Article
Automation & Control Systems
Xianzhi Wang, Shubin Si, Yongbo Li
Summary: A fault diagnosis scheme based on multiscale diversity entropy (MDE) and extreme learning machine (ELM) is proposed in this article, which quantifies dynamical complexity and provides a comprehensive feature description for pattern identification of rotating machinery. The effectiveness of the proposed MDE method is verified through simulated signals and experimental signals from bearing test and dual-rotator of aeroengine test, showing superior classification accuracy compared with existing approaches.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2021)
Article
Engineering, Mechanical
Yongbo Li, Shun Wang, Yang Yang, Zichen Deng
Summary: The paper introduces a method called Symbolic Fuzzy Entropy (SFE) based on symbolic dynamic filtering and fuzzy entropy to extract fault features and eliminate noise, effectively improving calculation efficiency. By extending SFE to multiscale analysis to form MSFE, experimental results demonstrate that MSFE outperforms three other methods in extracting weak fault characteristics.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Engineering, Multidisciplinary
Yanli Ma, Junsheng Cheng, Ping Wang, Jian Wang, Yu Yang
Summary: A new entropy method, MMFDE, is proposed by modifying the calculation process of MFDE, which can extract fault features from multivariate signals and consider the within- and crosschannel correlations. By applying MMFDE to simulation signals, the results demonstrate its stability and reliability.
Article
Engineering, Mechanical
Fei Chen, Liyao Zhang, Wenshen Liu, Tingting Zhang, Zhigao Zhao, Weiyu Wang, Diyi Chen, Bin Wang
Summary: In this study, a fault diagnosis method for rotating machinery based on improved multiscale attention entropy and random forests is proposed. The experimental results show that the proposed method achieves the optimal diagnostic performance on two different fault datasets and exhibits strong adaptability in practical applications.
NONLINEAR DYNAMICS
(2023)
Article
Acoustics
Fuming Zhou, Jun Han, Xiaoqiang Yang
Summary: This paper proposes a novel fault diagnosis model that combines MHMFDE, MCFS, and GWO-KELM for efficient and accurate fault diagnosis of rotating machinery. By extracting fault features, selecting sensitive features, and performing quantitative analysis, the presented approach demonstrates excellent diagnostic performance, especially for compound faults of rotating machinery.
Article
Engineering, Electrical & Electronic
Yuqing Zhou, Hongche Wang, Gonghai Wang, Anil Kumar, Weifang Sun, Jiawei Xiang
Summary: In recent years, deep learning-based methods have made remarkable achievements in the intelligent fault diagnosis of rotating machinery. However, the lack of labeled and large unlabeled samples in actual industrial scenes affects the performance of supervised learning methods. This paper proposes a novel semi-supervised fault diagnosis method based on multiscale permutation entropy (MPE) enhanced contrastive learning (CL). Experimental results in gearbox and milling tool fault diagnosis experiments show that the proposed MPE-CL method outperforms other benchmark methods with classification accuracy above 95.4% and 96.0% when the labeled training dataset size is 50/class, respectively.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Computer Science, Information Systems
Zhuang Tang, Jie Liu, Chaofeng Li
Summary: This paper proposes an improved method for vibration signal feature extraction by combining Improved Multivariate Multiscale Dispersion Entropy (IMMDE) with Hierarchical Entropy (HE). The method overcomes the limitations of traditional coarse-grained calculation and improves stability. It extracts deep fault frequency information from different frequency components of the signal using HE, and uses Max-Relevance Min-Redundancy (mRMR) to optimize the extracted features. Finally, Support Vector Machine (SVM) is used to determine the degree and type of fault.
