Heatwaves have a significant impact on urban areas, driving efforts to mitigate the urban heat island (UHI) effect through green infrastructure and sustainable planning. By integrating computational fluid dynamics (CFD) with digital twin technology, this study evaluates the effectiveness of climate adaptation infrastructures in urban areas. However, applying digital twin technology for UHI analysis and integrating data into actionable insights faces challenges due to long simulation times and focus of analysis. This study aimed to mitigate the societal impacts of urban heat islands and address the gaps in existing research and technology. A new machine learning model was developed to improve the urban thermal environment by optimizing green spaces and combating urban heat islands in densely populated cities, by integrating artificial intelligence (AI) and digital twin technology. Combining the strengths of Random Forest and XGBoost, the model was trained and tested on a dataset derived from CFD simulations to identify effective strategies for urban green spaces allocation. The primary results of the study are divided into three parts. First, a high-precision model for data augmentation and green space optimization was developed using machine learning. Second, the developed model reduced the time required for CFD simulation analysis from over 400,000 h to less than 1 h. Finally, the study found that the strategic placement of green spaces could result in approximately 1 % of the total urban area temperature. The results highlight the importance of strategic planning in the distribution of urban green space for effective mitigation of heat islands. The proposed model can be used as an efficient tool for sustainable urban development and is consistent with the overall goal of creating more livable and climate-resilient cities.

中文翻译：

---

使用基于自然的解决方案开发人工智能模型，用于 CFD 数据增强和改善城市地区的热环境


热浪对城市地区有重大影响，促使人们努力通过绿色基础设施和可持续规划来减轻城市热岛效应 （UHI）。通过将计算流体动力学 （CFD） 与数字孪生技术相结合，本研究评估了城市地区气候适应基础设施的有效性。然而，由于仿真时间长且分析重点突出，将数字孪生技术应用于 UHI 分析并将数据集成到可操作的见解中面临挑战。本研究旨在减轻城市热岛效应的社会影响，并解决现有研究和技术中的差距。开发了一种新的机器学习模型，通过集成人工智能 （AI） 和数字孪生技术，优化绿色空间和对抗人口稠密城市的城市热岛效应，从而改善城市热环境。结合 Random Forest 和 XGBoost 的优势，该模型在源自 CFD 模拟的数据集上进行了训练和测试，以确定城市绿地分配的有效策略。该研究的主要结果分为三个部分。首先，使用机器学习开发了一个用于数据增强和绿色空间优化的高精度模型。其次，开发的模型将 CFD 仿真分析所需的时间从超过 400,000 小时缩短到不到 1 小时。最后，研究发现，绿地的战略布局可能导致城市区域总温度的 1% 左右。结果强调了战略规划在城市绿地分配中对有效缓解热岛效应的重要性。 拟议的模型可以用作可持续城市发展的有效工具，并且与创建更宜居和气候适应能力更强的城市的总体目标一致。