Chao He, Xuming Li, Xiaoxin Li, Xingbang Qian, Zhiyao Tian, Quanmei Gong, Shunhua Zhou
Train-induced environmental vibration has become an increasingly critical issue in recent years, driven by rapid urbanization and the large-scale expansion of urban transportation infrastructure. For accurate dynamic analysis of rail transit systems, it is necessary to account for both the inherent periodicity of track structures and the spatiotemporal variability of soil-layer parameters, thereby improving the fidelity of environmental vibration prediction. To this end, this paper proposes a thin layer method-perfectly matched layer framework for evaluating ground vibrations to periodic moving harmonic loads while explicitly considering soil-parameter variability. A periodic wavenumber domain fundamental solution for the ground vibrations under periodic moving harmonic loading is derived. The proposed method is then validated against the stiffness matrix method, showing good agreement. Meanwhile, the proposed approach delivers a substantial improvement in computational efficiency, making it well suited to large-scale parametric analyses, including random-field simulations. A systematic parametric study is further conducted to investigate ground vibrations under combined effects of periodic moving harmonic loads and soil random fields. The results indicate that periodic moving harmonic loads are more sensitive to soil-parameter variations than non-periodic moving harmonic loads, leading to greater response uncertainty, particularly in the high-frequency range. Overall, the proposed model and findings support the transition from deterministic prediction to probabilistic, uncertainty-aware assessment of environmental vibrations.