Abstract
In China's heavy-haul railways, coal transportation accounts for nearly half of the total freight volume. Coal dust frequently spills from open-top wagons and accumulates on the ballast bed, which significantly undermines track stability and drainage performance. However, existing studies mainly focused on ballast fouling assessment, coal dust layers have not been systematically investigated as an independent detection target. This study proposes a rapid method for detecting coal dust thickness using high-frequency ground penetrating radar (GPR), integrating numerical modeling, advanced signal processing, and field validation to analyze the time- and frequency-domain response characteristics of coal dust layers. A finite-difference time-domain (FDTD) model was established to analyze how coal dust thicknesses and coverage lengths affect GPR time-frequency signatures. A dedicated signal processing protocol was developed to enhance coal dust feature localization. A Freight train mounted three-air-horn 2 GHz antenna was employed to conduct field experiments. Numerical results demonstrated that the proposed method achieves a thickness estimation accuracy within ±0.4 cm. Field experiments confirmed that coal dust contamination replaces the hyperbolic diffraction patterns of sleepers with continuous reflection layers, consistent with simulation findings. The proposed approach provides a novel, nondestructive, and data-driven solution for detecting coal dust contamination in freight railways, contributing to more precise and intelligent ballast maintenance strategies.