Lesson 6.1Lesson 6.1 · BIM in Design
Coordination & Clash Detection
BIM's most famous trick — and the honest truth about what it can and cannot catch
The duct runs straight through the beam. Better to find that on a screen than on a Tuesday, on site.
This is the demonstration that sells BIM. Overlay the architect's model, the structural engineer's model and the services engineer's model into one federated model, run clash detection, and the software instantly finds every place two disciplines have unknowingly claimed the same cubic metre — a duct through a beam, a pipe through a column, a light fitting where a sprinkler must go. Problems that used to surface on site, mid-construction, at ruinous cost, now surface on a screen, weeks early, fixable with a mouse.
It is genuinely transformative, and it is the single most cited reason clients adopt BIM. It is also the source of BIM's most dangerous myth — that it catches *everything*. This lesson gives you both halves honestly: the real, large power of clash detection, and the equally real limit that a clean clash report is never, by itself, proof of a good building.
'Zero clashes' is a great sentence with a hidden footnote: '...among the things I remembered to model, correctly.'
How coordination works: the federated model and its clashes
Coordination in BIM rests on the federated model from Module 2 — the discipline models overlaid at shared coordinates so they can be checked against one another. Against that federation, clash detection software (Navisworks, Solibri, BIMcollab and others) automatically finds conflicts, of which there are a few honest kinds.
A hard clash is two solid elements occupying the same space — a duct passing through a beam. A soft (or clearance) clash is elements too close for a real-world need — a pipe with no room for its insulation, or a valve with no access to maintain it; nothing overlaps, but the building would not work. And a workflow (or 4D) clash is the time-space conflict of Module 2 — two trades needing the same space in the same week. Found clashes are not just a list: they are triaged, grouped and assigned, and exchanged as BCF issues (Module 4) so each discipline resolves its share in its own tool. Coordination is a rhythm — detect, review, assign, resolve, re-run — not a one-off scan.
The power is real — and so is the relief
It is worth being clear about how large this benefit is, because it is not marketing. On a complex building, the services alone involve thousands of runs of duct, pipe, cable tray and conduit threading through a structure and around architecture. Coordinating that by overlaying 2D drawings by eye is slow, incomplete and error-prone; the misses become site clashes, and site clashes mean stopping work, cutting, re-routing, arguing over cost, and delay.
Clash detection turns that from a site problem into a design-office task. Finding the duct-through-beam on screen costs minutes; finding it on site costs days and money and goodwill. This is why coordination is the value most clients feel first and trust most — and why, on India's large, services-heavy projects (hospitals, airports, metros), it repays the modelling effort many times over. None of the honesty that follows takes this away: clash detection genuinely prevents a large, expensive class of problems, and that alone would justify BIM on many projects.
A clash found on a screen costs a mouse-click. The same clash found on site costs a week and an argument.
The honest limit: it only checks what you modelled
Now the truth that keeps clash detection from becoming a dangerous myth. It can only check what you actually modelled, to the quality you modelled it. It finds a duct running through a beam because both were modelled; it cannot find a duct you *forgot* to model, a wall with the wrong fire rating, a beam sized on a bad assumption, or a design that is coordinated but simply not good. It sees geometric collisions between modelled objects — nothing more.
This is 'garbage in, garbage out' in its most seductive form, because a clean clash report *feels* like proof of a sound building. It is not. A model full of omissions and errors will happily produce a clash report with zero clashes and coordinate its own mistakes perfectly. So clash detection must be paired with what it cannot do: model-quality checking (lesson 6.4), human design review, and the judgement to know that 'no clashes' means only 'the things I modelled do not overlap'. Treat a clean report as a *milestone*, not a *certificate* — the geometry does not conflict, which is necessary and valuable, but nowhere near sufficient for a building that is actually right.
What clash detection catches — and what it misses
Tap each problem — would clash detection catch it?
3 of 6 caught
The pattern in the misses: clash detection is blind to what you didn’t model, got wrong in data, or simply designed badly. A clean report is a milestone, not a certificate.
Three altitudes on the same idea
Read the band that fits you — or all three.
Learn both halves. Clash detection overlays the discipline models (the federated model) and automatically finds where they collide: a hard clash (two solids overlap, e.g. duct through beam), a soft/clearance clash (too close for access or insulation), or a workflow clash (two trades, same space, same time). Its power is real — it moves a whole class of expensive problems from the site to the screen. Its limit is equally real: it only checks *what you modelled, to the quality you modelled it*. A clean clash report means 'the things I modelled don't overlap' — not 'the building is good'. Garbage in, garbage out.
Run coordination as a rhythm, and never over-trust the report. Federate the models to shared coordinates, run clash detection, then triage: filter noise, group related clashes, set clearance tolerances sensibly, assign issues via BCF, and re-run. Distinguish hard from soft clashes and prioritise by real-world impact. But treat 'zero clashes' as one check among several, not a sign-off: pair it with model-quality validation and human review, because the report is blind to what you left out, mis-modelled, or simply designed badly. Your coordination is only as trustworthy as the models fed into it.
Buy the coordination value, and govern against the false confidence. Clash detection removes a large, measurable cost (site rework, delay, claims), and on services-heavy projects it repays the modelling effort many times over — a genuine business case. But manage the myth actively: require model-quality checks and design review alongside clash detection, define clash processes (matrix, tolerances, ownership, cadence) in the BEP, and never let a clean clash report be treated as proof of a sound building. The costliest failure is a well-coordinated disaster — a model whose errors and omissions were coordinated perfectly. Coordinate the geometry; verify the rest separately.
“If clash detection reports zero clashes, the design is coordinated and sound.”
Do it yourself
Sort real problems into 'clash detection catches this' and 'it cannot' — the point is to feel the boundary.
- 1Write six problems: (a) a duct runs through a beam; (b) a pipe has no room for its insulation; (c) a duct was never modelled at all; (d) a fire door is modelled with the wrong rating; (e) two trades are scheduled in the same room the same week; (f) the corridor is coordinated but too narrow to be pleasant.
- 2Mark which clash detection would catch: (a) hard clash — yes; (b) soft/clearance clash — yes; (e) workflow clash — yes (with 4D). And which it would miss: (c) an omission, (d) a wrong parameter, (f) a design judgement.
- 3Notice the pattern in the misses: clash detection is blind to what you did not model, got wrong in data, or simply designed poorly.
- 4Finally, write one line: if a project reports 'zero clashes', what three other checks would you still insist on before believing the design is sound?
The one line to carry out
Coordination checks whether the parts fit. The next question is whether the building will actually perform — will it be comfortable, efficient, safe? Next: using the model for design analysis, and the honesty simulation demands.
The author
Amogh N P
Architect, interior designer, and creative polymath. Studio Matrx began in his notebooks — his vision of design made honest, useful, and open to everyone. Its Academy is written and taught in his memory, and free, forever.
More about Amogh →