CFD Modelling of the acoustic response of sprays

在可预见的未来，燃烧化石燃料将是能源生产和运输部门，特别是航空部门的关键。减轻这种情况对环境的影响是一个关键的优先事项。为了满足对NOx排放的重要且日益严格的法规限制，“稀燃”燃烧器被视为至关重要的一步。贫燃燃烧器在燃烧区使用更多的空气来降低火焰温度和污染物的形成。这样做的缺点是火焰的“热声”稳定性降低。放热的波动导致压力波，压力波影响进入燃烧室的燃料和空气的流动，这又会影响放热。这种双向耦合可能导致反馈回路，从而导致大的、破坏性的压力振荡。为了成功地开发稀燃技术，必须控制这种热声不稳定性。数值模拟工具在这方面发挥着至关重要的作用。如果设计的稳定性或其他方面可以在计算模拟中预测，而不需要测试物理原型，那么可以节省大量的资源。精确的数值模拟也可以比单独的实验更深入地了解所涉及的物理过程。大多数燃气轮机和所有的航空发动机都使用液体燃料。燃料被分解成喷雾，喷雾与旋流混合并蒸发成旋流。在这种类型的燃烧器中的热声行为的准确模拟的主要障碍之一是如何模拟声压力波对喷雾的形成和传输的影响。目前，必须使用假设和经验关系来定义随时间变化的空气和燃料入口条件，这些假设和经验关系的适用性值得怀疑。特别是，它们没有考虑气流对压力波的空气动力学响应中的延迟或“相移”，这对于捕获燃烧系统的真实不稳定行为至关重要。该建议的目的是克服这一障碍，通过开发适当验证的计算方法来模拟复杂的旋流喷雾在声波的影响下，这种方法的验证需要访问适当的实验数据。拉夫堡大学（Loughborough University，LU）最近已经制作了这样的数据。一个典型的稀燃燃料雾化器，其中燃料喷射之间的两个共环形旋流空气流，已被测试与声学强迫。数据可用于随时间变化的燃料液滴的大小，数量和速度在几个位置的流动。这些数据将以以下方式使用：1.将开发用于一系列声激励频率的精确喷雾和空气入口条件的方法。这些将通过与靠近喷射器的实验数据进行比较以及与当前推导喷雾入口条件的简化方法进行比较来进行测试。将系统地研究与所有方法中的假设相关的误差。通过使用LU开发的可压缩计算流体动力学方法，将改进对喷油器单独气流通道中压力波的气流响应预测。2.将使用提案第一部分中开发的入口条件，对存在声强迫的喷雾传输过程进行模拟。这些模拟，验证了进一步下游采取的实验数据，将是一个彻底的测试的入口条件和建模方法。它们还将提供对所涉及的流动物理学的更深入的理解。这将包括湍流混合和声学强迫对不同尺寸的液滴的影响以及喷射器中不同气流的相对重要性。这种理解，加上改进的边界条件，将是未来设计低排放燃烧室的重要价值。

For the foreseeable future burning fossil fuels will be essential in the energy generation and transport sectors, particularly in aviation. Mitigating the environmental impacts of this is a key priority. In order to meet important, and increasingly stringent, regulatory limits on NOx emissions "lean burn" combustors are seen as a vital step. Lean burn combustors use more air in the combustion zone to reduce the flame temperature and the formation of pollutants. The downside to this is that the "thermoacoustic" stability of the flame is reduced. Fluctuations in heat release lead to pressure waves which affect the flow of fuel and air into the combustor which in turn can affect the heat release. This two-way coupling can lead to a feedback loop causing large, and damaging, pressure oscillations. For lean burn technology to be successfully developed this thermoacoustic instability must be controlled. Numerical simulation tools have a vital part to play in this. If the stability or otherwise of a design can be predicted in a computational simulation, without the need for testing a physical prototype then a great deal of resources can be saved. Accurate numerical simulation can also give a deeper understanding of the physical processes involved than experiments alone. Most gas turbines and all aero engines use liquid fuel. The fuel is broken up into a spray which mixes with and evaporates into a swirling air flow. One of the major hurdles standing in the way of accurate simulation of thermoacoustic behaviour in burners of this type is how to model the effect of acoustic pressure waves on the formation and transport of the spray. Currently assumptions and empirical relations have to be used to define the time varying air and fuel inlet conditions and these are of questionable applicability. In particular they take no account of the delay or "phase shift" in the aerodynamic response of the airflow to a pressure wave, which is vital for capturing the true unsteady behaviour of a combustion system. This proposal aims to overcome this hurdle by developing properly validated computational methods for simulating sprays in complex swirling flows under the influence of acoustic waves.Validation of this method requires access to appropriate experimental data. Such data has recently been produced at Loughborough University (LU). A typical lean burn fuel atomizer, in which fuel is injected between two co-annular swirling air streams, has been tested with acoustic forcing. Data is available for the time varying fuel droplet size, number and velocity at several locations in the flow. This data will be used in the following ways:1. Methods of deriving accurate spray and air inlet conditions for a range of acoustic excitation frequencies will be developed. These will be tested by comparison with the experimental data from close to the injector and also compared to the current simplified method of deriving the spray inlet conditions. The errors associated with the assumptions in all methods will be studied systematically. The prediction of air flow response to a pressure wave in the separate airflow passages of the fuel injector will be improved by the use of a compressible CFD method developed at LU. 2. Simulations of the spray transport process in the presence of acoustic forcing will be carried out using the inlet conditions developed in the first part of the proposal. These simulations, validated against experimental data taken further downstream, will be a thorough test of the inlet conditions and of the modelling methods employed. They will also provide a deeper understanding of the flow physics involved. This will include the effects of turbulent mixing and acoustic forcing on droplets of different sizes and the relative importance of the different airflow streams in the injector. This understanding, coupled with improved boundary conditions will be of great value in designing low emission combustors of the future.