A constrained subgrid-scale model for passive scalar turbulence

In this paper, a constrained large-eddy simulation (C-LES) model is used to simulate passive scalar in turbulent flows. The coefficients of the dynamic mixed nonlinear model (MNLM) are calculated by imposing a physical constraint that is based on the scale-invariance of the scalar variance flux in the large-eddy simulation (LES) of passive scalar. The constrained subgrid-scale (SGS) model proposed in this work is compared with different SGS models for passive scalar transported by incompressible forced isotropic turbulence. In addition, the models are compared to filtered direct numerical simulation (fDNS) results. A priori test demonstrates that the C-LES model performed the best among the SGS models we considered in predicting the direction and magnitude of the SGS scalar flux. Moreover, the C-LES model reduces the variation of SGS scalar variance flux in the inertial-convective range and at the largest scales in the viscous-diffusive range. In the a posteriori test, compared with the traditional dynamic Smagorinsky (DSM) and MNLM models, the C-LES model is found to perform significantly better in predicting the evolution of the scalar spectrum and internal intermittency. These results indicate that imposing physical constraints in SGS models is a promising approach for developing advanced SGS models in LES of passive scalar turbulence.