>a</b>) The waveform of the stator current signal. (<b>b</b>) The stator current. (<b>c</b>) The simulation of the stator current with a short-circuit turn ratio of 0.2.</p> Full article ">Figure 9
<p>The accuracy in the training process. (<b>a</b>) The accuracy in the training process by unimproved VMD. (<b>b</b>) The accuracy in the training process by improved VMD.</p> Full article ">Figure 10
<p>The confusion matrix. (<b>a</b>) The confusion matrix in the training process by the unimproved VMD. (<b>b</b>) The confusion matrix in the training process by the improved VMD.</p> Full article ">Figure 11
<p>The 2D scatter plot. (<b>a</b>) The 2D scatter plot in the training process by the unimproved VMD. (<b>b</b>) The 2D scatter plot in the training process by the improved VMD.</p> Full article ">Figure 12
<p>The accuracy of the training set using the improved multiscale feature extraction.</p> Full article ">Figure 13
<p>The confusion matrix of using the improved multiscale feature extraction.</p> Full article ">Figure 14
<p>The 2D scatter plot using the improved multiscale feature extraction.</p> Full article ">Figure 15
<p>The accuracy of the training set using the improved Bi-LSTM network.</p> Full article ">Figure 16
<p>The confusion matrix using the improved Bi-LSTM network.</p> Full article ">Figure 17
<p>The 2D scatter plot using the improved Bi-LSTM network.</p> Full article ">
16 pages, 1764 KiB  
Article
Reliability Optimization Design of Diesel Engine System Based on the GO Method
by Yuhang Cui, Huina Mu, Xiaojian Yi and Shijie Wei
Appl. Sci. 2023, 13(6), 3727; https://doi.org/10.3390/app13063727 - 15 Mar 2023
Viewed by 1310
Abstract
A reliability optimization design method based on the goal-oriented (GO) method is proposed in this study to tackle the problem of engine reliability optimization design. This proposed method considers a V-type diesel engine as the research object. Firstly, the reliability modeling and evaluation [...] Read more.
A reliability optimization design method based on the goal-oriented (GO) method is proposed in this study to tackle the problem of engine reliability optimization design. This proposed method considers a V-type diesel engine as the research object. Firstly, the reliability modeling and evaluation of diesel engines are conducted by the GO method. Secondly, the functional reliability is assigned according to the difference in diesel engine function. Finally, a three-objective diesel reliability optimization design model is constructed with the optimization objectives of maximizing robustness and reliability and minimizing cost. Then, an NSGA-II-PSO hybrid algorithm based on the GO method is designed to handle the problem, and the design interval of unit reliability is obtained. The results of the case study demonstrate that this method not only meets the requirements of reliability design but also achieves the purpose of reliability optimization design, providing a reference for other types of equipment. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Process of diesel reliability optimization design based on the GO method.</p> Full article ">Figure 2
<p>Operating principle of the diesel engine system.</p> Full article ">Figure 3
<p>The first-level GO diagram model of the diesel engine system. (<b>a</b>) First-level GO diagram model of the cooling system; (<b>b</b>) First-level GO diagram model of the lubrication system; (<b>c</b>) First-level GO diagram model of the fuel supply system; and (<b>d</b>) First-level GO diagram model of the inlet, exhaust, and turbocharge system.</p> Full article ">Figure 4
<p>GO diagram model of the diesel engine system.</p> Full article ">Figure 5
<p>Flowchart of NSGA-II-PSO algorithm.</p> Full article ">Figure 6
<p>Pareto solution of the diesel engine reliability optimization design.</p> Full article ">
15 pages, 10039 KiB  
Article
Dynamic Simulation Model-Driven Fault Diagnosis Method for Bearing under Missing Fault-Type Samples
by Junqing Ma, Xingxing Jiang, Baokun Han, Jinrui Wang, Zongzhen Zhang and Huaiqian Bao
Appl. Sci. 2023, 13(5), 2857; https://doi.org/10.3390/app13052857 - 23 Feb 2023
Cited by 6 | Viewed by 1555
Abstract
Existing generative adversarial networks (GAN) have potential in data augmentation and in the intelligent fault diagnosis of bearings. However, most relevant studies only focus on the fault diagnosis of rotating machines with sufficient fault-type samples, and some rare fault-type samples may be missing [...] Read more.
