Keywords
How to Cite
Abstract
This study elucidates a methodology for autonomously reconstructing the 3D spatial geometry of buildings from limited 2D static images to execute advanced climatic simulations. By integrating deep learning with computational fluid dynamics, solar radiation, wind aerodynamics, and energy consumption were analyzed. Empirical observations confirmed that AI-formulated topological data possess high validity in predicting microclimates (r = 0.94). The conceptual findings provide an automated express-diagnostic platform for optimizing energy consumption in urban planning.
References
1. Mildenhall B, Srinivasan PP, Tancik M. NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis. Communications of the ACM. 2021;65(1):99-106.
2. Gao J, Chen W, Xiang T. Integrating 3D Deep Learning with CFD for Urban Microclimate Assessment. Building and Environment. 2023;228:109785.
3. Wang P, Liu Y, Chen Z. NeuS: Learning Neural Implicit Surfaces for 3D Aerodynamic Analysis. Advances in Neural Information Processing Systems. 2022;34:1120-1135.
4. Chen X, Xu L, Yang Y. Deep Learning for 3D Reconstruction: A Survey. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2022;44(12):9889-9907.
5. Al-Sallami H, Rahmat M. Predicting Building Energy Load using Deep Learning on Architectural Facades. Energy and Buildings. 2021;245:111020.
6. Qulmamatov OS. Raqamli shaharsozlikda sun'iy intellekt texnologiyalarini qo'llash istiqbollari. Axborot texnologiyalari va innovatsiyalar jurnali. 2024;7(4):21-29.
7. Blocken B. Computational Fluid Dynamics in Urban Physics: Applications and Challenges. Journal of Wind Engineering and Industrial Aerodynamics. 2018;181:226-258.