Direct and large Eddy simulation of turbulent premixed combustion

提出的研究是利用大涡和直接数值模拟（LES和DNS）来研究湍流预混燃烧。预混燃烧发生在燃料和氧化剂在燃烧之前混合的地方，然后燃烧的特征是局部变化的湍流火焰速度，这使得系统建模变得困难。它的应用范围从用于发电的固定式燃气轮机到火花点火集成电路发动机。它在精益预混燃烧中尤其重要，这是一种减少排放的新方法，但它可能导致燃烧不稳定，损坏发动机或导致停机。火焰的局部行为和整个燃烧过程都受湍流-火焰相互作用的细节控制，因此需要更好地理解和建模这些相互作用的能力。这里的总体目标是：1。从本质上更好地理解预混火焰-湍流相互作用；最终开发一个能够模拟燃气轮机燃烧室内燃烧的LES代码，DNS和LES将在本工作中使用。DNS完全解决了反应流的控制方程，不需要建模。它包括湍流的所有尺度，因此计算非常密集——提议的研究包括使用升级的Sharcnet设备来观察SI发动机和大型燃气轮机中常见的简单火焰的部分。DNS运行的结果不仅将用于观察火焰-湍流的相互作用，而且还将用于开发大涡模拟的模型。LES解决了流动中的大部分湍流；然而，与DNS不同的是，它不计算湍流的最小尺度，这些小尺度的“亚网格尺度”特征必须建模。由此产生的LES能够同时考虑高雷诺数和真实燃烧室的复杂几何形状，同时仍然包括真实的火焰湍流行为。

The proposed research is on the use of Large Eddy and Direct Numerical Simulations (LES and DNS) to study turbulent premixed combustion.  Premixed combustion occurs where the fuel and oxidiser are mixed before they are burnt, and the combustion is then characterised by a locally varying turbulent flame speed which makes modelling the system difficult.  It is used in applications ranging from stationary gas turbines used for power generation to spark ignition IC engines.  It is particularly important in lean premixed combustion, a new approach to reducing emissions, but one that can lead to combution instabilities, damaging engines or causing shutdowns.  Both the local behaviour of the flames and the overall combustion process are controlled by the details of the turbulence-flame interactions, so a better understanding and modelling capability of these interactions are required.  The overall objectives here are to:   1. gain an inherently better understanding of premixed flame-turbulence interactions, and   2. ultimately develop an LES code capable of simulating the combustion within gas turbine combustors Both DNS and LES will be used in this work.  DNS solves the governing equations of the reacting flow in their entirety with no modelling required. It includes all of the scales of the turbulence and is consequently very computationally intensive -the proposed research involves using the upgraded Sharcnet facilities to look at parts of simple flames in regimes commonly seen in SI engines and large-scale gas turbines.  The results of the DNS runs will be used, not only to look at flame-turbulence interactions, but also to help develop models for the large eddy simulations.  LES resolves most of the turbulence in the flow; however, unlike DNS, it does not calculate the smallest scales of the turbulence and these small-scale "subgrid scale" features must then be modelled.  The resulting LES is then able to consider both the high Reynolds numbers and the complex geometries of real combustors while still including the real flame-turbulence behaviour.