Voids at the tunnel–soil interface for calculation of ground vibration from underground railways

Abstract

Voids at the tunnel–soil interface are not normally considered when predicting ground vibration from underground railways. The soil is generally assumed to be continuously bonded to the outer surface of the tunnel to simplify the modelling process. Evidence of voids around underground railways motivated the study presented herein to quantify the level of uncertainty in ground vibration predictions associated with neglecting to include such voids at the tunnel–soil interface. A semi-analytical method is developed which derives discrete transfers for the coupled tunnel–soil model based on the continuous Pipe-in-Pipe method. The void is simulated by uncoupling the appropriate nodes at the interface to prevent force transfer between the systems. The results from this investigation show that relatively small voids ( 4 m × 90 ∘ ) can significantly affect the rms velocity predictions in the near-field and moderately affect predictions in the far-field. Sensitivity of the predictions to void length and void sector angle are both deemed to be significant. The findings from this study suggest that the uncertainty associated with assuming a perfect bond at the tunnel–soil interface in an area with known voidage can reasonably reach ± 5 dB and thus should be considered in the design process.