Article
Engineering, Mechanical
Shun Wang, Yongbo Li, Khandaker Noman, Dong Wang, Ke Feng, Zheng Liu, Zichen Deng
Summary: The paper proposes a new entropy measure called cumulate spectrum distribution entropy (CSDEn), which can capture frequency-domain information of fault features. The method is evaluated using synthetic signals and experimental data, showing superior performance in detecting dynamic changes and measuring signal complexity compared to other methods.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2024)
Article
Automation & Control Systems
Zhe Chen, Shiqing Tian, Xiaotao Shi, Huimin Lu
Summary: This article proposes a novel multiscale shared-learning network architecture to extract and classify the fault features inherent to multiscale factors of vibration signals. Experimental results demonstrate the superiority of this method in fault diagnosis for bearings and gearboxes.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Engineering, Mechanical
Yanli Ma, Junsheng Cheng, Ping Wang, Jian Wang, Yu Yang
Summary: An improved multivariate multiscale fuzzy distribution entropy (IMMFDE) is proposed to extract fault features from multivariate vibration signals of the rotating machinery under different speeds. This method, based on multivariate empirical mode decomposition, can adaptively determine the maximum scale, eliminate frequency aliasing, and avoid the loss of potentially useful information. The trait of IMMFDE is verified by calculating the sequences and amplitude spectrums of simulated multivariate signals at each scale. A fault diagnosis method is further proposed for the rotating machinery under different speeds, which utilizes statistical parameters and IMMFDE as the fault feature set and support vector machine for fault diagnosis. The results show that the proposed method can achieve better fault diagnosis results.
NONLINEAR DYNAMICS
(2023)
Article
Automation & Control Systems
Jinyang Jiao, Hao Li, Jing Lin, Hui Zhang
Summary: To address fault diagnosis in rotating machinery under complex working conditions, unsupervised domain adaptation technology has been explored and a novel entropy-oriented domain adaptation (EODA) model has been developed. The EODA model utilizes entropy optimization strategies and achieves intelligent diagnosis of rotating machinery. Experimental results demonstrate the superior performance of the proposed method.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Engineering, Mechanical
Jinde Zheng, Wanming Ying, Jinyu Tong, Yongbo Li
Summary: The entropy-based complexity measurement tools are widely used in extracting fault characteristics of rolling bearings. However, the fault information is usually hidden in both time and frequency domains, making one-dimensional entropy insufficient to fully extract comprehensive fault information from measured vibration signals of rolling bearings. To address this issue, a novel entropy-based complexity evaluation method called three-dimensional Holo-Hilbert spectral entropy (HHSE3D) is developed, which utilizes Holo-Hilbert spectral analysis to expand the one-dimensional signal to a three-dimensional relationship among time domain information, amplitude-modulated, and frequency-modulated features. Furthermore, the proposed HHSE3D method is extended into a multiscale framework through coarse-graining process to obtain a comprehensive nonlinear dynamic feature description in different scales, resulting in the multiscale HHSE3D (MHHSE3D) method. The robustness and effectiveness of MHHSE3D are verified using simulated signals and experimental bearing data, showing that the proposed method exhibits superior feature extraction ability and diagnostic accuracy compared to the other four traditional entropy-based diagnosis methods.
NONLINEAR DYNAMICS
(2023)
Article
Automation & Control Systems
Xiaogang Xu, Ruijun Wang, Zhixia Fan, Xu Ma, Zeren Zhao, Huijie Wang
Summary: In this article, a multiscale dense residual network is proposed to address the issue of inefficient diagnosis of mechanical equipment in complex operating environments. The model improves performance from three aspects and demonstrates excellent generalization ability in the state recognition of different mechanical equipment.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Chemistry, Multidisciplinary
Shunming Li, Yuzhe Hou, Jiantao Lu, Mengqi Feng
Summary: A fault feature extraction method based on refined multiscale symbolic dispersion entropy (RMSDE) is proposed, which combines symbolic coding and dynamic complexity evaluation for multidimensional feature extraction. The method shows superior performance in complexity estimation, anti-noise interference, and computational efficiency through evaluation with simulation signals. When combined with a self-organizing fuzzy logic classifier (SOF), the proposed RMSDE-SOF method achieves high-precision fault diagnosis of rotating machinery. The effectiveness and robustness of the method are verified using real-world datasets, and the comprehensive performance is better than similar methods.
