Soil-structure interaction in tunnel lining analyses

In general, tunnel analyses are complex engineering tasks. This is mainly due to non-linear effects determining the overall loadbearing behaviour (even under service load conditions) because no clear separation between action and resistance is possible (safety concept) plus a series of interdependencies and interaction effects that must not be neglected in most cases. In order to achieve a realistic simulation of the actual behaviour and to gain reliable results, non-linear effects in the ground as well as physical and geometrical non-linearities in the structural behaviour of the tunnel lining, and the interaction of the soil and structures, need to be modelled carefully. When simplifications are to be introduced into the design, it is important that the simplified approach be conservative, that deformation results are in a realistic order of magnitude and, finally, that the analyses lead to adequate solutions regarding costs and workability on site.

Over the past decade, the finite element analysis (FEA) method has become more and more established internationally as a general numerical tool suitable for these diverging requirements. Where FEA is used in an appropriate way, the method allows almost all-embracing detailed modelling and may provide comprehensive and realistic results taking into account all relevant effects, including soil-structure interaction. The crucial factor in this is the determination of a realistic ground-structure stiffness ratio, especially if non-linear characteristics are to be taken into account, because the reaction forces (bedding) and the distribution of the load portions to be carried by the tunnel lining and the supporting ground are mainly determined by the stiffness and interaction parameters. Inadequate estimation of these governing parameters may lead to an unrealistic assessment regarding the loadbearing behaviour of the tunnel and thus - depending on the magnitude and direction of the deviation - to a wrong and uneconomic or unsafe design. During the authors' many years in design management, detailed design and value engineering and performing independent checking and design reviews for large international tunnel projects, their experience was that FE modelling may in numerous cases lead to improper results if the basic relations and plausibility checks were not considered in an adequate way. In the light of this, the case studies presented in section 5 illustrate some typical practical examples.

The present paper is intended to point out the most relevant relations, interdependencies and typical problems in soil-structure interaction analyses which should be well considered when performing tunnel analyses utilizing the FEA method. The main aim is to provide valuable recommendations and advice on the formulation of reliable, practice-oriented and robust numerical models in order to improve the overall quality of tunnel analyses.