Development of Efficient and Clean Turbulent Stratied Mixture Combustion Technologies for Future Combustors using High-Fidelity Simulations

根据美国能源部的数据，化石燃料的燃烧占世界能源使用量的80%，他们的预测显示，到2050年，这一比例将降至68%。燃烧产生的排放导致了气候变化和人口密集城市的空气质量差。因此，污染控制法规变得越来越严格，这给工程师们提出了一个重大挑战，即设计出一种具有可接受的发动机效率的环保型内燃机。在一些工程应用中，燃烧发生在燃料和氧化剂在燃烧前既不均匀混合也不完全分离的配置中。这种燃烧通常被称为分层装药燃烧或部分预混燃烧。这种不均匀的混合使得整体燃料空气比更低，这有可能显著减少污染物排放，同时提高发动机效率。该技术目前被用于现代车辆的直喷和均质增压点火发动机，以及飞机上的预混增压燃气涡轮发动机以及许多其他应用。在大多数情况下，这类引擎仍处于起步阶段。燃烧行为的不确定性，当与预混和非预混的对应物进行比较时，阻碍了它们的广泛采用。以汽车发动机为例，不适当的混合气制备会导致压力过大引起的预燃，或者由于火焰没有消耗全部反应物而导致发动机爆震。这两种情况都大大降低了发动机的效率。因此，为了实现可靠、清洁、高效的分层混合气燃烧，需要进一步研究合理的混合气制备方法。本课题将在国家级高性能计算设备上对紊流层状混合气燃烧进行直接数值模拟。与预混合和非预混合情况相比，分层混合燃烧的研究数量很少。直接数值模拟完全解决了湍流反应流动，而不需要湍流模型（即近似）。模拟数据可以作为实验数据处理，具有很高的分辨率。作为该项目的第一步，三维直接数值模拟将用简单的化学方法进行，以提高对分层装药燃烧过程的热方面的基本理解。简化的化学减少了要解决的方程的数量，从而使模拟时间大大缩短。在此基础上，将进行详细的化学三维模拟。这些模拟将为湍流混合的不完全理解的物理提供重要的物理见解，并将进一步了解清洁，高效燃烧的最佳混合物和湍流特性。由于这些模拟的计算成本很高，无法用于常规的工程计算。因此，从该模拟数据库中获得的物理洞察力将被转移到基于Reynolds平均Navier-Stokes和大涡模拟框架的工业求解器的精确湍流模型中。这项工作的主要受益者是从事开发设计低污染、高效率汽车发动机和燃气轮机的新概念的工业部门。这些应用中的设计过程取决于工程计算的预测能力。

According to the U.S. Department of Energy, the combustion of fossil fuels accounts for 80% of the worlds energy usage, and their projections show this gure reducing to only 68% in 2050. The emissions resulting from combustion lead to climate change and poor air quality in densely populated cities. Consequently, pollution control regulations are becoming ever stricter, posing a major challenge for engineers to design an environmentally friendly combustion engine withacceptable engine efficiency.In some engineering applications, combustion takes place in a conguration where the fuel and oxidiser are neither homogeneously mixed nor completely separated from each other prior to combustion. This kind of combustion is often referred to as stratified charge combustion or partially premixed combustion. This inhomogeneous mixture allows a leaner overall fuel-air ratio to be used, which has the potential to signicantly reduce pollutant emissions while increasingengine efficiency. The technology is currently being used in direct injection and homogeneous charge compression ignition engines in modern vehicles, and in lean premixed prevapourised gas turbine engines in aircraft as well as many other applications. Such engines are still in their infancy in most cases. Uncertainties regarding combustion behaviour, when compared with premixed and non-premixed counterparts, is impeding their widespread adoption. Taking automotive engines as an example, improper mixture preparation and can lead to pre-ignition caused by excessive pressure, or engine knocking due to the flame not consuming all of the available reactants. Both of these drastically reduce engine efficiency. Thus, further research is required into proper mixture preparation for reliable, clean and efficient stratified mixture combustion.In this research project, Direct Numerical Simulations of turbulent stratied mixture combustion will be performed on national high-performance computing facilities. The number of such studies of stratified mixture combustion is scarce when compared with premixed and non-premixed cases. Direct Numerical Simulations resolve the turbulent reacting flow entirely, without the need for turbulence models (i.e. approximations). The simulation data can be treated as experimental data with a very high resolution. As the first step of this project, three-dimensional Direct Numerical Simulations will be performed with simplied chemistry to improve the fundamental understanding of the thermal aspect of the stratified charge combustion process. Simplified chemistry reduces the number of equations to be solved resulting in much quicker simulation times. Based on this understanding, detailed chemistry 3D simulations will be carried out. These simulations will provide important physical insight into the not-well-understood physics of turbulent mixing and will further knowledge regarding optimal mixture and turbulence properties for clean, efficient combustion. As these simulations are of such high computational cost, they cannot be afforded for routine engineering calculations. Therefore, the physical insight obtained from this simulation database will be transferred into accurate turbulence models for industrial solvers based on the Reynolds Averaged Navier-Stokes and Large Eddy Simulation frameworks.The major beneciaries of this work are the industrial sectors engaged in developing new concepts for designing low-pollution high-efficiency automotive engines and gas turbines. The design process in these applications depends on the predictive capability of engineering calculations.