Scale-resolving Simulations of Multicomponent Nozzle Flows

该项目详细研究了空气中氢气的喷射和混合过程，因为它们发生在直接喷射系统中。由于进口压力过高，形成欠膨胀超声速流动，激波系统与湍流之间发生复杂的相互作用。这影响了混合位置和物种分布；同样，热力学参数的显著影响是可以预期的。流入或喷嘴几何形状的不确定性导致了复杂的非线性行为。这不仅影响燃烧，而且可能是间歇性循环变化的因果链的一部分。将这些混合过程的特征描述为已确定参数的函数是本子项目的目标。为此，进行了低膨胀氢注入的高分辨率大涡模拟。在空间和时间上的能量携带尺度的分辨率是捕捉多尺度多物理场效应和间歇现象的绝对必要的。在这里，湍流区域和激波系统的共存对数值学提出了巨大的挑战。此外，即使在高性能计算机上，发生尺度的差异（例如，喷嘴喉部和直径之间存在超过3个数量级）也无法用全局恒定网格间距的计算网格表示。因此，本文采用最先进的具有自适应能力的高阶方法，使局部网格分辨率适应发生的尺度。这些方法在本项目中扩展到多组件流。利用这些方法，可以以前所未有的精度解决喷嘴流动以及随后的混合，并与实验结果进行比较。在此基础上，结合数值和实验合作项目，进一步研究了热力学参数（状态方程、扩散法）的影响。得到的高分辨率流场以及被动平流示踪粒子的粒子轨迹将与燃烧模拟代码耦合，同时考虑到并行化要求。一方面，这将关闭从喷嘴到燃烧的模拟链，另一方面，它将允许对发生的变化进行联合调查（通过相关分析和示踪剂）。采用多级和多保真度方法，根据喷嘴流量的变化来验证结果。总体而言，该项目应有助于可靠和高度精确地绘制氢气注入和混合的地图，从而有助于预测周期波动并更好地了解其原因。

The project investigates in detail the injection and mixing processes of hydrogen in air, as they occur in direct injection systems. Due to the high inlet pressure, an underexpanded supersonic flow is established, in which a complex interaction of shock systems and turbulence occurs. This influences the mixing location and species distribution; likewise, a significant effect of thermodynamic parameters is to be expected. Uncertainties in inflow or nozzle geometry contribute to the complex, nonlinear behavior. This not only affects combustion but may also be part of the causal chain of intermittent cyclic variations. The characterization of these mixing processes as a function of the identified parameters is the goal of this subproject. For this purpose, high-resolution large eddy simulations of underexpanded hydrogen injections are performed. The resolution of the energy-carrying scales in space and time is absolutely necessary to capture the multiscale-multiphysics effects as well as the intermittent phenomena. Here, the coexistence of turbulent regions and shock systems pose great challenges to numerics. Additionally, the disparity of the occurring scales (e.g., more than 3 orders of magnitude lie between nozzle throat and diameter) cannot be represented with computational grids with globally constant grid spacing, even on high-performance computers. Therefore, state-of-the-art high-order methods with adaptation capabilities are used here, which can adapt the local grid resolution to the occurring scales. These methods are extended in this project for multi-component flows. With these methods, the nozzle flow as well as the subsequent mixing can be resolved with unprecedented accuracy and compared to the experimental results. Based on this methodology, the influence of the thermodynamic parameters (equation of state, diffusion approach) is further investigated in cooperation with numerical and experimental partner projects. The resulting high-resolution flow fields as well as the particle trajectories of passively advected tracer particles will be coupled to the combustion simulation code, taking into account the parallelization requirements. On the one hand, this will close the simulation chain from the nozzle to the combustion, and on the other hand, it will allow a joint investigation of the occurring variations (by correlation analysis as well as tracers). Multi-level and multi-fidelity methods are used to validate the results against changes in the nozzle flow. Overall, this project should contribute to a reliable and highly accurate mapping of hydrogen injection and mixing and thus to a contribution to the prediction of cyclic fluctuations and to a better understanding of their causes.