The multi-physics problems in fluid mechanics are computationally solved using direct numerical simulation (DNS) and large eddy simulation (LES). The range of the multi-physics reaches: (i) laminar-turbulent transition, (ii) self-sustained turbulence, (iii) viscoelastic fluid and (iv) two-phase fluid effects. The complex wall-geometry, which includes the (local) non-zero pressure gradient, is considered using either the curvilinear grid or the immersed boundary method.
Investigation of the turbulent boundary layer over an airfoil/cascade with compressibility
Numerical method
RANS for the industrial application
Eddy-resolving simulation (i.e., LES or hybrid RANS/LES) for examination of flow structures
High-order DG method (collaborated with NASA Ames Research center)
Physical interest
Flow separation in the downstream and/or blade-hub corner
Shock formation
Laminar-to-turbulent transition
Industrial interest
Blade loading
Efficiency and loss
High-fidelity database for the reference of RANS modeling
Related papers: Lee et al. (Stanford CTR summer program 2018), Min et al. (TurboExpo 2018), Xia et al. (TurboExpo 2020)
(figure from JHU)
Modeling of the Johns Hopkins' axial compressor (originally, NASA low-speed rig) with the axial casing groove
Sliding interface between the rotor and casing
Unsteady RANS
Hybrid RANS/LES using DDES model
Integration of the experimental data and DNS data for defining the RANS boundary conditions
Performance prediction of the newly-designed compressor at the research center
Funded by ONR
Collaborated with:
Prof. J. Katz, Johns Hopkins University
NASA Glenn Research Center
(figure from RTX Pratt & Whitney)
Modeling and analysis of the flow-induced noise in (proprietary projects)
Pratt & Whitney engines
Collins Aerospace HVAC system
Analysis of the broadband noise in a combustor
Funded by FAA
Optimization of the k-w RANS model for transonic compressor cascade
Training the RANS model using experimental data
The Gaussian Process Regression with spatial kernels
Learning the error term from flow variables
Related work: Lee et al. (UTC Analytic Conference 2018)
Laminar-turbulence transition by homogeneous isotropic turbulent in free stream
Passive scalar transports in the fully-turbulent boundary layer
Temperature-dependent viscosity
Adverse pressure gradient
Related papers: Lee et al. (JFM 2013, IJHFF 2014, JoT 2010)
Curvilinear coordinate system for the wavy surface
Immersed boundary method for the square-rod roughness
Related papers: Lee et al. (PoF 2016), Nadeem et al. (IJHFF 2015)
Finite difference method
MPI-OpenMP hybrid parallelization
Related papers: Lee et al. (JFM 2014, PoF 2015), Hwang et al. (JFM 2016)
Simulation in the cylindrical coordinates (radial grid)
Comparison to the channel flow at the same Reynolds number
Related papers: Lee et al. (PoF 2015), Ahn et al. (PoF 2015)
Counter-intuitive characteristics to the ideal Newtonian fluid
Delay of the laminar-turbulence transition
Drag reduction of turbulent flows
Mathematical modeling by the FENE-P closure
Related papers: Page et al. (ICTAM 2016), Lee et al. (in preparation)
Coupled level-set and volume-of-fluid method for the interface tracking
Analysis of turbulent structures near the deformable interface
Froude number effect
Related papers: Lee et al. (JoT 2012a)
Cylindrical coordinates with the multigrid pressure solver
Inflow conditions from an auxiliary pipe flow simulation
Related papers: Lee et al. (in preparation)
Generalized cylindrical coordinates
Three angles of the divergence from 2 deg. to 8 deg.
Related papers: Lee et al. (JoT 2012b)
3-D unsteady incompressible Navier-Stokes equation solver using the spectral method
Decaying turbulence from the initial random noise
For the free-stream turbulence of the by-pass transition simulation
Cylindrical coordinates (radial grid)
Turbulent flows by the counter-rotating cylindrical walls