The incorporation of physical environmental performance as a primary consideration in building design can facilitate the harmonization of the built environment with the surrounding site and climate, enhance the building’s environmental adaptability and environmental friendliness, and contribute to the achievement of energy-saving and emission-reduction objectives through the integration of natural lighting and ventilation. General computational fluid dynamics (CFD) can help architects make accurate predictions and effectively control the building’s wind environment. However, CFD integration into the design workflow in the preliminary stages is frequently challenging due to program uncertainty, intricate parameter settings, and substantial computational expenses. This study offers a methodology and framework based on machine learning to overcome the complexity and computational cost barriers in simulating outdoor wind environments of buildings. In this framework, the machine learning model is trained using an automated CFD simulation system based on Butterfly and implemented within the Rhino and Grasshopper environment. This framework provides real-time simulation feedback within the design software and exhibits promising accuracy, with a Structural Similarity Index Measure (SSIM) ranging from 90–97% on a training dataset of 1200 unique urban geometries in Xinjiekou Area of Nanjing, China. Furthermore, we programmatically integrate various parts of the simulation and computation process to automate multiscenario CFD simulations and computations. This automation saves a significant amount of time in producing machine-learning training sets. Finally, we demonstrate the effectiveness and accuracy of the proposed working framework in the design process through a case study. Although our approach cannot replace CFD simulation computation in the later design stages, it can support architects in making design decisions in the preliminary stages with minimal effort and immediate performance feedback.

The wind pressure on the surface of buildings can be critical, especially for high-rise buildings. To overcome the limitations in applying time-consuming CFD simulations to early design stage, we introduce a novel approach for generating real-time surface wind pressure in high-rise buildings through Deep Learning (DL). The DL model is trained on datasets of hundreds of buildings geometries generated by Grasshopper plugin EvoMass, and their corresponding surface wind pressure is simulated by RhinoCFD. Moreover, regarding the current performance prediction using DL, we have noted that the common labeling map and DL model may hinder the efficiency of training. Thus, we proposed a new labeling approach and a modified DL model to assess the accuracy of the prediction. To demonstrate the validity of the approach, the experiments of simulation predictions were conducted, and the results shows that the presented method can greatly enhance the accuracy.