Building isolation against train-induced vibrations considering the effects of soil and structure: A numerical and experimental study

Base isolation is probably the most prevalent method used to protect buildings from disruptive train-induced vibrations. While the isolators' performance has been the subject of several studies, certain vital limitations remain. Firstly, the optimal effectiveness of isolators with respect to the characteristics of the superstructure and underlying soil is unclear. Secondly, there is a notable lack of a prediction model assessing the isolators' efficacy before any detailed analysis. Lastly, the dearth of empirical evidence that testifies to the performance of isolators in real situations is palpable. This study was devoted to addressing these shortcomings. In order to meet this goal, a 2D numerical model of a building, soil medium, and under-foundation resilient layer (UFRL) was created. Extensive parametric analysis of the model demonstrated that the performance of UFRLs was significantly (up to 15 dB) affected by the features of the soil and building. The conditions in which UFRLs have optimal performance were identified. Based on the results obtained, a prediction model was proposed to estimate the efficacy of UFRLs. Afterward, a 3D numerical and physical simulation of a five-story building was developed to evaluate the performance of UFRLs and validate the proposed prediction model in practical scenarios. A comparison of predicted and computed insertion loss values showed that the mean absolute error for the sixteen scenarios examined here is 1.2 dB.