Foundation design software calculates bearing capacity, settlement and structural adequacy for shallow footings, rafts and piles, and produces the output a design has to be issued with. The category is broad, the products differ more than their feature lists suggest, and most selection mistakes are made by comparing the wrong things.
A feature list states that two programs both calculate bearing capacity. It does not state which formulation of which method, whether the shape and depth factors are the ones the reviewer expects, whether the settlement analysis reaches the layer that governs, or whether the result can be read by someone who has to approve it.
The decision is better made by asking what the office actually produces, and working backwards.
Software is chosen on capability and regretted on workflow.
What Does Foundation Design Software Have to Do?
| Task | What it requires |
|---|---|
| Ground model | Boreholes, layers, groundwater, parameter derivation |
| Bearing capacity | Recognised methods with their shape, depth and inclination factors |
| Settlement | Immediate and consolidation components, with time where it governs |
| Soil–structure interaction | Subgrade reaction or spring stiffness for the structural model |
| Pile capacity | Shaft and base resistance, group effects, negative skin friction |
| Code checks | The partial factor or safety factor framework in force |
| Output | Reports a checker accepts, and drawings or schedules if the office issues them |
Every product in the category claims the first six. The seventh is where they diverge, and it is also the one that consumes most of the engineer’s week.
Why Does Having Shallow, Raft and Pile in One Model Matter?
Because foundation selection is iterative, and the alternatives have to be compared on the same ground.
A typical sequence starts with pad footings, finds settlement excessive in a compressible layer, moves to a raft, finds the differential settlement acceptable but the bearing pressure marginal at the edges, and ends with a piled raft. If each step means rebuilding the soil profile in a different module or a different program, the comparison is expensive enough that it usually is not made properly, and the first workable option is adopted rather than the best one.
Where the same soil profile carries all three foundation types, the comparison costs an afternoon. That is a workflow property, not a capability one, and it does not appear on feature lists.
Which Methods Should Be Implemented?
The established ones, stated by name and by author.
For bearing capacity that means Terzaghi (1943), Meyerhof (1963) and Vesić (1973), with the bearing capacity factors and correction factors each of them defines. These are not interchangeable, and a program that offers a single unattributed method leaves the engineer unable to match the reviewer’s expectation.
For settlement, elastic methods for the immediate component and one-dimensional consolidation for the long-term, with stress distribution by an accepted solution rather than a simplified spread. For piles, shaft resistance by the α, β or λ approaches, base resistance, and group efficiency.
What matters is not the length of the list but whether each method is identified well enough to be reproduced by hand.
How Should Settlement Be Handled?
As the check that usually governs, rather than as a secondary output.
For most structures on compressible ground, the allowable pressure is set by settlement rather than by bearing failure. A tool that computes bearing capacity thoroughly and settlement approximately has optimised the wrong half of the problem.
Three things are worth confirming in a trial. That the stress distribution continues to the depth where the increase becomes negligible, rather than stopping at a fixed multiple of the foundation width. That consolidation parameters can come from an oedometer test rather than only from a typed value. And that the time course is available where construction sequence matters, since a settlement that completes over five years is a different engineering problem from the same magnitude in five months.
What Does Code Compliance Mean in a Tool?
That the factor framework is implemented, not that a standard is named on the website.
EN 1997 defines design approaches that apply partial factors to actions, materials and resistances in different combinations. Applying the right factors to the right quantities is where implementations differ, and it is not visible from the output unless the intermediate values are shown. TBDY 2018 adds seismic requirements including liquefaction assessment. TS 500 governs reinforcement minimums that determine quantities.
The practical test is whether a report states which clause each check follows. A tool that produces a utilisation ratio without the clause behind it has moved the compliance argument from the software to the engineer.
Where Do Tools Actually Differ?
| Dimension | What separates products |
|---|---|
| Analysis depth | Classical methods against nonlinear finite element continuum models |
| Scope | Foundations only, or foundations with retaining, slope and liquefaction |
| Deliverables | Numbers only, or numbers with reports, drawings and quantities |
| Transparency | Results only, or results with the equations and intermediate values |
| Code framework | Implemented and clause-referenced, or left to the engineer |
| Speed of iteration | Minutes per option, or hours |
| Licence model | Perpetual, annual, or per-module |
The first row is the one most often treated as a ranking. It is not. A nonlinear finite element package is more capable and slower, and on a standard footing it answers a question nobody asked. The correct row to optimise depends on what the office does most days.
Why Do Two Programs Give Different Bearing Capacities?
Four reasons, none of them a defect.
Different method: Terzaghi, Meyerhof and Vesić produce different factors for the same friction angle. Different correction factors: shape, depth and inclination factors vary between references, and not all programs apply all of them. Different treatment of groundwater: the position of the water table relative to the foundation changes the effective unit weight in the third term, and the interpolation between cases differs. Different safety framework: a global factor of safety and a partial factor set do not give the same design value.
A difference of twenty per cent between two programs on the same input is normal and explainable. A difference that cannot be explained means one of the four is undocumented, and that is the finding.
