A Multisource Dense Adaptation Adversarial Network for Fault Diagnosis of Machinery

Huang, Ziling; Lei, Zihao; Wen, Guangrui; Huang, Xin; Zhou, Haoxuan; Yan, Ruqiang; Chen, Xuefeng

doi:10.1109/tie.2021.3086707

Cited by 73 publications

(14 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Accuracy (%) Uncertainty (%) Proposed method 99.93 5.95 SDANN [34] 93.84 / MDAAN [35] 95.45 / CWT-CNN-gcForest [28] 99.80 / improved DCGAN [29] 98.99 / SDANN [31] 98.97 / The limitations of this study should be addressed in future studies. Given the accuracy and robustness, the VGGarchitecture was chosen as the basic classifier.…”

Section: Methodsmentioning

confidence: 99%

Trustworthy Fault Diagnosis With Uncertainty Estimation Through Evidential Convolutional Neural Networks

Zhou

Wen-he

Cheng

et al. 2023

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

Deep neural networks (DNNs) have been widely used for intelligent fault diagnosis under the closed world assumption that any testing data is within classes of the training data. However, in reality, out-of-distribution (OOD) cases such as new fault conditions can happen after the original trained model is deployed. Most of the current DNNs are deterministic which can misclassify with high confidence in the open-world scenario. This overconfident behavior would not guarantee the reliability and robustness of fault diagnosis results in practice. Therefore, trustworthy intelligent fault diagnosis with uncertainty estimation is crucial for real applications. In this paper, we develop a novel convolutional neural network integrating evidence theory to achieve fault classifications with prediction uncertainty estimation. The estimated prediction uncertainty can identify potential OOD samples. This approach allows a minimal modification of the state-of-the-art DNN model by using a riskcalibrated evidential loss function and Dirichlet distribution that replaces the classification probabilities. The experimental results show that the proposed approach can not only achieve accurate classification of known classes but also detect unknown classes effectively. The proposed method shows significant potential in detecting OOD patterns and provides trustworthy fault diagnosis in open and non-stationary environments.

show abstract

Section: Methodsmentioning

confidence: 99%

Trustworthy Fault Diagnosis With Uncertainty Estimation Through Evidential Convolutional Neural Networks

Zhou

Wen-he

Cheng

et al. 2023

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

show abstract

“…Unsupervised feature extraction with encoder. In the unsupervised feature extraction part of the encoder, the idea of the convolutional fusion network and the data-feature composite fusion framework theory are introduced (Huang et al, 2021). The fusion convolutional layer is composed of three-layer standard convolution and two types of atrous convolution, which is used as the first layer of the encoder network to synthesize the input data, expand the perceptual field, and extract features of many different types of scales.…”

Section: The Adversarial Fusion Convolutional Autoencodermentioning

confidence: 99%

Bearing performance degradation assessment using adversarial fusion convolutional autoencoder based on multi-source information

Wang,

Zhou,

Wen

et al. 2023

Transactions of the Institute of Measurement and Control

Self Cite

View full text Add to dashboard Cite

Bearing operation states will directly determine the performance of the equipment; thus, monitoring operation status and degradation indicators is the key to ensuring continuous and healthy operation of the equipment. However, most of the research uses single-source information data, which makes it difficult to model when dealing with multi-source information, complex data distribution, and noise. In this paper, a bearing performance degradation assessment method based on multi-source information is proposed to comprehensively utilize the data signals of different structures, spaces, types, and sources. First, the adversarial fusion convolutional autoencoder is constructed for obtaining the degradation index of the bearing, while the adversarial learning strategy is applied to achieve the effect of enhancing the robustness and sensitivity of the degradation indicators extracted by the network. Then the degradation index is input into the support vector data description to determine the fault anomalies of the degradation index adaptively and the fuzzy c-means algorithms to obtain the final rolling bearing performance degradation evaluation results. Through the verification results of two experiment datasets, it is found that the proposed model can achieve accurate evaluation and quantitative analysis of the performance degradation process of bearings. As a result, the entire network ensures the reconstruction accuracy of normal samples while simultaneously stretching the reconstruction error of abnormal samples to achieve accurate monitoring of degradation onset.

show abstract

“…Convolutional neural networks have received a lot of attention for their powerful feature learning capabilities. It has achieved great success in the fields of fault diagnosis [17][18][19][20], life medicine [21][22][23], and target detection [24][25][26][27]. By combining CNN with post-weld quality inspection, the literature [28] used a combination of CNN and long short-term memory (LSTM) to achieve an implicit mapping of molten pool images to weld defects for accurate identification of weld defects.…”

Section: Introductionmentioning

confidence: 99%

Online identification of burn-through and weld deviation in sheet lap MIG welding based on YOLOv5

Wang,

Zhang,

Qin

et al. 2023

Meas. Sci. Technol.

Self Cite

View full text Add to dashboard Cite

Online monitoring of the welding process is very important to the development of the modern manufacturing industry. However, monitoring welding defects in real-time and with high accuracy is challenging because welding is inherently a dynamic, non-linear process. In this paper, the forming defects generated during the melt Inert-gas welding of sheet steel as the object of research. We propose a multi-task simultaneous monitoring system for welding defects based on the YOLOv5 model, which can achieve highly accurate simultaneous detection of burn-through and weld deviation defects. Firstly, a passive vision system has been designed that can filter out strong arc light interference. In combination with a self-made camera bracket with flexible adjustment and repeatable positioning, clear images of the molten pool can be acquired. Compared to common methods that require a laser for weld deviation detection, all monitoring tasks can be performed with just one charge coupled device camera, significantly reducing the cost of deploying the system. Secondly, to obtain experimental data closer to the industrial field environment. We simulated different degrees of local and continuous burn-through by changing three parameters: welding current, plate thickness, and lap width. Four different deviations: left, right, and tilt were designed to obtain rich weld deviation image data. Finally, based on the above data, we compared several aspects such as detection speed, recognition accuracy, and loss function. The YOLOv5s model with the smallest model parameters was finally selected as the base model. The results show that our proposed multi-task simultaneous monitoring system achieves an average accuracy of 98.84% for the identification of four molten pool states. Burn-through defect and weld deviation detection with more than 50 frames per second can be achieved in the online detection state. This paper can provide some guidance for the online monitoring of the welding manufacturing process.

show abstract

A Multisource Dense Adaptation Adversarial Network for Fault Diagnosis of Machinery

Cited by 73 publications

References 33 publications

Trustworthy Fault Diagnosis With Uncertainty Estimation Through Evidential Convolutional Neural Networks

Trustworthy Fault Diagnosis With Uncertainty Estimation Through Evidential Convolutional Neural Networks

Bearing performance degradation assessment using adversarial fusion convolutional autoencoder based on multi-source information

Online identification of burn-through and weld deviation in sheet lap MIG welding based on YOLOv5

Contact Info

Product

Resources

About