Accepted Manuscript

Turbocharger turbine and exhaust manifold flow computation with the

Space–Time Variational Multiscale Method and Isogeometric
Analysis

Yuto Otoguro, Kenji Takizawa, Tayfun E. Tezduyar,

Kenichiro Nagaoka, Sen Mei

PII: S0045-7930(18)30262-7
DOI: 10.1016/j.compfluid.2018.05.019
Reference: CAF 3904

To appear in: Computers and Fluids

Received date: 9 April 2018

Accepted date: 18 May 2018

Please cite this article as: Yuto Otoguro, Kenji Takizawa, Tayfun E. Tezduyar, Kenichiro Nagaoka,
Sen Mei, Turbocharger turbine and exhaust manifold flow computation with the Space–Time
Variational Multiscale Method and Isogeometric Analysis, Computers and Fluids (2018), doi:
10.1016/j.compfluid.2018.05.019

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service
to our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and
all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Highlights

• Challenges in turbocharger turbine and exhaust man-

ifold flow analysis are addressed.

• ST-VMS method used in the computation provide

higher-order accuracy.
• ST-IGA provides accurate geometry representation
and increased solution accuracy.

• New stabilization parameters and element lengths

are used in the STVMS and ST-SI.

T
• Successful flow analysis for a full intake/exhaust cy-
cle has been presented.

IP
CR
US
AN
M
ED
PT
CE
AC

1
 ACCEPTED MANUSCRIPT

Turbocharger turbine and exhaust manifold flow computation with the Space–Time
Variational Multiscale Method and Isogeometric Analysis

Yuto Otoguroa , Kenji Takizawaa,∗, Tayfun E. Tezduyarb,a , Kenichiro Nagaokaa , Sen Meib
a Department of Modern Mechanical Engineering, Waseda University
3-4-1 Ookubo, Shinjuku-ku, Tokyo 169-8555, Japan
b Mechanical Engineering, Rice University – MS 321

6100 Main Street, Houston, TX 77005, USA

T
IP
Abstract
We address the computational challenges encountered in turbocharger turbine and exhaust manifold flow analysis. The
core computational method is the Space–Time Variational Multiscale (ST-VMS) method, and the other key methods

CR
are the ST Isogeometric Analysis (ST-IGA), ST Slip Interface (ST-SI) method, ST/NURBS Mesh Update Method
(STNMUM), and a general-purpose NURBS mesh generation method for complex geometries. The ST framework,
in a general context, provides higher-order accuracy. The VMS feature of the ST-VMS addresses the computational
challenges associated with the multiscale nature of the unsteady flow in the manifold and turbine, and the moving-mesh

US
feature of the ST framework enables high-resolution computation near the rotor surface. The ST-SI enables moving-
mesh computation of the spinning rotor. The mesh covering the rotor spins with it, and the SI between the spinning
mesh and the rest of the mesh accurately connects the two sides of the solution. The ST-IGA enables more accurate
representation of the turbine and manifold geometries and increased accuracy in the flow solution. The STNMUM
AN
enables exact representation of the mesh rotation. The general-purpose NURBS mesh generation method makes it easier
to deal with the complex geometries we have here. An SI also provides mesh generation flexibility in a general context
by accurately connecting the two sides of the solution computed over nonmatching meshes. That is enabling us to use
nonmatching NURBS meshes here. Stabilization parameters and element length definitions play a significant role in
M

the ST-VMS and ST-SI. For the ST-VMS, we use the stabilization parameters introduced recently, and for the ST-SI,
the element length definition we are introducing here. The model we actually compute with includes the exhaust gas
purifier, which makes the turbine outflow conditions more realistic. We compute the flow for a full intake/exhaust cycle,
ED

which is much longer than the turbine rotation cycle because of high rotation speeds, and the long duration required is
an additional computational challenge. The computation demonstrates that the methods we use here are very effective
in this class of challenging flow analyses.
Keywords: Turbocharger, Turbine, Exhaust manifold, Space–Time Variational Multiscale method, ST-VMS, ST Slip
PT

