Challenging modeling strategies for LES of non-adiabatic turbulent stratified combustion
Challenging modeling strategies for LES of non-adiabatic turbulent stratified combustion
复制标题
DOI：
10.1016/j.combustflame.2015.07.036
复制
发表时间：
2015-11-01
影响因子：
4.4
通讯作者：
Kempf, A.
中科院分区：
文献类型：
--
作者：
Fiorina, B.;Mercier, R.;Kempf, A.
五种不同的低马赫数大涡模拟进行了比较，在达姆施塔特技术大学（TUD）进行的湍流分层火焰实验。模拟由TUD、亚琛的燃烧技术研究所（ITV）、隆德大学（LUND）、巴黎中央理工学院的EM 2C实验室和杜伊斯堡-埃森大学（UDE）提供。燃烧模型由预混小火焰制表与局部火焰增厚（TUD），预混小火焰进展变量的方法耦合到水平集方法（ITV），4步机制结合隐式LES（隆德），F-TACLES模型，是基于过滤预混小火焰制表（EM 2C），和火焰表面密度的方法（UDE）。一个广泛的比较模拟和实验数据的前两个时刻的速度，温度，混合物分数，和主要物种的质量分数。热损失的重要性进行了评估，通过比较模拟绝热和等温边界条件在燃烧器壁。绝热计算预测锚定在燃烧器唇上的火焰，而非绝热模拟显示出一半的飞行员直径的火焰升空和温度和物种浓度的实验证据更好的协议。大多数模拟同意的平均火焰刷的位置，但很明显，亚网格湍流必须考虑到实现正确的湍流火焰速度。火焰的瞬时快照的定性比较表明，在解决的火焰燃烧模式的大小的差异。这些差异是由LES燃烧模型对火焰动力学的影响和网格、入口条件和数值模拟策略的不同造成的。模拟是用针对不同目标优化的方法进行的，例如低计算成本，或者在另一种情况下，短周转时间。(C)2015燃烧研究所爱思唯尔公司出版All rights reserved.
Five different low-Mach large eddy simulations are compared to the turbulent stratified flame experiments conducted at the Technical University of Darmstadt (TUD). The simulations were contributed by TUD, the Institute for Combustion Technology (ITV) at Aachen, Lund University (LUND), the EM2C laboratory at Ecole Centrale Paris, and the University of Duisburg-Essen (UDE). Combustion is modeled by a premixed flamelet tabulation with local flame thickening (TUD), a premixed flamelet progress variable approach coupled to a level set method (ITV), a 4-steps mechanism combined with implicit LES (LUND), the F-TACLES model that is based on filtered premixed flamelet tabulation (EM2C), and a flame surface density approach (UDE). An extensive comparison of simulation and experimental data is presented for the first two moments of velocity, temperature, mixture fraction, and major species mass fractions. The importance of heat-losses was assessed by comparing simulations for adiabatic and isothermal boundary conditions at the burner walls. The adiabatic computations predict a flame anchored on the burner lip, while the non-adiabatic simulations show a flame lift-off of one half pilot diameter and a better agreement with experimental evidence for temperature and species concentrations. Most simulations agree on the mean flame brush position, but it is evident that subgrid turbulence must be considered to achieve the correct turbulent flame speed. Qualitative comparisons of instantaneous snapshots of the flame show differences in the size of the resolved flame wrinkling patterns. These differences are (a) caused by the influence of the LES combustion model on the flame dynamics and (b) by the different simulation strategies in terms of grid, inlet condition and numerics. The simulations were conducted with approaches optimized for different objectives, for example low computational cost, or in another case, short turn around. (C) 2015 The Combustion Institute. Published by Elsevier Inc. All rights reserved.