CONDITIONAL MOMENT CLOSURE MODEL FOR SIMULATION OF TURBULENT DIFFUSION COMBUSTION AND NITRIC OXIDES FORMATION
条件矩模型模拟湍流扩散燃烧

Modeling turbulent combustion is a formidable task, since there is a strong interaction between turbulent mixing and chemical reaction, which results in the difficulties to express the averaged reaction rates in terms of the averaged values of the scalars. Most of presently developed engineering models in the commercial CFD codes are highly simplified and cannot accurately simulate chemical reactions in turbulent flows, such as NOX formation in turbulent flows. The PDF transport equation method can exactly simulate the chemistry-turbulence interaction, but computationally is very expensive. The conditional moment closure model (CMC) proposed by Klimenko and Bilger makes the reaction source term closed at the first conditional moment level, since the conditional averaging in terms of the mixture fraction suppresses fluctuations of most interesting scalar quantities (composition, temperature, enthalpy, etc.), which means that the fluctuations about the conditional average are so small that the covariance can be neglected. The CMC method is directly derived from the conservation equations for the mass fraction of species treated in a purely stochastic manner without any simplifications. Therefore, the CMC model has a rather sound basis and can be applicable to turbulent reacting flows in which species formation time scales are of the same order as that of local mixing times.A primary study of the simulation of turbulent diffusion combustion with conditional moment closure (CMC) model is presened in this paper. Experimental results show that the conditional averages have a very weak dependence on the across stream position in a steady axisymmetric jet flame. And Klimenko also proves it by asymptotic analysis. This weak dependence allows us to simplify the CMC equations by integrating them across the flow. The comparison of simulation results with experimental data given by Sandia laboratories, USA, shows that the performances of the prediction of temperature and main species concentrations are satisfactory, while the pre-dictions of NO are in agreement with measurement in tendency. The error analysis indicates that the improvement of computational accuracy of scalar dissipation rate as well as the employment of the second-order CMC model will be useful to the development of CMC model.