Studies On Sparse-Lagrangian Simulation Based On Mmc Model

Turbulent combustion widely exists in gas turbines and engines. Since the fuel consumption and the pollutants emission are mostly controlled by combustion, the issues on combustion are still very important in the energy and environment research. Combustion in the real combustors mainly appears in the form of turbulent combustion. Therefore, it is very important to study how to control turbulent combustion.Numerical simulation becomes one of the main approaches to investigate turbulent combustion in recent years.The LES/PDF method combines LES with PDF method to simulate turbulent combustion using LES to solve the velocity field PDF method to solve the scalar field. The LES/PDF method have advantages of both the LES and PDF, and that made it to be research focus in recent years.Large eddy simulation (LES), which can directly solve the large scale turbulence combining with the the subgrid scale models to take into account the turbulent small scales, attracts increasing attention. However, the chemical source terms in the filtered equations in LES still need to be modeled by appropriate combustion model. The method has been widely adopted because its cost is far less than the DNS method while its accuracy is much higher than the RANS method.PDF method solves the probability density function(PDF) transport equations to obtain the one-point,one-time statistical information. To the turbulent combustion, the chemical source terms are represented exactly in PDF transport equations, i.e. these terms no longer need to be modeled. So PDF method is a relatively generalized turbulent combustion model, and it can simulate any detailed chemical kinetics processes in theory.However, the conventional PDF method employs an intensive-Lagrangian simulation, which is high-cost and not easily used. Sparse-Lagrangian simulation based on Multiple Mapping Conditioning (MMC) can dramatically decrease the computational cost for PDF method. However, there isn’t a appropriate method to obtain the characteristic scales of reference space in MMC model.Considering the high sensitivity of MMC in the shear layer with large scalar gradient, and the effect of particle number, a generalized model of sparse-Lagrangian MMC is developed to model the characteristic scales of reference space in MMC model. Large eddy simulations of several round jet cases are performed using the improved model. The results predicted by the improved model have the same accuracy as those in the previous study which used the approach of the specific constant characteristic value in the reference space. Meanwhile, the production and variation of stochastic noise of RMS statistics is also investigated in the sparse particle field. The results indicate that the stochastic noise depends on the access method from the scalar value of particles to the corresponding LES grid. Weaker noise is produced on the condition of selection of more particles.