Rotor Fault Diagnosis Using Domain-Adversarial Neural Network with Time-Frequency Analysis

Xu, Yongjie; Liu, Jingze; Wan, Zhang; Zhang, Dahai; Jiang, Dong

doi:10.3390/machines10080610

Cited by 17 publications

(4 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Different from the above references, domain adversarial learning builds a domain discriminator to minimize the discrepancy between different domains in a latent feature space [18,23,24]. The short-time Fourier transform was adopted to map both source domain data and target date into the same time-frequency feature space, and then a domain adversarial learning-based network was proposed to realize distribution alignment [25]. Li et al [26] proposed a domain adversarial graph CNN for DA, the data structure was considered to improve the performance of transfer diagnosis tasks.…”

Section: Introductionmentioning

confidence: 99%

A two-stage multi-scale domain adversarial transfer learning method and its application in fault diagnosis

Zhang,

Huang,

Wang

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Domain adaptation (DA) based intelligent fault diagnosis (IFD) methods have attracted great attention in recent years. The key motivation of DA methods is to extract the domain-invariant features. In most cases, the one-dimensional convolutional neural networks (CNN) are adopted as the feature extractor, in which the kernels are usually single and fixed. However, the monitoring data for IFD usually involves various scale information, the feature representation extracted by above models may be incomprehensive. Moreover, the target domain data is only used for narrowing distribution discrepancy in an unsupervised way, which may lead to the ignorance of class information of target domain. To address these issues, in this paper, a two-stage multi-scale domain adversarial fault diagnosis method is proposed. A multi-scale feature extractor with different kernel sizes is designed to acquire more discriminative domain-invariant features. Meanwhile, the pseudo label learning is adopted for providing transfer learning process with the pseudo labels of target domain, which are generated by a pre-train network in the first stage and then are optimised through iterations in the second stage. The maximum mean discrepancy (MMD) is also adopted to enhance the ability of model for marginal distribution alignment, which can make the model more robust. Thirty-eight transfer tasks from two different datasets were conducted to evaluate the effectiveness of the proposed method. The experimental results demonstrated that the proposed method achieved higher average diagnosis accuracy compared with several popular methods. The superiority of our proposed method was further explained by visualization of learned features.

show abstract

Section: Introductionmentioning

confidence: 99%

A two-stage multi-scale domain adversarial transfer learning method and its application in fault diagnosis

Zhang,

Huang,

Wang

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…The axial motion [19][20][21] was also taken into consideration. The effects of misalignment of the rotor-bearing system on the fault diagnosis was also a hot topic [22,23]. However, there is less research concentrating on the comprehensive effects, including axial motion and radial and circumferential disturbances of the rotor.…”

Section: Introductionmentioning

confidence: 99%

Model and Algorithm for a Rotor-Bearing System Considering Journal Misalignment

Zhao,

Ma,

Liu

et al. 2023

Coatings

View full text Add to dashboard Cite

Disturbances caused as a result of the misalignment and axial motion of the journal affect the characteristics of the rotor-bearing system. This paper aims to propose an algorithm for the theoretical analysis of a rotor-bearing system that considers these disturbances. A theoretical model for a journal bearing considering disturbances is given. The dynamic equations for a rigid rotor-bearing system are introduced. A detailed algorithm that can simultaneously solve the rotor-dynamic equations and the Reynolds equation is proposed. The static performance, such as the bearing attitude angle and the fluid film pressure, are given, and dynamic characteristics such as the nonlinear dynamic responses and the axial orbits of a rigid rotor-bearing system are presented. The hydrodynamic effect of the bearing is enhanced by the axial disturbance. Disturbances in the circumferential and radial directions lead to variations in the fluid film thickness distribution in the axial direction and the offset of the fluid film pressure distribution in the axial direction. When these disturbances work together, the variation trend is more obvious and affects the capacity and dynamic characteristics of the bearing. When the L/D value of the bearing increases, the clearance between the journal and the bearing decreases rapidly. When the value reaches a certain limit, contact and collision might occur. The theoretical analysis method and the algorithm proposed for a rotor-bearing system considering several disturbances could enhance the design level for a bearing and rotor-bearing system.

show abstract

“…Compared with the manual The associate editor coordinating the review of this manuscript and approving it for publication was Chuan Li. observation method, infrared camera monitoring method and mark-and-recapture method, the bird sounds recognition method is efficient and stable. With the rapid development of artificial intelligence [14], [15], the use of deep learning technology to identify birds has become the mainstream of technology [16], [17], [18].…”

Section: Introductionmentioning

confidence: 99%

Wild Bird Species Identification Based on a Lightweight Model with Frequency Dynamic Convolution

et al. 2023

View full text Add to dashboard Cite

The sound recognition technology is used to count bird speciespopulation to provide accurate data for population ecology and conservation biology research. Bird species can be identified by spectrogram of bird's sounds. The existing Convolutional Neural Network(CNN) is not good at mining the relationship between features in time-frequency tasks, and the existing CNN has more parameters and complicated calculation, so existing models are not suitable for deployment in an actual field environment. In order to fill this gap, this paper proposes a lightweight model with frequency dynamic convolution for bird species identification. We use frequency dynamic convolution innovatively to better capture features of bird sounds at different frequencies. First of all, we replaced two-dimensional convolution with frequency-dynamic convolution in order to achieve not-shift invariance of the bird sound's spectrogram, so that we can effectively capture the feature differences of spectrogram in different frequency bands. Then, we replaced part of the Squeeze and Excitation(SE) attention mechanism with the Coordinate Attention(CA) attention mechanism in order to get more comprehensive global information. Finally, the feature fusion module was used to fuse the local and global features. In addition, we built a datasets containing 160 bird sounds, improving the generalization ability of the model. The experimental results show that our model has good generalization ability and is superior to the existing Lightweight CNN, and obtained better results in top1 accuracy and top5 accuracy.

show abstract

Rotor Fault Diagnosis Using Domain-Adversarial Neural Network with Time-Frequency Analysis

Cited by 17 publications

References 34 publications

A two-stage multi-scale domain adversarial transfer learning method and its application in fault diagnosis

A two-stage multi-scale domain adversarial transfer learning method and its application in fault diagnosis

Model and Algorithm for a Rotor-Bearing System Considering Journal Misalignment

Wild Bird Species Identification Based on a Lightweight Model with Frequency Dynamic Convolution

Contact Info

Product

Resources

About