采用基于火焰面的燃烧模型研究部分预混燃烧

Study of Partially-Premixed Combustion Using Combustion Models Based on Flamelet Concept

关键词:  火焰面思想  反应进度变量  部分预混  燃烧模型  三岔火焰

Key words:  Flamelet concept  Progress variable  Partially-premixed  Combustion model  Triple flame

摘要:

考虑到火焰面模型的优点，采用基于RANS的稳态火焰面模型，稳态火焰面/反应进度变量模型和非稳态火焰面/反应进度变量模型对部分预混燃烧室进行了数值计算，并将这三种燃烧模型的计算结果和实验值进行对比研究。发现稳态和非稳态火焰面/反应进度变量模型均成功地预测到了部分预混燃烧中的三岔火焰结构和火焰抬举现象，分析了三岔火焰的形成机理及抬举高度。文中稳态火焰面模型计算部分预混燃烧完全失效，无法捕捉到火焰抬举现象，稳态火焰面/反应进度变量方法计算的火焰抬举高度仅仅为10，基于反应进度变量为水、二氧化碳、一氧化碳和氢气的质量分数之和的非稳态火焰面/反应进度变量方法计算的火焰抬举高度仅为20，和实验值之间的误差为42.8%，而基于反应进度变量为一氧化碳和二氧化碳质量分数之和的非稳态火焰面/反应进度变量方法计算的火焰抬举高度大致32，和实验测量值35非常接近，误差仅为8%。对燃烧热力学标量温度和组分的计算，可以发现非稳态火焰面/反应进度变量方法和实验结果吻合最好，其次是稳态火焰面/反应进度变量模型，最差的是稳态火焰面模型。

Abstract:

Considering advantage of flamelet model，steady flamelet model，steady flamele/progress variable approach and unsteady flamelet/progress variable approach based on RANS were used to simulate partially-premixed combustor，and comparing the results of the three combustion models with experiment result. It is indicated steady flamelet/progress variable approach and unsteady flamelet/progress variable approach both captured triple flame and lifted flame phenomenon in partially-premixed combustion，analyzing mechanism of triple flame and life-height. Steady flamelet model is failed for partially-premixed combustion and can not capture lifted flame at all，the lift-off height using steady flamelet/progress variable approach is only 10，the lift-off height is 20 using unsteady flamelet/progress variable approach in definition of progress variable using sum of mass fraction of water，carbon dioxide，carbon moNO_xide and hydrogen，error is 42.8%，and the lift-off height is 32 using unsteady flamelet/progress variable approach in definition of progress variable using sum of mass fraction of carbon dioxide，carbon moNO_xide，it is closely to experiment which is 35，error is only 8%. According to thermochemical quantities of temperature and species mass fraction，it is indicated that results of unsteady flamelet/progress variable approach are most close to experiments，second is steady flamelet/progress variable approach，and the worse is steady flamelet model.