Existing generative adversarial networks (GAN) have potential in data augmentation and in the intelligent fault diagnosis of bearings. However, most relevant studies only focus on the fault diagnosis of rotating machines with sufficient fault-type samples, and some rare fault-type samples may be missing in training in practical engineering. To address those deficiencies, this paper presents an intelligent fault diagnosis method based on the dynamic simulation model and Wasserstein generative adversarial network with gradient normalization (WGAN-GN). The dynamic simulation model of bearing faults is constructed to obtaining simulation signals to replace and complement the missing fault samples, which are combined with the measured signals as training data and then input into the proposed WGAN-GN model for expanding and enhancing the data. To test the effectiveness of the simulated samples, a fault classification model constructed by stacked autoencoders (SAE) is used to classify the enhanced dataset. According to the results, the proposed model performs well when used to diagnose faults under missing samples and is preferable to other methods. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Diagrammatically representing the simulated rotor-bearing system.</p> Full article ">Figure 2
<p>Schematic of the local defect area: (<b>a</b>) Outer race fault, (<b>b</b>) Inner race fault, (<b>c</b>) Ball fault.</p> Full article ">Figure 3
<p>Time−domain waveforms of simulation data.</p> Full article ">Figure 4
<p>Generative adversarial networks.</p> Full article ">Figure 5
<p>Structure of WGAN-GN.</p> Full article ">Figure 6
<p>Fault diagnosis model design.</p> Full article ">Figure 7
<p>Rotating machinery fault diagnosis experimental device.</p> Full article ">Figure 8
<p>Comparison between raw and generated signals for bearings: (<b>a</b>) Raw signals, (<b>b</b>) Generated signals.</p> Full article ">Figure 9
<p>Extracted features of the generator for NC samples: (<b>a</b>) the 1st layer, (<b>b</b>) the 2nd layer, (<b>c</b>) the 3rd layer, (<b>d</b>) the output layer, (<b>e</b>) the raw sample.</p> Full article ">Figure 9 Cont.
<p>Extracted features of the generator for NC samples: (<b>a</b>) the 1st layer, (<b>b</b>) the 2nd layer, (<b>c</b>) the 3rd layer, (<b>d</b>) the output layer, (<b>e</b>) the raw sample.</p> Full article ">Figure 10
<p>Accuracies of the two datasets using the three methods: (<b>a</b>) Dataset A, (<b>b</b>) Dataset B.</p> Full article ">Figure 11
<p>Dimensionality reduction visualization results on Dataset A: (<b>a</b>) GAN, (<b>b</b>) WGAN. (<b>c</b>) WGAN-GN.</p> Full article ">Figure 12
<p>Dimensionality reduction visualization results on dataset B: (<b>a</b>) GAN. (<b>b</b>) WGAN. (<b>c</b>) WGAN-GN.</p> Full article ">Figure 13
<p>Comparison between real and generated signals for dataset C: (<b>a</b>) Raw signals, (<b>b</b>) Generated signals.</p> Full article ">Figure 14
<p>Extracted features of generator for RF samples: (<b>a</b>) The first layer, (<b>b</b>) The second layer, (<b>c</b>) The third layer, (<b>d</b>) The output layer, (<b>e</b>) The raw sample.</p> Full article ">Figure 14 Cont.
<p>Extracted features of generator for RF samples: (<b>a</b>) The first layer, (<b>b</b>) The second layer, (<b>c</b>) The third layer, (<b>d</b>) The output layer, (<b>e</b>) The raw sample.</p> Full article ">Figure 15
<p>Accuracies of the datasets C using the three methods.</p> Full article ">Figure 16
<p>Dimensionality reduction visualization results on Dataset C: (<b>a</b>) GAN. (<b>b</b>) WGAN. (<b>c</b>) WGAN-GN.</p> Full article ">
16 pages, 4470 KiB  
Article
An Experimental Setup to Detect the Crack Fault of Asymmetric Rotors Based on a Deep Learning Method
by Chongyu Wang, Zhaoli Zheng, Ding Guo, Tianyuan Liu, Yonghui Xie and Di Zhang
Appl. Sci. 2023, 13(3), 1327; https://doi.org/10.3390/app13031327 - 19 Jan 2023
Cited by 10 | Viewed by 1667
Abstract
Crack is a common fault of rotor systems. The research on crack fault detection methods is mainly divided into numerical and experimental studies. In numerical research, the current fault detection algorithms based on deep learning are mostly applied to bearings and gearboxes, and [...] Read more.
