Trench Base Boiling in Sand Over Clay
A 40 m long, 6 m deep sewer trench is being excavated through loose to medium fine sand with groundwater at 2.5 m. Steel sheet piles are driven to 7.0 m, stopping 1 m above a stiff clay layer at 8.0 m. The contractor is dewatering with a sump at the trench base, holding the water 1 m below the base. After two days the trench bottom became spongy and began to boil near the sump, and a survey shows 40 mm of settlement at a road 12 m away. The inspector wants to know whether to stop work.
1. Is the trench base stable? Show the check and name the mechanism.
Read the worked answer
Stop work. The base is at or past the piping condition.
Mechanism: upward seepage through the sand plug between the trench base and the clay, with the sheet piles terminating above the clay so the flow passes under the toe.
- Head difference, h = 7.0 m (water level held) - 2.5 m (outside water table) = 4.5 m
- Thickness of the sand plug below the base, D = 8.0 - 6.0 = 2.0 m
- Buoyant unit weight of sand, γ' ≈ 9.5 kN/m3
FS against heave/piping = γ' x D / (γw x h) = (9.5 x 2.0) / (9.81 x 4.5) = 19.0 / 44.1 = 0.43
Equivalently, exit gradient = h/D = 4.5/2.0 = 2.25 against a critical gradient of about 0.97; FS = 0.43. Either way it fails by a wide margin, which is exactly what the spongy, boiling base looks like.
The condition is worst near the sump, where the flow lines converge, so local boiling starts there and spreads.
2. Why is a road 12 m away settling, and how is that connected to the sump?
Read the worked answer
The sump is the cause.
- The sump lowers the water level inside the trench, creates a head difference of 4.5 m, and draws water horizontally toward the trench from the surrounding sand. The drawdown cone extends well beyond 12 m in fine sand.
- The road settles by two mechanisms:
- Consolidation and densification. Lowering the water table increases effective stress in the sand and in any silt or clay lenses beneath the road. Even a small amount of silt or organic material in the profile can produce tens of millimeters.
- Fines migration and internal erosion. Boiling and piping near the trench wash fines out of the sand, which both loosens the sand and causes loss of ground. This is often the larger and more damaging component.
- The settlements compound: the road is at the edge of a drawdown cone whose center is failing. If the boiling continues, the loss of ground can become a sinkhole or a sheet-pile failure.
Note also that the sheet piles stop above the clay, so they provide no cut-off; the whole head difference is being carried by seepage through a 2 m sand plug.
3. What do you direct the contractor to do, and what would you monitor?
Read the worked answer
Immediate actions
- Stop excavation and stop deepening the sump. Backfill the trench bottom with granular material or place a tremie concrete seal to hold the base down.
- Do not simply pump harder - that increases the gradient and accelerates the failure.
Permanent fix, in order of preference
- Cut-off to the clay. Drive the sheet piles into the stiff clay at 8 m, lengthening the flow path so the head is dissipated before it reaches the base. This is the most reliable fix given the 2 m sand plug.
- Wellpoint or deepwell dewatering around the trench, installed before excavation, lowering the piezometric level uniformly below the base rather than drawing water through the base. This also produces a flatter drawdown cone and reduces off-site settlement.
- A tremie seal designed for the full uplift on the plug, with the slab or pipe buoyancy checked.
Monitoring
- Piezometers outside and inside the trench, with a trigger level tied to the base stability calculation.
- Inclinometers on the sheet piles and the adjacent road; survey points on the road, the curb, and nearby structures.
- A turbidity or fines check on pumped water - increasing turbidity means internal erosion is ongoing.
- Re-run the FS calculation for the as-built pile toe and the actual water levels before excavation resumes; anything below 1.5 is not acceptable in this ground.