The geometric analysis and parameterisation of array of obstacles undergoing high Reynolds number flows

城市环境中的空气污染具有巨大的经济和人力成本，每年在英国导致4万人死亡和200亿英镑的损失（RCP和RCPCH，2016）。气候变化和人口增长使这一问题更加严重。科学上的挑战是，在复杂的城市环境中，污染物的扩散是由多尺度湍流控制的。它需要创新的数学模型，以更好地为政策制定者提供信息并改善我们社会的福祉。该项目的目的是采用分形结构的集群来模拟城市景观，并使用新的数据驱动方法来研究城市流的独特属性。其结果可能对政策制定和城市规划产生潜在影响。其目的是采用分形结构的集群来模拟城市景观，并使用新的数据驱动的方法来研究城市流的独特属性。目标1：了解分形几何对污染物扩散的影响目标2：开发基于物理的数据驱动方法，以有效预测污染物扩散研究方法：1。采用格子玻尔兹曼方法（LBM）模拟污染物在分形之间的扩散，阐明其独特性2。使用机器学习技术识别关键物理过程，并使用分形几何和场参数来参数化污染物扩散。3.使用优化来找到最能减轻污染影响的几何图形。4.发展以LBM为基础的数据同化（DA）计划以预测污染物扩散5.交叉验证的方法与智利和法国的研究人员的实验数据。该项目符合，例如，与环境变化的生活主题，以及该主题下的各个研究领域，如基础设施和城市系统，人工智能技术，流体动力学，统计学和应用概率。

Air pollution in urban environments has huge economical and human costs, leading to 40k deaths and £20bn loss in UK annually (RCP and RCPCH, 2016). The problem is being worsen by climate change and population growth. The scientific challenge is that pollutant dispersion is governed by multi-scale turbulent flows in a complex urban environment. It calls for innovative mathematical models to better inform policy makers & improve the well-being of our society. The aim of the project is to employ clusters of fractal structures to model urban landscape and investigate the unique properties of urban flows using novel data-driven methods. The outcome can have have potential impacts on policy making and urban planning. The aim is to employ clusters of fractal structures to model urban landscape and investigate the unique properties of urban flows using novel data-driven methods. Objective 1: Understand the effects of fractal geometries on pollutant dispersionObjective 2: Develop physics-based data-driven methods for efficient prediction of pollutant dispersionResearch methodology:1. Employ the Lattice Boltzmann Method (LBM) to simulate pollutant dispersion between fractals, elucidate its unique features2. Use machine learning techniques to identify key physical processes and parametrize pollutant dispersion with the fractal geometry and field parameters. 3. Use optimisation to find geometries that best mitigate pollution effects. 4. Develop LBM-based data assimilation (DA) scheme to predict pollutant dispersion5. Cross validate the methods with experimental data from researchers in Chile and France. The project is aligned with, e.g., the living with environmental change theme, and various research areas under the theme, such as infrastructure and urban systems, artificial intelligence technologies, fluid dynamics, statistics and applied probability.