Coordination · 8 min read

Visual coordination or automatic clash detection: which to choose?

By Mickael Quinart · 9 October 2026

Partager :
Visual coordination or automatic clash detection: which to choose?

Visual model review or automatic clash detection: strengths, limits and a combined method for efficient coordination.

As Mickael Quinart, BIM/CAD Consultant at PALLADION, I invite you to enrich your understanding of BIM coordination, an essential pillar for the success of your projects.

Coordination is at the heart of the BIM process. It aims to identify and resolve potential interferences and inconsistencies between the various technical and architectural packages of a project, even before physical construction begins. Effective coordination helps avoid delays, cost overruns, and errors on site, by leveraging digital models. It primarily relies on two complementary approaches.

Two Ways to Coordinate

Visual coordination involves reviewing the federated model in meetings to identify inconsistencies. This method requires experts from each trade to examine the 3D model, simulate operational or maintenance scenarios, and assess the feasibility or relevance of installations. It is intuitive and allows for the detection of problems that are not purely geometric.

Automated clash detection uses geometric rules to list interferences between objects. Specialized software analyses models and identifies all areas where elements intersect or are too close, according to predefined tolerances. This is a systematic approach that ensures comprehensive geometric checks.

The two are complementary. The first brings human intelligence and experience; the second, computational power and rigour.

Figure 7 — General BIM Process. Source: M. Quinart's thesis.
Figure 7 — General BIM Process. Source: M. Quinart's thesis.

Comparison

CriterionVisual CoordinationAutomated Detection
ExhaustivenessLowHigh
RelevanceHigh (human judgment)Variable (many false positives)
Preparation TimeLowRules and tolerances to be configured
Problems DetectedLogic, access, maintenanceGeometric clashes
Ideal forConcept Design / Developed Design, reviewsTechnical Design / Production Information, clash detection

The Pitfall of Thousands of Clashes

An initial detection run often produces thousands of results. Most are duplicates or acceptable contacts (partition against slab, duct in a penetration). Without sorting, the tool can discourage teams. This digital "noise" is the main reason for abandoning or underutilising automated detection tools. It is imperative to understand that the machine identifies all contact, but does not immediately distinguish a real problem from a harmless contact or a false positive.

The Combined Method

To get the most out of these two approaches, a structured method is essential. It ensures the efficiency of the process and the relevance of corrective actions.

  1. Check the quality of the models first (levels, duplicates, incorrectly classified objects). A poor-quality model will inevitably generate an astronomical number of false positives and obscure real problems. The use of standardised classification schemas (e.g., well-populated IFC properties) is crucial.
  2. Define targeted test matrices (structure / services, services / services). It is not always relevant to test all possible combinations. Focus on critical interfaces and areas with high element density.
  3. Set realistic tolerances and exclude acceptable contacts. For example, a 0mm tolerance is rarely appropriate. Allow for minimal clearance for insulation or installation. Clash detection software (such as Autodesk Navisworks or Solibri Model Checker) allows for the definition of specific clearances and exclusions.
  4. Group clashes by zone and by cause. Do not treat each clash individually. Analyse groups of clashes that stem from the same design decision or geographic area. This facilitates arbitration and decision-making.
  5. Present them in a visual review for arbitration. It is at this stage that human expertise takes over. Grouped clashes are shown to experts from the various disciplines concerned, who decide on the best solution.
  6. Disseminate decisions via BCF with responsible party and deadline. The BCF (BIM Collaboration Format) is an open standard (ISO 19650-2:2018 Annex D) that allows problems, their locations, screenshots, and comments to be communicated directly between BIM software without exchanging the full model. Each decision must be assigned to a responsible party with a resolution deadline.

Real-world Case Study in a Design Office

Imagine a complex renovation project for a city-centre office building. The engineering team consists of architects, structural engineers, HVAC, plumbing, and electrical engineers, each working on their own discipline model.
During the technical coordination phase (equivalent to the Technical Design phase), the BIM coordinator aggregates all models into a federated model. They begin by running automated clash detections. An initial Navisworks report reveals thousands of "conflicts" between services (ducts, pipes) and the structure (beams, walls), as well as between the different services themselves.

