Related references
Note: Only part of the references are listed.
Article
Engineering, Multidisciplinary
Ling Xiang et al.
Summary: A novel transfer unsupervised learning method called CCDAN is proposed in this paper to solve the problem of data without failure samples. The CCDAN extracts transfer characteristics from simulated samples by a theory model for experimental rotor fault diagnosis. The proposed method uses a dynamical model of Jeffcott rotor with crack fault to generate simulation samples and adopts multilayer convolutional network and MK-MMD to extract similar characteristics between source and target domains. Two independent classifiers are designed to constrain the features near the decision boundary in network learning, improving accuracy and fault identifying generalization.
Article
Engineering, Industrial
Hui Wang et al.
Summary: Digital twin is the embodiment of the most advanced achievements of the current simulation technology theory development and the direction of intelligent development in the future. However, it is a great challenge to really integrate it into practical project application. Motivated by DT, an application method combining numerical simulation model and machine learning classification is proposed to show the advantages of digital twin.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Industrial
Yifei Ding et al.
Summary: The success of deep learning and transfer learning has expanded the scope of fault diagnosis, especially in improving diagnosis accuracy under multiple working conditions. However, most existing approaches do not account for the diversity of fault mode distributions and weaken the generalization to imbalanced domain adaptation scenarios. This work proposes a novel deep imbalanced domain adaptation framework for fault diagnosis, which overcomes class-imbalanced label shift and improves cross-domain generalization for IDA tasks.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Multidisciplinary
Anil Kumar et al.
Summary: A non-parametric complementary ensemble empirical mode decomposition (NPCEEMD) method is proposed for identifying bearing defects using weak features. Unlike existing decomposition methods, NPCEEMD does not require defining the ideal SNR of noise and the number of ensembles each time while processing the signals. Simulation results show that NPCEEMD has less mode mixing compared to existing decomposition methods. After conducting in-depth simulation analysis, the proposed method is applied to experimental data. The steps of the NPCEEMD method include obtaining the raw signal, decomposing the obtained signal, computing the mutual information (MI) of the raw signal with NPCEEMD-generated IMFs, selecting IMFs with MI above 0.1, combining them to form a resulting signal, and computing the envelope spectrum of the resulting signal to confirm the presence of defect.
Article
Engineering, Mechanical
Ke Feng et al.
Summary: Gearbox is widely used as a power transmission system in various applications due to its compact structure, stable transmission capability, and high transmission efficiency. However, it usually operates in harsh and non-stationary working environments, causing gear surface degradation. To ensure reliable operation, it is essential to assess the progression of gear surface degradation. This paper develops a digital twin-driven intelligent health management method to monitor and assess the gear surface degradation progression, which can accurately predict the remaining useful life (RUL) of the gearbox.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Engineering, Industrial
Yongchao Zhang et al.
Summary: Fault diagnosis of rolling bearings is crucial for ensuring the reliability, safety, and efficiency of mechanical systems. However, traditional data-driven methods require a dataset of full failure modes, which may not always be available, limiting their practicality. This study proposes a digital twin-driven approach for fault diagnosis of rolling bearings with insufficient training data, using a dynamics-based virtual representation and a Transformer-based network for diagnostics. Experimental results demonstrate high diagnostic performance even with unlabeled real-world data and unknown health conditions, highlighting the significant benefits of the proposed method for health management of critical rolling bearings.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Industrial
Jinrui Wang et al.
Summary: With the advance of digital twin, machine health management has become the focus of equipment monitoring upgrading. The digital twin model can generate system performance data close to reality, providing a new way for cyber-physical integration of equipment monitoring. This paper proposes a digital twin aided intelligent fault diagnosis model for a triplex pump. The experimental results show that the proposed method is effective and superior to other advanced transfer learning methods.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Computer Science, Artificial Intelligence
Jian Lin et al.
Summary: This paper proposes a generalized model-agnostic meta-learning method for few-shot fault diagnosis of bearings driven by heterogeneous signals under various operating conditions. The proposed method utilizes a channel interaction feature encoder with multi-kernel efficient channel attention to focus on mutual fault information and effectively extract general diagnostic knowledge for multiple diagnostic meta-tasks. Additionally, a flexible weight guidance factor is designed to adjust the training strategy and optimize the inner loop weights for different diagnostic meta-tasks, improving the overall generalization performance. The method is validated through the analysis of acceleration and acoustic signals of bearings, demonstrating its extensiveness and effectiveness in various few-shot cross-domain scenarios.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Hardware & Architecture
Mingzhi Chen et al.
