Article
Energy & Fuels
Xinghua Huang, Yuanyuan Li, Yi Chai
Summary: This paper proposes a CNN-based MSDFN model for fault diagnosis of wind turbines planetary gearboxes, achieving high accuracy through preprocessing vibration signals and utilizing MSFF and FoM modules for feature fusion and classification. The effectiveness of the method is verified through experimental results, with the MSFF and FoM modules playing a positive role in fault diagnosis.
FRONTIERS IN ENERGY RESEARCH
(2021)
Article
Chemistry, Multidisciplinary
Zhinong Li, Zedong Li, Yunlong Li, Junyong Tao, Qinghua Mao, Xuhui Zhang
Summary: A machine fault intelligent diagnosis method based on federated learning is proposed in this paper, which establishes local fault diagnosis models and fuses global model parameters to recognize newly added fault types. The method is validated on bearing data with an accuracy of 100%.
APPLIED SCIENCES-BASEL
(2021)
Article
Engineering, Multidisciplinary
Yiming Ma, Guojun Wen, Siyi Cheng, Xin He, Shuang Mei
Summary: This study proposes a multimodal neural network model that combines continuous wavelet transform and symmetrized dot pattern graphs for information fusion, resulting in improved fault diagnosis performance. Experimental results demonstrate that this model outperforms traditional single-modal CNN structures.
MEASUREMENT SCIENCE AND TECHNOLOGY
(2022)
Article
Engineering, Electrical & Electronic
Yadong Xu, Xiaoan Yan, Beibei Sun, Zheng Liu
Summary: Deep learning has been widely used in mechanical fault diagnosis due to its powerful feature extraction capabilities. However, traditional deep learning models lack the ability to extract multiscale discriminative information from mechanical vibration signals. In this article, a hierarchical multiscale dense network (HMSDN) is proposed to address this issue. The HMSDN incorporates a hierarchical procedure into the CNN structure and uses a multiscale dense connection structure to learn discriminative features from measured signals. Experimental results on two electromechanical datasets demonstrate that the proposed method achieves state-of-the-art performance.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Engineering, Mechanical
Dongying Han, Jinghui Tian, Peng Xue, Peiming Shi
Summary: This paper proposes a neural network model for fault diagnosis of rotating machinery, which utilizes multi-sensor information for multi-level fusion, improving the reliability and accuracy of diagnosis.
JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY
(2021)
Review
Acoustics
Su Jiang, Jianping Xuan, Jian Duan, Jianbin Lin, Hongfei Tao, Qi Xia, Ruizhen Jing, Shoucong Xiong, Tielin Shi
Summary: A dual attention dense convolutional network was proposed to address the gradient vanishing issue in deep neural networks and enhance network performance using dense connections and attention mechanisms. The method showed higher accuracy and faster convergence under complex operational conditions in comparison with other traditional and deep learning approaches.
JOURNAL OF VIBRATION AND CONTROL
(2021)
Article
Computer Science, Artificial Intelligence
Jinghui Tian, Dongying Han, Lifeng Xiao, Peiming Shi
Summary: The paper proposes a multi-scale deep coupling convolutional neural network (MDCN) to fuse fault information from heterogeneous sensor data, achieving the identification of device failure states.
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2021)
Article
Automation & Control Systems
Tingli Xie, Xufeng Huang, Seung-Kyum Choi
Summary: In this article, a novel intelligent fault diagnosis method based on multisensor fusion and convolutional neural network is explored. The proposed method converts multisignal data into RGB images and uses an improved CNN for classification, resulting in higher accuracy in fault diagnosis.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Chemistry, Analytical
Han Dong, Jiping Lu, Yafeng Han
Summary: This study proposes a new fault diagnosis model based on convolutional neural network, which can utilize the information contained in vibration signals effectively. The model achieves high accuracy in experiments and has positive implications for multiple inputs and information fusion.
Article
Green & Sustainable Science & Technology
Zifei Xu, Xuan Mei, Xinyu Wang, Minnan Yue, Jiangtao Jin, Yang Yang, Chun Li
Summary: This study developed a MSCNN-BiLSTM model for fault diagnosis of bearings in wind turbines, which leverages a weighted majority voting rule to fuse information from multiple sensors for improved extrapolation, showing superior performance compared to existing methods in experimental data analysis.
Article
Computer Science, Artificial Intelligence
Zifei Xu, Musa Bashir, Wanfu Zhang, Yang Yang, Xinyu Wang, Chun Li
Summary: This study develops an intelligent fault diagnosis model that can process multi-scale information and solves the problems of multi-scale models in complex environments using multi-attention capabilities. The model demonstrates superior performance in experiments, with a 27% higher F1 value compared to existing multi-scale CNN models in similar environments.
INFORMATION FUSION
(2022)
Article
Computer Science, Artificial Intelligence
Qiuyu Kong, Jin Tang, Chenglong Li, Xin Wang, Jian Zhang
Summary: The study demonstrates that utilizing the complementary properties of different CNNs can improve visual tracking performance. By jointly inferring candidate location, predicted location, and confidence score, the importance of different CNNs is identified, and the adaptive fusion of prediction scores enhances tracking robustness.
