TI Addenbrooke, Imperial College, University of London. This paper was first published in GE’s September 1994 edition.
To support excavations in urban areas where construction space is limited and excavation depths large (deep building substructures, or cut and cover transportation tunnels) it is common practice to use vertical retaining walls supported internally by multi level propping. Construction is usually carried out in close proximity to existing roads, buildings and services.
The analysis of such an excavation should therefore examine:
(1) The possibility of collapse of the retaining structure, either as a result of excessive prop loads or failure of the soil at the bottom of the excavation.
(2) The predicted movements caused by excavation and dewatering, particularly with regard to their effects on nearby structures (Clayton & Milititsky, 1986).
(3) The displaced shape of the wall, as large bending moments may be induced (Potts & Fourie, 1985).
This paper considers points (2) and (3), movements and induced bending moments. It is rare for a multi propped excavation to fail due to structural problems. More commonly ‘failure’ of a support system is because of unacceptable movements (Clough et al, 1989). Buildings sited adjacent to deep cuts are generally less tolerant of the subsequent excavation —induced differential settlements than similar structures settling under their own weight (Boscardin & Cording, 1989).
Potts & Day (1990) state that both experimental work (eg Rowe, 1952) and more recent numerical work (eg Potts & Fourie, 1985) indicate that under the same operating conditions stiffer walls attract larger bending moments than more flexible walls. The stresses imposed by the soil are free to redistribute through a more flexible structure thus reducing the structural forces imposed on the wall. This in itself is beneficial, but occurs at the expense of larger wall and soil movements. They observe that there is therefore a compromise between reduced bending moments and increased movements as the flexibility of the wall increases.
If greater movements cannot be tolerated, then more props are required if a more flexible wall is to be employed. The engineer needs a framework within which bending moment reduction can still be considered, balanced against the required increase in the number of propping levels. This paper introduces a new flexibility number for multi propped retaining wall design, the ‘displacement flexibility’. The displacement flexibility number permits the engineer to confidently consider cost and construction variations within a displacement controlled design framework. Its applicability is justified using finite element predictions of movements for various specific, but representative support systems.
Two sets of analyses are presented. Set A models different retaining wall systems supporting an excavation in dry soil, while Set Bmodels systems supporting a rapid excavation followed by consolidation in a water bearing soil.