Deep Transfer Learning in Mechanical Intelligent Fault Diagnosis: Application and Challenge

Qian, Chenhui; Zhu, Junjun; Shen, Yehu; Jiang, Quan; Zhang, Qingkui

doi:10.1007/s11063-021-10719-z

Cited by 48 publications

(21 citation statements)

References 86 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Qian et al [46] combined the SDA method with the adaptive probability distribution to reduce the distribution difference between the source domain and the target domain. Zhang et al [47] fault features embedded in the popular subspace for feature transformation and then fault classification to reduce the domain displacement phenomenon.…”

Section: Domain Adaptation In the Fault Diagnosis Fieldmentioning

confidence: 99%

Mechanical fault diagnosis based on deep transfer learning: a review

Yang

Zhang

Yong-zheng

2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Mechanical fault diagnosis is an important method to accurately identify the health condition of mechanical equipment and ensure its safe operation. With the advent of the era of "big data", it is an inevitable trend to choose deep learning for mechanical fault diagnosis. At the same time, to improve the generalization ability of deep learning applications in different scenarios of fault diagnosis, mechanical diagnosis based on transfer learning has also been proposed and become an important branch in the field of mechanical fault diagnosis. This paper introduces the principle of transfer learning, summarizes the research and application of transfer learning in the field of fault diagnosis, discusses the shortcomings of transfer learning in the field of fault diagnosis, and discusses the future research direction of transfer learning in the field of fault diagnosis.

show abstract

Section: Domain Adaptation In the Fault Diagnosis Fieldmentioning

confidence: 99%

Mechanical fault diagnosis based on deep transfer learning: a review

Yang

Zhang

Yong-zheng

2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…The module takes the hidden layer output ℎ 𝑡−1 of the past moment and the input 𝑥 𝑡 of the current moment as inputs. Then, the output data 𝑍 𝑡 of the "update gate" are used to describe the degree to which the hidden layer state information of the past moment is brought into the hidden layer state of the current moment, which can be calculated by (12). The gate threshold is determined by the Sigmoid function.…”

Section: B Gated Recurrent Unit (Gru)mentioning

confidence: 99%

“…With the continuous growth in the scope of fault diagnosis objects, the number of sensors installed on the mechanical apparatus to be diagnosed, and the increased sampling frequency, the obtained data samples have also increased exponentially. Consequently, the fault diagnosis field has joined the big-data era [12]. When the traditional fault diagnosis method faces huge amounts of data, it usually experiences problems such as insufficient model generalization, poor robustness, the difficulty of human operation, and largescale work content.…”

Section: Introductionmentioning

confidence: 99%

Fault Diagnosis Method for Imbalanced Data of Rotating Machinery Based on Time Domain Signal Prediction and SC-ResNeSt

et al. 2023

View full text Add to dashboard Cite

In an actual engineering environment, some rotating machines are usually in normal operation, but their time in a fault state is very short, which leads to a serious imbalance in the fault diagnosis datasets for rotating machinery, and gives the traditional network model the shortcomings of poor stability and low accuracy in practical engineering applications. To solve this problem, we propose a fault diagnosis method based on the combination of a new Dual-stage Attention-based Recurrent Neural Network (DA-RNN) and depth residual dispersion self-calibration convolution network (SC-ResNeSt). Firstly, a novel DA-RNN network with a gated cycle unit (GRU) as a coding-decoding unit was designed, and the network was used to predict and expand the scarce fault signals. Secondly, to make full use of the time domain information of vibration signals, a new image coding method, namely, Gram Angle Product Field (GAPF), was proposed. Then, because the traditional convolution layer lacks a dynamic receptive field to extract more representative features, self-calibrated convolution modules were introduced on the basis of the distraction network (ResNeSt), and a new network model, SC-ResNeSt, was established. Finally, the expanded vibration signal is converted into GAPF, which is used as the input for the SC-ResNeSt network to classify the fault types. To check the performance of the model, the Case Western Reserve University rolling bearing dataset and planetary gearbox dataset were used for testing. Ultimately, good results were obtained in a prediction experiment for bearing fault samples and a subsequent fault diagnosis experiment for bearings and gears, which verified the feasibility and practicability of the model.

show abstract

“…The emergence of Deep Learning (DL) can solve the problem of relying on the workforce [ 11 ]. The DL model represented by the Convolution Neural Network (CNN) and Long Short-Term Memory (LSTM) neural network is used in fault diagnosis, with promising results because of the powerful feature extraction capabilities [ 12 , 13 , 14 , 15 , 16 , 17 , 18 ]. Among most research, applying DL to a fault diagnosis requires two preconditions: (1) Test samples and samples participating in model training have the same label space.…”

Section: Introductionmentioning

confidence: 99%

Cross-Domain Open Set Fault Diagnosis Based on Weighted Domain Adaptation with Double Classifiers

Wang

Tong

et al. 2023

Sensors

View full text Add to dashboard Cite

The application of transfer learning in fault diagnosis has been developed in recent years. It can use existing data to solve the problem of fault recognition under different working conditions. Due to the complexity of the equipment and the openness of the working environment in industrial production, the status of the equipment is changeable, and the collected signals can have new fault classes. Therefore, the open set recognition ability of the transfer learning method is an urgent research direction. The existing transfer learning model can have a severe negative transfer problem when solving the open set problem, resulting in the aliasing of samples in the feature space and the inability to separate the unknown classes. To solve this problem, we propose a Weighted Domain Adaptation with Double Classifiers (WDADC) method. Specifically, WDADC designs the weighting module based on Jensen–Shannon divergence, which can evaluate the similarity between each sample in the target domain and each class in the source domain. Based on this similarity, a weighted loss is constructed to promote the positive transfer between shared classes in the two domains to realize the recognition of shared classes and the separation of unknown classes. In addition, the structure of double classifiers in WDADC can mitigate the overfitting of the model by maximizing the discrepancy, which helps extract the domain-invariant and class-separable features of the samples when the discrepancy between the two domains is large. The model’s performance is verified in several fault datasets of rotating machinery. The results show that the method is effective in open set fault diagnosis and superior to the common domain adaptation methods.

show abstract

Deep Transfer Learning in Mechanical Intelligent Fault Diagnosis: Application and Challenge

Cited by 48 publications

References 86 publications

Mechanical fault diagnosis based on deep transfer learning: a review

Mechanical fault diagnosis based on deep transfer learning: a review

Fault Diagnosis Method for Imbalanced Data of Rotating Machinery Based on Time Domain Signal Prediction and SC-ResNeSt

Cross-Domain Open Set Fault Diagnosis Based on Weighted Domain Adaptation with Double Classifiers

Contact Info

Product

Resources

About