AI for early-stage zoning and massing compliance: a lightweight prototype for the Bangladesh National Building Code (BNBC 2020)

Adhikary, Gautam and ZamaniGoldeh, Erfan 2026. AI for early-stage zoning and massing compliance: a lightweight prototype for the Bangladesh National Building Code (BNBC 2020). Presented at: The 31st International Conference on CAADRIA, Hsinchu, Taiwan, 26 April - 2 May 2026. Published in: Makki, Mohammed, Chen, Jielin, Nik Chien, Sheng-Fen, Liu, Jie, Wang, Likai and Wang, Shih-Yuan eds. Proceedings of the 31st International Conference of the Association for Computer-Aided Archetectural Design Research in Asia (CAADRIA) 2026. , vol.1 Hong Kong: Association for Computer-Aided Archetectural Design Research in Asia (CAADRIA), pp. 51-60.

Full text not available from this repository.

Abstract

This research presents an AI-assisted framework for early stage building code compliance using the Bangladesh National Building Code (BNBC 2020) as a case study. The system integrates Large Language Models with a deterministic rule engine inside Rhino Grasshopper to provide real-time feedback on zoning and massing. It bridges unstructured code text and geometric modelling through a transparent, lightweight workflow. Validation showed high reliability, with semantic queries achieving around 90% F1 accuracy and deterministic checks reaching full precision. The approach demonstrates a scalable, human-centred method to improve design efficiency and regulatory literacy in low-BIM, resource-limited contexts. The study specifically addresses the research question of how to reconcile textual building codes with fluid geometric exploration before fundamental design decisions are permanently locked in. By providing a platform-agnostic and modular architecture, the framework enables architects to navigate complex regulatory landscapes without the technical overhead of high-fidelity BIM environments, significantly reducing late-stage rework.

Item Type:	Conference or Workshop Item - published (Paper)
Status:	Published
Schools:	Schools > Architecture
Publisher:	Association for Computer-Aided Archetectural Design Research in Asia (CAADRIA)
ISBN:	978-988-78918-8-8
Related URLs:	Organisation
Last Modified:	14 Sep 2026 10:29
URI:	https://orca.cardiff.ac.uk/id/eprint/189550

Actions (repository staff only)

Edit Item