Open Access

Dynamic analysis of railroad track faults based on neural network algorithms

Yue Lyu

Yue Lyu

China Railway Xi’an Group Co., LtdXi’an, China
Search for this author on
Sciendo|Google Scholar
Lyu, Yue
, 
Jian Sun

Jian Sun

China Railway Xi’an Group Co., LtdXi’an, China
Search for this author on
Sciendo|Google Scholar
Sun, Jian
, 
Yifei Cao

Yifei Cao

China Railway Xi’an Group Co., LtdXi’an, China
Search for this author on
Sciendo|Google Scholar
Cao, Yifei
, 
Jie Sun

Jie Sun

China Railway Xi’an Group Co., LtdXi’an, China
Search for this author on
Sciendo|Google Scholar
Sun, Jie
, 
Meng Tian

Meng Tian

China Railway Xi’an Group Co., LtdXi’an, China
Search for this author on
Sciendo|Google Scholar
Tian, Meng
, 
Hongchang Wang

Hongchang Wang

China Railway Xi’an Group Co., LtdXi’an, China
Search for this author on
Sciendo|Google Scholar
Wang, Hongchang
 and 
Feilong Wu

Feilong Wu

Xi’an Jiaotong UniversityXi’an, China
Search for this author on
Sciendo|Google Scholar
Wu, Feilong
  
Mar 17, 2025
Dynamic analysis of railroad track faults based on neural network algorithms's Cover Image
About this article
Previous Article
Next Article
Cite
Download Cover
Published Online: Mar 17, 2025
Received: Oct 17, 2024
Accepted: Feb 07, 2025
DOI: https://doi.org/10.2478/amns-2025-0300
Keywords
© 2025 Yue Lyu et al., published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Figure 1.

Schematic diagram of rolling bearing motion model
Schematic diagram of rolling bearing motion model

Figure 2.

Dynamic model of rolling bearing
Dynamic model of rolling bearing

Figure 3.

Flow chart of fault diagnosis
Flow chart of fault diagnosis

Figure 4.

The diagnosis of the nuclear size and number of different convolution
The diagnosis of the nuclear size and number of different convolution

Figure 5.

Model training time and test time
Model training time and test time

Figure 6.

Raw data
Raw data

Figure 7.

After the attention mechanism is added
After the attention mechanism is added

Figure 8.

Add the attention mechanism and the BiLSTM
Add the attention mechanism and the BiLSTM

Figure 9.

Calibration curve contrast
Calibration curve contrast

Error analysis of error characteristic frequency simulation

Fault element Speed of revolution(rpm) Simulation frequency(Hz) Theoretical frequency(Hz) Error/%
Outer ring 300 25.3 25.36 0.237
600 50.58 50.71 0.257
900 75.84 76.09 0.330
1200 101.07 101.41 0.336
1500 126.42 126.75 0.261
1800 151.61 152.16 0.363
Inner ring 300 39.35 39.67 0.813
600 78.71 79.22 0.648
900 118.12 118.88 0.643
1200 157.51 158.56 0.667
1500 196.86 198.19 0.676
1800 236.23 237.79 0.660
Scroll body 300 21.39 21.58 0.888
600 42.88 43.28 0.933
900 64.3 64.91 0.949
1200 85.75 86.58 0.968
1500 107.17 108.19 0.952
1800 128.7 129.87 0.909

The average accuracy of different modules added

Model 735W 1470W 2205W Mean value/%
F1 99.12 99.56 99.83 99.50
F2 95.45 95.68 95.89 95.67
F3 96.58 96.89 97.12 96.86
F4 91.53 92.13 92.45 92.04
Language:
English
Publication timeframe:
1 times per year
Journal Subjects:
Life Sciences, Life Sciences, other, Mathematics, Applied Mathematics, General Mathematics, Physics, Physics, other
Journal RSS Feed