Top-down hay Bottom-up? Chọn biện pháp thi công tầng hầm thế nàoTop-down or bottom-up? How to choose a basement construction method

For high-rise buildings in dense cities, the basement is usually the riskiest phase of the whole project: deep excavation in soft ground, right next to existing houses, below the water table. The first question a temporary-works engineer must answer is: build from the bottom up, or from the top down?

This article summarises how I look at that question — not to give a universal answer, but to make clear which factors really drive the choice.

Conceptual sketch, not to scale.

1. Two approaches

Bottom-up (conventional). Once the retaining wall is in place (diaphragm wall, sheet piles, bored piles…), the soil is excavated layer by layer and temporary supports — steel struts, walers or ground anchors — are installed at each level. At formation level the foundation is cast, then the basement slabs are built upwards; each level of struts is removed once the slab above it has gained strength.

Top-down. The diaphragm wall and bored piles are installed first, with steel king posts plunged into the piles. The ground-floor slab is cast first, then soil is excavated beneath it through openings, and the basement slabs are cast one by one going down. The slabs themselves act as the lateral support for the wall. In full top-down, the superstructure can rise at the same time.

2. Quick comparison

3. What really decides

1. Depth and number of basements. The deeper the pit, the more strut levels, the higher the cost and deflection — this is where top-down starts to win.
2. Neighbouring structures. Shallow-founded houses, roads and utilities next door are the strongest constraint. When allowable movements are tiny, the stiffness of concrete slabs is a major advantage.
3. Ground and groundwater. Thick soft clay and a high water table increase deflection and the risk of basal heave or leakage through wall joints.
4. Site footprint. Wide pits mean long struts that need intermediate supports; narrow sites make laydown areas and openings for top-down harder.
5. Programme and client cash flow. If an early handover of the superstructure (e.g. a retail podium) matters, top-down allows two fronts in parallel.
6. Contractor capability. Top-down demands tight control of king-post verticality, excavation sequence and joint detailing. A method that looks good on paper but exceeds what the site can control becomes a risk.

Lesson: there is no absolute right answer. Many projects use a hybrid (semi top-down) — for instance open excavation in one zone and slab propping next to sensitive neighbours — to balance safety, cost and time.

4. Monitoring: the engineer’s eyes underground

Whatever the method, the design model is only a forecast. Soil parameters are highly uncertain, so excavation must always be monitored alongside construction:

• Inclinometers in the wall: lateral deflection with depth after each excavation stage.
• Settlement points on neighbouring buildings and the surrounding ground.
• Standpipes / piezometers inside and outside the pit.
• Strut load measurement (strain gauges, load cells) for bottom-up.
• Tiltmeters and crack monitoring on adjacent buildings.

Each quantity has an alert level and an action level. What matters is not just measuring, but comparing measurements with predictions after every stage — in time to adjust the sequence, add support or pause work.

5. In short

• Bottom-up suits moderately deep pits on roomy sites with less sensitive neighbours.
• Top-down pays off for deep basements, sensitive neighbours, or when the overall programme must be shortened.
• The final decision should rest on deformation analysis, cost, programme and site capability, always paired with a good monitoring plan.

In upcoming articles I will go deeper into each part: designing excavation strutting, controlling king-post verticality, and reading inclinometer data.