LN-MRSCAE: A novel deep learning based denoising method for mechanical vibration signals

Du, Wenliao; Yang, Lingkai; Wang, Hongchao; Gong, X. G.; Zhang, Lianying; Li, Chuan; Lian-qing, JI

doi:10.1177/10775463231151721

Cited by 9 publications

(5 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hu [28] designs a multi-scale attention residual network to achieve environmental sound classification, which alleviates the problem of gradient explosion and gradient disappearance. Du [29] proposes a noise learning method based on multi-level residual convolutional networks, which solves the problem of gradient disappearance in network learning by designing multi-level residual structures; Chen [30] establishs a residual structure graph neural network RSGNN, which constructs residual links on local subgraphs to alleviate the problem of excessive smoothing.…”

Section: Deepgcn Networkmentioning

confidence: 99%

A graph residual generation network for node classification based on multi-information aggregation

Liang,

Jia,

Han

et al. 2024

Preprint

View full text Add to dashboard Cite

The key to improving the performance of graph convolutional networks (GCN) is to fully explore the correlation between neighboring and distant information. Aiming at the over-smoothing problem of GCN, in order to make full use of the relationship among features, graphs and labels, a graph residual generation network based on multi-information aggregation (MIA-GRGN) is proposed. Firstly, aiming at the defects of GCN, we design a deep initial residual graph convolution network (DIRGCN), which connects the initial input through residuals, so that each layer node retains part of the information of the initial features, ensuring the localization of the graph structure and effectively alleviating the problem of over-smoothing. Secondly, we propose a random graph generation method (RGGM) by utilizing graph edge sampling and negative edge sampling, and optimize the supervision loss function of DIRGCN in the form of generation framework. Finally, applying RGGM and DIRGCN as inference modules for modeling hypotheses and obtaining approximate posterior distributions of unknown labels, an optimized loss function is obtained, we construct a multi-information aggregation MIA-GRGN that combines graph structure, node characteristics and label joint distribution. Experiments on benchmark graph classification datasets show that MIA-GRGN achieves better classification results compared with the benchmark models and mainstream models, especially for datasets with less dense edge relationships between nodes.

show abstract

Section: Deepgcn Networkmentioning

confidence: 99%

A graph residual generation network for node classification based on multi-information aggregation

Liang,

Jia,

Han

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Using a model-driven approach offers numerous advantages, but it is associated with relatively high simulation costs [39,40]. Artificial intelligence algorithms typically demand a significant volume of data for effective training and learning [41,42]. Insufficient data can potentially result in a deterioration of model performance.…”

Section: Optimization Of Molding Process 41 Regression Modelmentioning

confidence: 99%

Simulation and Experimental Study on the Precision Molding of Irregular Vehicle Glass Components

Chen,

Hu,

Zhang

et al. 2023

Micromachines

View full text Add to dashboard Cite

The high level of stress and dimension deviation induced by glass molding are the main causes of the low yield rate of large, irregular glass components on vehicles. To solve this issue, a numerical model of large glass component molding was established in this study, which aimed to analyze the dominant factors of molding quality and achieve a synergistic balance between quality characteristics and energy consumption. The results show that molding temperature is the dominant factor affecting the energy consumption and residual stress, and the molding pressure is the main factor affecting the dimension deviation. Furthermore, the NSGA-II optimization algorithm was used to optimize the maximum residual stress, dimension deviation, and energy consumption with the numerical results. The combination of a heating rate of 1.95 °C/s, holding time of 158 s, molding temperature of 570 °C, molding pressure of 34 MPa, and cooling rate of 1.15 °C/s was determined to be the optimized scheme. The predictive error of the numerical result, based on the optimized scheme, was experimentally verified to be less than 20%. It proved the accuracy of the model in this study. These results can provide guidance for the subsequent precision molding of large, irregular glass components.

show abstract

“…Vibration signal analysis is pivotal in diagnosing rolling bearing faults due to the plethora of fault-related features inherent in such signals [6,7]. However, signals collected from rolling bearings in real-world industrial settings often suffer from significant noise interference, presenting substantial challenges [8]. Over the years, considerable attention has been devoted to reducing noise interference and extracting meaningful fault features [9][10][11].…”

Section: Introductionmentioning

confidence: 99%

Bearing Fault Vibration Signal Denoising Based on Adaptive Denoising Autoencoder

Lu,

Zhou,

2024

Electronics

View full text Add to dashboard Cite

Vibration signal analysis is regarded as a fundamental approach in diagnosing faults in rolling bearings, and recent advancements have shown notable progress in this domain. However, the presence of substantial background noise often results in the masking of these fault signals, posing a significant challenge for researchers. In response, an adaptive denoising autoencoder (ADAE) approach is proposed in this paper. The data representations are learned by the encoder through convolutional layers, while the data reconstruction is performed by the decoder using deconvolutional layers. Both the encoder and decoder incorporate adaptive shrinkage units to simulate denoising functions, effectively removing interfering information while preserving sensitive fault features. Additionally, dropout regularization is applied to sparsify the network and prevent overfitting, thereby enhancing the overall expressive power of the model. To further enhance ADAE’s noise resistance, shortcut connections are added. Evaluation using publicly available datasets under scenarios with known and unknown noise demonstrates that ADAE effectively enhances the signal-to-noise ratio in strongly noisy backgrounds, facilitating accurate diagnosis of faults in rolling bearings.

show abstract

LN-MRSCAE: A novel deep learning based denoising method for mechanical vibration signals

Cited by 9 publications

References 29 publications

A graph residual generation network for node classification based on multi-information aggregation

A graph residual generation network for node classification based on multi-information aggregation

Simulation and Experimental Study on the Precision Molding of Irregular Vehicle Glass Components

Bearing Fault Vibration Signal Denoising Based on Adaptive Denoising Autoencoder

Contact Info

Product

Resources

About