Estimations of Vertical Rail Bending Moments from Numerical Track Deflection Measurements Using Wavelet Analysis and Radial Basis Function Neural Networks

A method for estimating rail bending moments from relative vertical track deflection data measured by a train-mounted measurement system is presented in this paper. The novelty of the current study is that complete estimations of rail positive and negative bending moments from track deflection measurements are conducted by using a wavelet multiresolution analysis in conjunction with the radial basis function neural network considering the effects of varying track modulus. The simulation results show that the proposed framework can effectively employ vertical track deflections to estimate both maximum positive and negative bending moments in rails, with the average estimation error being 6.22% (i.e., 2.82 kNm). Moreover, the study confirms the capability of the train-mounted vertical track deflection measurement system (commercially known as MRail) in evaluating the rail bending moments over long distances.

• Record URL:
  https://doi.org/10.1061/JTEPBS.0000489
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/issn/24732907
• Supplemental Notes:
  • © 2020 American Society of Civil Engineers.
• Authors:
  • Do, Ngoan Tien
  • Gül, Mustafa
• Publication Date: 2021-2

Language

• English

Media Info

• Media Type: Web
• Features: References;
• Pagination: 04020154
• Serial:
  • Journal of Transportation Engineering, Part A: Systems
  • Volume: 147
  • Issue Number: 2
  • Publisher: American Society of Civil Engineers
  • ISSN: 2473-2907
  • EISSN: 2473-2893
  • Serial URL: http://ascelibrary.org/journal/jtepbs

Subject/Index Terms

• TRT Terms: Bending moments; Deflection; Estimating; Instruments for measuring deformation or deflection; Neural networks; Rail (Railroads); Train track dynamics; Wavelets
• Subject Areas: Railroads; Vehicles and Equipment;

Filing Info

• Accession Number: 01760225
• Record Type: Publication
• Files: TRIS, ASCE
• Created Date: Nov 20 2020 3:04PM