Modelling construction phases of bored tunnels with respect to internal lining forces

Abstract

Areas are getting more and more populated causing new infrastructure lines to be constructed below the surface. A bored tunnel is one of the possibilities to create this subsurface infrastructure, but the construction process of a bored tunnel is a complicated one. Many loads and aspects are present in this construction process that can be divided into six phases. In each of these phases, different loads and aspects are acting on the tunnel lining or the surrounding soil which can cause the lining to deform.

The increasing complexity and demands of problems have led to the use of the finite element method. A computer based method which allows one to model the problem.

For finite element modelling, numerous programs are available of which several claim to be able to model bored tunnels. However, it is not yet clear what the exact differences between the different programs are. With many aspects to be modelled, many differences between programs occur, either in the soil, the tunnel lining or a combination of both.

This research has focussed on the possibilities of modelling the different construction phases of bored tunnels in two widely used programs: DIANA and Plaxis. Simple two dimensional (2D) models were created to which the different construction phases were added before continuing with three dimensional (3D) modelling. This approach has led to a good assessment of the possibilities and limitations within these two programs.

DIANA is not yet suitable for modelling the construction process of bored tunnels completely. The construction phases are modelled undrained to account for the relative short time they are acting. A consolidation phase in which the pore pressure can dissipate cannot be modelled in DIANA, which is essential for modelling the construction phases.

Plaxis, on the contrary, is not able to model joints in the segmental lining appropriate for 3D. In 3D, Plaxis only allows to model a joints as "fixed" or "free". In DIANA different theories can be applied to the joints, including Janssens. For 2D, both programs have a rotational springs besides the free and fixed connection for modelling the joints.

For the model in which the material models were changed, the difference with the main impact between the two programs occurred, especially for the bending moment. This means the Modified Mohr-Colomb material model in DIANA is different than the Hardening Soil model in Plaxis.

Including the construction phases leads to more favourable internal lining forces for tunnels, something of which clients should be convinced. However, not until the models have been benchmarked with measured data from a tunnel project.

While 3D models have been investigated in this research, they should be extended in order have a better understanding of the different 3D phenomena that are present in the construction of bored tunnels. This will both assess the possibilities of modelling this process and more potential differences between programs can be investigated.

Besides extending the 3D models, other programs should be investigated on their capabilities too. In order to come to a proper assessment of the possibilities, program experience is strongly recommended. These programs should also be compared with measured data for benchmark purposes.