Hydro-chemo-mechanical modelling of tunnels in sulfated rocks

Abstract

Permission is granted by ICE Publishing to print one copy for personal use. Any other use of these PDF files is subject to reprint fees.A modelling procedure to address the tunnel–anhydritic rock interaction is described in this paper. The model incorporates the basic physico-chemical phenomena involved in rock swelling, often observed during excavation and subsequent operation of tunnels. It includes (a) a provision for rock damage during tunnel excavation, (b) the precipitation of gypsum crystals in discontinuities and (c) a stress-dependent relationship between swelling strains and mass of gypsum precipitation. The model includes hydro-mechanical coupling and the transport of sulfate salts dissolved in the massif water. Rock damage is described by the development of a network of fractures that increases permeability and allows gypsum crystal growth. Field information, laboratory data and monitoring records available for Lilla tunnel, located in the province of Tarragona, Spain and excavated in Tertiary anhydritic claystone, were selected as a convenient benchmark case to test model capabilities. Predictions and measurements (swelling records of the unlined tunnel floor and swelling pressures against a structural invert) were found to agree reasonably well.Peer ReviewedPostprint (published version

Reactive Transport Modeling of Swelling Processes in Clay‐sulfate Rocks

“…The inversely determined parameter b of 2.5 is similar to the estimate by Ramon et al () and yields a bulking coefficient γ of 0.15 ± 0.11 (For further details on γ , refer to the supporting information). Thus, according to the theoretical basis of γ a large part of the gypsum crystals precipitates in already existing cavities and therefore does not cause significant deformation.…”

“…Consequently, the solution volume V in contact with the mineral equals the immobile porosity ( im ) instead of the total porosity. It should be noted that this assumption is different from the model of Ramon et al (2017), who assumed that gypsum crystal growth occurs solely in fractures. The focus of their study, however, was on mechanical considerations such as fracture opening.…”

“…In order to relate stress to strain and derive the vertical deformation from the mass of precipitated gypsum crystals determined by the coupled flow, transport and reaction simulations in the swelling zone, a bulking (swelling) coefficient was used, following the work by Oldecop and Alonso (), and Ramon et al (): where ϵ i is the deformation (strain) and γ i the bulking coefficient in the i direction, respectively, ρ G is the gypsum density (kg/m 3 ) and m G the mass of precipitated gypsum crystals (kg). The geomechanical stress ‐ strain relationship is calculated posterior to the completed MODFLOW and PHT3D model simulation using a Python script.…”

“…The effect of stress (expressed by γ ) is approximated by an exponential decay function: where γ max is the bulking coefficient for zero pressure stress which was set to 2 as in Ramon et al (), b is a model fitting parameter estimated during the inversion process, and is the effective stress.…”

“…In order to relate stress to strain and derive the vertical deformation from the mass of precipitated gypsum crystals determined by the coupled flow, transport and reaction simulations in the swelling zone, a bulking (swelling) coefficient was used, following the work by Oldecop and Alonso (2012), and Ramon et al (2017):…”