Article
Computer Science, Hardware & Architecture
Xiaojun Bai, Zhenxi Ma, Wei Chen, Shenhang Wang, Yanfang Fu
Summary: A new method for diagnosing faults in laser gyroscopes using Kernel Extreme Learning Machine (KELM) is proposed in this study. The method efficiently extracts signal features using Wavelet Packet Decomposition (WPD) and optimizes parameters using the Improved Dung Beetle Optimizer (IDBO) algorithm, resulting in improved diagnostic accuracy.
COMPUTERS & ELECTRICAL ENGINEERING
(2023)
Article
Energy & Fuels
Hui Zhang, Cunhua Pan, Yuanxin Wang, Min Xu, Fu Zhou, Xin Yang, Lou Zhu, Chao Zhao, Yangfan Song, Hongwei Chen
Summary: This study proposes a kernel extreme learning machine diagnosis model for typical faults in coal mills operation process. The model is based on variational model feature extraction and kernel principal component analysis. By decomposing the collected signals of vibration and loading force using variational model decomposition, the eigenvectors consisting of feature energy and sample entropy in the intrinsic model functions are obtained. Kernel principal component analysis is then applied for noise removal and dimensionality reduction. The results show that the proposed model improves the input features compared to the single eigenvector model, and the kernel principal component analysis method effectively reduces information redundancy and correlation of eigenvectors, enhancing the prediction performance of the model.
Article
Engineering, Electrical & Electronic
Zhongqiang Wu, Xueqin Lu
Summary: A microgrid fault diagnosis method based on whale algorithm optimizing extreme learning machine (ELM) is proposed. The method analyzes the three-phase fault voltage using wavelet packet decomposition, and uses a whale algorithm to optimize the extreme learning machine to establish a diagnostic model for identifying and diagnosing microgrid faults. The proposed method achieves faster learning speed, stronger generalization ability and higher recognition accuracy compared to other neural network models.
JOURNAL OF ELECTRICAL ENGINEERING & TECHNOLOGY
(2023)
Article
Mathematics
Hanen Chaouch, Samia Charfeddine, Sondess Ben Aoun, Houssem Jerbi, Victor Leiva
Summary: In this study, a multiscale monitoring method for nonlinear processes was developed using a machine learning tool based on kernel PCA and discrete wavelet transform. The proposed method decomposes multivariate data into wavelet coefficients and applies kernel PCA for defect detection. It combines multiscale analysis and kernel PCA to monitor nonlinear processes by considering only the scales with prediction errors exceeding control limits in the data reconstruction phase.
Article
Engineering, Electrical & Electronic
Binbin Li, Zhu Zhang, Lijian Ding
Summary: The article proposes a self-error detection method based on PCA-WPD, which maps errors to SPE statistics for online monitoring based on a comparison standard established on the characteristics of symmetrical three-phase operation, accurately detecting measurement errors.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Automation & Control Systems
Guoliang Lu, Xin Wen, Guangshuo He, Xiaojian Yi, Peng Yan
Summary: In this article, a new dynamic modeling approach called GMWPCs is proposed for health monitoring of rolling element bearings (REBs) by integrating wavelet packet decomposition (WPD) and graph theory. The GMWPCs can enhance the analysis of WPD and enable early detection and fault identification in REBs, demonstrating effectiveness for real engineering applications.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Automation & Control Systems
Guoliang Lu, Xin Wen, Guangshuo He, Xiaojian Yi, Peng Yan
Summary: This article introduces a new dynamic modeling approach GMWPCs, which integrates WPD and graph theory to extract correlation information for early warning detection and fault identification. Experimental results validate the effectiveness and suitability of the proposed framework.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Energy & Fuels
YuPeng Wu, WenBing Wu
Summary: The wavelet packet decomposition method can divide signals into multiple frequency bands, which helps improve the accuracy of fault diagnosis for mechanical vibration signals. Signals with obvious differences in specific frequency bands are easier to distinguish, leading to a more effective engine fault diagnosis rate.
Article
Ergonomics
Yongkui Sun, Yuan Cao, Peng Li
Summary: This paper proposes a non-contact fault diagnosis method for train plug doors based on sound signals. The method utilizes empirical mode decomposition and signal reconstruction to process raw sound signals. It also introduces the concept of fractional calculus and weight to develop a novel feature named weighted fractional wavelet packet decomposition energy entropy. The proposed method achieves the best accuracy of 97.87% in fault diagnosis of train plug doors, as verified through field-collected data and compared with different fault diagnosis methods.
ACCIDENT ANALYSIS AND PREVENTION
(2022)
Article
Engineering, Chemical
Hairong Fang, Wenhua Tao, Shan Lu, Zhijiang Lou, Yonghui Wang, Yuanfei Xue
Summary: This paper proposes a new two-step dynamic local kernel principal component analysis method, which can handle the nonlinearity and the dynamic features simultaneously.
Article
Automation & Control Systems
Guang Wang, Jinghui Yang, Yucheng Qian, Jingsong Han, Jianfang Jiao
Summary: In this article, a kernel principal component analysis (KPCA)-based canonical correlation analysis (CCA) model is proposed for nonlinear process monitoring in quality-related fault detection and diagnosis (QrFDD). The KPCA is used to eliminate nonlinear coupling among the variables by extracting kernel principal components (KPCs) of original variables data. The KPCs and output are then used for CCA modeling, establishing a linear regression model between process and quality variables based on the proportional relationship between process variables sample and kernel sample under the Gaussian kernel. This nonlinear QrFDD method outperforms existing kernel-based CCA methods in terms of algorithmic complexity and interpretability, as demonstrated by simulation results.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Ying Xie, Fanchao Hu, Xuewei Liu, Lirong Zhai
Summary: This paper proposes a process monitoring approach based on locally weighted probabilistic kernel principal component analysis (LWPKPCA), which builds an accurate time-varying model by selecting normal process data with high similarity to the test samples and weighting the data of different importance. Experimental results show that the proposed method performs well in dealing with time-varying data and improving fault detection performance.
