Back to Search
Start Over
Artificial neural network-based investigation on high-speed train-induced embankment vibration in frozen regions.
- Source :
-
Soil Dynamics & Earthquake Engineering (0267-7261) . Oct2023, Vol. 173, pN.PAG-N.PAG. 1p. - Publication Year :
- 2023
-
Abstract
- High-speed trains (HSTs) induce vibrations that deteriorate the embankments of ballastless tracks, particularly in frozen regions, necessitating vibration assessments for existing high-speed railways (HSRs). This study developed an indirect thermal-dynamic coupled two-dimensional (2D) finite element (FE) model to investigate the dynamic responses of the embankment of ballastless tracks to various subgrades and subsoils, train speeds, and frost depths in frozen regions. The results indicate that subgrade is the key parameter to analysing the dynamic response of the train–track–ground system considering train speeds. A fast and high-accuracy prediction method for the vibration responses was developed using an artificial neural network (ANN) and used to predict the vibration level under different conditions. Prediction results determined train speeds under 400 km/h safe for HSRs. This study can improve knowledge of ballastless track substructure design and extend the applicability of the current railway code to higher train speed conditions in frozen regions. • Proposed an accurate and reliable artificial neural network-based prediction model. • Train speeds under 400 km/h are safe for high-speed railways in frozen regions. • Subgrade is key to analysing the dynamic response of train–track–ground systems. • Substructure layer reinforcement is recommended for higher speeds. [ABSTRACT FROM AUTHOR]
- Subjects :
- *HIGH speed trains
*EMBANKMENTS
*SUBSOILS
*PREDICTION models
Subjects
Details
- Language :
- English
- ISSN :
- 02677261
- Volume :
- 173
- Database :
- Academic Search Index
- Journal :
- Soil Dynamics & Earthquake Engineering (0267-7261)
- Publication Type :
- Academic Journal
- Accession number :
- 169814951
- Full Text :
- https://doi.org/10.1016/j.soildyn.2023.108093