A Two-Stage Transfer Adversarial Network for Intelligent Fault Diagnosis of Rotating Machinery With Multiple New Faults

Li, Jipu; Huang, Ruyi; He, Guolin; Liao, Yixiao; Wang, Zhen; Li, Weihua

doi:10.1109/tmech.2020.3025615

Cited by 139 publications

(41 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The UMU could be used for fault diagnosis of rotating machinery. Typically, for example accelerometers mounted to the gearbox are utilized to gather data from machines, and traditional signal processing or the increasingly popular machine learning techniques are used to identify the fault type [23]- [27].…”

Section: B Benefits Of the Umumentioning

confidence: 99%

On-Shaft Wireless Vibration Measurement Unit and Signal Processing Method for Torsional and Lateral Vibration

Koene

Haikonen

Tiainen

et al. 2022

IEEE/ASME Trans. Mechatron.

View full text Add to dashboard Cite

Microelectromechanical systems accelerometers have opened new possibilities for vibration monitoring of a rotating machinery. They enable mounting accelerometers directly to the rotating component of the machine, e.g., shaft. This enables not only the measurement of a lateral vibration but also a torsional vibration of the machine. This increases the vibration data gathered from the machine by one measurement instrument. This article presents an onshaft wireless universal measurement unit (UMU) with innovative combination of features, such as a high measurement range and easy mounting. The UMU has a two-sensor configuration where two accelerometers are mounted to the opposite sides of a shaft. This enables to utilize a novel signal processing method to separate torsional and lateral vibration from the data. The signal processing method combines and modifies methods presented in the published literature. The results presented in this article demonstrate that the UMU together with the presented signal processing method can measure frequencies of torsional and lateral vibration accurately. However, the amplitude comparison between the UMU and reference sensors cannot be done adequately because they are measuring either different components of the machine or different physical properties.

show abstract

Section: B Benefits Of the Umumentioning

confidence: 99%

On-Shaft Wireless Vibration Measurement Unit and Signal Processing Method for Torsional and Lateral Vibration

Koene

Haikonen

Tiainen

et al. 2022

IEEE/ASME Trans. Mechatron.

View full text Add to dashboard Cite

show abstract

“…Li et al. [27] proposed a multiple‐layer CNN combined with adversarial learning strategy for fault diagnosis with multiple new fault types. One‐dimension CNN and ensemble learning were used by Huang et al.…”

Section: Introductionmentioning

confidence: 99%

“…Zhang et al [26] proposed a convolutional neural network with two dropout layers and two fully connected (FC) layers for fault diagnosis, which was effective compared with other machine learning methods. Li et al [27] proposed a multiple-layer CNN combined with adversarial learning strategy for fault diagnosis with multiple new fault types. One-dimension CNN and ensemble learning were used by Huang et al [28] for compound fault decoupling and diagnosis through multisensory data.…”

mentioning

confidence: 99%

Intelligent fault diagnosis of rotating machinery using lightweight network with modified tree‐structured parzen estimators

Liang

Liao

Chen

et al. 2022

IET Collab Intel Manufact

Self Cite

View full text Add to dashboard Cite

Deep learning-based methods have been widely used in the field of rotating machinery fault diagnosis. It is of practical significance to improve the calculation speed of the model on the premise of ensuring accuracy, so as to realise real-time fault diagnosis. However, designing an efficient and lightweight fault diagnosis network requires expert knowledge to determine the network structure and adjust the hyperparameters of the network, which is time-consuming and laborious. In order to design fault diagnosis networks considering both time and accuracy effortlessly, a novel lightweight network with modified treestructured parzen estimators (LN-MT) is proposed for intelligent fault diagnosis of rotating machinery. Firstly, a lightweight framework based on global average pooling and group convolution is proposed, and a hyperparameter optimisation (HPO) method based on Bayesian optimisation called tree-structured parzen estimator is utilised to automatically search the optimal hyperparameters for the fault diagnosis task. The objective of the HPO algorithm is the weighting of accuracy and calculating time, so as to find models that balance both time and accuracy. The results of comparison experiments indicate that LN-MT can achieve superior fault diagnosis accuracies with few trainable parameters and less calculating time. K E Y W O R D S convolutional neural network, fault diagnosis, hyperparameter optimisation, neural architecture search | INTRODUCTIONModern electromechanical equipment is playing an important role in manufacturing and industry. Gears and bearings are key components of rotating machinery [1,2], the failure of which will greatly affect the performance of the machine, resulting in serious security risk and economic loss [3][4][5]. In order to ensure the continuous and stable operation of the equipment, an algorithm that can timely diagnose the failure of rotating machinery is required [6][7][8], so as to reduce economic losses. Hence, it is of great significance to develop a fault diagnosis algorithm for rotating machinery [9].

show abstract

“…To further improve the automation and intelligence of data-driven monitoring technology, deep transfer learning [26][27][28][29] 31 proposed a multi-task structural health monitoring method based on deep learning network, and transferred the high-level features of damage degree monitoring task to damage location recognition task. Feng et al 32 proposed a damage image detection method based on convolutional neural network with transfer learning, and used high-definition cameras to collect images for transfer training of the INCEP-V3 network.…”

Section: Introductionmentioning

confidence: 99%

“…To further improve the automation and intelligence of data-driven monitoring technology, deep transfer learning 26–29 is emerging to solve the problem of the universality and generalization of monitoring model in different scenarios. Deep transfer learning can utilize the existing knowledge in the source domain to solve different but related target domain problems, and has been effectively verified in machine vision and fault diagnosis.…”

Section: Introductionmentioning

confidence: 99%

Distribution adaptation deep transfer learning method for cross-structure health monitoring using guided waves

Zhang

Hong

Liu

2021

Structural Health Monitoring

View full text Add to dashboard Cite

Deep learning algorithm can effectively obtain damage information using labeled samples, and has become a promising feature extraction tool for ultrasonic guided wave detection. But it is difficult to apply the monitoring expertise of structure A to structure B in most cases due to the differences in the dispersion and receiving modes of different waveguides. For multi-structure monitoring at the system level, how to transfer a trained structural health monitoring model to another different structure remains a major challenge. In this article, a cross-structure ultrasonic guided wave structural health monitoring method based on distribution adaptation deep transfer learning is proposed to solve the feature generalization problem in different monitoring structures. First, the joint distribution adaptation method is employed to adapt both the marginal distribution and conditional distribution of the guided wave signals from different structures. Second, convolutional long short-term memory network is constructed to learn the mapping relationship from adapted training samples in source domain. Batch normalization layer is implemented to balance the input tensors of each sample to the same distribution. Finally, the multi-sensor damage indexes are utilized to visually present the damage by probability imaging. The experimental results show that proposed method can utilize the single-sensor monitoring data in one structure to implement the multi-sensor damage monitoring in another structure and achieve the damage imaging visualization. The imaging performance is significantly superior to the existing principal component analysis, transfer component analysis, and other state-of-art comparison methods.

show abstract

A Two-Stage Transfer Adversarial Network for Intelligent Fault Diagnosis of Rotating Machinery With Multiple New Faults

Cited by 139 publications

References 35 publications

On-Shaft Wireless Vibration Measurement Unit and Signal Processing Method for Torsional and Lateral Vibration

On-Shaft Wireless Vibration Measurement Unit and Signal Processing Method for Torsional and Lateral Vibration

Intelligent fault diagnosis of rotating machinery using lightweight network with modified tree‐structured parzen estimators

Distribution adaptation deep transfer learning method for cross-structure health monitoring using guided waves

Contact Info

Product

Resources

About