Basement Excavation in Lahore: Wall Movement Beyond Trigger and a Cracked Neighbour
A 7.5 m deep basement is being excavated in Lahore for a commercial plaza. The support system is a 600 mm diameter secant pile wall at 750 mm centres, propped by two rows of ground anchors grouted into a stiff clayey silt, with the water table 3.0 m below existing ground. A load-bearing brick masonry building from the 1940s stands 4.0 m from the wall crest; a pre-construction survey recorded a pre-existing crack of 0.3 mm. The monitoring plan sets an alert at 12 mm, a trigger at 15 mm and an action level at 20 mm of lateral wall movement. Six weeks into the works, excavation has reached 6.5 m. The inclinometer shows 14 mm on day 40 and 19 mm on day 52. A load cell on the lower anchor row reads 10% below its lock-off value, and the crack in the adjacent building has widened to 1.2 mm. The contractor wants to finish the last 1.0 m and cast the base slab this week, arguing that the slab will itself act as a strut.
1. Interpret the monitoring data. Is the contractor's argument that the base slab will act as a strut a sufficient reason to continue excavation?
Read the worked answer
Reading the data. Movement has already passed the trigger (15 mm) and is approaching the action level (20 mm). More importantly, the rate is adverse: 5 mm in 12 days, about 0.4 mm/day, at a time when the excavation depth increment was small. Movement that is still accelerating as restraint is added is the pattern that precedes a stability problem, not the pattern that stabilises.
The crack widening from 0.3 mm to 1.2 mm is a separate indicator: in masonry, 1 mm is commonly the threshold at which tensile cracking becomes visible and serviceability concerns begin. Two independent instruments - inclinometer and crack gauge - are telling the same story. The 10% loss of lower-anchor load suggests the anchors are not holding the load they were designed for, whether through lock-off loss, creep in the grout bulb, or the wall simply shedding load into the soil as it deflects.
The base slab argument. A base slab does act as a strut once it has gained strength and once it is cast against a wall that is not still moving. It provides no restraint during the one to two weeks before it cures. Using a future slab as justification for excavating the last metre now is exactly the reasoning that leads to collapse. Stop excavation below the current level.
2. State your immediate instructions on site, in sequence, and the notifications you would issue.
Read the worked answer
Immediate actions, in order:
- Stop excavation below 6.5 m. No dewatering changes without approval. Instruct a controlled berm or backfill against the toe of the most affected wall panels if the rate does not arrest within 24 hours.
- Escalate monitoring to daily inclinometer and crack-gauge readings, plus precise levelling of the adjacent building and any ground settlement points. Install tell-tales across the crack.
- Survey and photograph the adjacent building internally and externally, and inform the owners in writing the same day. This protects both the project and the occupants.
- Check the anchors: verify lock-off loads, re-stress the lower row to design load if the tendon and bond length allow, and review the anchor test records for creep.
- Inspect for water: leaking services, broken drains, or groundwater inflow softening the clayey silt is a common accelerator of movement.
- Notify the client and the design engineer of record immediately, and notify the local authority if the building is occupied and the action level is reached.
None of these are optional courtesies. Under an observational-method approach, the pre-agreed response to exceeding a trigger is a contractual obligation, and failing to act on it removes the defence that the design relied on.
3. Once the wall is stable, how would you decide whether to redesign the support before completing the excavation?
Read the worked answer
Back-analysis first. Re-run the soil-structure interaction with the parameters that reproduce the observed 19 mm, not the original design parameters. The fact that measured movement exceeded prediction, and anchor loads fell short, suggests the original model was optimistic - perhaps undrained strength too high, anchor bond in the silt overestimated, or the wall stiffness modelled as higher than a secant pile wall with joint discontinuities actually is.
Then test options against the remaining 1.0 m:
- Add a third anchor row or a steel strut at the current excavation level, designed for the load the back-analysis indicates.
- Excavate the last metre in short panels with immediate casting of a blinding layer, accepting higher cost for lower exposure.
- Locally widen or thicken the wall at the panels nearest the building.
Revise the monitoring plan with new, tighter trigger values justified by the back-analysis, and define in advance what happens at each level. A revised plan that has no pre-agreed action is not an observational method.
Communicate the revised design and the reasoning to the client and the adjacent owner before work resumes. Also check whether the original design was reviewed by the Authority having jurisdiction, and whether a revised submission is required.