Rather than tackling each clash individually, the coordinator applies the combined method:
1. Model Quality: They check that all objects are correctly classified (e.g., `IfcPipeSegment`, `IfcDuctSegment`) and that levels are aligned. They identify some misoriented services or duplicated elements which are corrected upstream.
2. Test Matrices: They prioritise tests: structure/HVAC, structure/plumbing, HVAC/electrical, plumbing/electrical. They exclude, for example, "façade/plumbing" tests, as these interfaces are less critical at this stage.
3. Tolerances: They set tolerances of 5 cm for services relative to structural elements (to allow space for insulation and fixings) and 2 cm between services (for spacing and maintenance). They automatically exclude legitimate contacts, such as a duct passing through a purpose-made penetration in a slab.
4. Grouping: They notice that many clashes between a ventilation duct and beams are on the same level and follow the same route. They group these 30 clashes into a single major "conflict point": "Main duct run on Level 3 too low". Similarly, they group repeated crossings of plumbing and electrical services in a plant room.
5. Visual Review: During the coordination meeting, the coordinator projects the federated model and presents the most impactful clash groups. The HVAC engineer, structural engineer, and architect discuss solutions live: rerouting the duct to a less constrained path, planning a lower suspended ceiling, or locally modifying a beam if structurally acceptable.
6. Dissemination of Decisions (BCF): For each resolved or outstanding issue, the coordinator creates an "issue" in BCF format, including a screenshot, a detailed comment on the decision, and assigns it to the HVAC engineer with a 3-day deadline to update their model. This information is then easily importable by the various stakeholders' software, allowing for precise tracking.

This approach transforms a daunting report into a manageable list of issues to address, ensuring efficient and collaborative coordination.

Best Practices / Common Mistakes

Best PracticesCommon Mistakes
Initial quality control of models.Running detection on unchecked models.
Defining targeted clash matrices.Testing all combinations indiscriminately.
Configuring realistic tolerances.Using a 0mm tolerance everywhere.
Excluding acceptable contacts.Treating every "clash" as a problem.
Grouping clashes by logic or zone.Analysing clashes one by one.
Arbitration in collaborative visual meetings.Sending raw clash reports to teams.
Using BCF for communication.Communicating by email with isolated screenshots.
Involving trade experts in arbitration.Leaving the BIM coordinator to arbitrate technical solutions alone.
Regularly updating resolved models.Not verifying that problems have been corrected.
Documenting decisions and non-resolutions.No record of arbitration and reasons for choices.

Step-by-Step Method

Here are the key steps to conduct effective BIM coordination:

1. Data Preparation (Models):
* Ensure all discipline models are up to date, correctly oriented, and positioned (shared geolocation).
* Check object quality (IFC classification, populated information). Use auditing tools like Solibri Model Checker or free online BIM analysis sites to validate the IFC structure before integration.
* Federate models in an aggregation and clash detection software (e.g., Autodesk Navisworks, Solibri Model Checker, Trimble Connect).
2. Clash Detection Configuration:
* Identify critical interfaces and define relevant clash matrices (e.g., Structure vs Services, Architecture vs Structure, etc.).
* Configure specific clash rules, including tolerances (minimum clearance) and exclusions (acceptable objects or contacts, such as predefined penetrations).
3. Detection Launch and Analysis:
* Execute clash detection tests.
* Analyse results by grouping similar clashes by location or main cause. Filter and sort to identify major issues.
4. Collaborative Review and Arbitration:
* Organise regular BIM coordination meetings with representatives from each discipline.
* Visually present clash groups on the federated model.
* Discuss and make collective decisions on proposed solutions.
* Document each decision: who is responsible for the modification, what is the agreed solution, and what is the deadline.
5. Resolution Communication and Monitoring:
* Use the BCF format to communicate identified issues, decisions made, and responsibilities assigned to design teams. These BCF issues can be directly imported into modelling software (Revit, ArchiCAD, etc.) to guide modifications.
* Ensure regular monitoring of resolution progress. Run new detections to verify that problems have been corrected and that no new ones have appeared.
6. Arbitration Documentation:
* Maintain a register of decisions made, reasons for arbitrations (including clashes intentionally unresolved because they were deemed acceptable or manageable on site). This documentation is essential for traceability and in case of dispute.

Points of Vigilance

  • Human involvement is key: The software only detects. The judgment and experience of professionals are essential for qualifying and resolving problems.
  • Model quality upstream: Poor model quality (incorrectly classified objects, invalid geometries, missing information) is the primary source of coordination failure. Invest in model verification from the outset.
  • Clear definition of roles: Who is responsible for modelling? Who is responsible for coordination? Who arbitrates clashes? These roles must be established at the beginning of the project, often in the BIM Execution Plan (BEP).
  • Non-compliance does not always mean a clash: A detection can indicate a deviation from a rule, but it does not always constitute a blocking problem. The project context is paramount.
  • Time management: Clash detection and coordination meetings require time. Plan them regularly and early enough in the project to allow teams time to make corrections.

Before any coordination, check for duplicates and the structure of your IFCs with the free BIM analysis tool on the website.

Key Takeaways

  • Visual coordination brings judgment, automated brings exhaustiveness.
  • Undiscriminating detection creates unnecessary noise.
  • Model quality, targeted tests, grouping, review, and BCF are the pillars of successful coordination.

Article based on the professional thesis "BIM transition and optimised deployment, applied in an engineering firm".

Mickael Quinart, PALLADION

clash detectioncoordinationsynthèseinterférences
Partager :