Crack is a common fault of rotor systems. The research on crack fault detection methods is mainly divided into numerical and experimental studies. In numerical research, the current fault detection algorithms based on deep learning are mostly applied to bearings and gearboxes, and there are few studies on rotor fault diagnosis. In experimental research, the rotors used in an experiment are mostly single-span rotors. However, there are complex structures such as multi-span rotor systems in the actual industrial field. Thus, the fault detection algorithms that have been successfully applied on single-span rotors have not been verified on complex rotor systems. To obtain a fault signal close to the actual asymmetric shaft system of an asymmetric rotor system and validate the fault detection method, the crack fault detection platform is designed and built independently. We measure the vibration signals of three channels under five working conditions and establish an intelligent detection method for crack location based on a residual network. The factors that influence fault detection performance are analyzed, and the influence laws are discussed. Results show that the accuracy and anti-noise performance of the proposed method are higher than those of the commonly used machine learning. The average accuracy is 100% when SNR (signal-to-noise ratio) is greater than or equal to −2 dB, and the average accuracy is 98.2% when SNR is −4 dB. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Schematic diagram of the experimental platform for crack fault detection in asymmetric shafts.</p> Full article ">Figure 2
<p>Experimental platform for crack fault detection in asymmetric shafts.</p> Full article ">Figure 3
<p>Measuring system.</p> Full article ">Figure 4
<p>Diagram of asymmetric shaft system; (<b>a</b>) dual-disk rotor; (<b>b</b>) single disk rotor; (<b>c</b>) generator rotor.</p> Full article ">Figure 5
<p>Physical diagram of a healthy rotor and a cracked rotor: (<b>a</b>) healthy rotor; (<b>b</b>) cracked rotor.</p> Full article ">Figure 6
<p>Overlap sampling.</p> Full article ">Figure 7
<p>Fault detection model based on a convolutional neural network.</p> Full article ">Figure 8
<p>Comparison of different algorithms in crack location detection.</p> Full article ">Figure 9
<p>Confusion matrix of this model.</p> Full article ">Figure 10
<p>Visualization of crack location detection process: (<b>a</b>) residual layer 1; (<b>b</b>) residual layer 2; (<b>c</b>) residual layer 3; (<b>d</b>) residual layer 4.</p> Full article ">Figure 11
<p>Influence of SNR on accuracy.</p> Full article ">Figure 12
<p>Confusion matrix under different SNR: (<b>a</b>) SNR = −6; (<b>b</b>) SNR = −4.</p> Full article ">Figure 13
<p>Influence of training set sample number on accuracy.</p> Full article ">Figure 14
<p>Confusion matrix under different numbers of training samples: (<b>a</b>) training samples = 300; (<b>b</b>) training samples = 500.</p> Full article ">
15 pages, 4887 KiB  
Article
Blade Crack Diagnosis Based on Blade Tip Timing and Convolution Neural Networks
by Guangya Zhu, Chongyu Wang, Wei Zhao, Yonghui Xie, Ding Guo and Di Zhang
Appl. Sci. 2023, 13(2), 1102; https://doi.org/10.3390/app13021102 - 13 Jan 2023
Cited by 2 | Viewed by 1985
Abstract
The diagnosis of blade crack faults is critical to ensuring the safety of turbomachinery. Blade tip timing (BTT) is a non-contact vibration displacement measurement technique, which has been extensively studied for blade vibration condition monitoring recently. The fault diagnosis methods based on deep [...] Read more.
The diagnosis of blade crack faults is critical to ensuring the safety of turbomachinery. Blade tip timing (BTT) is a non-contact vibration displacement measurement technique, which has been extensively studied for blade vibration condition monitoring recently. The fault diagnosis methods based on deep learning can be summarized as studying the internal logical relationship of data, automatically mining features, and intelligently identifying faults. This research proposes a crack fault diagnostic method based on BTT measurement data and convolutional neural networks (CNNs) for the crack fault detection of blades. There are two main aspects: the numerical analysis of the rotating blade crack fault diagnosis and the experimental research in rotating blade crack fault diagnosis. The results show that the method outperforms many other traditional machine learning models in both numerical models and tests for diagnosing the depth and location of blade cracks. The findings of this study contribute to the real-time online crack fault diagnosis of blades. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>The complete closed block diagram.</p> Full article ">Figure 2
<p>The lumped parameter model.</p> Full article ">Figure 3
<p>Method for obtaining blade tip arrival time.</p> Full article ">Figure 4
<p>Convolutional neural network structure.</p> Full article ">Figure 5
<p>BTT test system structure.</p> Full article ">Figure 6
<p>Test blades and crack locations.</p> Full article ">Figure 7
<p>Prediction accuracy of blade crack location by different deep learning models (case 1: No system deviation and No noise; case 2: System deviation and No noise; case 3: Noise and No system deviation; case 4: System deviation and Noise).</p> Full article ">Figure 8
<p>Prediction accuracy of blade crack depth by different deep learning models (case 1: No system deviation and No noise; case 2: System deviation and No noise; case 3: Noise and No system deviation; case 4: System deviation and Noise).</p> Full article ">Figure 9
<p>Confusion matrix of blade crack location detection results.</p> Full article ">Figure 10
<p>Regression curve of blade crack degree detection results.</p> Full article ">Figure 11
<p>Campbell diagram of modal frequency of model blade.</p> Full article ">Figure 12
<p>Campbell diagram of variable speed strain signal based on wavelet analysis.</p> Full article ">Figure 13
<p>The first three orders of natural frequency of the blade at different crack depths.</p> Full article ">Figure 14
<p>Prediction accuracy of different deep learning models.</p> Full article ">Figure 15
<p>Detection results of crack location and depth of 1500 rpm blade.</p> Full article ">
20 pages, 9361 KiB  
Article
Research on Interference Mechanism of 25 Hz Phase Sensitive Track Equipment from Unbalanced Current
by Lei Wang, Benchao Zhu, Jingcao Chen, Mingchao Zhou, Zhengyan Liu, Jinyan Li and Chunmei Xu
Appl. Sci. 2023, 13(2), 1033; https://doi.org/10.3390/app13021033 - 12 Jan 2023
Cited by 1 | Viewed by 1493
Abstract
In a 25-Hz phase-sensitive track circuit, traction backflow is unevenly distributed in the two rails, resulting in interference caused by the 50 Hz unbalanced current, which leads to misoperation of relays and other equipment in the circuit. Focusing on the mechanism of unbalanced [...] Read more.
