城市化的不断推进导致一系列的“城市病”，特别是热岛和空气污染对城市可持续 发展带来严峻挑战。通风对空气污染物扩散和缓解城市热岛效应起着至关重要的作用， 已有研究表明建筑形态可显著影响通风效率。然而，街区尺度建筑形态与通风内在关 系仍不十分清楚。因此，深入探索建筑形态与风场间的关系，对城市规划和环境治理 具有一定理论与实践价值。前人研究表明，计算流体力学（Computational Fluid Dynamics，CFD）是研究复杂城市风环境的有效方法之一。然而，目前 CFD 研究中的 湍流模型选择以及数值模拟结果验证仍存在一定问题。尤其是特定区域的研究结果缺 少必要的泛化性评价，限制了其在客观场景的应用和规划策略的落地。 基于此，本研究综合利用 CFD、地理信息系统（Geographic Information System， GIS）和随机森林（Random Forest，RF），在街区尺度上研究建筑形态对城市通风的影 响，旨在为城市街区规划设计提供科学依据。首先，利用遥感（Remote sensing，RS） 和 GIS 提取建筑形态参数，并通过野外实测获取试验区内风速数据。第二，使用野外 实测数据进行CFD 模型选择以及模拟结果精度验证。此外，对CFD模型进行计算域尺 寸和网格无关性检验，以确定适合本研究的合理的计算域尺寸和网格大小，同时利用 已有的风洞实验结果进一步验证Standard k-ε模型的可行性。第三，使用CFD模拟试验 区内的风场分布并分析建筑形态对风场的影响。第四，使用统计分析和RF重要性排序， 明确建筑形态参数与风速比之间的关系，筛选出对风场影响较大的建筑形态参数，探 究建筑形态对风场影响结论的科学性。最后，根据建筑形态与风场之间的关系，提出 街区规划设计策略。本文的主要结论如下： （1）试验区内建筑形态具有显著的异质性，试验区北侧建筑密度、建筑高度和容 积率较低，街区内开敞。试验区南侧建筑密度、容积率与建筑高度较高、街区对外封 闭。 （2）本文证明CFD可用于城市街区尺度风环境模拟，且其在分析复杂建筑形态风 场方面具有一定可行性。Standard k-ε 模型的 CFD 模拟与野外实测数据拟合较好， Standard k-ε 模型是本研究中模拟风场的最优方法，可有效预测街区尺度的风场。计算 域大小和网格无关性检验结果表明，本研究设置的网格大小（0.8m），以及计算域大小 （5Hmax×5Hmax×6Hmax）可以满足需要。 （3）在城市街区尺度上，建筑形态显著影响城市通风。其中，围合度、建筑密度 和容积率是三个显著影响城市行人高度风速的重要建筑形态参数。此外，来流方向对 气流扰动具有显著影响，进一步影响建筑群内部的风场变化。建筑形态对风场的影响 基于通用的空气动力学原理，包括风流模式、压力分布和湍流效应。通过控制实验， 本研究的结论在一定程度上得到了验证, 研究结果可为类似城市街区尺度风环境研究提 供一定参考。 （4）通过调控围合度、建筑密度和容积率，能够显著改善行人高度风环境。建议 在街区规划和设计中，通过增加广场等开放空间，减少建筑密度，增加建筑间的间距 以降低围合度、建筑密度和容积率，为建筑之间提供更多的风道，为改善城市风环境 提供一定借鉴。

The continuous advancement of urbanization has led to a series of "urban diseases", especially heat island and air pollution, which have brought serious challenges to the sustainable development of cities. Ventilation plays an important role in diffusing air pollutants and alleviating urban heat island effect. Previous studies have shown that architecture morphology can significantly affect ventilation efficiency. However, the internal relationship between block scale architecture morphology and ventilation is still unclear. Therefore, the in-depth exploration of the relationship between architecture morphology and wind field has certain theoretical and practical value for urban planning and environmental governance. Previous studies have shown that Computational Fluid Dynamics (CFD) is one of the effective methods to study the complex urban wind environment. However, there are still some problems in the selection of turbulence model and the verification of numerical simulation results in CFD research. In particular, the research results in specific areas lack the necessary generalization evaluation, which limits its application in objective scenarios and the landing of planning strategies. Based on this, this study comprehensively utilizes CFD, Geographic Information System (GIS) and Random Forest (RF) to study the impact of architecture morphology on urban ventilation at the block scale. The aim is to provide scientific basis for urban block planning and design. Firstly, Remote sensing (RS) and GIS were used to extract architecture morphology parameters. In addition, wind speed data in the test area were obtained through field measurement. Secondly, the field measurement data were used to select the CFD model and verify the accuracy of the simulation results. In addition, the computational domain size and grid independence of the CFD model were tested to determine the reasonable computational domain size and grid size suitable for this study. At the same time, the feasibility of the Standard k-ε model was further verified by the existing wind tunnel experimental results. Thirdly, CFD was used to simulate the wind field distribution in the test area and analyze the influence of architecture morphology on the wind field. Fourthly, statistical analysis and RF importance ranking were used to clarify the relationship between architecture morphology parameters and wind speed ratio, identifying the key architecture morphology parameters that significantly influence the wind field and exploring the scientific validity of the conclusions on the impact of architecture morphology on the wind field. Finally, according to the relationship between architecture morphology and wind field, the block planning and design strategy was proposed. The main conclusions of this paper are as follows: (1) The architecture morphology in the test area has significant heterogeneity. The building density, building height and floor area ratio in the north of the test area are low, and the blocks are open. The south block of the test area is compact, the floor area ratio and building height are high, and the block is closed to the outside world. (2) This paper proves that CFD can be used to simulate urban block scale wind environment, and it has certain feasibility in analyzing wind field of complex architecture morphology. The CFD simulation of Standard k-ε model fits well with the field measurement data, and the Standard k-ε model is the best method to simulate the wind field in this study, which can effectively predict the wind field at the block scale. The results of calculation domain size and grid independence test show that the grid size (0.8m) and calculation domain size (5Hmax×5Hmax×6Hmax) set in this study can meet the needs. (3) At the urban block scale, architecture morphology significantly affects urban ventilation. Among them, the Degree of Enclosure (DE), Building Coverage Ratio (BCR), and Floor Area Ratio (FAR) are three important architecture morphology parameters that significantly affect the wind speed of urban pedestrian height. In addition, the direction of incoming flow has a significant effect on airflow disturbance, which further affects the variation of wind field inside the building clusters. The effect of architecture morphology on the wind field is based on general aerodynamic principles, including airflow patterns, pressure distribution and turbulence effects. Through the control experiment, the conclusion of this study has been verified to a certain extent, and the research results can provide a certain reference for similar urban block scale wind environment research. (4) The wind environment of pedestrian height can be significantly improved by adjusting DE, BCR, and FAR. It is suggested that in the planning and design of blocks, building density should be reduced by increasing the open space such as squares, and the spacing between buildings should be increased to reduce DE, BCR, and FAR, so as to provide more air passages between buildings and provide some reference for improving the urban wind environment.