Dynamic Performance of 400 km/h High-Speed Railway Subgrade under Filler Spatial Variability

The non-uniform deterioration of existing subgrade fillers constitutes a critical constraint for the comprehensive speed increase of China's high-speed railway network. To characterize the evolution of uneven settlement and determine the lower limit of spatial variability deterioration for 350 km/h-designed subgrades to meet 400 km/h requirements, this study establishes a data-driven random field framework for CRTS III ballastless track-subgrade systems, integrating Rayleigh wave detection and Monte Carlo simulation. The framework reveals the attenuation and distribution patterns of dynamic indicators under spatial variability of surface and underlayer fillers, elucidates the uneven settlement mechanism induced by fill variability, and compares long-term performance evaluations based on dynamic shear strain and equivalent linear model criteria. Results shows that: graded aggregates exhibit horizontal/vertical autocorrelation distances of 17m/1.21m, while Group A/B fills measure 4.93m/0.56m. When filler variability changes, exceedance probabilities for dynamic stress/deformation remain below 0.01%, indicating limited evaluation capacity of conventional strength/deformation criteria. The proposed variability coefficient thresholds, derived from dynamic shear strain and equivalent linear models, provide theoretical foundations for speed-up assessments. These robustness-based findings offer theoretical support for both new high-speed railway subgrade design and condition evaluation of existing infrastructure