Broad interests within fluid mechanics, flow modeling, high-order numerical methods, and uncertainty quantification:
• High-order accurate numerical methods for compressible turbulence: LAD scheme, secondary conservative scheme, DG method
Compressible turbulent flows that interact with shock waves, contact surfaces, material interface and phase changes are key features in many
engineering and scientific problems.
We have been developing wide range of high-order accurate numerical methods for compressible turbulence, such as
localized artificial diffusivity (LAD) method, secondary conservative scheme, discontinuous Galerkin (DG) method,
to study compressible fluid mechanics of shock waves, contact surfaces, material interfaces, turbulence and their interactions.
• Kawai, Shankar and Lele, J. Comput. Phys., 229 (5), 2010.
• Kawai and Lele, AIAA J., 48 (9), 2010.
• Kawai and Lele, J. Comput. Phys., 227 (22), 2008.
||Multicomponent interfacial flows
• Kawai and Terashima, IJNMF, 66 (10), 2011.
• Terashima, Kawai and Koshi, Comput. Fluids, 88, 2013.
• Terashima, Kawai and Yamanishi, AIAA J., 49 (12), 2011.
• Wall-modeling in large-eddy simulation for realistic high Reynolds number flows
High-fidelity numerical simulations have received increased attention in recent years,
as a tool to study the flow physics.
Its most successful applications, however, have still been for moderate Reynolds numbers due to the computational cost required to
resolve the broadband scales of turbulence although many engineering applications are indeed at high Reynolds numbers.
We have proposed a simple method to bypass the the inevitable presence of numerical and subgrid modeling errors in
the grid points closest to the wall (Phys. Fluids paper (2012)
), and then based on that method
we proposed a dynamic non-equilibrium wall model where convection and pressure gradient terms are not neglected
(Phys. Fluids paper (2013)
). We are currently extending these approaches to
more practical flows and high heat flux flows.
• Numerical methods on hierarchical Cartesian grid suitable for exa-scale post-k supercomputer
Our group has been working on MEXT post-K computer project as a social and scientific priority issue 8
Sub-issue D: Research and development of core technologies to innovate aircraft design and operation.
Our group specifically works on developing a highly-accurate secondary conservative scheme on hierarchical Cartesian grid for the massive-parallel exa-scale post-k supercomputer environment.
• LES of flow around an aircraft across the whole flight envelope: Transonic buffet and stall phenomena
We have been developing a large-eddy simulation (LES) methodology to make the LES as a next generation aircraft aerodynamic design tool to predict the
aerodynamics across the whole flight envelope including the boundary of the flight envelope, such as transonic buffet and stall phenomenon,
at realistic high Reynolds number conditions.
The LES methodology is built based on our achievements of high-order accurate numerical method (LAD approach) and wall-modeling in LES.
The present wall-modeled LES successfully predicts the transonic buffet onset and also turbulence statistics without the use of ad hoc corrections,
something that existing studies fail to do robustly.
• DNS and physical modeling of supercritical turbulent flows in liquid rocket engine
Unique turbulent heat transfer mechanisms under transcritical and supercritical environments are important phenomena
in liquid-rocket engine and high-pressure turbine applications.
The mechanisms of turbulent heat transfer across the critical point under transcritical and supercritical environments are investigated
by solving compressible Navier-Stokes equations with direct numerical simulations.
Here we have been studying the fundamental physics of non-linear interactions between the thermal/transport properties and turbulence in
heated turbulent boundary layers under transcritical and supercritical environments,
and based on the DNS database we also have been studying RANS modeling of supercritical turbulent flows.
• Uncertainty quantification in CFD
The behaviors of realistic performance of designed vehicles and the underlying physics are not well described by deterministic approaches
due to limited information regarding their operating conditions and
lack of knowledge about the governing physical laws, physical transport properties, physical models, etc.
Thus it is crucial that we understand the intrinsic uncertainties and their influence on the quantities of
interest for realistic analyses and design of complex systems.
We devised a non-intrusive dynamic adaptive sampling method based on
the Kriging-model predictors for accurate and effective uncertainty quantification, and successfully applied to
quantify the uncertainty operating conditions in transonic flow around an airfoil.
We believe that the uncertainty quantification can be an essential part of validation, provide a rigorous measure of confidence,
indicate sensitivities and priorities, support fine understanding of physics and physical model development,
and also contribute to perform robust design and optimization.
Uncertainty quantification of transonic airfoil under uncertainty in freestream Mach number.
• Magnetohydrodynamics turbulence: shock waves and magnetic reconnection
Understanding the magnetohydrodynamics (MHD) of compressible flows is a topic of interest for a wide range of research fields,
including astrophysics, magnetospheric and heliospheric physics, and engineering.
A significant challenge in the field of compressible MHD simulations is to construct a numerical method
that is high-order accurate and simultaneously captures shock waves while
numerically satisfying the physical constraint of a divergence-free magnetic field.
We proposed a new high-order numerical algorithm for compressible MHD flows
that inherently maintains a divergence-free magnetic field, while robustly capturing shocks
(Kawai, J. Comput. Phys. 2013
This method is now used to study turbulent magnetic reconnection in magnetospheric plasma physics.
Orszag-Tang MHD shock-vortex interaction:
left, temperature; right, div. of magnetic fields
Fast magnetic reconnection
Past major research topics may be found in
Past Major Research