Running a clash test is trivial. Producing a report that a trade contractor will act on is not. The difference between the two is the reason many coordination programs generate thousands of clashes each week and resolve very few of them.
A clash report is not a deliverable. Resolution is the deliverable. This article describes a workflow built around that distinction.
Why raw clash counts are meaningless
Point a clash test at two full discipline models with default settings and you will get tens of thousands of results. A large proportion of them are the same physical condition reported repeatedly, elements the model was never intended to coordinate, or geometric intersections that are entirely acceptable in construction — a pipe passing through a wall that will be sleeved, or a hanger touching a ceiling grid.
Presenting that output to a project team destroys confidence in the process within one meeting. The workflow exists to convert raw geometric intersections into a short, ordered list of decisions someone must make.
Step 1 — Prepare the models
Coordination quality is determined before any test runs.
- Shared coordinates verified. Every model must be published on the same coordinate system with a confirmed survey point. A single misaligned model produces a clash storm that masks every real issue.
- Appropriate LOD. Coordination requires LOD 350 for the trades being coordinated. At LOD 300, supports, hangers, insulation, and connections are absent, so the model cannot represent the condition that actually conflicts.
- Model hygiene. Purged, audited, no unplaced elements, no linked models nested unintentionally.
- Consistent element categorization. If ductwork is modeled with generic models in one file and duct categories in another, search sets cannot be built reliably.
Step 2 — Build search sets, not selection sets
Search sets are rule-based and update automatically when the model changes. Selection sets are static lists and become wrong the moment a model is refreshed. Every recurring coordination program should be built on search sets.
A workable structure organizes by discipline, then system, then level:
- Structure → Beams, Columns, Slabs, Braces, Foundations
- Architecture → Walls, Floors, Ceilings, Doors, Vertical Circulation, Shafts
- Mechanical → Supply Duct, Return Duct, Exhaust Duct, Hydronic Pipe, Equipment, Hangers
- Plumbing → Domestic Water, Sanitary, Storm, Gas, Equipment
- Electrical → Cable Tray, Conduit, Busway, Panels, Lighting
- Fire Protection → Sprinkler Main, Branch, Heads, Standpipe
Building these once, correctly, is the highest-return hour in the whole process.
Step 3 — Define the clash matrix
The clash matrix states which system pairs are tested, at what tolerance, and with what priority. Without it, coordination is arbitrary.
A representative matrix for a commercial fit-out:
| Test | Tolerance | Priority |
|---|---|---|
| Structure vs Mechanical Duct | 0 mm | Critical |
| Structure vs Hydronic Pipe | 0 mm | Critical |
| Structure vs Sprinkler Main | 0 mm | Critical |
| Mechanical Duct vs Sprinkler | 0 mm | High |
| Mechanical Duct vs Cable Tray | 25 mm | High |
| Hydronic Pipe vs Sprinkler Branch | 25 mm | Medium |
| Ceiling vs All MEP | 0 mm | High |
| Architecture Walls vs Duct | 0 mm (sleeved exceptions) | Medium |
| Cable Tray vs Conduit | 50 mm | Low |
Negative tolerance — testing for insufficient clearance rather than physical intersection — is under-used and valuable. Testing duct against structure at −50 mm surfaces conditions where insulation and hangers will not fit even though the geometry does not intersect.
Step 4 — Group before you report
Grouping is the step that separates a useful report from an unusable one. A single duct running through eight beams produces eight clashes and represents one decision. Group by the offending element, or by zone and system pair, so each group corresponds to one conversation.
Well-grouped reports typically reduce raw clash counts by 80–95%. A weekly coordination meeting can work through 30 grouped issues. It cannot work through 4,000 individual hits.
Step 5 — Assign, track, and close
Every grouped issue needs four attributes: a responsible party, a required-by date, a status, and a resolution note. Navisworks alone tracks status poorly across model versions; a coordination platform such as BIM Track, Revizto, or ACC Issues is what keeps the audit trail intact.
