Time series modelling: applications for groundwater control in urban tunnelling

Water ingress to tunnels may result in pore pressure drawdown and consolidation settlements in areas above tunnels founded on soft soil deposits, potentially causing damage to buildings and infrastructure. To limit pore pressure drawdown, requirements are set on water ingress to bedrock tunnels. To meet these requirements, pre-excavation grouting is often performed to reduce the hydraulic conductivity of the rock mass surrounding the tunnel. Real-time pore pressure monitoring may be used to document pore pressure drawdown during construction. However, the effect of tunnel water ingress can be difficult to distinguish from natural pore pressure fluctuations. This paper presents a tunnel case in Oslo, Norway, where time series modelling was applied to local pore pressure data using the transfer function model framework. The input to the models was daily meteorological data considering precipitation and evapotranspiration, and the output was simulated pore pressure levels with impulse response functions. The models were optimised with data from before tunnel excavation, and simulations were run during the tunnel excavation period. Simulated pore pressure levels were compared with observed pore pressure levels to assess tunnelling-induced drawdown. Model uncertainty ranges were used to produce upper, lower, and best estimates of the drawdown. The findings show that time series modelling with transfer function models may be used in tunnel projects to continuously assess the impact on the local groundwater environment, for better evaluation of the pre-grouting performance, and for quantifying both the temporary and long-term drawdown with increased accuracy.