Designing high-speed train railway embankments using finite element analysis

Abstract

Design and construction of high-speed train railway (HSTR) infrastructure is different from that for normal trains, since high geometric standards and material properties are involved. For example, HSTR embankments are designed to limit total-differential settlements to minute amounts. In this study, a typical 'slab-track' type HSTR embankment was used to investigate the replaceability of one fill strata known as 'uncemented-prepared-subgrade layer' (U-PSL), constructed by using a locally obtained medium sand, as opposed to various 'cemented-prepared-subgrade layers' (C-PSL), whose mixes were prepared at two water-to-cement (w/c) ratios and three cement contents (c). Three size cylindrical samples were cast, 7-28 days water cured and were tested to obtain the unconfined compressive stresses, strains, elasticity moduli and Poisson's ratios at failure. Test results were then fed into the Plaxis-FEM program to find the maximum total settlements of individual layers and compared with the requirements. Only three C-PSL mixes having cement contents (c) of 20, 25, 30 % and water-to-cement (w/c) ratios of 0. 4 and 0. 5 met the strict settlement criteria. The study showed that the original (h = 2 m) thick U-PSL can be replaced with 0. 3h-m thick C-PSL at w/c = 0. 5 (i. e. h = 0. 6 m). Likewise, the original (h = 2 m) thick U-PSL can be replaced with 0. 2h-m thick C-PSL at w/c = 0. 4 (i. e. h = 0. 4 m). Also, the extra effort of doing in situ soil compaction and testing in layers is reduced or eliminated. This would give not only alternative ways to HSTR embankment designers/constructors, but also substantial savings in construction time and costs.