Automated MEP coordination and routing with multi-model clash avoidance

Part of the Bonsai10D Vision - 10-dimensional construction coordination integrating 3D geometry, time, cost, compliance, and ERP systems.

This module enhances BlenderBIM's clash detection capabilities for large-scale MEP coordination projects. Following IfcOpenShell best practices, it provides bbox-based pre-broadphase filtering to optimize performance when working with multiple discipline models.

Design Philosophy: Enhance existing BlenderBIM tools, don't replace them.

• ✅ Automated conduit routing with obstacle avoidance
• ✅ Multi-discipline clash detection (90K+ elements)
• ✅ IFC4 cable carrier segment generation
• ✅ Federation-based spatial queries (<100ms)
• ✅ Discipline-grouped clash reporting
• 🚧 BCF export for Navisworks (Phase 2C - planned)

Based on community feedback from Dion Moult (IfcOpenShell maintainer), this module implements a three-component architecture:

┌─────────────────────────────────────────┐
│ 1. Preprocessing (One-time)            │
│    IfcPatch Recipe: ExtractSpatialIndex│
│    → Creates bbox database              │
└─────────────────────────────────────────┘
                ↓
┌─────────────────────────────────────────┐
│ 2. Runtime Optimization                 │
│    Federation Module: Spatial Index     │
│    → 95% memory reduction               │
│    → 20x faster queries                 │
└─────────────────────────────────────────┘
                ↓
┌─────────────────────────────────────────┐
│ 3. User Interface                       │
│    MEP Engineering: Routing + Validation│
│    → Blender integration                │
└─────────────────────────────────────────┘

Performance: Validated on Terminal 1/2 project (7 disciplines, 93K elements, 302MB):

• Memory: 93% reduction (30GB → 2GB)
• Time: 20x faster (10 min → 30 sec)
• Accuracy: 100% (no false negatives)

1. Blender 4.2+ with Bonsai addon installed
2. Python dependencies:

   # In Blender's Python environment
   /path/to/blender/python -m pip install rtree

This module depends on the Federation module for spatial indexing and multi-model queries.

cd src/bonsai/bonsai/bim/module/

# Install federation module (dependency)
git clone https://github.com/red1oon/federation.git

# Install MEP engineering module
git clone https://github.com/red1oon/mep_engineering.git

Add to src/bonsai/bonsai/bim/__init__.py:

modules = {
    # ... existing modules ...
    "federation": None,        # ← Add this
    "mep_engineering": None,   # ← Add this
}

Restart Blender to load the modules.

Option A: Via Bonsai UI (Recommended for most users)

1. Open Blender with any IFC project
2. Go to Properties → Scene → Quality Control tab
3. Expand "Multi-Model Federation" panel
4. Click "Add File" for each discipline IFC file
5. Set Database Path: /path/to/project_federation.db
6. Click "Preprocess Federation" button
7. Wait ~7 minutes for completion (progress shown in console)

Option B: Standalone CLI (For automation/scripting)

# Extract bounding boxes from discipline IFC files
cd /path/to/project

python federation_preprocessor.py \
  --files ARC.ifc ACMV.ifc STR.ifc ELEC.ifc \
  --output project_federation.db \
  --disciplines ARC ACMV FP SP STR ELEC

See Federation module documentation for detailed CLI usage.

Output: SQLite database with spatial index (~50MB for 90K elements)

1. Open Blender with your IFC project
2. Go to Properties → Scene → Quality Control tab
3. Expand "Multi-Model Federation" panel
4. Click "Load Federation Index"

✅ Status: "Federation Active - 93,000 elements from 4 disciplines"

1. Go to Properties → Scene → Quality Control tab
2. Expand "MEP Engineering" panel
3. Set routing points:
   • Position 3D cursor at start → Click "Set from Cursor"
   • Position 3D cursor at end → Click "Set from Cursor"
4. Configure settings:
   • Clearance Distance: 0.5m (500mm default)
   • Conduit Diameter: 0.05m (50mm default)
   • Obstacle Disciplines: STR,ACMV,ARC,FP,SP,CW
5. Click "Route Conduit"