COGNITIVE COMPUTATION
(2022)
Article
Automation & Control Systems
Ziling Huang, Zihao Lei, Guangrui Wen, Xin Huang, Haoxuan Zhou, Ruqiang Yan, Xuefeng Chen
Summary: In this study, a novel fault diagnosis method is proposed that utilizes multisource information fusion and classification label information. By extracting features and fusing them, along with a joint loss function, the method effectively addresses fault diagnosis under polytropic working conditions. Experimental results demonstrate the method's great potential.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2022)
Article
Engineering, Mechanical
Chao Zhang, Qixuan Huang, Chaoyi Zhang, Ke Yang, Liye Cheng, Zhan Li
Summary: This article proposes a new intelligent fault diagnosis method that can better handle fault diagnosis tasks under different working conditions and noise interference through an improved CNN model and a residual network feature calibration and fusion block.
Article
Engineering, Mechanical
Zhiwu Shang, Wanxiang Li, Maosheng Gao, Xia Liu, Yan Yu
Summary: This paper proposes a multi-scale deep feature fusion intelligent fault diagnosis method based on information entropy, using a variety of autoencoders to construct a deep neural network feature extraction structure and employing deep belief network probability model as the fault classifier. Experimental results show that compared to traditional methods, this approach obtains higher accuracy features from raw data.
CHINESE JOURNAL OF MECHANICAL ENGINEERING
(2021)
Article
Computer Science, Artificial Intelligence
Jinyang Jiao, Ming Zhao, Jing Lin, Kaixuan Liang, Chuancang Ding
Summary: In this paper, a mixed adversarial adaptation network (MAAN) based intelligent framework for cross-domain fault diagnosis of machinery is presented. Differences in marginal distribution and conditional distribution are reduced together through an adversarial learning strategy with a simple adaptive factor to dynamically weigh the relative importance of two distributions. Experimental results show that the proposed model outperforms popular deep learning and deep domain adaptation diagnosis methods.
JOURNAL OF INTELLIGENT MANUFACTURING
(2022)
Article
Engineering, Multidisciplinary
Chuancang Ding, Ming Zhao, Jing Lin, Kaixuan Liang, Jinyang Jiao
Summary: This article introduces a method called kernel ridge regression-based chirplet transform (KRR-CT) for analyzing non-stationary signals. The KRR-CT can provide a stable solution, generate an energy concentrated TF plane, and provide more precise IF information even in the presence of noise. The proposed iterative algorithm shows good performance in machine fault detection.
Article
Engineering, Multidisciplinary
Mingkuan Shi, Chuancang Ding, Juanjuan Shi, Xingxing Jiang, Weiguo Huang, Zhongkui Zhu
Summary: This study proposes a similarity balance discriminant projection (SBDP) algorithm, which incorporates an optimized support vector machine (SVM) to develop a fault diagnosis model for rolling bearing fault diagnosis. The application results demonstrate that SBDP is effective in extracting features representing the intrinsic information of faults, and the optimized SVM is successful in identifying fault types with high accuracy.
MEASUREMENT SCIENCE AND TECHNOLOGY
(2022)
Article
Automation & Control Systems
Jinyang Jiao, Kaixuan Liang, Chuancang Ding, Jing Lin
Summary: This article introduces a domain adaptation method for intelligent fault diagnosis of machinery, which uses minimum class confusion and maximum nuclear norm-based constraints to improve accurate diagnosis results.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Mingkuan Shi, Chuancang Ding, Rui Wang, Qiuyu Song, Changqing Shen, Weiguo Huang, Zhongkui Zhu
Summary: This study proposes an efficient fault diagnosis method based on deep hypergraph autoencoder embedding (DHAEE) to effectively process large-scale unlabeled industrial data and obtain good fault diagnosis results. The unlabeled vibration signals are converted into hypergraphs by a designed hypergraph construction method, and a hypergraph convolutional extreme learning machine autoencoder (HCELM-AE) is designed to mine the higher-order structural information and subspace structural information of the original unlabeled data. By stacking multiple HCELM-AE modules in a deep learning framework, the DHAEE and its fault diagnosis method are constructed, which combines the high computational efficiency of ELM-AE with the strong representational learning ability of deep learning methods.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Engineering, Multidisciplinary
Mingkuan Shi, Chuancang Ding, Hongbo Que, Chengpan Wu, Juanjuan Shi, Changqing Shen, Weiguo Huang, Zhongkui Zhu
Summary: The health condition monitoring of rolling bearings is crucial for the safe operation of electromechanical systems. In order to improve the accuracy of bearing performance degradation prediction, a novel Graph embedded ELM autoencoder (GEELM-AE) is constructed by combining the graph embedding framework, and a Multilayer-Graph-embedded ELM (MGEELM) method is developed by stacking multiple GEELM-AEs. The proposed MGEELM accurately predicts the performance degradation trend of rolling bearings and reduces training time and improves prediction efficiency.
