Urban-Scale Meshing for City Energy Simulation

From urban data to multi-level watertight meshes, ready for building energy simulation at city scale and beyond.

Identity Card

Authors

  • Christophe Prud’homme (Unistra–Cemosis)

  • Vincent Chabannes (Unistra–Cemosis)

  • Javier Cladellas (Unistra–Cemosis)

  • Pierre Alliez (INRIA Sophia Antipolis) + Geometry Factory

Collaboration

  • CoE HiDALGO2

  • ExaMA/NumPEx (UNISTRA, INRIA-CA)

  • Ktirio & CGAL projects

Date: September 2025

Context and Objective

Urban energy, wind, and heat modeling requires closed and coherent geometries, while input data (buildings, roads, terrain, hydrology, vegetation) are heterogeneous and often non-watertight.

Objective: Reconstruct watertight urban meshes at LoD-0/1, interoperable and enriched with physical model attributes for simulation.

Key Result and Innovation

Development of a complete GIS → Mesh processing pipeline:

Pipeline Components

  • Ingestion: Mapbox data integration

  • Robust Operations: Ktirio-Geom with CGAL

  • Boolean Operations: Multi-layer processing

  • Topological Control: Closure verification

Performance Gains

  • Multi-km² scenes → closed LoD-1/2 solids

  • Buildings, roads, terrain, rivers, vegetation, roofs

  • Preparation: weeks → minutes

  • Numerical stability improvements

Outputs compatible with: Urban Building Model (Ktirio-UBM) and energy/CFD solvers.

Grenoble City Reconstruction

Reconstruction of the city of Grenoble over a 10 km radius with roads, rivers, water bodies, buildings with roofs, and topography — with a single click.

Ktirio Interface
Figure 1. Ktirio application interface for urban data selection
Grenoble Overview
Figure 2. Bird’s eye view of Grenoble 10km mesh
Grenoble City Center
Figure 3. City center with buildings, roads and vegetation
Grenoble Mountains
Figure 4. Mountain terrain with urban area
Grenoble Detailed View
Figure 5. Detailed view: buildings and infrastructure
Grenoble River Detail
Figure 6. Close-up: river and urban structures

Construction of a watertight (conformal) mesh ready for HPC (EuroHPC) for building energy simulation.

Impact and Next Steps

Current Impact

  • Quick access to urban use cases faithful to energy/microclimate modeling

  • Dataset sharing in CoE HiDALGO2 ecosystem

  • Tools for urban heat island analysis

Next Steps

  • Publication of reference datasets (watertight + scripts) via CKAN

  • CFD & energy simulation coupling

  • Extended urban heat island studies

Valorization

Demonstrations

Deployment

  • Cloud deployment of graphical application (coming soon)

  • Valorization via SATT Conectus

Related Work Packages

WP1

Discretization — Mesh quality and adaptivity

WP7

Showroom & Benchmarking — Software deployment and integration

← Neural Operators | Back to Results | HPC Benchmarking →