In a 25-Hz phase-sensitive track circuit, traction backflow is unevenly distributed in the two rails, resulting in interference caused by the 50 Hz unbalanced current, which leads to misoperation of relays and other equipment in the circuit. Focusing on the mechanism of unbalanced current generation, this paper probes the causes of track circuit equipment interference and innovatively analyzes the mechanism of the choke transformer and relay affected, in order to find a method to suppress the interference of the 25 Hz phase-sensitive track equipment. Firstly, the mechanism of unbalanced current generation is explained, and the influence of the unbalanced impulse current on the choke transformer and binary two-bit relay is analyzed. Secondly, the DC magnetic bias, the second side voltage of the choke transformer and the excitation current, flux density, core loss of choke transformer and relay under a different unbalance impulse current are simulated. Then, the unbalanced current simulation test, unbalanced current test during driving and grounding wire test are carried out. Finally, it is concluded that the unbalanced impulse current causes magnetic saturation of the choke transformer, then affects voltage sag of the relay coil, resulting in misoperation of equipment. The conclusions of this paper can play an important guiding role in studying the influence of unbalanced current and restraining the interference of the 25 Hz phase-sensitive track circuit. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Traction current wheel rail circuit.</p> Full article ">Figure 2
<p>Magnetic field coupling in adjacent orbit.</p> Full article ">Figure 3
<p>Signal current and traction current.</p> Full article ">Figure 4
<p>Equivalent circuit of unbalanced impulse current invading choke transformer.</p> Full article ">Figure 5
<p>DC bias of choke transformer. (<b>a</b>) The change of magnetic flux curve when the impulse current flows into the transformer; (<b>b</b>) The working characteristic curve of the transformer under normal working condition; (<b>c</b>) The change of excitation waveform when the impulse current flows into the transformer.</p> Full article ">Figure 6
<p>Choke transformer circuit and equivalent circuit.</p> Full article ">Figure 7
<p>Waveform of excitation current under different DC interference sources.</p> Full article ">Figure 8
<p>Excitation current of choke transformer under different DC components.</p> Full article ">Figure 9
<p>Magnetic flux density nephogram under different DC components.</p> Full article ">Figure 10
<p>Loss of choke transformer core under different DC components.</p> Full article ">Figure 11
<p>Secondary voltage waveform under different DC components.</p> Full article ">Figure 12
<p>Impulse current interference source waveform.</p> Full article ">Figure 13
<p>Relay coil flux density.</p> Full article ">Figure 14
<p>Relay core loss.</p> Full article ">Figure 15
<p>Unbalanced current simulation test circuit.</p> Full article ">Figure 16
<p>Choke transformer waveform without interference.</p> Full article ">Figure 17
<p>Receiver choke transformer waveform in case of interference and misoperation. (<b>a</b>) Receiver choke transformer waveform in case of interference; (<b>b</b>) Receiver choke transformer waveform in case of misoperation.</p> Full article ">Figure 18
<p>Receiver relay waveform in case of interference and misoperation. (<b>a</b>) Receiver relay waveform in case of interference; (<b>b</b>) Receiver relay waveform in case of misoperation.</p> Full article ">Figure 19
<p>Layout of field measurement equipment.</p> Full article ">Figure 20
<p>Unbalanced current in track circuits under different conditions. (<b>a</b>) Monitoring road section waveform when small unbalanced current is generated in track circuit; (<b>b</b>) Adjacent track section waveform when small unbalanced current is generated in track circuit; (<b>c</b>) Monitoring road section waveform when transient impulse unbalance current exists; (<b>d</b>) Adjacent track section waveform when transient impulse unbalance current exists; (<b>e</b>) Monitoring road section waveform when steady-state unbalanced current exists; (<b>f</b>) Adjacent track section waveform when steady-state unbalanced current exists.</p> Full article ">Figure 21
<p>Field probe layout diagram.</p> Full article ">Figure 22
<p>Choke transformer and relay grounding wire test waveform. (<b>a</b>) The track circuit is not grounded; (<b>b</b>) The position of grounding wire is close to the monitored position; (<b>c</b>) The grounding wire is far from the monitored position; (<b>d</b>) Waveform distortion during grounding wire hanging.</p> Full article ">
19 pages, 4046 KiB  
Article
Feature Mining and Sensitivity Analysis with Adaptive Sparse Attention for Bearing Fault Diagnosis
by Qinglei Jiang, Binbin Bao, Xiuqun Hou, Anzheng Huang, Jiajie Jiang and Zhiwei Mao
Appl. Sci. 2023, 13(2), 718; https://doi.org/10.3390/app13020718 - 4 Jan 2023
Cited by 5 | Viewed by 1817
Abstract
Bearing fault diagnosis for equipment-safe operation has a crucial role. In recent years, more achievements have been made in bearing fault diagnosis. However, for the fault diagnosis model, the representation and sensitivity of bearing fault features have a great influence on the diagnosis [...] Read more.
