The inherent uncertainties in the ground model and influencing factors, in most cases, do not allow a reliable direct design of a tunnel. Empirical tunnel design methods do not allow proofing stability or serviceability, thus are not acceptable according to common understanding.
To allow for construction of safe and economical tunnels, the uncertainties have to be dealt with by observing the behaviour of the structure, and adjust construction measures to the real ground conditions and behaviours ("observational method").
After characterizing the ground, potential hazards are identified, the possible range of behaviours is assessed and appropriate construction measures to the expected behaviours are assigned. A monitoring plan needs to be established together with a safety management plan including contingency measures for inacceptable predicted behaviour deviations.
The paper with the help of a few worked examples illustrates the influence of the rock mass structure and fault zones on the system behaviour.
Optimization of underground structures design has been a topic widely studied by researchers and the industry. New approaches, guidelines and techniques have been developed replacing standard classification systems which tried to simplify/generalize the ground conditions with the inherent shortcomings (Daller et al. 1994, Palmström & Broch 1996, Riedmüller & Schubert 1999a, Riedmüller & Schubert 1999b). New approaches attempted to consider biases and uncertainties that arise from the multiple influencing factors which govern the ground behaviour and focus on achieving an appropriate technical and economical design (Radoncic et al. 2009a, Clayton et al. 1982, Head 1986, Oliveira 1992). Yet each design phase entails uncertainties that have to be considered during construction (Einstein 2001).
This paper presents a description on how to achieve a proper design and connect it, through the observational approach, to the construction phase. The first part deals with the design stage and how to attain expected system behaviours and tunnelling classes based on ground types, boundary conditions, and influences. The second part presents some examples showing how to link the work done during the design stage with the observational approach during construction.