Multi-label fault diagnosis of rolling bearing based on meta-learning

Yu, Chongchong; Ning, Yaqian; Qin, Yong; Su, Weijun; Zhao, Xia

doi:10.1007/s00521-020-05345-0

Cited by 57 publications

(14 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Metalearning is referred to as "learningto-learn," which is generally utilized for tackling few-shot image classification. In recent years, metalearning research can be generally divided into three different categories, including optimizing the initialization parameters of the metalearner to quickly adapt to new tasks [43][44][45], generating the metric-learning network by judging the feature similarity between sample pairs [21,46], and learning a recurrent neural network model with memory storage function [47]. Different from traditional deep learning methods, metalearning is a flexible framework that learns prior experience from multiple relevant tasks, which relies on the obtained experience to improve its performance on target tasks without training from scratch [42].…”

Section: Related Research Workmentioning

confidence: 99%

“…Model agnostic metalearning (MAML) is a metalearning method based on parameter optimization proposed by Finn et al [43], which had demonstrated excellent generalization ability in image recognition for processing new tasks under a small number of training samples. Yu et al [44] proposed a metalearning fault diagnosis model based on gradient optimization [45], which optimized the initial parameters of the model network through the scenario training mechanism, so that the model can also perform fault diagnosis efficiently and quickly under the condition of limited training data. Zhang et al [21] applied Siamese network to fault diagnosis with data imbalanced.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An Improved Metalearning Framework to Optimize Bearing Fault Diagnosis under Data Imbalance

Man

Yang

et al. 2022

Journal of Sensors

View full text Add to dashboard Cite

The intelligent diagnosis of rotating machinery with big data has been widely studied. However, due to the variability of working conditions and difficulty in marking fault samples, it is difficult to obtain enough high-quality fault marking data for training bearing fault diagnosis models in practical industrial application scenarios. Aiming at the problem of training data imbalance caused by lack of fault samples, a novel metalearning fault diagnosis method (MOFD) is proposed to get the bearing fault diagnosis solution under data imbalance. Firstly, in order to enhance the variety of fault samples, a Feature Space Density Adaptive Synthetic Minority Oversampling Technique (FSDA-SMOTE) is proposed in this paper, which takes the density difference of minority samples in the spatial domain within the class as the constraint of local neighbor similarity to generate new fault samples for data augmentation. In addition, in order to strengthen the model’s learning ability and diagnosis performance under limited fault samples, a residual-attention convolutional neural network (RA-CNN) was constructed to identify the deep features of fault signals, and a metalearning strategy based on parameter gradient optimization was applied to RA-CNN for refining the learning process of the diagnosis model. Finally, the reliability of the proposed method is verified through experimental analysis of public bearing dataset.

show abstract

Section: Related Research Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

An Improved Metalearning Framework to Optimize Bearing Fault Diagnosis under Data Imbalance

Man

Yang

et al. 2022

Journal of Sensors

View full text Add to dashboard Cite

show abstract

“…There are many achievements in fault diagnosis based on meta-learning methods recent years [9]. Yu et al [10] combined the model-agnostic meta learning (MAML) with multilabel CNN, demonstrated the practicability of MAML in solving the few-shot fault diagnosis problems. Zhang et al [11] proposed a few-shot learning approach for bearing fault diagnosis based on Siamese neural network which is a kind of metric-based meta-learning.…”

Section: Introductionmentioning

confidence: 99%

Transfer multiscale adaptive convolutional neural network for few-shot and cross-domain bearing fault diagnosis

Wang

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Bearing intelligent fault diagnosis has been researched comprehensively in recent years. However, the scarcity of labelled training samples and various working conditions seriously hinder the widespread application of deep learning based fault diagnosis methods. To address this problem, we propose a transfer multiscale adaptive convolutional neural network (TMACNN), which significantly enhances the performance of deep learning based methods on few-shot and cross-domain bearing fault diagnosis in terms of network architecture and transfer strategy. On the one hand, we design a novel multiscale adaptive convolutional neural network (MACNN) framework that effectively improves the feature extraction and generalization abilities for bearing fault diagnosis by introducing mega-scale convolutions and continuous stacked multiscale convolutions. On the other hand, we propose an innovative transfer strategy for the MACNN that freezes the six stacked multiscale convolutional feature extraction units and fine-tunes the mega-scale convolution unit and the classifier, which are more suitable for few-shot transfer learning. In experiments on the CWRU dataset and Paderborn dataset, our proposed TMACNN outperforms various advanced baseline models for few-shot and cross-domain bearing fault diagnosis.

show abstract

“…Georgoulas et al [ 14 ] applied the problem transformation strategy in multi-label learning to appropriately represent the faults that occurred concurrently in an induction motor, and combined with the classifier to achieve an efficient diagnosis. Yu et al [ 15 ] came up with a multi-label convolutional neural network incorporating meta-learning to address few-shot fault diagnosis.…”

Section: Introductionmentioning

confidence: 99%

Exploring the essence of compound fault diagnosis: A novel multi-label domain adaptation method and its application to bearings

Chu¹,

Li²,

Yang³

et al. 2023

Heliyon

View full text Add to dashboard Cite

Multi-label fault diagnosis of rolling bearing based on meta-learning

Cited by 57 publications

References 39 publications

An Improved Metalearning Framework to Optimize Bearing Fault Diagnosis under Data Imbalance

An Improved Metalearning Framework to Optimize Bearing Fault Diagnosis under Data Imbalance

Transfer multiscale adaptive convolutional neural network for few-shot and cross-domain bearing fault diagnosis

Exploring the essence of compound fault diagnosis: A novel multi-label domain adaptation method and its application to bearings

Contact Info

Product

Resources

About