Comparison between 2D and 3D considerations of a railway embankment

Student thesis: Master thesis (including HD thesis)

  • Britta Seidenkrantz S√łndergaard
4. term, Structural and Civil Engineering, Master (Master Programme)
This master of science project considers 2D and 3D representations of railway embankments with a focus on constituting relationships that make it possible to simplify the computationally expensive 3D cases into equivalent 2D cases. As a result the problem formulation is stated as:

- Which studies have so far been carried out concerning the simplification from 3D to 2D representations of railway embankments?
-How can 3D representations of a loaded railway embankment be simplified to equivalent 2D representations based on numerically determined lengths of influence in the service limit state and ultimate limit state?

A state of the art survey has been carried out to investigate existing studies within the field. From the survey it was seen that the safety factor in 3D is generally larger than the safety factor of a corresponding 2D case. The influence of geometric variations in the third dimension have also been investigated for turns and corners and these effects generally amplify the difference between the safety factors. Other studies also consider the influence of the height and the width of the embankment.

In order to simplify the 3D states to equivalent 2D states a 2D linear elastic finite element program has been developed in MATLAB. A model was set up in the MATLAB program to determine the load distribution in the direction perpendicular to the cross-section plane of the embankment. Based on the MATLAB-model and models created in the commercial FE programs PLAXIS 2D and 3D it was possible to establish correlations between the length of influence for two load cases in the service limit state. In the first load case a single load of 250 kN was applied on each rail while 4 loads of the same size were applied on each rail in the second load case. Furthermore, it was verified that superposition principles were applicable to determine the corresponding deformations arising from other load cases than the two specific load cases considered.

In the same way it was sought to establish a correlation in the ultimate limit state. However, it was not possible to obtain satisfactory results of the ultimate load in PLAXIS 3D due to the way that these loads are determined in PLAXIS. Consequently, it is necessary to determine the ultimate load in a different manner to obtain reliable results of the ultimate loads but this lies within a future study.
Publication date8 Jun 2015
Number of pages122
External collaboratorCOWI Aalborg
Head of Section Kim Andre Larsen
Information group
ID: 213964376