现有标量小尺度混合模型由于没有考虑化学反应对反应标量混合的影响，导致概率密度函数方法不能准确预测湍流预混燃烧中的局部熄火/再燃和NOx/CO生成。反应标量小尺度混合和化学反应交互作用机理是发展适用于湍流预混燃烧的混合模型亟需研究的关键科学问题。本项目从直接数值模拟数据出发，运用包含湍流扩散的低维火焰理论和方法，揭示小尺度湍流对火焰结构和反应标量耗散率影响机制，建立反应标量小尺度混合、湍流混合和化学反应之间的内在关联，发展一个适用于各个预混燃烧状态的混合模型，提升概率密度函数方法对预混燃烧的预测精度。同时完善详细化学反应动力学高效计算的理论和方法，与大涡模拟/概率密度函数方法结合，解决最具挑战的三个模拟难点：湍流、湍流/化学反应相互作用和详细化学反应动力学，实现高效、高精度的湍流预混燃烧模拟，并用于揭示局部熄火/再燃及污染物生成的物理机制，为发展超低排放的预混燃烧技术提供理论和模拟方法。

Currently Probability Density Function (PDF) methods can not accurately predict local extinction/re-ignition and NOx/CO formation in turbulent premixed combustion due to the fact that current scalar micro-mixing models do not account for the effect of chemical reactions on reactive scalar mixing. Further investigation of the interaction between scalar micro-mixing and chemical reactions is the key to the development of mixing models applicable for turbulent premixed combustion. This project employs direct numerical simulation (DNS) data and the method of low-dimensional flames with turbulent diffusion to investigate the effect of the small-scale turbulence on flame structure and reactive scalar dissipation rates; to build the underlying connection between scalar micro-mixing, turbulent mixing and chemical reactions; finally to develop a scalar micro-mixing model applicable to all the regimes of turbulent premixed combustion and therefore significantly improve the predictability of PDF methods for premixed combustion. Furthermore acceleration methods for detailed chemical kinetics are investigated and then integrated into large eddy simulation/probability density function (LES/PDF) to form the high-fidelity simulation approach that can address the three most challenging modelling difficulties: turbulence, turbulence-chemistry interaction, and detailed chemical kinetics. LES/PDF simulations with detailed chemistry are performed to investigate the local-extinction/re-ignition and emission formation in turbulent premixed flames and therefore to gain physical insights on the development of premixed combustion technologies with ultra-low emission.