Article
Engineering, Mechanical
Shudong Ou, Ming Zhao, Sen Li, Tao Zhou
Summary: A framework of shock sensing for rotating machinery using encoder signal is constructed in this paper, and a rank-estimated online RPCA method is proposed to capture the weak feature introduced by shock in encoder signal. Experimental results demonstrate that the proposed scheme can effectively sense the external shock, providing a new idea for intelligent perception of rotating machinery.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Engineering, Multidisciplinary
Sen Li, Ming Zhao, Shudong Ou, Dexin Chen, Yiyang Wei
Summary: As the core power transmission component, planetary gearboxes often experience faults due to harsh working environments and heavy loads. Therefore, condition monitoring is crucial to ensure safe operation and reduce maintenance costs. However, current unsupervised anomaly detection methods face challenges in dealing with the complex monitoring signal of planetary gearboxes.
Article
Engineering, Electrical & Electronic
Xuan Li, Jiaqing Huang, Chuancang Ding, Ran Guo, Weilong Niu
Summary: This study developed a more effective and realistic lumped parameter dynamic model for RV reducers by considering the tooth profile modification of cycloid gears and system errors. The system's differential equations were derived by analyzing the relative displacement relationships between each component, and a numerical method was used to solve the equations. The effects of errors such as machining errors, assembly errors, and bearing clearances on the dynamic behaviors and transmission precision were investigated by comparison.
Article
Automation & Control Systems
Xiao Zhang, Weiguo Huang, Rui Wang, Yi Liao, Chuancang Ding, Jun Wang, Juanjuan Shi
Summary: Due to the lack of labeled data, deep learning-based methods perform poorly in fault diagnosis. To address this issue, we propose a data augmentation method called Multi-Stage Distribution Correction (MSDC) for few-shot fault diagnosis. This method corrects the training data by clustering and extracting Gaussian statistics, effectively expanding the training dataset and improving classification performance.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Engineering, Industrial
Rui Wang, Weiguo Huang, Yixiang Lu, Xiao Zhang, Jun Wang, Chuancang Ding, Changqing Shen
Summary: This study proposes a novel domain generalization network for machinery fault diagnosis where interest data are completely unavailable during model training. Multiple domain-specific auxiliary classifiers are designed to effectively learn domain-specific features, and a convolutional auto-encoder module is used to remove the learned domain-specific features. A domain-invariant classifier with inter-domain alignment strategy is designed to learn generalization diagnostic knowledge among different source domains. Experiments validate the effectiveness of the proposed network, showing its promising potential for fault diagnosis tasks in practical scenarios.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Multidisciplinary
Chuancang Ding, Weiguo Huang, Changqing Shen, Xingxing Jiang, Jun Wang, Zhongkui Zhu
Summary: The fault diagnosis of rotating machine is crucial for ensuring operational safety and preventing catastrophic accidents. This paper presents a unique time-frequency analysis method called synchroextracting frequency synchronous chirplet transform (SEFSCT) for vibration signal analysis and fault diagnosis of rotating machinery. The SEFSCT generates high-quality time-frequency representation, enabling accurate identification of time-varying instantaneous frequencies and mechanical failures.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2023)
Article
Engineering, Electrical & Electronic
Juntao Ma, Weiguo Huang, Yi Liao, Xingxing Jiang, Chuancang Ding, Jun Wang, Juanjuan Shi
Summary: The diagnosis of early bearing faults is crucial for machine condition monitoring. The existing sparse low-rank (SLR) methods have limitations in accurately estimating amplitude and approximating singular values (SVs). To address this, a novel SLR matrix estimation method with nonconvex enhancement (SLRNE) is proposed in this article. The method extracts fault transients from observed noisy signals, leveraging their sparse and low-rank properties in the time-frequency domain. Simulated and experimental signals confirm the effectiveness of SLRNE, and contrast experiments demonstrate its superiority.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Engineering, Industrial
Mingkuan Shi, Chuancang Ding, Rui Wang, Changqing Shen, Weiguo Huang, Zhongkui Zhu
Summary: “This paper proposes a graph embedding based deep broad learning system (GEDBLS) for addressing the problem of imbalanced data fault diagnosis in rotating machinery. GEDBLS utilizes category and structure information for reconstruction and learns high-level abstract features of vibration signals through a progressive encoding and decoding mechanism. Additionally, GEDBLS considers category weights and intra-class tightness for imbalanced data classification.”
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Electrical & Electronic
Zhipeng Ma, Ming Zhao, Chao Gou
Summary: Residual signal analysis is a promising tool for health monitoring of rotating machinery. The challenge lies in establishing an accurate time series model that can highlight fault features. The probabilistic TSM (PTSM) approach addresses this challenge by introducing recursive Gaussian process regression (GPR) and an improved multiple kernel learning (MKL) method. The results demonstrate the capability of recovering failure symptoms even under strong noise interference.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)