(142l) Inverse Approach for Turbulent Models Using Optimized Lattice Boltzmann Method
Turbulent flow plays an important role in many areas of applied science and engineering. The simulation of turbulent flow is a challenging subject, and it is still remained as an unsolved problem due to the difficulties associated with accurately analyzing chaotic fluid motion at high Reynolds number. Here, we utilize the lattice Boltzmann method (LBM)  as an alternative methodology in resolving issues in computational fluid dynamics (CFD) where the governing equations can be derived through the Chapman-Enskog multi-scale expansion via the Boltzmann transport equation (BTE). LBM, a descritized form of BTE, is a powerful tool for fluid flow simulations as well as multiscale modeling methodology due its advantages of algorithmic simplicity, transient nature, and ease in the treatment of complex geometrical boundaries, in addition to its amenability to parallel computing platforms. Furthermore, the methodology of LBM is “rule based,” and one may be able to find appropriate physical models based on experimental findings as shown in this study via what is termed an inverse approach.
We developed an extensive single-phase flow analysis tool based on LBM with non-reflecting boundary conditions to avoid spurious disturbances at the boundaries with the Smargorinsky large eddy simulation (LES) model for turbulent flows . Our model describes most of the essential characteristics of the complex flows without losing the critical competitive edge of computational resource efficiency. In this study, we exploit the abovementioned rule-based characteristics to obtain an accurate LBM model, by minimizing the disparity between our numerical results and the experimental results of high Reynolds number flows for various geometries (square cylinder with rounded and square corners at various orientations towards the incoming flow) by a novel inverse modeling approach. This will result in an accurate performance comparison of a broad range of turbulence models via an “optimized” set of multiple relaxation time parameters, as well as the capability to adjust the model automatically when the experimental data is inconsistent with simulation. This preliminary test case will help us to observe the robustness of LBM, unlike conventional CFD approaches which have no easy implementation for this scenario. Furthermore, we will extend our model towards complicated flow systems consisting of multi-physical phenomena, which will lead to deeper insight into the area of turbulence simulation..
 S. Chen and G.D. Doolen, Annu. Rev. Fluid Mech., 30, 329 (1998).
 D. Kim, H. M. Kim, M. S. Jhon, S. J. Vinay III, and J. Buchanan, Chin. Phys. Lett., 25, 1964, (2008)