Bearing fault diagnosis for equipment-safe operation has a crucial role. In recent years, more achievements have been made in bearing fault diagnosis. However, for the fault diagnosis model, the representation and sensitivity of bearing fault features have a great influence on the diagnosis output results; thus, the attention mechanism is particularly important for the selection of features. However, global attention focuses on all sequences, which is computationally expensive and not ideal for fault diagnosis tasks. The local attention mechanism ignores the relationship between non-adjacent sequences. To address the respective shortcomings of global attention and local attention, an adaptive sparse attention network is proposed in this paper to filter fault-sensitive information by soft threshold filtering. In addition, the effects of different signal representation domains on fault diagnosis results are investigated to filter out signal representation forms with better performance. Finally, the proposed adaptive sparse attention network is applied to cross-working conditions diagnosis of bearings. The adaptive sparse attention mechanism focuses on the signal characteristics of different frequency bands for different fault types. The proposed network model achieves better overall performance when comparing the cross-conditions diagnosis accuracy and model convergence speed. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>LSTM memory unit.</p> Full article ">Figure 2
<p>(<b>a</b>) Attention structure. (<b>b</b>) Adaptive sparse attention structure.</p> Full article ">Figure 3
<p>A deep learning architecture for fault diagnosis based on the adaptive sparse attention mechanism.</p> Full article ">Figure 4
<p>Flowchart of fault diagnosis method.</p> Full article ">Figure 5
<p>CWRU bearing test bench.</p> Full article ">Figure 6
<p>Test results of each test. (where cps is the diagnostic accuracy with parameter Convolution parameter sharing, cpi is the diagnostic accuracy with parameter Convolution parameter independence, cps-cross is the diagnostic accuracy with parameter cps and across working conditions, and cpi-cross is the diagnostic accuracy with parameter cpi and across working conditions).</p> Full article ">Figure 7
<p>Cross-condition diagnostic performance of different methods under three input data.</p> Full article ">Figure 8
<p>Loss trend graph of the training set.</p> Full article ">Figure 9
<p>The average attention of each method to the bearing envelope spectrum.</p> Full article ">Figure 10
<p>Average attention of each working condition.</p> Full article ">Figure 11
<p>Visualization of attention weights for each working condition.</p> Full article ">
18 pages, 4506 KiB  
Article
Structural Damage Detection Based on One-Dimensional Convolutional Neural Network
by Zhigang Xue, Chenxu Xu and Dongdong Wen
Appl. Sci. 2023, 13(1), 140; https://doi.org/10.3390/app13010140 - 22 Dec 2022
Cited by 3 | Viewed by 1736
Abstract
This paper proposes a structural damage detection method based on one-dimensional convolutional neural network (CNN). The method can automatically extract features from data to detect structural damage. First, a three-layer framework model was designed. Second, the displacement data of each node was collected [...] Read more.