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2023)
Article
Multidisciplinary Sciences
Qixin Wang, Kun Qin, Binbin Lu, Huabo Sun, Ping Shu
Summary: The study proposes a Time-Feature Attention-based Convolutional Auto-Encoder (TFA-CAE) network model to extract essential flight features from QAR data. The results show that the TFA-CAE model outperforms traditional or similar approaches in extracting representative flight features, can recognize flight patterns corresponding to different runways, and effectively identify anomalous flights.
SCIENTIFIC REPORTS
(2023)
Article
Engineering, Electrical & Electronic
Tao Li, Yongli Li, Xiaolong Chen
Summary: This paper introduces a fault diagnosis algorithm using wavelet packet transform and principal component analysis, applied to the inverter-side fault diagnosis of multi-terminal hybrid HVDC network, significantly improving the speed and accuracy of fault diagnosis.
JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY
(2021)
Article
Computer Science, Information Systems
Yonghui Xu, Ruotong Meng, Zixuan Yang
Summary: This paper introduces a method to address the accuracy issue in fault detection and fault diagnosis of gas sensor arrays. The method, which combines the advantages of principal component analysis and kernel principal component analysis, achieves higher accuracy in fault detection and isolation.
Article
Computer Science, Information Systems
He Jun, Yong Chen, Qing-Hua Zhang, Guoxi Sun, Qin Hu
Article
Engineering, Electrical & Electronic
Aisong Qin, Qin Hu, Yunrong Lv, Qinghua Zhang
IEEE SENSORS JOURNAL
(2019)
Article
Engineering, Mechanical
Qin Hu, Xiao-Sheng Si, Qing-Hua Zhang, Ai-Song Qin
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2020)
Article
Chemistry, Analytical
Jun He, Ming Ouyang, Chen Yong, Danfeng Chen, Jing Guo, Yan Zhou
Article
Engineering, Electrical & Electronic
Qin Hu, Qi Zhang, Xiao-Sheng Si, Guo-Xi Sun, Ai-Song Qin
IEEE SENSORS JOURNAL
(2020)
Article
Engineering, Multidisciplinary
Ai-Song Qin, Han-Ling Mao, Qin Hu
Summary: This study proposed a novel cross-domain fault diagnosis method for rolling bearings, which extracts features, selects discriminative features, and classifies unlabeled samples to achieve cross-domain fault diagnosis successfully.
Article
Acoustics
Jun He, Xiang Li, Yong Chen, Danfeng Chen, Jing Guo, Yan Zhou
Summary: The study introduces a deep transfer learning method based on 1D-CNN for rolling bearing fault diagnosis, which addresses the high computational complexity issue in interdomain distribution discrepancy measurement by minimizing the marginal distribution discrepancy between the source and target domains. Experimental results demonstrate the superior performance of the proposed method.
SHOCK AND VIBRATION
(2021)
Article
Engineering, Electrical & Electronic
Aisong Qin, Hanling Mao, Kuangchi Sun, Zhengfeng Huang, Xinxin Li
Summary: This study proposes a novel cross-domain fault diagnosis method based on improved multi-scale fuzzy measure entropy and enhanced joint distribution adaptation. It addresses the inconsistent data distribution between the source and target domains. The method generates discriminative and similar features, reduces the distribution discrepancy, and classifies unlabeled samples in the target domain using a statistical classifier.
IEEE SENSORS JOURNAL
(2022)
Article
Engineering, Electrical & Electronic
Qin Hu, Xiaosheng Si, Aisong Qin, Yunrong Lv, Mei Liu
Summary: This study proposes a novel fault diagnosis method based on enhanced multi-scale sample entropies and balanced adaptation regularization based transfer learning to address the issue of inconsistent distribution between training and testing data in fault diagnosis. The method enhances feature discriminability and similarity of fault information, and uses balanced adaptation regularization based transfer learning to learn an adaptive classifier for cross-domain fault diagnosis.
IEEE SENSORS JOURNAL
(2022)
Article
Multidisciplinary Sciences
Jun He, Zheshuai Zhu, Xinyu Fan, Yong Chen, Shiya Liu, Danfeng Chen
Summary: This study proposes a deep learning-based method for few-shot bearing fault diagnosis, which addresses the issue of limited labeled training samples by using pseudo-labels and kernel principal component analysis. Experimental results demonstrate that the proposed method achieves higher classification accuracy compared to existing methods.
Article
Engineering, Electrical & Electronic
Jun He, Ming Ouyang, Zhiwen Chen, Danfeng Chen, Shiya Liu
Summary: In real industry, domain shift is a common problem that affects diagnostic performance due to changes in operating conditions and system differences. Insufficient labeled or unlabeled samples also limit the adaptability of fault diagnosis methods. This article proposes a deep transfer learning method based on WGAN and minimum singular value to address these problems. The method utilizes a domain critic network to improve domain adaptation ability and incorporates the minimum singular value to capture effective category information measurement.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Engineering, Electrical & Electronic
Kuangchi Sun, Zhenfeng Huang, Hanling Mao, Aisong Qin, Xinxin Li, Weili Tang, Jianbin Xiong
Summary: In this article, a multi-scale cluster-graph convolution neural network with multi-channel residual network (MR-MCGCN) is proposed for machine fault diagnosis, which effectively extracts weak features in the signal and achieves high diagnosis results even under variable load conditions. The experimental results demonstrate the effectiveness of the proposed MR-MCGNN.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)