 |
Center for Nonlinear Studies |
Pulsatile flow, which is commonly present in the human body, is characterized by complex cyclic transitions between laminar and turbulent states. Most existing large eddy simulations (LES) are designed for fully developed turbulence, making them unsuitable for pulsatile flow. To address this gap, we propose a new LES approach that self-regulates between laminar and turbulent modeling. This presentation will focus on the development of sub-grid scale (SGS) model within the proposed LES framework. We have derived a new kinetic-based discrete dynamical system (DDS) from the volumetric lattice Boltzmann method (VLBM). Through extensive numerical experiments, this DDS has proven capable of capturing a range of turbulent flow behaviors, such as noisy subharmonic n-period, noisy quasi-periodic, and broadband, among others. The validity of the DDS is further confirmed by comparing the predicted velocity time series from DDS with results from direct numerical simulations, which demonstrates its potential as an SGS model in LES for both pulsatile and turbulent flows. Additionally, other application studies of the GPU-accelerated VLBM including image-based blood flow and pore-scale diffusion-advection in porous structures will also be discussed.
Host: Min Wang & Duan Zhang T-3