Interface method, ST-SI, ST Isogeometric Analysis, ST-IGA

1. Introduction also by other researchers, with approaches ranging from

CE

using a single blade with spatially-periodic boundary con-

Turbocharger turbine and exhaust manifold flow analy- ditions (see, e.g., <1; 2; 3; 4; 5; 6>) to “sliding interfaces”
sis is computationally challenging. The challenges include (see, e.g., <7; 8; 9>). The precursors of the work pre-
unsteady flow through a complex geometry with multi- sented in this article were reported in <10; 11; 12>. In
AC

ple inlets, the need for high-resolution flow representation
near the rotor surface, high Reynolds numbers, and mul-
tiscale flow behavior. The flow unsteadiness comes from
the intake/exhaust cycle and the flow in the manifold and
turbine. An additional challenge is that the time scale of
the intake/exhaust cycle is much larger than that of the
turbine because of high turbine rotation speeds and this
requires long-duration computations in the turbine time
scale. The parts of these challenges faced in turbine com-
putations in a more general context have been addressed
∗ Correspondingauthor. Tel.: +81-3-5286-2143.
Email address: Kenji.Takizawa@tafsm.org (Kenji Takizawa)
this article, we target more realistic flow analysis con-
ditions, including the turbine inflow and outflow condi-
tions, increased accuracy in the flow solution, more de-
tailed turbine performance analysis, and covering the full
intake/exhaust cycle.
The core computational method is the Space–Time Vari-
ational Multiscale (ST-VMS) method <13; 14; 15>, and
the other key methods are the ST Isogeometric Analysis
(ST-IGA) <13; 16; 10>, ST Slip Interface (ST-SI) method
<17; 18>, ST/NURBS Mesh Update Method (STNMUM)
<16; 19; 20; 21>, and a general-purpose NURBS mesh
generation method for complex geometries <11; 12>.
 ACCEPTED MANUSCRIPT
1.1. ST-VMS and ST-SUPS
The ST-VMS is the VMS version of the Deforming-
Spatial-Domain/Stabilized ST (DSD/SST) method <22;
23; 24>. The DSD/SST was introduced for computation
of flows with moving boundaries and interfaces (MBI), in-
cluding fluid–structure interaction (FSI). In MBI compu-
tations the DSD/SST functions as a moving-mesh method.
Moving the fluid mechanics mesh to track a fluid–solid in-
terface enables mesh-resolution control near the interface
and, consequently, high-resolution representation of the
boundary layer. Because the stabilization components of
the DSD/SST are the Streamline-Upwind/Petrov-Galerkin
(SUPG) <25> and Pressure-Stabilizing/Petrov-Galerkin
(PSPG) <22> stabilizations, the method is also called
“ST-SUPS.” The VMS components of the ST-VMS are
from the residual-based VMS (RBVMS) method <26; 27;
28; 29>. The ST-VMS has two more stabilization terms
beyond those the ST-SUPS has, and these additional terms
give the method better turbulence modeling features. Con-
versely, we can see the ST-SUPS as a reduced version of
the ST-VMS. The ST-SUPS and ST-VMS, because of the
higher-order accuracy of the ST framework (see <13; 14>),
are desirable also in computations that do not involve
MBI.
The Arbitrary Lagrangian–Eulerian (ALE) method is
US
AN
an earlier and more commonly used moving-mesh method.
The ALE finite element method was introduced in 1981
<30>. The ALE-VMS method <31; 32; 33; 34; 35; 36>
is the VMS version of the ALE. It was introduced after
M
the ST-SUPS <22> and ALE-SUPS <37> methods and
preceded the ST-VMS. The ALE-VMS and RBVMS are
often supplemented with special methods, such as those
for weakly-enforced no-slip boundary condition <38; 39;
ED
40>, “sliding interfaces” <41; 42> and backflow stabi-
lization <43>. They have been successfully applied to
many classes of FSI, MBI and fluid mechanics problems.
The classes of problems include wind-turbine aerodynam-
PT
ics and FSI <7; 44; 45; 46; 47; 48; 8; 49>, more specifi-
