Version control for asynchronous BIM collaboration: Model merging through graph analysis and transformation

Sebastian Esser, Simon Vilgertshofer, André Borrmann

Research output: Contribution to journalArticlepeer-review

Abstract

The design process in construction projects is iterative and multi-disciplinary in nature. In today's industry practice, several discipline experts concurrently author multiple versions and design variants of BIM models and share them at frequent intervals. Applying a sound version control methodology can significantly enhance automation, enabling the coordination and combination of these model versions into consistent overall models with less extensive manual effort. This paper introduces a diff-and-patch mechanism for transferring changes between model versions, facilitating object-level change tracking using graph representations of BIM model data, and specifically focuses on merging diverging versions through the application of graph transformations. The mechanisms for executing branching and merging of model versions are thoroughly explained and showcased through various illustrative scenarios. The presented method adheres to the established principles of federated BIM collaboration but equips the participating parties with additional means to automate the combination of various model versions, allowing them to focus on the relevant conflicts. The proposed methodology of graph-based version control unlocks the potential of analyzing interdisciplinary dependencies across partial models and enables the more efficient resolution of conflicting model versions.

Original languageEnglish
Article number105063
JournalAutomation in Construction
Volume155
DOIs
StatePublished - Nov 2023

Keywords

  • BIM Level 3
  • Common Data Environment
  • Graph transformation
  • Merging
  • Model Coordination
  • Model synthesis
  • Patching
  • Time-evolving BIM models

Fingerprint

Dive into the research topics of 'Version control for asynchronous BIM collaboration: Model merging through graph analysis and transformation'. Together they form a unique fingerprint.

Cite this