The Development of Unstructured Mesh Technology for Viscous High Speed Flows

航空航天器的传统设计涉及广泛使用风洞来测试不同的配置并最终完成设计。然而，这是一个昂贵而漫长的过程，还需要使用为特定流速设计的专业测试设施。随着计算机的出现，在过去的20年里出现了一种新的技术，它为辅助空气动力学设计提供了一个强大的工具。支配空气运动的方程式早在几个世纪前就已为人所知。然而，对于一般的流，它们的解并不服从于经典的数学求解技术。随着高性能计算机的出现，出现了一种新的技术，称为计算模拟，或者更广泛地说，科学模拟，它基于在计算机上求解这些复杂的方程。模拟气流所涉及的基本概念很简单。对控制气流的方程中的未知数进行了近似，将几个高度复杂的方程转化为数百万个简单的方程。然后，使用计算机通过算法方法来求解这些方程。在现实中，飞机周围的区域被细分成几个小单元，每个单元内的流动变量以某种适当且一致的形式进行近似。这种细分空间的过程称为网格生成。求解方程并进而产生未知流动变量(如压力、密度等)的算法称为求解算法，这些算法的构造是为了确保从通常具有许多处理器的高性能计算机获得最大效率。然后使用计算机图形处理计算结果，并提取重要的定量数据，如升力和阻力。这项技术现在所有主要的航空航天公司都在常规使用。这项技术在不冗余使用风洞的同时，使设计者能够探索新的和创新的设计，并确保更少的几何图形需要进行昂贵的风洞分析。尽管高速流动的计算机模拟的基本概念很简单，但准确预测关键空气动力学参数的要求代表着巨大的技术和智力挑战。精确地表示飞机的几何形状需要创新的方法来表示三维表面，而生成飞机周围的元素使求解算法能够捕捉所有复杂的物理仍然是一个挑战。虽然流体流动的方程可以精确地写出来，但可用计算能力的限制，即使考虑到世界上最大的计算机的能力，也需要研究人员进行近似计算，就像湍流的模拟一样。在某些情况下，这些近似不能捕获流的细节，因此预测不能准确地代表现实。该项目的目标是致力于进一步的技术开发，以提高复杂气动外形(如完整的飞机外形)的高速流动的精度，同时确保计算可以在满足设计中遇到的真实项目截止日期的时间尺度内进行。特别是，该项目将专注于增强我们准确预测空气动力学参数(如升力和阻力)的能力，以及模拟飞机在起飞和降落时产生的高度复杂的流场，在这种情况下，地面影响可能会很大。当这些开发完成后，计算机预测将与实际测试数据进行比较，以确保这些技术得到适当的验证。

Traditional design of aerospace vehicles has involved the extensive use of wind tunnels to test different configurations and to finalise design. However, this is an expensive and lengthy process that also requires the use of specialist test facilities designed for particular flow speeds. With the advent of the computer a new technology has emerged over the last 20 years that provides a powerful tool to aid aerodynamic design. The equations that govern the movement of air have been known for several centuries. However, for general flows, their solution is not amenable to classical mathematical solution techniques. With the advent of high performance computers, a new technology, termed computational simulation or, more generally, scientific simulation, that is based upon solving these complicated equations on the computer, has emerged. The basic concepts involved in simulating airflow are straightforward. Approximations to the unknowns in the equations that govern airflow are made that transforms the few highly complicated equations into millions of simple equations. The computer is then used to solve these equations using an algorithmic approach. In reality, the region around an aircraft is subdivided into small elements and within each element the flow variables are approximated in some appropriate and consistent form. This process of subdividing the space is termed mesh generation. The algorithms that solve the equations and in turn produce the unknown flow variables (such as pressure, density etc) are called the solution algorithms and these are structured to ensure that maximum efficiency can be obtained from high performance computers that will, in general, have many processors. The results of the calculations are then processed using computer graphics and important quantitative data such as lift and drag can be extracted.This technology is now used routinely in all major aerospace companies. Whilst not making the use of the wind tunnel redundant, the technology has enabled designers to explore new and innovative designs and ensure that fewer geometries need to be subjected to costly wind tunnel analysis.Whilst the basic concepts of computer simulation for high speed flows are simple, the requirement to predict accurately key aerodynamic parameters represents a significant technical and intellectual challenge. Representing the geometry of an aircraft accurately demands innovative ways of representing three-dimensional surfaces and the generation of the elements around the aircraft that will enable the solution algorithm to capture all the complex physics still remains a challenge. Whilst the equations of fluid flow can be written exactly, the restrictions in available computing power, even taking into account the capabilities of the World's largest computers, require researchers to make approximations, as is the case for the simulation of turbulent flow. For some cases, these approximations do not enable the details of the flow to be captured and hence the predictions do not accurately represent reality. This project is aimed at focusing on further technical developments that will increase the accuracy of high speed flows for complicated aerodynamic shapes, such as complete aircraft configurations, whilst ensuring that the computations can be performed in a time scale that meets real-world project deadlines encountered in design. In particular, the project will focus on enhancing our capability to predict aerodynamic parameters accurately, such as lift and drag, and to simulate highly complicated flowfields generated when an aircraft is in take-off and landing configuration where ground effects can be significant. When these developments have been completed, computer predictions will be compared with real test data to ensure appropriate validation of the techniques.