A Fully Automated DM-BIM-BEM Pipeline Enabling Graph-Based Intelligence, Interoperability, and Performance-Driven Early Design

TL;DR

The paper tackles the challenge of applying AI to early-stage building design by bridging unstructured boundary-representation geometry with graph-based BIM and energy models. It introduces a fully automated, reversible DM→BIM→BEM pipeline with modular components for face classification, geometry cleansing, 1LSB/2LSB space topology generation, and ontology-aligned BIM-to-BEM transformations. Validation across diverse datasets demonstrates strong topology accuracy ($nGED\approx0.87$ in CCR), high energy-model fidelity (RMSE ~ $0.5\ \mathrm{kWh/m^2}$ and $R^2$ up to $0.987$), and robust workflow performance with rapid analysis times (0.29 s for full transformation) within the MOOSAS environment. The framework enables AI-driven design exploration, surrogate modeling, and LLM-based reasoning on concept-to-performance data, significantly expanding the applicability of KG/NG-based intelligence in early design and interconnected construction workflows. Open-source availability and ontology-aligned interoperability position this work as a foundational automation layer for next-generation digital construction and performance-driven design.

Abstract

Artificial intelligence in construction increasingly depends on structured representations such as Building Information Models and knowledge graphs, yet early-stage building designs are predominantly created as flexible boundary-representation (B-rep) models that lack explicit spatial, semantic, and performance structure. This paper presents a robust, fully automated framework that transforms unstructured B-rep geometry into knowledge-graph-based Building Information Models and further into executable Building Energy Models. The framework enables artificial intelligence to explicitly interpret building elements, spatial topology, and their associated thermal and performance attributes. It integrates automated geometry cleansing, multiple auto space-generation strategies, graph-based extraction of space and element topology, ontology-aligned knowledge modeling, and reversible transformation between ontology-based BIM and EnergyPlus energy models. Validation on parametric, sketch-based, and real-world building datasets demonstrates high robustness, consistent topological reconstruction, and reliable performance-model generation. By bridging design models, BIM, and BEM, the framework provides an AI-oriented infrastructure that extends BIM- and graph-based intelligence pipelines to flexible early-stage design geometry, enabling performance-driven design exploration and optimization by learning-based methods.

Figures (14)

• Figure 1: Potential applications with this framework.
• Figure 2: Modules in the transformation framework. Each module is a stand-alone method and allows substitution following the API documentation on the project
• Figure 3: visualization of 6 cleansing methods
• Figure 4: ASG methods description. a) The view factors to link elements, each element has 2 surfaces; b) Split the connection network on elements and calculate the envelope; c) The topology and space relations to identify by BTG; d) The fusiform structure of two spaces; e) The recursively division logic of the CCR method; f) Boolean and location of the ceilings and grounds.
• Figure 5: space network as a graph composed by SPACE and VOID