MEP Clash Detection: How BIM Coordination Saves Rework

A 600 mm duct meets a post-tensioned beam. Found in the model, that conflict costs an hour of an engineer's time and a routing change. Found on site, it costs a stalled crew, a variation claim, an RFI chain, and — in the worst cases — core-drilling a structural member or dropping a ceiling by 100 mm across an entire floor. That difference is the entire business case for MEP clash detection.

This guide covers what clash detection actually checks, how a BIM coordination workflow runs in practice with Revit and Navisworks, the level of development that makes coordination meaningful, and the habits that separate a productive clash process from a report with 4,000 unread items.

What Clash Detection Is — and the Three Kinds of Clash

Clash detection is the automated comparison of discipline models — structural, architectural, mechanical, electrical, plumbing and fire — to find places where they conflict. The conflicts come in three types:

• Hard clashes: two objects occupying the same space — the duct through the beam, a cable tray through a pipe. Unambiguous and must be resolved.
• Soft (clearance) clashes: objects that do not touch but violate required clearances — no room to pull a filter, insulation squeezed to nothing, code-required access space in front of a panel blocked. These cause as much site pain as hard clashes and are easier to miss.
• Workflow (4D) clashes: sequencing conflicts — equipment that cannot physically reach its room once the structure is closed, or trades scheduled into the same zone at the same time.

The Coordination Workflow in Practice

On a typical project the cycle looks like this:

1. Each discipline models in its own file — in Revit, usually workshared models per discipline, developed against a shared grid and levels.
2. Models are federated — linked together, or appended into Navisworks, into a single combined model of the building.
3. Clash tests run against defined pairs — ducts vs structure, pipes vs ducts, cable containment vs everything, each with sensible tolerance settings so trivia does not drown the signal.
4. Clashes are grouped, filtered and assigned — one misrouted duct can generate dozens of raw clashes that are really a single issue; grouping by zone and system turns noise into a worklist.
5. Coordination meetings resolve them — each issue gets an owner and a resolution, tracked through BCF or the clash report, and verified closed in the next test run.
6. Zones are signed off — once an area passes clean, it is frozen for coordination and the next zone proceeds.

Revit's interference check handles quick intra-model checks during design; Navisworks (or equivalent checkers) carries the heavy multi-discipline federation. Where consultants work on different platforms, IFC exchange keeps every discipline inside the same federated model, and BCF carries the issues between tools without anyone retyping clash lists. The tool matters less than the discipline of the cycle: test, assign, resolve, re-test — on a fixed cadence, zone by zone, with the clash count trending visibly towards zero before each design freeze.

Why LOD 300–350 Is the Coordination Sweet Spot

Clash detection is only as good as the geometry it checks. Below LOD 300, elements are placeholders — generic shapes at approximate positions — and a "clash-free" result means nothing. Around LOD 300–350, elements carry accurate size, location and the connection-level detail that makes spatial coordination meaningful: this is where the duct is really 600×400 with 50 mm insulation, routed at a real elevation, with flanges and hangers acknowledged.

Pushing far beyond LOD 350 before coordination, on the other hand, wastes modelling effort on fabrication detail that may be rerouted anyway. Coordinate at 300–350; let shop modelling take it further once the routing is settled.

What Clashes Cost When They Reach Site

Industry studies consistently put rework at a meaningful percentage of construction cost, and field-discovered coordination errors are among its biggest drivers. The asymmetry is stark: a clash resolved in the model costs design hours; the same clash on site costs demolition, material, delay, and the knock-on resequencing of every following trade. On congested projects — hospitals, laboratories, high-rise cores — a single ceiling void conflict can cascade into weeks.

Who Runs Coordination — and When It Should Start

Clash detection fails most often not technically but organisationally: nobody owns it. Successful projects settle three things in the BIM execution plan before modelling starts:

• A named coordinator — one party (often the MEP designer or a dedicated BIM coordinator) runs the federation, owns the clash tests, chairs the meetings and tracks closure. Coordination by committee is coordination by no one.
• A model breakdown and schedule — which models exist, who authors each, shared coordinates, naming conventions, and a cadence of coordination cycles tied to design milestones, zone by zone.
• Decision authority — when the duct and the beam cannot both move, someone must be empowered to rule. Without that, clashes bounce between disciplines for weeks.

Timing matters as much as ownership. Coordination that starts at tender stage is documentation of problems; coordination that runs alongside design development is prevention. The cheapest clash of all is the one avoided because the structural engineer saw the main duct route before sizing the transfer beams.

For smaller consultancies and contractors, the practical constraint is capacity: running federated models, clash tests and coordination cycles is a specialist workflow that does not justify a full-time hire on every project. That is exactly the kind of scope that outsources well — the models and clash reports travel perfectly over a file link, and the coordination meetings happen on a call.

Habits That Make Coordination Actually Work

Teams that get clean models by handover tend to share the same working habits, none of which depend on software licences:

• Route gravity systems first. Drainage cannot negotiate; give it priority, then ducts, then pressurised pipe, then cable containment.
• Set tolerances deliberately. A 1 mm overlap between insulation and a wall finish is not the same severity as a duct through a column.
• Group before you count. Report issues, not raw clash numbers — 4,000 ungrouped clashes is a spreadsheet, not a plan.
• Reserve maintenance space as geometry. Model access zones and pull spaces as objects so soft clashes surface automatically.
• Freeze coordinated zones. Uncontrolled late changes to signed-off areas are how coordinated projects become uncoordinated ones.

Conclusion

BIM coordination is not about owning the right software — it is a management process that happens to use one. Model the disciplines to a meaningful level of development, federate them, run the clash cycle rigorously, and the arguments happen in a meeting room instead of a ceiling void. We produce coordinated Revit MEP models and drawings for consultants and contractors worldwide; if you are deciding what to send out versus keep in-house, start with our outsourcing checklist.