"Clash-free" gets used loosely in this industry — sometimes it means "we ran a clash test," sometimes it means "we resolved everything we found," and sometimes it's just a phrase on a proposal with nothing behind it. When we say a model is clash-free, it means something specific: every hard clash between modelled systems has been identified, assigned, resolved, and verified resolved, on a model detailed enough for that resolution to hold up in construction.

Here's what that actually involves, and where the guarantee does and doesn't apply.

THE SHORT VERSION

What "clash-free" actually covers

HARD CLASHES

Physical overlap

Two modelled elements occupying the same space. These get fully resolved before sign-off.

CLEARANCE ISSUES

Access & maintenance

No physical overlap, but a code or service clearance violation. Tracked separately, resolved to the same standard.

SCOPE

LOD 400 only

The guarantee applies to construction-ready models. Design-stage coordination doesn't carry it — see below.

HARD CLASH VS. CLEARANCE

Two different problems, tracked differently

A hard clash is straightforward: two pieces of modelled geometry physically occupy the same space. A duct running through a structural beam. A pipe intersecting a cable tray. These are unambiguous — either the geometry overlaps or it doesn't.

A clearance issue is different. Nothing physically overlaps, but a required clearance is violated — a fire damper without the code-mandated service access, a valve installed too close to a wall to be operated or maintained, a panel without its required working clearance. These matter just as much as hard clashes, but they get missed by a naive clash test that only checks for physical intersection. Both categories get resolved to the same standard before we call a model clash-free, but they're tracked as distinct issue types because the fix for each is different.

THE PROCESS

How resolution actually happens

Clash detection isn't a single pass. It's a cycle, run in Navisworks or Revizto, repeated until the list gets to zero:

  1. Test setup — clash rules configured by discipline pair, tolerance and severity, so the test is checking the right things at the right sensitivity.
  2. Run and group — results grouped by location and discipline instead of reviewed as one flat list of hundreds of items, most of which are duplicates of the same underlying conflict.
  3. Assign and prioritise — every clash gets a named owner and a priority, in a joint review rather than left as an unassigned report.
  4. Resolve and re-test — the model gets updated, and the same test runs again. This repeats until the list is empty, not until it's "smaller."
A clash report with zero items on the day it was generated isn't the same as a model that's actually clash-free — the model has to stay clean after every subsequent change, not just on test day.
WHY LOD MATTERS HERE

Why this only means something at LOD 400

A clash test run on a design-stage LOD 300 model can absolutely return a clean result — and that result is still useful, it tells you the current layout doesn't conflict at a coordination level. But the components in that model are generic. Real fittings, real supports, real insulation thickness aren't there yet. A clean LOD 300 clash report describes the design intent, not the thing that's going to get fabricated and installed.

That's why the clash-free guarantee is scoped to LOD 400 construction-ready engagements specifically. At that level, the geometry being tested is the geometry that's actually going to be built — manufacturer-specific fittings, real support brackets, real clearance envelopes. A clean result at LOD 400 is a claim about the real world, not the design.

WHAT'S NOT COVERED

Where the guarantee stops

A clash-free guarantee covers modelled systems against each other and against structure/architecture as coordinated. It doesn't cover: scope that wasn't modelled (existing conditions not captured, systems outside the contracted discipline set), changes made after sign-off by someone other than the coordination team, or field conditions that differ from what was surveyed or documented. This isn't a loophole — it's the same boundary any legitimate coordination guarantee has to have, since a model can only be checked against what's actually in it.

HOW WE APPLY THIS

Part of construction-ready delivery, not a separate add-on

Clash detection isn't something we bolt onto a project at the end — it's run continuously through LOD 400 modelling and formalised through BIM coordination, with the clash-free sign-off as the gate before shop drawings get produced. If a project is still at design stage, we run the same disciplined process without the formal guarantee attached, since the model itself isn't yet detailed enough for that guarantee to mean what it should.