APPLIED SCIENCES-BASEL
(2023)
Article
Automation & Control Systems
Xianzhi Wang, Shubin Si, Yongbo Li
Summary: This article proposes a feature extraction method for synchronous fault feature extraction using multichannel vibration signals. A novel fault diagnosis scheme for condition monitoring of large-scale machinery is developed by combining the proposed method with a random forest classifier. Experimental results show that the proposed method has better performance in multichannel feature extraction.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2022)
Article
Engineering, Mechanical
Haiyang Pan, Jinde Zheng, Yu Yang, Junsheng Cheng
Summary: Traditional time-frequency analysis methods have limitations in handling nonlinear signals. A new algorithm called nonlinear sparse mode decomposition (NSMD) is proposed, which utilizes a singular local linear operator to extract local narrowband signal components and adaptively decompose the signal with good robustness and adaptability. Analysis results show that NSMD method is effective for raw vibration signals.
MECHANISM AND MACHINE THEORY
(2021)
Article
Engineering, Mechanical
Jian Cheng, Yu Yang, Niaoqing Hu, Zhe Cheng, Junsheng Cheng
Summary: An adaptive weighted symplectic geometry decomposition (AWSGD) method is proposed for noise reduction, which avoids the defect of traditional methods by energy size. The AWSGD method can improve the accuracy and efficiency of noise reduction.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Automation & Control Systems
Jian Cheng, Yu Yang, Haidong Shao, Haiyang Pan, Jinde Zheng, Junsheng Cheng
Summary: This paper proposes an effective method for composite fault diagnosis of rolling bearings. It utilizes the theory of Ramanujan sum to generate precise periodic components and employs the SOSO-MAIHND method to reduce noise and enhance relatively weak periodic impulses.
Article
Computer Science, Artificial Intelligence
Xin Li, Yu Yang, Niaoqing Hu, Zhe Cheng, Haidong Shao, Junsheng Cheng
Summary: This paper presents a fault diagnosis framework for rotating machinery using multi-channel feature fusion and Riemannian manifold. The proposed approach calculates a multi-channel fusion covariance matrix (MFCM) as the representation of fault features and utilizes a maximum margin Riemannian manifold-based hyperdisk (MMRMHD) classifier for fault classification. Experimental results demonstrate the effectiveness and superiority of the proposed framework.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Automation & Control Systems
Jian Cheng, Yu Yang, Xiaowei Wu, Jian Wang, Zhantao Wu, Junsheng Cheng
Summary: This paper proposes a new method called symplectic Ramanujan mode decomposition (SRMD) for compound-fault diagnosis of rolling bearings. By combining symplectic similarity transformation with Ramanujan subspace theory, SRMD can effectively extract the periodic impulses and diagnose the compound faults in bearings.
Article
Engineering, Industrial
Xin Li, Yu Yang, Zhantao Wu, Ke Yan, Haidong Shao, Junsheng Cheng
Summary: A high-accuracy gearbox health state recognition model, named graph sparse RVFLN (GSRVFLN), is proposed in this paper to fully leverage the label sparsity and raw data's manifold structure information. The model introduces a sparse constraint term and a discriminative adjacency graph to capture the inherent geometry structure and discriminative information of data. An effective solution is derived with the ADMM framework, showing great convergence and outperforming other state-of-the-art models in gearbox health state recognition.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2022)
Article
Engineering, Mechanical
Yanli Ma, Junsheng Cheng, Ping Wang, Jian Wang, Yu Yang
Summary: Multi-channel signal processing method can improve the accuracy and confidence level of fault diagnosis. However, existing methods have problems, hence a new method called Lanczos quaternion singular spectrum analysis (LQSSA) is proposed in this paper. The method reduces calculation time by using Lanczos method, improves signal purity by introducing Lagrange multiplier, and breaks the restriction between different channels using periodic similarity.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Engineering, Mechanical
Jie Zhou, Yu Yang, Xin Li, Haidong Shao, Junsheng Cheng
Summary: A novel multivariate signal decomposition method MLCD and 1.5-dimensional empirical envelope spectrum 1.5D EES are proposed to process multi-channel signals and highlight gear fault characteristic frequency effectively.