What Does the Software Not Decide?
The parameters, and whether the answer is reasonable.
Characteristic values are selected by the engineer from scattered data. The decision to treat a layer as drained or undrained is an engineering judgement. Whether the ground model reflects the site between boreholes is a matter of interpretation, not computation. A program applies the method it is given to the numbers it is given, and it will do so precisely with parameters that are wrong.
It also does not decide the foundation type. That is a judgement about cost, programme, buildability and risk, and the calculation informs it rather than making it.
How Should a Trial Be Run?
On a finished project, not on a tutorial example.
The useful trial takes a job the office has already delivered, rebuilds it, and compares. It answers questions a feature list cannot: how long the ground model took to enter, whether the results match what was issued, whether a reviewer would accept the report as printed, and how long a design change takes to propagate.
Three specific checks are worth including. Change a layer thickness and see what has to be redone. Export the report and read it as a checker would. Take one bearing capacity result and reproduce it by hand from the printed equations. If the third is not possible, the first two matter more than they appear to.
Common Errors in Selecting Foundation Design Software
| Error | Consequence |
|---|---|
| Comparing feature lists rather than workflows | Capability bought, hours not saved |
| Evaluating on a tutorial example | Data entry effort and revision cost never tested |
| Treating analysis sophistication as a ranking | Effort spent where ground parameters dominate uncertainty |
| Ignoring report output during the trial | The largest recurring time cost is discovered after purchase |
| Assuming a named standard means an implemented framework | Compliance argument quietly returns to the engineer |
| Choosing per-module licensing without mapping actual use | Cost scales with the modules, not with the work |
| Not testing a design change | Revision cost, the dominant cost on real projects, stays unknown |
Where SETAF2018 Fits
SETAF2018 is built for offices whose work is repeated standard foundation and excavation design rather than research or exceptional structures, and the case for it follows from that.
Shallow, raft, continuous, pile, micropile and ground improvement foundations are designed in one model on one soil profile, so the comparison between foundation types is an afternoon rather than a project. Bearing capacity follows Terzaghi, Meyerhof and Vesić with SPT-based methods in sand; settlement combines elastic and one-dimensional consolidation with Terzaghi construction-period correction; subgrade reaction is derived from settlement, from the Vesić equation or from a plate load test. Pile capacity uses the α, λ and β methods with Converse–Labarre and Terzaghi block group efficiency, with negative skin friction and the neutral plane where compressible layers apply.
Design checks follow EN 1997 Design Approach 2, TBDY 2018 including Annex 16B liquefaction, and TS 500 minimum reinforcement, with the clause visible beside the check. Calculation reports print the equations and intermediate values, which is what makes an independent hand check possible.
What the software does not provide is nonlinear continuum finite element analysis, Hardening Soil behaviour or probabilistic assessment. For exceptional structures those remain the right tools. The foundation analysis workflow is aimed at the standard structures that make up most of the work, and a trial version is the only reliable way to establish which case an office is in. A survey of the wider category is set out in the geotechnical software buyer’s guide, and the cost question in the pricing guide.
Frequently Asked Questions
What is foundation design software?
It is software that calculates bearing capacity, settlement and structural adequacy for shallow footings, rafts and piles from a ground model, applies the relevant code framework, and produces the reports, drawings or schedules the design is issued with.
What should foundation design software include as a minimum?
A ground model built from boreholes with groundwater, recognised bearing capacity methods with their correction factors, immediate and consolidation settlement, subgrade reaction for the structural model, pile shaft and base resistance with group effects, an implemented code framework, and report output a checker accepts.
Why do two programs give different bearing capacity results?
Because they use different methods, apply different shape, depth and inclination factors, treat the groundwater position differently, or work in a different safety framework. A difference of around twenty per cent is normal. A difference that cannot be traced to one of those four is the real finding.
Is finite element software better for foundation design?
Not for standard foundations. Nonlinear continuum analysis adds capability that irregular geometry and unusual soil behaviour require, at a cost in time and parameter derivation. On a routine footing or raft, classical methods answer the design question and are easier to check.
Does settlement or bearing capacity govern foundation design?
Settlement, for most structures on compressible ground. Bearing failure governs on weak soils under heavy concentrated loads. Because settlement governs more often, a tool that treats it as a secondary output has optimised the wrong half of the problem.
How should foundation design software be evaluated?
By rebuilding a completed project during a trial rather than following a tutorial. That tests data entry effort, whether results match what was issued, whether the report is acceptable as printed, and how long a design change takes to propagate, which is the cost that dominates real projects.
What does code compliance mean in a design tool?
That the partial factor or safety factor framework is implemented and that each check names the clause it follows. Naming a standard on a website is not the same as applying the right factors to the right quantities, and the difference is invisible unless intermediate values are shown.
Can foundation design software choose the foundation type?
No. It calculates the consequences of a choice. Foundation type is decided on cost, programme, buildability and risk, with the calculation informing the decision. The software also does not select characteristic parameters, which remains an engineering judgement.