This paper proposes a structural damage detection method based on one-dimensional convolutional neural network (CNN). The method can automatically extract features from data to detect structural damage. First, a three-layer framework model was designed. Second, the displacement data of each node was collected under the environmental excitation. Then, the data was transformed into the interlayer displacement to form a damage dataset. Third, in order to verify the feasibility of the proposed method, the damage datasets were divided into three categories: single damage dataset, multiple damage dataset, and damage degree dataset. The three types of damage dataset can be classified by the convolutional neural network. The results showed that the recognition accuracy is above 0.9274. Thereafter, a visualization tool called “t-SNE” was employed to visualize the raw data and the output data of the convolutional neural network. The results showed that the feature extraction ability of CNN is excellent. However, there are many hidden layers in a CNN. The outputs of these hidden layers are invisible. In the last section, the outputs of hidden layers are visualized to understand how the convolutional neural networks work. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>The overall framework.</p> Full article ">Figure 2
<p>The model and the number of beams, columns, and joints.</p> Full article ">Figure 3
<p>Sample generation.</p> Full article ">Figure 4
<p>Convolution operation diagram.</p> Full article ">Figure 5
<p>Activation function.</p> Full article ">Figure 6
<p>Training accuracy and valid accuracy.</p> Full article ">Figure 7
<p>The two-dimensional display of data. (<b>a</b>) raw data (<b>b</b>) output data of CNN.</p> Full article ">Figure 8
<p>Training accuracy and valid accuracy.</p> Full article ">Figure 9
<p>The two-dimensional display of data. (<b>a</b>) raw data (<b>b</b>) output data of CNN.</p> Full article ">Figure 10
<p>Training accuracy and valid accuracy.</p> Full article ">Figure 11
<p>The two-dimensional display of data. (<b>a</b>) raw data (<b>b</b>) output data of CNN.</p> Full article ">Figure 12
<p>The visualization of convolution kernels in the first convolutional layer. (<b>a</b>) Kernel 1 (<b>b</b>) Kernel 2.</p> Full article ">Figure 13
<p>The curve of each channel. (<b>a</b>) Kernel 1 (<b>b</b>) Kernel 2.</p> Full article ">Figure 13 Cont.
<p>The curve of each channel. (<b>a</b>) Kernel 1 (<b>b</b>) Kernel 2.</p> Full article ">Figure 14
<p>The visualization of convolution kernels in the second convolutional layer. (<b>a</b>) Kernel 1 (<b>b</b>) Kernel 2.</p> Full article ">Figure 15
<p>The curve of the first twelve channels. (<b>a</b>) Kernel 1 (<b>b</b>) Kernel 2.</p> Full article ">Figure 15 Cont.
<p>The curve of the first twelve channels. (<b>a</b>) Kernel 1 (<b>b</b>) Kernel 2.</p> Full article ">Figure 16
<p>Input data.</p> Full article ">Figure 17
<p>Output of the first convolutional layer.</p> Full article ">Figure 18
<p>Output of the first pooling layer.</p> Full article ">
18 pages, 12233 KiB  
Article
An Accelerated Degradation Durability Evaluation Model for the Turbine Impeller of a Turbine Based on a Genetic Algorithms Back-Propagation Neural Network
by Xiaojian Yi, Zhezhe Wang, Shulin Liu, Xinrong Hou and Qing Tang
Appl. Sci. 2022, 12(18), 9302; https://doi.org/10.3390/app12189302 - 16 Sep 2022
Cited by 3 | Viewed by 1396
Abstract
Durability evaluation plays an important role in product operation and maintenance during the design stage. In order to ensure a long life, high reliability, and short development cycle, an accelerated degradation durability evaluation model for the turbine impeller of a turbine based on [...] Read more.
Durability evaluation plays an important role in product operation and maintenance during the design stage. In order to ensure a long life, high reliability, and short development cycle, an accelerated degradation durability evaluation model for the turbine impeller of a turbine based on a genetic algorithms back-propagation neural network is established. Based on the proposed model, we discuss two types of practical problems. One is the matching problem of the component strengthening test and whole machine system test. The other is the design problem of two kinds of bench tests. All in all, this work not only proposes a durability evaluation model to effectively solve the current turbine durability evaluation problems, but it also provides a feasible research idea for similar problems. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Stress distribution of turbine blades under centrifugal loads and thermal loads.</p> Full article ">Figure 2
<p>The 120-h turbocharger standard structure assessment test profile.</p> Full article ">Figure 3
<p><span class="html-italic">C</span> − <span class="html-italic">R</span><sup>2</sup> relation curve.</p> Full article ">Figure 4
<p>The 120 h test endurance limit–time curve.</p> Full article ">Figure 5
<p>The 350 h test endurance limit–time curve.</p> Full article ">Figure 6
<p>Optimal individual fitness curve.</p> Full article ">Figure 7
<p>Training error curve.</p> Full article ">Figure 8
<p>Model’s GOF.</p> Full article ">
15 pages, 3383 KiB  
Article
Data Augmentation in 2D Feature Space for Intelligent Weak Fault Diagnosis of Planetary Gearbox Bearing
by Rui Yang, Zenghui An, Weiling Huang and Rijun Wang
Appl. Sci. 2022, 12(17), 8414; https://doi.org/10.3390/app12178414 - 23 Aug 2022
Cited by 3 | Viewed by 1364
Abstract
Quickly detecting and accurately diagnosing early bearing faults is the key to ensuring the stable operation of high-precision equipment. In actual industrial applications, it is common to face the issues of big data and poor fault identification accuracy. To accurately and automatically realize [...] Read more.
