Article
Computer Science, Software Engineering
Ren-wang Song, Lei Yang, Lin-ying Chen, Zeng-shou Dong
Summary: The construction of health indicators for rolling bearings is important for monitoring their health conditions and predicting residual life. This paper focuses on the frequency band sparse optimization algorithm to construct the rolling bearing health index. The method based on machine learning lacks interpretability, and using statistical features directly from the original signal as a health indicator is not effective due to interference components. To solve these issues, this paper proposes a method based on sparse frequency band characterization. It utilizes wavelet packet decomposition to obtain independent sub-bands, constructs a sparse model using the sub-band amplitudes and decomposed matrices, and calculates the spectral amplitude sum of the sensitive frequency band as a health indicator. The method's effectiveness is confirmed using published bearing accelerated life experimental data.
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
(2023)
Article
Computer Science, Information Systems
Chuang Liang, Changzheng Chen, Ye Liu, Xinying Jia
Summary: This study proposes an intelligent fault diagnosis method for rolling bearings based on compressive sensing and a neural network, which effectively extracts multi-granularity high-level features to reduce data volume, improve classification accuracy, and validate its effectiveness and high diagnostic accuracy.
Article
Computer Science, Information Systems
Yi Zhu, Lei Li, Xindong Wu
Summary: The research proposes a semi-supervised deep learning framework to address the issue of insufficient labeled image data, utilizing stacked layers, convolutional approach, and sparse auto-encoder to learn feature representations. The framework also includes an algorithm to handle data redundancy and encodes label information using a Softmax regression model.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2021)
Article
Automation & Control Systems
Jinde Zheng, Haiyang Pan, Jinyu Tong, Qingyun Liu
Summary: Extracting failure-related information from vibration signals is crucial for vibration-based fault detection in rolling bearings. This article proposes a new nonlinear dynamic parameter to enhance the measurement of data complexity and compares it with existing algorithms. Furthermore, a novel fault diagnosis approach is introduced, which achieves the highest identifying rate and the best performance among the comparative approaches.
Article
Automation & Control Systems
Hua Li, Tao Liu, Xing Wu, Qing Chen
Summary: The study introduces an enhanced SVD method E-SVD to address the issues with SVD, achieving superior signal reconstruction and noise reduction effects through the combination of ISVD and IWPT. Additionally, an evaluation indicator is introduced to assess the performance of the results.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2021)
Article
Engineering, Multidisciplinary
Junchao Guo, Zhanqun Shi, Dong Zhen, Zhaozong Meng, Fengshou Gu, Andrew D. Ball
Summary: This article proposes a novel fault diagnosis scheme based on optimized wavelet packet denoising and modulation signal bispectrum to accurately diagnose bearing faults by enhancing transient impulses with wavelet packet denoising and demodulating modulation signal bispectrum. The method effectively purifies the signal and extracts the modulation components contained in the transient impulses, determining the type of bearing faults with superior performance in extracting the fault feature of incipient defects on different bearing components compared to other methods.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2022)
Article
Engineering, Electrical & Electronic
Xiaoli Zhao, Minping Jia, Zheng Liu
Summary: The article introduces an intelligent fault diagnosis method for rotating machinery based on semisupervised deep sparse auto-encoder (SSDSAE) with local and nonlocal information. Vibration spectrum signals are fed into the SSDSAE algorithm for fault feature extraction, and the extracted sparse discriminant features are used for fault diagnosis with a back-propagation (BP) classifier. The method utilizes weighted cross-entropy (WCE) techniques to improve the generalization performance of the fault diagnosis model and is validated with experimental data.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Engineering, Electrical & Electronic
Ning Jia, Yao Cheng, Yunyang Liu, Youyuan Tian
Summary: A bearing health monitoring and defect diagnostic model based on variational mode decomposition, continuous wavelet transform, and stacked denoising auto-encoder optimized by sparrow search algorithm is proposed in this paper. By mapping the fault characteristic information to different local positions in time and scale, the model enhances the feature extraction capabilities and achieves optimal parameters through global combination and adaptive selection.
IEEE SENSORS JOURNAL
(2022)
Article
Engineering, Multidisciplinary
Zhilin Dong, Dezun Zhao, Lingli Cui
Summary: In this paper, a time-frequency joint metric feature extraction technique named non-negative wavelet matrix factorization (NWMF) is proposed to extract more effective features from raw signals. Based on this technique and convolutional neural network (CNN), an intelligent diagnosis framework is constructed to detect bearing fault.
MEASUREMENT SCIENCE AND TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Jinyu Cai, Shiping Wang, Wenzhong Guo
Summary: The paper proposes a deep stacked sparse embedded clustering method that considers both local structure preservation and input sparsity. The deep learning approach jointly learns clustering-oriented features and optimizes cluster label assignments by minimizing both the reconstruction and clustering loss. Comprehensive experiments validate the effectiveness of introducing sparsity and preserving local structure in the proposed method.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Acoustics
HongChao Wang, WenLiao Du
Summary: The proposed method in the article aims to extract the latent fault components of rolling bearing based on self-learned sparse atomics and frequency band entropy. The method involves applying sparse atomics method on the vibration signal, selecting atomics with larger kurtosis values for reconstruction, analyzing the reconstructed signal with frequency band entropy, and filtering the signal with optimal band-pass filter for fault feature extraction. The feasibility and superiority of the method are demonstrated using vibration data from accelerated fatigue life test of rolling bearing.
JOURNAL OF VIBRATION AND CONTROL
(2021)
Article
Engineering, Multidisciplinary
Haiquan Song, Wengang Ma, Zhonghe Han, Xiaoxun Zhu
Summary: In this paper, an adaptive sparse contractive auto-encoder and an improved gray wolf optimization unsupervised extreme learning machine model were proposed for diagnosing unbalanced rolling bearing faults. The method demonstrates high accuracy and efficiency in fault diagnosis by extracting multi-layer features from vibration signals.
Article
Computer Science, Information Systems
Maohua Xiao, Yabing Liao, Petr Bartos, Martin Filip, Guosheng Geng, Ziwei Jiang
Summary: A new fault diagnosis method based on back propagation neural network optimized by cuckoo search algorithm is proposed in this paper. By using wavelet packet decomposition for feature extraction of vibration signals, the method achieved high diagnostic accuracy rate in initial experiments.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Rui Fan, Huipeng Li, Tao Zhang, Hong Wang, Linhai Qi, Lina Sun
Summary: This paper proposes a voltage sag identification method that combines sparse auto-encoder and Attention Unet, achieving high accuracy in recognition by performing feature learning and extraction on high-dimensional data. It is of great significance for auxiliary decision-making in power quality management and governance.
Article
Computer Science, Artificial Intelligence
Zhu Yin, WuZhen Shi, Zhongcheng Wu, Jun Zhang
Summary: In this paper, we propose a novel multilevel wavelet-based hierarchical networks for image compressed sensing (MWHCS-Net). The proposed network utilizes multilevel wavelet transform for progressive measurement acquisition and hierarchical initial reconstruction for preserving finer texture details. Experimental results demonstrate that MWHCS-Net achieves state-of-the-art performance with efficient running speed, and outperforms existing deep learning-based image compressed sensing methods in terms of anti-noise performance in most cases.
PATTERN RECOGNITION
(2022)