Summary: This paper proposes an improved local fusion generative adversarial network to address the challenges of dealing with extremely limited training samples and effectively fusing representative and diverse information. The proposed method is applied to the analysis of planetary gearbox vibration signals and the results show that it can generate more similar and diverse samples compared to the existing mainstream GANs using only six training samples in each type, significantly improving the accuracy of fault diagnosis task.
IEEE TRANSACTIONS ON RELIABILITY
(2023)
Article
Engineering, Mechanical
Xingkai Chen et al.
Summary: Most of the existing research on unsupervised cross-domain intelligent fault diagnosis focuses on single-source domain adaptation, lacking the ability to utilize multiple source domains with diverse diagnostic information. This paper proposes a dual adversarial guided unsupervised multi-domain adaptation network (DAG-MDAN) to better extract common features and integrate multi-source domain knowledge. The DAG-MDAN includes an edge adversarial module (EA-Module) and an inner adversarial module (IA-Module) to enhance domain confusion and a multi-subnet collaborative decision module (MCD-Module) for better fusion decisions.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Engineering, Industrial
Shen Yan et al.
Summary: This paper proposes a new machinery anomaly detection method called full graph dynamic autoencoder (FGDAE) for complex operating conditions. It develops a full connected graph (FCG) to obtain global structure information and constructs a graph adaptive autoencoder (GAAE) model to aggregate multi-perspective feature information between channels. The method achieves better performance compared to other popular anomaly detection methods on machinery datasets.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Review
Engineering, Electrical & Electronic
Zhijun Ren et al.
Summary: The theoretical developments of data-driven fault diagnosis methods have been successful in the industry. However, most methods are developed based on the assumption of data balance, which is not applicable in engineering scenarios. The consequence of data imbalance has attracted attention, but a comprehensive review of achievements and research perspectives is still lacking.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Engineering, Multidisciplinary
Yizhen Peng et al.
Summary: This paper proposes a new unbalanced fault diagnosis framework for rolling bearings, which combines Wasserstein loss and hierarchical feature matching loss to improve data validity from both global and class-specific perspectives.
Article
Computer Science, Artificial Intelligence
Wei Li et al.
Summary: This study proposes a modified auxiliary classifier GAN (MACGAN) model to address the issue of insufficient samples in fault diagnosis. By improving the ACGAN framework and introducing techniques such as Wasserstein distance and spectral normalization, the proposed method can generate high-quality multi-mode fault samples more effectively, improving the accuracy and stability of fault diagnosis models.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Zhenyu Wu et al.
Summary: A deep adversarial transfer learning model (Deep Imba-DA) is proposed in this paper to tackle the issues of class imbalances and distributional discrepancies in bearing fault diagnosis.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Automation & Control Systems
Yaoxiang Yu et al.
Summary: In this paper, a method for industrial mechanical fault diagnosis is proposed to address the issues of imbalanced datasets and parameter determination. By utilizing Pareto-optimal adaptive loss residual shrinkage network, the candidate architectures are encoded and appropriate optimization targets are selected to highlight minority samples and optimize performance. Experimental results demonstrate the superior performance of the proposed method in imbalanced classification.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Automation & Control Systems
Hui Liu et al.
Summary: This article proposes a novel imbalanced data classification method based on weakly supervised learning. It introduces real-world samples to solve the class imbalance problem and uses bidirectional gated recurrent units to construct a fault diagnostic model. A new loss function is proposed to reduce the impact of noise.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Zhijun Ren et al.
Summary: This paper proposes a strategy to calculate sample costs adaptively to address the issue of imbalanced industrial data, demonstrating significant improvement in model performance under different sample sets. Analysis of training losses and relationships between parameters reveals the advantages of the proposed strategy.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Engineering, Mechanical
Xiaoli Zhao et al.
Summary: This paper proposes a class imbalanced fault diagnosis framework based on Normalized Conditional Variational Auto-Encoder with Adaptive Focal Loss (NCVAE-AFL), which enhances the feature learning ability of data and balances the diagnosis difficulty for mechanical systems with class imbalanced fault data.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Engineering, Mechanical
Tianfu Li et al.
Summary: Deep learning methods have advanced the field of Prognostics and Health Management, but handling irregular data in non-Euclidean space remains a challenge. Research has proposed a practical guideline for utilizing graph neural networks for intelligent fault diagnostics and prognostics, and established a framework based on GNN for this purpose.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Chemistry, Analytical
Elsie Fezeka Swana et al.
