Clash detection

What the software does

It compares the disciplines’ models and flags where elements intersect, or come closer to one another than a set distance.

The result is a list — often hundreds or thousands of items.

Why the list is not enough

Because most of the hits are not real problems:

• two elements touch deliberately,
• modelling inaccuracy causes an intersection,
• the clash is manageable in reality,
• or it is already known and decided.

So the value of the exercise rests on the filtering: which hit matters and which does not.

The useful sequence

1. Set the rules: which groups of elements are compared, with what tolerance.
2. Run it.
3. Filter: pick out the real clashes.
4. Assign an owner: who resolves it.
5. Record the decision: how it was resolved.
6. Re-run: check on the next round.

Steps 4 and 5 are what make the exercise useful — without them all you get is a list.

Where it brings most

Services and structure

Pipe penetrations, web holes, soil stacks. The most common and most expensive type of clash. Web holes and service openings.

Structure and architecture

Openings, ceiling heights, the position of finishes.

Disciplines against one another

Mechanical against electrical, sprinklers against ventilation.

What matters more than this

Clash detection finds the geometric problems. A significant share of design faults is not geometric:

• wrong actions taken,
• missing restraint,
• a requirement not met,
• a contradictory technical description.

The model does not warn you about these — they need professional review. The design checker.

Timing

It is worth most when it happens before manufacture and there is time to resolve. Run a week before manufacture it is only damage assessment.

The point

Clash detection is a tool, not a result. The machine finds the intersections; which one matters and who solves it is an engineering decision.