Density-Based Affinity Propagation Tensor Clustering for Intelligent Fault Diagnosis of Train Bogie Bearing

Self Cite

Applying methods such as deep learning improves the efficiency of bearing fault diagnosis and reduces trains’ operation and maintenance costs. However, in practical applications, the deficiency of historical data and the imbalance of data types often limit the effectiveness of the diagnosis. The variability between operating conditions also restricts the availability of transfer learning including domain adaptation. To address this challenge, a digital twin (DT) framework is established to fill the data for train fault diagnosis. A train bearing dynamics model is optimized using virtual-reality mapping in the DT framework with measured health data as a baseline to generate data closer to reality. Finally, the fault diagnosis uses a hybrid dataset that mixes measured and simulated data as a source domain for transfer learning. The Case Western Reserve University dataset is used as an example, and the accuracy reaches up to 99.40%, which verifies the method’s effectiveness.

Section: Multiscale Attention Modulementioning

confidence: 99%

Personalized fault diagnosis of rolling bearings in trains based on digital twin

Liu,

He,

Wei

et al. 2023

Self Cite

“…noise monitoring technology, infrared temperature measurement technology, and ray scanning, requiring a lot of human and financial resources. With rapid advancements in computer technology and deep learning, compared with the traditional fault diagnosis techniques [3][4][5], convolutional neural network models [6][7][8] have emerged as a promising alternative, offering additional feature information for more effective fault diagnostics. The influence of deep learning on intelligent fault diagnosis (IFD) is notable, with machine learning techniques increasingly integrating with artificial intelligence.…”

Section: Introductionmentioning

confidence: 99%

Fault diagnosis of rolling bearings based on dynamic convolution and dual-channel feature fusion under variable working conditions

Yao,

Zhou,

Yang

et al. 2024

Addressing the challenge of inconsistent data feature distribution and the difficulty of fault diagnosis in rolling bearings operating under variable conditions, a novel approach is proposed for bearings fault diagnosis. Dynamic convolution and dual-channel feature fusion are utilized in our method. In the shallow network layer, we employ a dual-channel convolutional structure, combining a large convolutional group with a small convolutional group to enhance the extraction of high-low frequency fault information from images. The improved GhostNetV2 bottleneck layer was used in the deeper layer of the network to obtain more informative features through the residual module and attention mechanism. Finally, fault classification and evaluation under variable working conditions was performed on the CWRU and DDS datasets. Our results showed that the methods and model used in this study can effectively handle the precision fault detection across various operational scenarios.

“…Due to the advantages of environmental protection and high reliability, rail transit plays an important role in daily travel [1,2]. However, the safety of trains needs to be improved.…”

Section: Introductionmentioning

confidence: 99%

An improved single-stage convolutional neural network for rail transit obstacle detection

Qin,

He,

Sun

et al. 2023

Self Cite

Obstacles that intrude into the rail area can lead to serious rail accidents, so obstacle detection technology is an essential guarantee for the safe operation of fully automatic trains. To meet the high-performance requirements of onboard obstacle detection, an efficient feature-aware convolutional neural network (EFA-Net) is proposed in this paper. The MA-FPN is designed as feature fusion network to extract multi-scale context information. In the detection head, the dynamic awareness block is used to refine the features. A joint representation branch and the generalized focal loss function are introduced to optimize the training effect. The experiments are based on the dataset of real-world rail transit environment. The results show that EFA-Net can achieve a detection accuracy of 90.4% mAP at a detection speed of 20.4 FPS, and the lightweight design significantly reduces the computational complexity of the proposed model. Compared with other classical detectors, EFA-Net has the best comprehensive performance.