Model coordination and clash detection

Model coordination is the most widely used construction-phase BIM use, and the one with the clearest return. By combining discipline and trade models and resolving conflicts before anything is fabricated or installed, a project team can avoid much of the rework that comes from ducts hitting beams, pipes passing through ceilings, or equipment with no room for maintenance. CON-120 introduced clash detection and CON-230 showed why MEP coordination in ceiling spaces matters. This lesson is about running coordination as a managed process: federating models, setting rules and tolerances, triaging results, running meetings, tracking issues and knowing when a zone is ready to release for fabrication.

Federated models

A federated model is a combination of separate discipline models, each still authored and owned by its originator, brought together in a coordination tool. Federation is different from merging: the architectural, structural, mechanical, plumbing, electrical, fire protection and specialty models remain separate files, and changes are made by their authors in their own software. The coordinator refreshes the federated model as new versions are shared through the CDE (Lesson 4).

Federation depends on shared coordinates, consistent levels and agreed units. If models do not align, every clash report is meaningless. The first step in any coordination cycle is a quick alignment check: grids, levels and a few known points in each model should coincide within a few millimeters.

Large models can be heavy. A full hospital might have dozens of model files totaling several gigabytes. Coordinators manage this by splitting models by zone and level, using lightweight exported formats for review, and loading only the models and zones relevant to the test being run. As a rough guide, file sizes grow with the number of elements and the detail of fabrication-level geometry; a LOD 400 MEP model of one floor can be larger than the whole LOD 300 design model.

Types of clashes

Not every conflict is a solid object passing through another. Coordinators normally distinguish three types.

Clearance clashes require modeling of the clearance zones themselves: maintenance access in front of equipment, code-required working space around electrical equipment, insulation thickness, door swings and ceiling access panels. Many conflicts that cause problems on site are clearance clashes, not hard clashes, because the model shows the pipe but not the insulation, or the panel but not the space needed to open it.

Electrical working space is a good example. The National Electrical Code (NFPA 70) requires minimum working space in front of electrical equipment, and owners often add maintenance zones in front of air handlers and pumps. Modeling these as transparent solid volumes lets the clash test find intrusions automatically.

Rules, tolerances and the clash matrix

A clash test compares one set of elements with another using a tolerance. A tolerance of zero reports every touching surface, which floods the results with trivial items, such as a pipe hanger touching the slab it hangs from. A larger tolerance ignores small overlaps that will be resolved in the field. Typical tolerances depend on the systems involved and the stage of design; they should be agreed in the BIM execution plan.

The clash matrix also sets priority. Gravity drainage, which must slope continuously, and large ductwork usually take precedence over pressurized piping and conduit, which can be rerouted more easily. Many teams follow a rough hierarchy: structure first, then gravity piping, then large ductwork, then sprinkler mains, then pressurized piping, cable trays and finally small conduit. Knowing the hierarchy speeds up decisions in meetings.

From raw clashes to issues

A first clash run between two complete trade models can report thousands of clashes. Most are duplicates or symptoms of a few root causes: one duct run that crosses forty joists is forty clashes but one issue. Coordinators group clashes by location, system and element before assigning them, and they mark clashes that are approved (acceptable as modeled, for example a pipe intended to pass through a sleeve) so they do not reappear.

The following script shows the idea with simulated data. It generates random clashes along duct runs, groups them by the run and the grid bay, and counts the issues that need decisions.

import random
from collections import Counter

random.seed(260)
runs = ["SA-1", "SA-2", "RA-1", "EA-1", "SA-3"]
bays = ["A-B/1-2", "A-B/2-3", "B-C/1-2", "B-C/2-3"]
clashes = []
for i in range(400):
    run = random.choice(runs)
    bay = random.choice(bays)
    other = random.choices(["beam", "joist", "pipe", "conduit"], weights=[3, 5, 2, 1])[0]
    depth = round(random.uniform(1, 180), 1)   # overlap in mm
    clashes.append((run, bay, other, depth))

significant = [c for c in clashes if c[3] > 25]          # beyond 25 mm tolerance
groups = Counter((r, b, o) for r, b, o, d in significant)
print("raw clashes:", len(clashes))
print("beyond tolerance:", len(significant))
print("issue groups (run, bay, element type):", len(groups))
for key, n in groups.most_common(3):
    print(key, n)

raw clashes: 400
beyond tolerance: 350
issue groups (run, bay, element type): 75
('EA-1', 'B-C/2-3', 'joist') 14
('SA-2', 'B-C/2-3', 'joist') 13
('EA-1', 'A-B/2-3', 'joist') 12

Of 400 raw clashes, the script keeps those beyond the 25 mm tolerance and groups them into a manageable list of issues by duct run, grid bay and element type. Real coordination tools do the same with rules, saved searches and grouping options. The goal of each cycle is not to drive the raw count to zero, but to resolve each real issue with a decision.

The coordination cycle

Coordination works best on a fixed weekly or biweekly cycle. Each trade uploads its updated model to the CDE by a set day; the coordinator federates and runs the tests; issues are grouped, assigned and circulated as BCF before the meeting; the meeting resolves issues and records decisions; and the trades update their models for the next cycle. Progress is tracked with a few simple measures.

Coordination meetings

A good coordination meeting is short and decisive. It focuses on the issues that need a decision from more than one party; simple fixes are handled by the responsible trade without discussion. The coordinator navigates the federated model on a large screen, the trade foremen or detailers attend in person or online, and the design team joins when a design change or clarification is needed. Decisions are recorded in the BCF issue, with an owner and a due date.

Coordination meetings often reveal design questions that only the engineer of record can answer: whether a beam can have a web opening, whether a duct can be reduced in size, whether a ceiling can drop. These should become RFIs (Lesson 7) promptly, because they are on the critical path to fabrication. The construction manager should watch for issues that stay open because nobody owns the decision.

Zone sign-off and release

When a zone is coordinated, each trade signs off that its model is coordinated with the others, and the zone is released for fabrication and installation. Sign-off typically requires that no unresolved clashes remain beyond tolerance, that all clearance zones are clear, that hangers and supports are modeled and coordinated, that sleeves and openings are confirmed with structure, and that the model matches approved shop drawings. Sign-off is a real commitment: once released, changes cost money.

After sign-off, the coordinated model drives the field. Trades print or display coordinated shop drawings, prefabricate spools and racks, and use layout tools to set inserts and sleeves from the model (CON-250 covers total station layout). Reality capture after installation (Lesson 8) checks that what was built matches what was coordinated.

Limits of coordination

Clash detection finds geometric conflicts between modeled elements. It cannot find conflicts with elements that were not modeled, such as small conduit, field-routed cable or temporary works. It cannot judge constructability, such as whether there is room to install a pipe in sequence or to swing a wrench. And it cannot check design intent: two systems may coexist without clashing while both violate the owner's requirements. Experienced superintendents and foremen remain essential in coordination meetings for exactly these reasons.