Quickly detecting and accurately diagnosing early bearing faults is the key to ensuring the stable operation of high-precision equipment. In actual industrial applications, it is common to face the issues of big data and poor fault identification accuracy. To accurately and automatically realize the diagnostics of rolling bearings, a convolutional neural network algorithm and fault feature enhancement method is proposed. A two-dimensional space feature extraction method based on the Cyclostationary theory and wavelet transform shows good results in noise suppression. Firstly, the cyclic demodulation of wavelet transform coefficients is performed on bearing vibration signals to convert one-dimensional vibration data into a two-dimensional spectrogram for enhancing the weak fault feature. Secondly, the image segmentation theory is introduced, which can obtain more data and improve the calculation accuracy and efficiency on the basis of data dimension reduction. Finally, the augmented 2D spectrograms are inputted into a convolutional neural network. Through the analysis of the actual planetary gearbox bearing data, and compared with other mainstream intelligence algorithms, the effectiveness and superiority of this method are verified. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>The fully connected layer.</p> Full article ">Figure 2
<p>Image segmentation based on the local overlap pattern.</p> Full article ">Figure 3
<p>The flowchart of the novel proposed algorithm.</p> Full article ">Figure 4
<p>Picture of the test rig.</p> Full article ">Figure 5
<p>Wavelet cycle frequency spectrum of each operating condition (<b>a</b>) Normal condition; (<b>b</b>) Inner-race Fault; (<b>c</b>) Outer-race Fault; (<b>d</b>) Ball Fault.</p> Full article ">Figure 6
<p>Partial enlarged view of wavelet cycle frequency spectrogram; (<b>a</b>) Normal condition; <span class="html-italic">f<sub>r</sub></span> is the rotation frequency of planet carrier, <span class="html-italic">f<sub>s</sub></span> is the rotation frequency of sun gear; (<b>b</b>) Inner-race Fault; <span class="html-italic">f<sub>ib</sub></span> is the planetary bearing inner race fault characteristic frequency; <math display="inline"><semantics> <mrow> <msubsup> <mi>f</mi> <mi>p</mi> <mrow> <mrow> <mo>(</mo> <mi>o</mi> <mo>)</mo> </mrow> </mrow> </msubsup> </mrow> </semantics></math> is the absolute rotation frequency of planetary gear; (<b>c</b>) Outer-race Fault; <span class="html-italic">f<sub>ob</sub></span> is the planetary bearing outer race fault characteristic frequency; (<b>d</b>) Ball Fault; <span class="html-italic">f<sub>rb</sub></span> is the planetary bearing rolling element fault characteristic frequency.</p> Full article ">Figure 7
<p>Weight iteration update times and classification accuracy graph; The blue, orange, and gray bars represent the results of executing the program three times at different iterations.</p> Full article ">Figure 8
<p>Visual graph after state feature learning by t-SNE.</p> Full article ">Figure 9
<p>Evaluate classifiers performance by comparison.</p> Full article ">Figure 10
<p>The learned features of two-dimensional visualization maps by comparison.</p> Full article ">
26 pages, 3837 KiB  
Article
Reliability Modeling and Analysis of a Diesel Engine Design Phase Based on 4F Integration Technology
by Meng Zhang, Shuangfeng Liu, Xinrong Hou, Haiping Dong, Chunsheng Cui and Yafen Li
Appl. Sci. 2022, 12(13), 6513; https://doi.org/10.3390/app12136513 - 27 Jun 2022
Cited by 6 | Viewed by 2110
Abstract
As one of the most important components within a vehicle, diesel engines have high requirements for reliability due to the harsh operating environments. However, previous studies have mainly focused on the reliability assessment of diesel engines, while less research has been conducted on [...] Read more.
