6.8 Deep excavations

Stability of slurry trench constructed in sandy ground

Many interpretations on the stability mechanism of slurry trenches have been proposed on the basis of either small scale model tests under 1 G or limited field observations. The former, of course, did not satisfy the similarity law in modeling. The only centrifuge test on the problem was reported by Bolton et al (1973). However, their test was limited to slurry trenches in clay.

In Japan, diaphragm walls approximately 100 m deep have been constructed, while some failures of slurry trenches, mostly those constructed in sandy ground under high ground-water table conditions, have occurred. Thus, a series of 2-dimensional and 3-dimensional centrifuge model tests were conducted to investigate the stability mechanism of the slurry trenches in sandy ground.

Photo 1 shows an over view of testing system. Model ground was prepared by sand pouring. After saturation of the model ground, container was put into a centrifugal acceleration of n G under the saturated condition to give the model ground a soil self-weight stress similar to that of the prototype (the value of n ranged from 50 thru 175).

After the flight was stopped, the ground-water table was lowered below the trench bottom, and the dummy plate, which had been installed beforehand, was removed while the vacated space was filled with slurry. When the required time for slurry to penetrate into the ground had elapsed, centrifugal acceleration was increased to the n G, and the ground-water table was gradually raised until the trench failed owing to the ground movement. As are shown in Photo 3, the failure was clearly observed.

Trench depth Z=20 m in prototype	Z=30 m in prototype

Photo 3 Failure pattern in 2-D models

After the 3-D model tests, model ground was sliced vertically to investigate the 3-dimensional shape of slip failure (Photo 4). These test results were employed to establish a new analytical method, using a 3-dimensional limit equilibrium method, for predicting the stability of slurry trench.

Bolton, M.D. et al. (1973) : Ground displacements in centrifugal models, Proc. of 8th ICSMFE, Vol. 1, pp. 65-70.

Tohda, J., Nagura, K., Kawasaki, K., Higuchi, Y., Yagura, T. and Yano, H. (1991) : Stability of slurry trench in sandy ground in centrifuged models, Proc. of CENTRIFUGE 91, pp. 75-82.

Stability and failure region of slurry trench

Similar tests were carried out at Nikken Nakase Geotechnical Institute. In the test, trench was modelled by rubber bags filled with saline water (Fig. 1).

1) sandy ground Dr = 50 %

2) sandy ground Dr = 50 % with partial surcharge 34 kPa

3) clay layer overlying sandy ground

4) sandy ground (after Higuchi et al. (1994))

Failure patterns with various ground conditions were investigated as shown in Photo 5-7.

2-D condition	3-D condition L=6 m	3-D condition L=3 m

Photo 5 Sandy ground

	side view	plane view
2-D condition	3-D condition L=6 m

Photo 6 Sandy ground with partial surcharge

plane view	plane view
side view	side view
3-D condition L=6 m (thick clay layer)	3-D condition L=6 m (thin clay layer)

Katagiri, Saitoh, Masuda, Aizawa and Ugai (1997a) : Shape effect on deformation behaviour and stability of slurry trench walls constructed in Sandy ground, Proc. of Deformation and Progressive Failure in Geomechanics, pp. 665 - 670.

Katagiri, Saitoh, Masuda, Aizawa and Ugai (1997b) : Centrifuge model tests on stability of slurry trenches, Tsuchi-to-kiso, Vol. 45, No. 10, pp. 13-16 (In Japanese).

Higuchi, Tohda, Nagura and Kawasaki (1994) : A new stability analysis method for slurry trenches constructed in sandy ground, Tsuchi-to-kiso, Vol. 42, No. 3, pp. 7-12 (In Japanese).