Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE TENTH INTERNATIONAL CONFERENCE ON CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING COMPUTING
Edited by: B.H.V. Topping
Enhancing Numerically-Predicted Ground Movement Patterns due to Shield Tunnelling in Clays
T.T. Abdel-Fattah+, A.Y. Akl*, H.A. Hodhod* and A.M. Abdel-Rahman+
+Department of Soil Mechanics and Foundation Engineering, Housing and Building Research Centre, Giza, Egypt
T.T. Abdel-Fattah, A.Y. Akl, H.A. Hodhod, A.M. Abdel-Rahman, "Enhancing Numerically-Predicted Ground Movement Patterns due to Shield Tunnelling in Clays", in B.H.V. Topping, (Editor), "Proceedings of the Tenth International Conference on Civil, Structural and Environmental Engineering Computing", Civil-Comp Press, Stirlingshire, UK, Paper 276, 2005. doi:10.4203/ccp.81.276
Keywords: shield tunnelling, deconfinement method, tunnel excavation, gap closure, ground movement, finite-element analysis.
Predictions from conventional numerical techniques for simulating shallow-tunnel excavations are normally found to be in poor agreement with the ground response measured. These techniques (e.g., deconfinement of the initial geostatic stress and shrinkage of the tunnel lining) usually predict ground settlement troughs that are relatively wider and flatter than those of the normal Gaussian distribution, proved to be in good agreement with the observed response. This is likely due to the use of techniques for simulating excavations that are not sufficiently capable of describing the unloading pattern associated with the excavation process.
A number of simplified procedures were introduced in the literature to obtain better numerical predictions for the ground response. Burghignol et al.  suggested the use of two different reduction factors for both the vertical and horizontal components of the initial geostatic loads acting on the tunnel perimeter prior to excavation. The results obtained were found in very good agreement with their measured counterparts for ground surface movements. However, the method used does not completely represent a realistic subsurface ground movement pattern.
Bloodworth  proposed the introduction of an additional external restraint to the tunnel lining at either the springline or invert. Although better estimates for the settlement trough were obtained, they were still relatively far from the field observations. Yet, the trough that resulted from imposing a fixity at the invert was found to be a better approximation to the Gaussian model.
Brinkgreve and Broere  argued that the flatter settlement troughs obtained from the finite element (FE) analyses are due to the underestimation of the soil stiffness in the small-strain region (region away from the tunnel). A slightly improved shape for the settlement trough was obtained when a finite element analysis was produced using a soil stiffness five times higher than the original stiffness. The use of this procedure resulted in a slightly improved shape for the settlement trough. Yet, a detailed investigation made by Burghignoli et al.  showed that even when adopting advanced constitutive modelling via which the small strain stiffness could be adequately estimated, only a slight improvement in the prediction of the settlement trough was obtained.
The objective of this paper is to introduce a technique in which the unloading pattern adopted is adequately representative to the states of stress release caused by excavation, and this may result a better estimates of the ground response from the finite-element analyses Abdel-Fattah .
A numerical simulation of the deconfinement of a tunnel constructed in clays using the shield tunnelling method by differential unloading in terms of tangential and radial components was carried out. The former is intended to simulate the effect of the machine driving through clays, whereas the later is intended to simulate the gap closure behind the shield tail.
In order to examine proposed procedure, the test problem previously solved by Burghignoli et al.  is resolved here. The results obtained from the proposed procedure using the finite element method were found in very good agreement with their measured counterparts, and superior to those obtained previously for the subsurface movements.
purchase the full-text of this paper (price £20)