Statuses should be minimal and unambiguous: New, Assigned, Resolved in Model, Verified Closed. Anything more granular is rarely maintained honestly.
The critical discipline is verification. An issue marked resolved by the trade must be re-tested in the next model cycle and closed only when the test confirms it. Self-certified resolution is the most common source of clashes reappearing on site.
Step 6 — Run a disciplined cycle
A functioning coordination rhythm on a typical commercial project:
- Monday — trades publish updated models to the common data environment
- Tuesday — federation, clash tests, grouping, report issue
- Wednesday — coordination meeting, assignment, decisions recorded
- Thursday–Friday — trades revise
- Following Monday — republish, verify prior issues, repeat
Predictability matters more than frequency. Teams meet the cycle they can plan around.
Metrics worth tracking
Three numbers tell you whether coordination is working:
Open critical issues by zone. This is the schedule risk indicator, and it should trend to zero ahead of each zone's installation date.
Average age of open issues. Rising age means decisions are not being made, regardless of how many clashes are being closed.
Recurrence rate. Issues closed and reopened indicate verification failure or a model exchange problem.
Total clash count, tracked alone, tells you almost nothing.
Common failure modes
Coordinating too late. Clash detection starting after shop drawings are underway converts coordination into rework.
No zone sequencing. Coordinating the whole building at once means the first zone to be built is not prioritized. Zone-based coordination aligned to the construction sequence is far more effective.
Excluding the ceiling and structure interfaces. Most costly field conflicts occur in the interstitial space, where every trade competes for the same 400 mm.
Treating the BIM coordinator as a report generator. The role requires the authority to escalate unresolved issues. Without it, the process documents failure rather than preventing it.
Soft clashes and clearance testing
Hard clashes — geometry physically intersecting — are the easy category. The conflicts that cause site problems are more often clearance failures, where nothing intersects but nothing can be installed or maintained either.
Clearance testing is done by modeling the required space as geometry and testing against it:
- Maintenance envelopes around equipment: filter withdrawal, coil pull space, valve access, panel door swing
- Installation clearance: the space a component must pass through to reach its final position, which is frequently larger than the component
- Code-required clearances: electrical working space, egress widths, fire-rated separation
- Insulation and lagging, tested at full installed thickness
- Structural deflection allowance on long spans, where a beam's deflected position differs from its modeled one
- Ceiling access: tile removal and access panel positions
Modeling these as explicit clearance solids in the coordination model, and adding them to the clash matrix, converts a category of problem that is normally discovered on site into one that is resolved in the office.
Reporting that gets read
The final output should be short and directive. A useful weekly coordination report contains:
- A one-page summary: open critical issues by zone, movement since last week, and anything at risk of a schedule date
- The grouped issue list, ordered by priority and zone, each with an owner and a date
- Screenshots with enough context to locate the issue — grid references and level, not just a floating close-up
- Verification results on issues closed since the last cycle
- An escalation list: issues past their required-by date, named
The screenshot detail matters more than it sounds. A clash image with no grid reference forces every recipient to spend several minutes locating it, multiplied across every issue and every attendee.
Frequently asked questions
When should clash detection start? As soon as two disciplines have modeled a shared zone to LOD 350 — typically well before construction documents are complete.
Is Navisworks still the standard? It remains the most widely used federation and clash engine, though Revizto, Solibri, and ACC Model Coordination are all in common use. The workflow principles are identical across them.
How many clashes are acceptable at handover to construction? Zero critical clashes in the zone being released. Low-priority clashes may remain open with agreed field resolution, provided they are documented.
Related reading: The BIM Execution Plan: What to Put In It and What to Leave Out · MEP Shop Drawings: Coordinating the Interstitial Space · LOD 100 to LOD 500 Explained: A Practical Reference
Vantage CAD Services provides BIM coordination, clash detection, and issue tracking as a managed service for AEC teams. Contact info@vantagecadservices.com or +1 (512) 543-0831.