Finite element modelling of blowouts in compressed air tunnels

Abstract

This study develops criteria for predicting blowouts in compressed air tunnels through an integrated design approach using empirical laboratory test data in a finite element model. Compressed air tunnelling is first reviewed in the context of temporary works, and the relative merits of compressed air as a support system are compared with that of other methods commonly used. Design methods for estimating the stability of tunnel faces in general are discussed along with the particular problems of stability against a blowout. The various mechanical processes that contribute to the formation of blowouts are then discussed, including unsaturated shear strength, desaturation processes, and seepage forces. A plausible mechanism which can be used analytically in studying blowouts is presented, whereby the How of air desaturates the soil causing mechanical failure, which is the result of serious reductions in effective stress due to increased pore pressures and seepage forces. The failure self propagates with changes in the soil properties. The permeability of soils with respect to compressible fluids is a complex property not only dependent on the soil properties but on fluid properties and pressures as well. Permeability with respect to air is presented as an empirical power function of pressure. Laboratory methods of measuring permeability with respect to air is discussed along with a description of all relevant apparatus. A finite element package developed by NAg software was used as the basic model of potential flow. The model was adapted to account for the progressive desaturation of the soil and its influence on permeability. The model was also modified to account for compressible flow using the Rayleigh-Janzen method of linearizing the compressible potential equation. The model was run for subaqueous tunnels of various diameters in sands of different silt contents beneath rivers of various depth. The amount of excess air pressure was varied and provision was made for a forepoling hood on the shield. Results show compressible flow and incompressible flow are essentially the same except at the crown of the tunnel face. Blowouts demonstrate the unique feature of propagating from the surface down as well as from the tunnel up, a phenomenon which is attributed to high seepage forces and negligible total stresses at the ground surface. The overall factor of safety is most sensitive under shallow rivers, and critical tunnel depth is independent of river depth. The data satisfy a linear relationship between critical depth and tunnel diameter, with smaller diameter tunnels requiring more cover relative to tunnel diameter than large diameter tunnels. The effect of a hood is most significant at high air pressures and in clean sands. The critical tunnel depth is directly proportional to air pressure and is slightly more sensitive for small tunnels than for large tunnels. The critical depth generally is not influenced by silt content. A significant exception is a large drop in critical depth for silt contents between 5% and 10%.The results of the study were summed up in a set of nomographs to provide practical aid to the design of tunnels against blowouts. They can be used to estimate safe cover depths and operating pressures, or to estimate factors of safety for a given set of operating conditions.