相关参考文献
注意：仅列出部分参考文献，下载原文获取全部文献信息。
Article
Computer Science, Artificial Intelligence
Ke Zhao et al.
Summary: This study develops a multi-source domain transfer learning approach to address the challenges of cross-domain fault diagnosis. The proposed method adopts an indirect latent alignment idea to achieve better feature alignment and designs an ingenious conditional weighting strategy to quantify the similarity of different source domains to the target domain. Experimental results demonstrate that the proposed method can sufficiently transfer knowledge from all source domains to the target domain and has extensive application prospects.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Pengfei Ge et al.
Summary: Unsupervised domain adaptation (UDA) aims to generalize the supervised model trained on a source domain to an unlabeled target domain. To address the conditional dependence between features and labels as well as negative transfer, this paper proposes a Deep Conditional Adaptation Network (DCAN) that aligns the conditional distributions using Conditional Maximum Mean Discrepancy and extracts discriminant information from the target domain using mutual information.
PATTERN RECOGNITION
(2023)
Article
Automation & Control Systems
Nagendra Singh Ranawat et al.
Summary: In this study, a test rig is developed to detect blockages in the pump, and different models and feature sets are trained to find the most accurate and efficient method for diagnosing pump blockages.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Automation & Control Systems
Anurag Choudhary et al.
Summary: This paper presents a vibro-acoustic fusion technique for accurate fault diagnosis of induction motors (IMs) under varying working conditions. The proposed method utilizes the Multi Input-Convolutional Neural Network (MI-CNN) to fuse the features of vibration and acoustic signals. Experimental results show that the suggested methodology is accurate and reliable for IMs and other rotating machine components.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Zhiheng Su et al.
Summary: The application of traditional deep learning methods for intelligent fault diagnosis is limited by the distribution discrepancy of unlabeled data. Transfer learning can overcome this limitation by generalizing a model trained on labeled data in the source domain to solve fault diagnosis in the target domain with unlabeled data. However, current transfer learning methods may face difficulties in complex and heterogeneous distributions and may cause incorrect alignment of data with the greatest distribution discrepancy across the domains. This paper proposes a deep transfer learning method with inter-domain decision discrepancy minimization (InDo-DDM) to address these issues and outperforms other widely used methods in experimental scenarios.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Engineering, Mechanical
Yongchao Zhang et al.
Summary: Cross-domain machinery fault diagnosis aims to transfer enriched diagnosis knowledge from a labeled source domain to a new unlabeled target domain. Existing methods often assume prior knowledge of fault modes in the target domain, which is rare in engineering practice. This study proposes a source-free domain adaptation method that can handle cross-domain fault diagnosis scenarios without source data and explicit assumptions about target fault modes.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Engineering, Mechanical
Quan Qian et al.
Summary: An improved joint distribution adaptation (IJDA) mechanism is proposed to enhance the distribution alignment and match the marginal distributions as well as conditional distributions of two domains. It combines maximum mean discrepancy and CORrelation Alignment (CORAL) to enhance domain confusion and constructs an improved conditional distribution alignment mechanism. In addition, a new I-Softmax loss is introduced to contribute to feature learning and learn more separable features. Experimental results on six cross-machine diagnostic tasks demonstrate that the proposed DDTLN achieves higher performance in transfer fault diagnosis compared to other typical domain adaptation methods.
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, Biomedical
Mei-yu Zhong et al.
Summary: This paper proposes a novel EEG-based emotion recognition framework using TQWT-features and HCRNN model to extract emotional information from multichannel EEG signals. The experimental results on the SEED dataset demonstrate its effectiveness.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Mathematics
Fuqiang Liu et al.
Summary: In this paper, an entropy-optimized method based on an unsupervised domain-adaptive technique is proposed to enhance fault diagnosis (FD) model training. Pseudosamples and labels are generated through data augmentation and self-training strategies to reduce the distribution discrepancy between the source and target domains. Experimental results show that our method achieves significant accuracy improvement in rolling bearing FD compared to other state-of-the-art algorithms.
Article
Thermodynamics
Guannan Li et al.
Summary: Timely and accurate fault diagnosis in building energy systems can improve energy efficiency and sustainable development. This study proposes high-adaptability fault diagnosis models using deep transfer learning strategies. Experimental results show that the network-based fine-tuning method achieves 93% accuracy, a 55% improvement compared to the benchmark model. The impacts of data volume and transfer learning tasks are analyzed, and practical application issues are discussed.
Proceedings Paper
Computer Science, Artificial Intelligence
Aadarsh Sahoo et al.
Summary: This paper proposes a novel framework called SLM for partial domain adaptation in computer vision. The framework consists of three modules: select, label, and mix, which aim to learn discriminative invariant feature representations. The proposed framework addresses the problems of negative transfer, lack of discriminability, and domain invariance in the latent space.
2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV)
(2023)
Article
Automation & Control Systems
Yanxu Liu et al.
Summary: In this article, a deep adversarial subdomain adaptation network is proposed to reduce the distribution discrepancy between the source domain and target domain. The effectiveness and superiority of the proposed method are demonstrated through experimental results.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Thermodynamics
Deqiang He et al.
Summary: The health of bearings in flywheel energy storage systems is crucial for effective energy recovery during train braking, but diagnosing faults based on complex vibration signals is challenging. This study proposes a fault diagnosis method based on VMD energy entropy, which optimizes parameters to extract feature vectors and utilizes a deep learning model for fault diagnosis, achieving high diagnostic accuracy.
Article
Automation & Control Systems
Zhenzhen Jin et al.
Summary: In this paper, a weak fault diagnosis method for train axle box bearings is proposed based on parameter optimization Variational Mode Decomposition (VMD) and improved Deep Belief Network (DBN). By optimizing algorithm parameters and extracting fault feature information, the diagnostic accuracy of the bearings can be improved.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2022)
Article
Engineering, Multidisciplinary
Lanjun Wan et al.
Summary: This paper proposes a deep convolution multi-adversarial domain adaptation model for rolling bearing fault diagnosis, which achieves cross-domain fault diagnosis through feature extraction, domain adaptation, and fault identification modules.
Article
Automation & Control Systems
Quan Qian et al.
Summary: In this study, a novel multiscale weight-selection adversarial network (MWSAN) is proposed to enhance the effect of partial domain adaptation (DA). By using a designed multiscale domain adversarial network (MDAN) and a multiscale weight-selection mechanism for instance and class, the proposed method can achieve partial DA to a larger degree.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Engineering, Multidisciplinary
Yaowei Shi et al.
Summary: This paper proposes a novel multiscale feature adversarial fusion network (MFAFN) for rotating machinery fault transfer diagnosis, which improves the model's feature extraction ability at different scales and enhances diagnostic performance.
Article
Automation & Control Systems
Te Han et al.
Review
Computer Science, Artificial Intelligence
Saptarshi Sengupta et al.
KNOWLEDGE-BASED SYSTEMS
(2020)
Article
Automation & Control Systems
Xiang Li et al.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2020)
Article
Automation & Control Systems
Siyu Shao et al.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2019)
Review
Engineering, Mechanical
Wade A. Smith et al.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2015)