MECHANISM AND MACHINE THEORY
(2022)
Article
Computer Science, Artificial Intelligence
Zuanyu Zhu, Yu Yang, Niaoqing Hu, Zhe Cheng, Junsheng Cheng
Summary: The study proposes a sparse random projection-based hyperdisk classifier model for fault diagnosis of bevel gearbox. The method efficiently screens out core samples, introduces slack variables and a dynamic penalty parameter to improve the boundary of the model, and develops a strategy to handle imbalanced training data, resulting in better performance and efficiency in fault diagnosis.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Engineering, Multidisciplinary
Hailong Liu, Yu Yang, Niaoqing Hu, Zhe Cheng, Junsheng Cheng
Summary: A soft-margin HD tensor machine (SHDTM) model is proposed to overcome the defects of a hard margin HD model, by adaptively adjusting the HD margin and extending the model input from vector data to tensor data. The results of fault diagnosis experiments show that the SHDTM model has excellent resistance to outliers and noise interference, and performs well in diagnosing unbalanced datasets.
MEASUREMENT SCIENCE AND TECHNOLOGY
(2022)
Article
Engineering, Mechanical
Jian Cheng, Yu Yang, Xin Li, Junsheng Cheng
Summary: This article introduces the problems of several commonly used signal decomposition methods in gear fault diagnosis and proposes a novel multi-layer decomposition method called symplectic geometry packet decomposition (SGPD). The SGPD method combines symplectic geometry theory and the multi-layer decomposition idea of wavelet packet to minimize noise and retain fault information during the decomposition of non-steady signals.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Engineering, Electrical & Electronic
Jie Zhou, Junsheng Cheng, Xiaowei Wu, Jian Wang, Yu Yang
Summary: This paper introduces a multivariate signal noise reduction method, adaptive quaternion multivariate local characteristic-scale decomposition (AQMLCD), for gear fault diagnosis. By adaptively reducing noise and effectively fusing fault information from each channel, AQMLCD shows improved noise robustness compared to other methods, while maintaining similar computational efficiency.
DIGITAL SIGNAL PROCESSING
(2022)
Article
Automation & Control Systems
Jian Cheng, Yu Yang, Zhantao Wu, Haidong Shao, Haiyang Pan, Junsheng Cheng
Summary: In this paper, a Ramanujan Fourier mode decomposition (RFMD) method is proposed for gear fault diagnosis. The RFMD method can accurately extract and identify the periodic components of gear faults, showing good noise robustness.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Engineering, Multidisciplinary
Jie Zhou, Junsheng Cheng, Xiaowei Wu, Jian Wang, Jian Cheng, Yu Yang
Summary: A new method called CAPMLCD is proposed in this paper for decomposing multichannel signals. It uses completely adaptive projection to optimize the multivariate local characteristic-scale decomposition. Experimental results show that CAPMLCD outperforms other methods in terms of suppressing mode mixing, decomposition efficiency, and decomposition accuracy.
Article
Engineering, Multidisciplinary
Xin Kang, Junsheng Cheng, Ping Wang, Jian Wang, Zuanyu Zhu, Yu Yang
Summary: This study proposes a novel ensemble model, the ensemble convex hull (CH)-based (EnCH) classification model, which improves the differentiation ability of classifiers, enhances generalization and robustness, and has excellent tolerance to noise and outliers.
MEASUREMENT SCIENCE AND TECHNOLOGY
(2023)