Foundation Engineering / Choosing a foundation
Shallow, deep, or improve the ground
Three options, and the questions that choose between them.
Lesson review: expert verification remains open
This lesson has received a full-context editorial review and a check of its technical explanations. Selected arithmetic has been checked separately. This is not independent engineering sign-off.
Checks needed for project use
- A geotechnical reviewer must confirm drainage and stress basis, bearing-factor formulation, correlations and settlement influence factors. Partial factors and project-specific ground conditions are outside these teaching checks.
- Investigation-depth and failure-mode rules are indicative. Specialist must check geometry, layering, load, soft layers outside initial influence estimate and any asserted strength-parameter reductions.
These are educational models and exercises. Check their assumptions before using the results. The full technical content and every diagram have not yet been independently verified.
Concrete and steel calculations retain their stated first-generation Eurocode teaching basis. Use the standards and amendments specified for your project; do not mix generations without an appropriate design basis.
Read the lesson register and outstanding review notesWhy this matters
Piles are expensive, slow, and require plant that does not fit on every site. They are also, quite often, the right answer. Knowing which situation you are in is worth more than knowing how to design either kind.
The choice is usually made by working down a short list of questions, in order.
Can a shallow foundation carry the load without excessive settlement? If yes, use one. Pads and strips are cheap, quick and understood by everyone on site.
Can it be made to work by going wider, or a little deeper? At a fixed total load, widening reduces average pressure; its effect on settlement must be calculated for the ground profile. Going deeper changes the strata, net pressure and construction stress path, and can help or create additional movement and groundwater problems.
Can the ground be changed instead of the foundation? Vibro compaction, stone columns, preloading and grouting all alter the soil rather than the structure. On a large site this is frequently the cheapest answer.
Only then: does the load need taking to deeper ground? That is when piles earn their cost.
Settlement, not capacity, is the usual reason to go deep. A raft on soft clay will very often satisfy bearing capacity comfortably and still be rejected, because the predicted settlement — or, more often, the differential settlement between one part of the structure and another — is more than the frame and its finishes will tolerate.
This is worth stating plainly because it changes what you look for in a ground investigation. If capacity governed, strength data would be enough. Because settlement governs, stiffness and compressibility data matter more, and those are harder to obtain and easier to leave out.
Check yourself
A single-storey warehouse with a tolerant portal frame is to be built on 8 m of soft alluvium over dense gravel. Loads are light. Which option deserves to be examined FIRST?
Try it
Gross, net and the compensated foundation
Excavating removes stress the ground was already carrying. Only the net increase reaches the soil below.
Reading it
- Net pressure is what the settlement calculation must use.
- At 100% compensation the ground sees no stress increase at all.
- Beyond 100% the foundation is a net unloading — and heave, not settlement, becomes the problem.
- Paid for by excavation
- Net increase on the ground
- Stress removed
- 29 kPa
- Net pressure qnet
- 152 kPa
- Degree of compensation
- 16 %
- Depth for full compensation
- 9.5 m
16% compensated. The settlement calculation should use 152 kPa, not 180.
Summary
- Work down the options in order of cost: shallow, then wider or deeper, then ground improvement, then piles
- Settlement rather than capacity is the usual reason a shallow foundation is rejected
- A pile carries load past bad ground to good ground; with no good ground in reach it helps far less
- What the structure can tolerate is part of the foundation decision, not a separate question