Integrated BIM–Machine Learning Models for Green and Healthy Building Performance at the Conceptual Design Stage

Integrated BIM–Machine Learning Models for Green and Healthy Building Performance at the Conceptual Design Stage

Authors

  • Mustika Sari Department of Civil and Environmental Engineering, Faculty of Engineering, Universitas Indonesia, Kampus Baru UI Depok 16424, Indonesia; Center for Sustainable Infrastructure Development, Faculty of Engineering, Universitas Indonesia, Kampus Baru UI Depok 16424, Indonesia https://orcid.org/0000-0003-3941-051X
  • Mohammed Ali Berawi Department of Civil and Environmental Engineering, Faculty of Engineering, Universitas Indonesia, Kampus Baru UI Depok 16424, Indonesia; Center for Sustainable Infrastructure Development, Faculty of Engineering, Universitas Indonesia, Kampus Baru UI Depok 16424, Indonesia https://orcid.org/0000-0002-4781-7494
  • Teuku Yuri Zagloel Center for Sustainable Infrastructure Development, Faculty of Engineering, Universitas Indonesia, Kampus Baru UI Depok 16424, Indonesia; Department of Industrial Engineering, Faculty of Engineering, Universitas Indonesia, Kampus Baru UI Depok 16424, Indonesia

DOI:

https://doi.org/10.37965/jait.2026.1762

Keywords:

Building Information Modeling, green building, healthy building, machine learning

Abstract

Green building (GB) and healthy building (HB) performance are commonly assessed through separate tools, while conventional simulation requires detailed inputs that may be unavailable during conceptual design. This study developed BIMLY, a Building Information Modeling (BIM) and machine learning (ML) integration architecture and interface prototype for early GB and HB performance assessment. The research comprised three stages: identification of predictive variables, development of GB and HB prediction models, and design of a BIM-based integration architecture. The study identified 12 GB indicators and 14 HB indicators, including 12 synthetically generated HB outputs. The GB model combined a feedforward artificial neural network for three water-consumption indicators with deterministic IF-ELSE rules for the remaining nine. The HB model used a multilayer perceptron and achieved a mean absolute error of 3.02. The proposed architecture connects an Autodesk Revit plug-in to both models through a Python REST API, with JSON used for data exchange. BIMLY provides an integration design and interface prototype for GB and HB during conceptual design.

Downloads

Published

2026-11-19

How to Cite

Sari, M., Berawi, M. A., & Zagloel, T. Y. (2026). Integrated BIM–Machine Learning Models for Green and Healthy Building Performance at the Conceptual Design Stage. Journal of Artificial Intelligence and Technology. https://doi.org/10.37965/jait.2026.1762

Issue

Section

Research Articles

Similar Articles

You may also start an advanced similarity search for this article.

Loading...