EPSRC-CBET:Turbulent flows over heterogeneous multiscale surfaces

在几乎所有的工程和环境流动中，湍流边界层（最接近给定表面的流动部分）都处于粗壁状态。典型的例子包括飞机和风力涡轮机叶片表面不规则的边界层、宏观生物污染的船体、森林或风力发电场的边缘、城市树冠、作物边界、河床和波涛汹涌的海面上的风。尽管经过数十年的持续研究，但仅根据表面的几何信息准确预测动量传递和/或表面摩擦阻力仍然很困难。这主要是因为在大多数情况下，表面粗糙度的地形是多尺度的，也就是说，它包含了各种各样的粗糙度长度尺度。此外，粗糙度长度尺度范围的变化和粗糙度特征的分布在整个表面上是不均匀的。目前的预测方法主要是针对均匀和单尺度粗糙度元素分布而设计的，既不能准确预测也不能深入了解多尺度非均质表面上的复杂物理流动。在这项合作研究中，将采用一种系统的方法来表征非均质多尺度表面上流动中的阻力和动量转移机制。这项研究将广泛适用于许多行业，在这些行业中，粗糙表面上的流动对性能至关重要。例如，在运输行业中，粗糙表面产生的阻力对运输效率及其环境足迹具有重要影响。这项研究对于理解和模拟与天气预报相关的大气流动也很重要。目前，大多数大气流动模型对复杂地形上的气流处理能力较差，需要改进预测模型。更好的预测模型对于了解城市地区和风力发电场的流量也很重要。在这个项目中，一系列高保真的计算机模拟——将在美国约翰霍普金斯大学进行——以及物理实验——将在英国南安普顿进行——将产生前所未有的在异质、多尺度表面上流动的数据。数值模拟将基于大涡模拟，该模拟使用高精度有限差分解算器实现的新型积分壁面模型，该解算器使用尖锐浸入边界法来求解更大规模的粗糙度元素。数值和实验将考虑三种不同的情况：(i)多尺度粗糙度性质的突变，（ii）多尺度粗糙度的有限斑块，以及（iii）多尺度粗糙度的重复变化。将对数据进行分析，并对模拟和实验进行比较。实验和数值数据以及获得的物理见解将用于测试现有的，并开发新的分析模型，这些模型仅基于多尺度非均质表面的地形信息，就能准确预测阻力和动量传递。该项目将加强研究生教育，因为参与该项目的博士生将获得计算方法，建模策略和国际实验合作方面的大量专业知识。这种培训将是非常宝贵的，因为这些方法被广泛认为是未来几十年将有大量增长的领域，在这些领域将最需要有经验的研究人员。

In almost all engineering and environmental flows, turbulent boundary layers (the part of the flow closest to a given surface) are in the rough-wall regime. Typical examples include boundary layers developing over surface irregularities on aircraft and wind turbine blades, macro bio-fouled ship hulls, edges of forests or wind-farms, urban canopies, crop boundaries, river-beds, and wind over rough seas. Despite decades of sustained research, accurate predictions of momentum transfers and/or skin-friction drag based on geometric information about the surface alone are difficult. This is primarily because in most cases, the topography of surface roughness is multi-scale, that is to say, it contains a wide variety of roughness length scales. Moreover, the variation in the range of roughness length scales and the distribution of the roughness features is heterogeneous across the surface. Current predictive approaches, designed mostly for homogeneous and single-scale roughness element distributions, can neither accurately predict nor offer insights into the complex physics of flow over multi-scale heterogeneous surfaces. In this collaborative research,a systematic approach to characterize drag and the mechanisms of momentum transfers in flows over heterogeneous multi-scale surfaces will be applied. This research will be broadly relevant to a large number of industries where flows over rough surfaces are critical for performance. In the transportation industry for example, the drag incurred by rough surfaces has important impact on transportation efficiency and its environmental footprint. This research is also important for understanding and modeling atmospheric flows, of relevance to weather prediction. The flows over complex terrain are currently poorly resolved in most atmospheric flow models and there is a need for improved predictive models. Better predictive models are also important for understanding flows in urban regions and wind farms.In this project, a series of high-fidelity computer simulations - to be carried out at Johns Hopkins in the US - and of physical experiments - to be performed at Southampton in the UK - will generate unprecedented data of flows over heterogeneous, multi-scale surfaces. Numerical modeling will be based on Large Eddy Simulation that uses a novel integral wall model implemented in a high-accuracy finite difference solver that uses sharp immersed boundary method to resolve larger-scale roughness elements. Three different cases will be considered both numerically and experimentally: (i) an abrupt change in nature of multi-scale roughness, (ii) finite patch of multi-scale roughness, and (iii) repeated changes in multi-scale roughness. The data will be analyzed and simulations and experiments compared. The experimental and numerical data as well as the physical insights obtained will be used to test existing, and develop new, analytical models that enable accurate prediction of drag and momentum transfers based only on available information about the topography of multi-scale heterogeneous surfaces. The project will strengthen graduate education, since the PhD student who will be a part of this project will gain substantial expertise in computational methods, modeling strategies and collaborating internationally with experimentalists. This training will be invaluable as these methodologies are widely recognized as areas of substantial growth in the coming decades, where experienced researchers will be most needed.