Numerical modeling of a ballastless track mockup based on asphalt

Tulika Bose, Eyal Levenberg*, Varvara Zania

*Corresponding author for this work

    Research output: Contribution to journalJournal articleResearchpeer-review

    Abstract

    A 3D mechanical model of a ballastless asphalt track mockup was developed within the general-purpose finite element software ABAQUS. The mockup (and model) consisted of three wide-base sleepers equipped with a geotextile at the bottom, supported on an asphalt pavement structure encapsulated in a large rigid box. The asphalt layer was modeled as linear viscoelastic; the underlying unbound granular layer was treated as stress-dependent nonlinear-elastic, implemented via a user-defined subroutine. The vast majority of model parameters were calibrated through laboratory element tests, while the remaining parameter values were based on technical literature or material specifications; only very few were calibrated using experimental data from the mockup itself. Implicit dynamic analysis was carried out under a loading history that simulated a train passage by sequentially exciting the three sleepers with a time delay. Vertical stresses at the bottom of the unbound granular layer, horizontal strains at the bottom of the asphalt layer, vertical accelerations at the track surface, and relative displacements at different locations were numerically evaluated. Subsequently, model predictions were validated by comparison against corresponding response-traces measured in the mockup. Overall, the predicted responses were in very good agreement with the experimental measurements. The peak vertical stresses below the unbound granular layer were moderately overestimated, while the peak horizontal asphalt strains were underestimated. In particular, the characteristic shape-features recorded by the stress and strain sensors at multiple locations in the mockup could be replicated. Pearson's correlation coefficients between measured and calculated response histories for stresses and strains were higher than 0.96. Predicted and measured vertical accelerations were of the same order of magnitude, and their corresponding frequency spectra exhibited a correlation value greater than 0.97. The validated model has verified that during a simulated train passage, the substructure of a ballastless track mockup experiences low-magnitude vertical deformations, 77% of which develop in the rail pad, 15% in the geotextile, and 6% within the unbound granular layer. Moreover, the peak vertical stress in the unbound granular layer (33 kPa) and the peak horizontal tensile strain in the asphalt layer (18με) are lower than limiting design values.
    Original languageEnglish
    Article number121852
    JournalConstruction and Building Materials
    Volume274
    Number of pages18
    ISSN0950-0618
    DOIs
    Publication statusPublished - 2021

    Keywords

    • Asphalt overlayment
    • Ballastless asphalt track
    • Finite element method
    • Nonlinear Dynamic analysis
    • Viscoelasticity
    • Model validation

    Fingerprint

    Dive into the research topics of 'Numerical modeling of a ballastless track mockup based on asphalt'. Together they form a unique fingerprint.

    Cite this