As one of the most important components within a vehicle, diesel engines have high requirements for reliability due to the harsh operating environments. However, previous studies have mainly focused on the reliability assessment of diesel engines, while less research has been conducted on the modeling of the diesel engine reliability analysis and its management. For this reason, this paper proposes a comprehensive method for reliability analysis and its management based on the use of 4F integration technology in the early stages of diesel engine design. First of all, an expert group used FEMCA (failure mode, effects and criticality analysis) and FHA (functional hazard analysis) to find the most harmful level of fault mode. At the same time, a new method for the repair of dynamic fault trees to find the weak links at the component level was developed. Finally, a FRACAS (fracture report analysis and corrective action system) was used during the above analysis process. By applying this method to the reliability assessment of a diesel engine in the design stage, the problems of failure information feedback and the reuse of failure information in the actual reliability assessment can be solved. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>The RDFTA OR logic gate.</p> Full article ">Figure 2
<p>The RDFTA AND logic gate.</p> Full article ">Figure 3
<p>The transformation of the CSP gate into a Markov model.</p> Full article ">Figure 4
<p>The transformation of the feedback gate into a Markov model.</p> Full article ">Figure 5
<p>The transformation of the priority AND logic gate into Markov models.</p> Full article ">Figure 6
<p>The reliability analysis process framework of the RDFTA method.</p> Full article ">Figure 7
<p>The 4F integration technology analysis process.</p> Full article ">Figure 8
<p>The FRACAS analysis process within 4F integration technology.</p> Full article ">Figure 9
<p>A schematic diagram of the FMECA–FHA integrated model.</p> Full article ">Figure 10
<p>A schematic diagram of the diesel engine composition system.</p> Full article ">Figure 11
<p>The corresponding relationships between the product function level and the structure level of diesel engines.</p> Full article ">Figure 12
<p>A schematic diagram of the pressurization and electronic control systems.</p> Full article ">Figure 13
<p>A comparison to the Monte Carlo algorithm.</p> Full article ">Figure 14
<p>A comparison to the Monte Carlo algorithm.</p> Full article ">
19 pages, 3327 KiB  
Article
Rolling Bearing Fault Diagnosis Based on Time-Frequency Compression Fusion and Residual Time-Frequency Mixed Attention Network
by Guodong Sun, Xiong Yang, Chenyun Xiong, Ye Hu and Moyun Liu
Appl. Sci. 2022, 12(10), 4831; https://doi.org/10.3390/app12104831 - 10 May 2022
Cited by 3 | Viewed by 2062
Abstract
The traditional rolling bearing diagnosis algorithms have problems such as insufficient information on time-frequency images and poor feature extraction ability of the diagnosis model. These problems limit the improvement of diagnosis performance. In this article, the input of the time-frequency image and intelligent [...] Read more.
The traditional rolling bearing diagnosis algorithms have problems such as insufficient information on time-frequency images and poor feature extraction ability of the diagnosis model. These problems limit the improvement of diagnosis performance. In this article, the input of the time-frequency image and intelligent diagnosis algorithms are optimized. Firstly, the characteristics of two advanced time-frequency analysis algorithms are deeply analyzed, i.e., multisynchrosqueezing transform (MSST) and time-reassigned multisynchrosqueezing transform (TMSST). Then, we propose time-frequency compression fusion (TFCF) and a residual time-frequency mixed attention network (RTFANet). Among them, TFCF superposes and splices two time-frequency images to form dual-channel images, which can fully play the characteristics of multi-channel feature fusion of the convolutional kernel in the convolutional neural network. RTFANet assigns attention weight to the channels, time and frequency of time-frequency images, making the model pay attention to crucial time-frequency information. Meanwhile, the residual connection is introduced in the process of attention weight distribution to reduce the information loss of feature mapping. Experimental results show that the method converges after seven epochs, with a fast convergence rate and a recognition rate of 99.86%. Compared with other methods, the proposed method has better robustness and precision. Full article
(This article belongs to the Special Issue Intelligent Fault Diagnosis and Health Detection of Machinery)
►▼ Show Figures

Figure 1

Figure 1
<p>Overall model architecture.</p> Full article ">Figure 2
<p>Time–frequency analysis and comparison of simulation signals: (<b>a</b>) time–domain waveform of simulation signal <math display="inline"><semantics> <mrow> <msub> <mi>x</mi> <mi>e</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </semantics></math>: (<b>b</b>–<b>f</b>) are STFT, TMSST and MSST time–frequency images of simulation signals, respectively; (<b>d</b>,<b>e</b>) are enlarged images of regions 1 and 2 in (<b>c</b>), respectively; (<b>g</b>,<b>h</b>) are respectively enlarged images of regions 1 and 2 in (<b>f</b>).</p> Full article ">Figure 3
<p>Structure of residual time–frequency mixed attention module.</p> Full article ">Figure 4
<p>Time–frequency images of vibration signal of rolling bearing: (<b>a</b>–<b>c</b>) are time–domain waveforms of normal, inner race fault and outer race fault samples respectively; (<b>d</b>–<b>f</b>) are STFT time-frequency images corresponding to (<b>a</b>–<b>c</b>) respectively; (<b>g</b>–<b>i</b>) are TMSST time-frequency images corresponding to (<b>a</b>–<b>c</b>); (<b>j</b>–<b>l</b>) are MSST time-frequency images corresponding to (<b>a</b>–<b>c</b>) respectively.</p> Full article ">Figure 5
<p>Training of RTFANet model under different inputs: (<b>a</b>) curve of training set loss; (<b>b</b>) curve of the accuracy of the test set.</p> Full article ">Figure 6
<p>Gradient–weighted class activation mapping for different fault samples: (<b>a</b>) normal; (<b>b</b>) inner race fault; (<b>c</b>) outer race fault.</p> Full article ">Figure 7
<p>Training of different combination models: (<b>a</b>) curve of training set loss; (<b>b</b>) accuracy curve of the test set.</p> Full article ">Figure 8
<p>RTFANet model confusion matrix on test set.</p> Full article ">Figure 9
<p>Experimental results of model performance test.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-17
Back to TopTop