Citation

Adegoke CW, Chang CS, Selig ET. Transp. Res. Rec. 1979; 733: 12-20.

Copyright

(Copyright © 1979, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing)

DOI

unavailable

PMID

unavailable

Abstract

Data on the dynamic responses of ballast, subballast, and subgrade of track sections at the Facility for Accelerated Service Testing track in Pueblo, Colorado, are compared with predictions from three available analytical models for track support systems. The response data include ballast strain, subballast strain, subgrade deflection, and subgrade stress. The analytical solutions are provided by (a) a model that combines Burmister's three-dimensional elasticity solution with a structural analysis model that solves for the tie-ballast reaction (MULTA), (b) a finite-element, three-dimensional model that has prismatic elements combined with a structural analysis model (PSA), and (c) a quasi-three-dimensional, finite-element model, in which a longitudinal two-dimensional analysis is followed by a transverse two-dimensional analysis (ILLI-TRACK). The results show that all three models can reasonably predict the behavior of the track system, provided that values for the material properties and model parameters are correctly specified. Each model has advantages and limitations compared with the others. ILLI-TRACK is the only model that can vary properties in the vertical, longitudinal, and transverse directions and also the only one having a nonlinear stress-strain representation. However, the accuracy of ILLI-TRACK is the only model that can vary properties in the vertical, longitudinal, and transverse directions and also the only one having a nonlinear stress-strain representation. However, the accuracy of ILLI-TRACK predictions is less certain because it depends on two empirical parameters, the effective tie-bearing length and the angle of distribution. The PSA model permits property variation in the transverse and vertical directions, but its computer costs are an order of magnitude greater than those for the other two models. The MULTA model is restricted to homogeneous layers of ballast and underlying materials, but it combines the features of both three-dimensionality and economy.