Summary: Data-driven methods play a prominent role in the development of condition monitoring systems for electrical machines, but they face challenges like data imbalance. This research compares different classifiers and resampling techniques for handling imbalanced fault data. The results show that a combination of synthetic minority oversampling technique and Tomek link outperforms other techniques in terms of performance.
Article
Computer Science, Artificial Intelligence
K. Ruwani M. Fernando et al.
Summary: This paper proposes a class rebalancing strategy based on a class-balanced dynamically weighted loss function to address class distribution imbalance in deep learning. Experimental results demonstrate the robust generalization of this strategy across different applications such as cyber intrusion detection and medical imaging.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Xuan Liu et al.
Summary: In this paper, a transfer residual network augmented with explicit weight self-assignment strategy based on meta data (TRN-EWM) is proposed for cross-domain fault diagnosis. By extracting depth features and conducting class imbalanced cross-domain transfer, this method effectively addresses the diagnosis difficulties in actual working conditions and achieves high diagnosis accuracy.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Engineering, Mechanical
Rujie Hou et al.
Summary: In practical mechanical fault diagnosis, obtaining fault data is difficult and there is a great imbalance between normal and fault data. To address this issue, this study proposes a two-stage training approach and utilizes a contrastive-weighted self-supervised model with augmented vision transformer to handle the imbalance learning strategies. The pre-training stage effectively learns a good initialized encoder for better classification of long-tailed data in downstream tasks.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Multidisciplinary Sciences
Meng-Meng Song et al.
Summary: This research develops a simulation data-driven deep transfer learning fault diagnosis method for planetary gearboxes, which utilizes a dynamic simulation model to collect various fault simulation data and uses a deep transfer learning network model to learn domain invariant features for fault diagnosis on real data.
SCIENTIFIC REPORTS
(2022)
Article
Engineering, Electrical & Electronic
Yaoxiang Yu et al.
Summary: In real industrial scenarios, the issue of data imbalance in machine fault diagnosis, where the number of healthy samples exceeds the number of fault ones, affects the accuracy and robustness of data-based fault diagnosis methods. To address this, a method based on data augmentation and a new GAN model called PCWGAN-GP are proposed. The PCWGAN-GP model generates high-quality faulty samples to balance the imbalanced dataset, improving the performance of fault diagnosis. Experimental results on a bearing dataset confirm the superiority of this proposed method.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Engineering, Multidisciplinary
Shaowei Liu et al.
Summary: A novel data augmentation method, named variational autoencoding generative adversarial networks with deep regret analysis, is proposed to improve the fault diagnosis ability of rolling bearings. The method integrates an encoder, discriminator, and feature matching module to enhance data generation quality and reduce over-fitting. The results show that the proposed method is more effective and robust compared to typical data synthesis based fault diagnosis methods.
Article
Computer Science, Artificial Intelligence
Yongchun Zhu et al.
Summary: This study introduces a deep subdomain adaptation network (DSAN) that aligns relevant subdomain distributions across different domains based on the local maximum mean discrepancy (LMMD). DSAN is simple but effective, does not require adversarial training, and converges quickly. It can be easily integrated into feedforward network models to achieve efficient adaptation via backpropagation.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Engineering, Electrical & Electronic
Zhibin Zhao et al.
Summary: Recent progress in intelligent fault diagnosis relies heavily on deep representation learning and labeled data. However, the unsupervised deep transfer learning-based IFD still faces many open and fundamental issues.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Engineering, Mechanical
James Xi Gu et al.
Summary: This study proposes a novel method for diagnosing the gradual deterioration of gears using modulation signal bispectrum (MSB) and vibration measurements. Experimental results show that small gears deteriorate faster at low-speed stages due to poorer lubrication conditions. The developed MSB-based method outperforms other methods in detecting early gear deterioration.
INTERNATIONAL JOURNAL OF HYDROMECHATRONICS
(2021)
Article
Engineering, Mechanical
Zhiyi He et al.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2020)
Article
Engineering, Electrical & Electronic
Andongzhe Duan et al.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2020)
Article
Automation & Control Systems
Siyu Shao et al.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2019)
Article
Engineering, Mechanical
Yong Yi et al.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2019)
Article
Engineering, Mechanical
Anil Kumar et al.
PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART J-JOURNAL OF ENGINEERING TRIBOLOGY
(2018)
Article
Engineering, Mechanical
Feng Jia et al.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2018)
Article
Materials Science, Characterization & Testing
Anil Kumar et al.
JOURNAL OF NONDESTRUCTIVE EVALUATION
(2017)
Article
Engineering, Mechanical
Anil Kumar et al.
TRIBOLOGY TRANSACTIONS
(2017)