Result: IFC4 IfcCableCarrierSegment elements created, avoiding obstacles

1. After routing, click "Validate Route" in MEP panel

2. Console output:

   ═══════════════════════════════════════════════════
   CLASH VALIDATION RESULTS
   ═══════════════════════════════════════════════════
   Total clashes: 67
   
   By responsible discipline:
     ARC: 31 clashes
     CW: 16 clashes
     FP: 11 clashes
     ACMV: 8 clashes
     STR: 1 clashes
   

3. Review grouped results by discipline

4. (Phase 2C) Export to BCF for Navisworks coordination

• Obstacle Detection: Uses federation spatial queries to identify conflicts
• Pathfinding: Orthogonal routing (Phase 1: direct paths, Phase 2: A* planned)
• IFC Generation: Creates schema-aware elements (IFC2X3 fallback supported)
• Multi-discipline: Queries across Architecture, Structural, MEP, Fire Protection

• Conservative Filtering: Bbox-based prefilter eliminates 95% of candidates
• Discipline Grouping: Results organized by responsible discipline
• False Positive Reduction: Sanitizes clashes based on clearance rules
• Console Reporting: Detailed clash summary with top conflicts

• Memory Efficient: 2GB vs 30GB (traditional approach)
• Fast Queries: <100ms per corridor query (vs 10 min full geometry load)
• Scalable: Tested on 93K element federations

• Standalone bbox preprocessing script
• Spatial index with R-tree queries
• BlenderBIM module integration

• Start/end point selection
• Direct path routing
• Obstacle detection via federation

• IFC4 cable carrier segments
• System assignment
• Property sets (dimensions, type)

• Federation-based clash detection
• Discipline filtering
• Console reporting

• Export clashes to BCF format
• Navisworks/BIM360 integration
• Viewpoint snapshots

• A* pathfinding algorithm
• Multi-path optimization
• Cost-aware routing

• Submit IfcPatch recipe to IfcOpenShell
• Integrate bbox prefilter with IfcClash
• Community-maintained enhancement

This module follows BlenderBIM best practices:

✅ Reuses Community Modules:

• Federation module for spatial queries
• IfcOpenShell API for IFC manipulation
• Bonsai UI patterns for consistency

✅ No Reinvention:

• Leverages existing clash detection (IfcClash)
• Uses standard property groups and operators
• Follows naming conventions

✅ Clean Architecture:

• Separation of concerns (preprocessing → runtime → UI)
• Minimal dependencies (rtree only)
• Backward compatible

• Strategic Guide: See MEP Coordination: Strategic & Technical Guide
• Validation Checklist: See Federation Module Validation
• OSArch Forum: Community discussion
• GitHub Issues: Report bugs or request features

• Federation Module - Multi-model spatial indexing (required dependency)
• Bonsai10D - 10D construction coordination vision
• IfcOpenShell - IFC toolkit and library
• BlenderBIM - OpenBIM Blender add-on

Validated on real-world project:

• Project: Terminal 1/2 Jetty Complex Expansion
• Scale: 7 disciplines, 93,000 elements, 302MB
• Results: 67 clashes detected correctly, grouped by discipline
• Performance: 30 seconds, 2GB RAM

Run validation tests:

# From Blender console
import bpy
props = bpy.context.scene.BIMmepEngineeringProperties

# Set test route
props.route_start_point = (-50428, 34202, 6)
props.route_end_point = (-50434, 34210, 6)

# Run routing
bpy.ops.bim.route_mep_conduit()

# Validate clashes
bpy.ops.bim.validate_conduit_route()

GPL-3.0-or-later

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

See LICENSE for full details.

Redhuan D. Oon (red1) - Lead Developer
Naquib Danial Oon - Contributor

We welcome contributions! This project aims to upstream enhancements to IfcOpenShell/Bonsai:

1. Fork the repository
2. Create a feature branch
3. Follow BlenderBIM code standards
4. Submit a pull request

For major changes, please open an issue first to discuss.

• Dion Moult - IfcOpenShell maintainer, architectural guidance
• OSArch Community - Feedback and support
• BlenderBIM Team - Foundation and ecosystem

Status: Active Development | Version: Phase 2B Complete (v0.2b)
Last Updated: 2025-01-14