cally, vertical-axis wind turbines <50; 51>, floating wind
turbines <52>, wind turbines in atmospheric boundary
CE
layers <53>, and fatigue damage in wind-turbine blades
<54>, patient-specific cardiovascular fluid mechanics and
FSI <55; 31; 56; 57; 58; 59; 60>, biomedical-device FSI
<61; 62; 63; 64; 65; 66>, ship hydrodynamics with free-
AC
surface flow and fluid–object interaction <67; 68>, hydro-
dynamics and FSI of a hydraulic arresting gear <69; 70>,
hydrodynamics of tidal-stream turbines with free-surface
flow <71>, and bioinspired FSI for marine propulsion <72;
73>.
The ST-SUPS and ST-VMS have also been success-
fully applied to many classes of FSI, MBI and fluid me-
chanics problems. The classes of problems include space-
craft parachute analysis for the main parachutes <74; 34;
75; 76; 77>, cover-separation parachutes <78> and the
drogue parachutes <79; 80; 81>, wind-turbine aerodynam-
ics for horizontal-axis wind-turbine rotors <7; 82; 83; 34>,
full horizontal-axis wind turbines <21; 84; 85; 8> and
3
vertical-axis wind turbines <17>, flapping-wing aerody-
namics for an actual locust <16; 19; 34; 86>, bioinspired
MAVs <20; 87; 84; 85> and wing-clapping <88; 89>,
blood flow analysis of cerebral aneurysms <90; 84>, stent-
blocked aneurysms <91; 92; 90>, aortas <93; 94> and
heart valves <88; 85; 95; 96; 97; 94>, spacecraft aero-
dynamics <78; 98>, thermo-fluid analysis of ground ve-
hicles and their tires <15>, thermo-fluid analysis of disk
brakes <18>, flow-driven string dynamics in turbomachin-
ery <99>, flow analysis of turbocharger turbines <10; 11;
12>, flow around tires with road contact and deformation
<100; 101>, ram-air parachutes <102>, and compressible-
T
flow parachute aerodynamics <103>.
In the flow analysis presented here, the ST framework
IP
provides higher-order accuracy in a general context. The
VMS feature of the ST-VMS addresses the computational
challenges associated with the multiscale nature of the un-
CR
steady flow in the manifold and turbine. The moving-mesh
feature of the ST framework enables high-resolution com-
putation near the rotor surface.
1.2. Discontinuity-capturing term
When the flow field has a shock or some other discon-
tinuity, stabilized methods are often supplemented with
a discontinuity-capturing (DC) term. We do not have a
DC term in the computations presented in this article.
The introduction segment we include here just serves as
background material for the introduction segment given in
Section 1.7 on stabilization parameters and element length
definitions.
Supplementing the SUPG method with a DC term goes
back more than three decades <104; 105>. In fact, the DC
term played a key role in the evolution of the compressible-
flow SUPG method <106; 107; 108>, which was originally
introduced in 1982 in the context of conservation variables.
That 1982 method is now called “(SUPG)82 .” At first
(SUPG)82 was not used with any DC (shock-capturing)
term, and the test computations clearly showed the need
for something extra at the shocks. Later (SUPG)82 was
recast in entropy variables, but also supplemented with a
DC term <109>. This resulted in better shock profiles.
In a 1991 ASME paper <110>, (SUPG)82 was supple-
mented with a very similar DC term. It was shown in
<110; 111> that, with the added DC term, (SUPG)82
was very comparable in accuracy to (SUPG)82 recast in
entropy variables. The stabilized methods and DC terms
introduced in <105> for the advection–diffusion–reaction
equation accounted for the interaction between the DC
and SUPG terms. Taking that interaction into account
precludes “compounding” (i.e. augmentation of the SUPG
effect by the DC effect when the advection and disconti-
nuity directions coincide).
1.3. ST-SI
The ST-SI was introduced in <17>, in the context
of incompressible-flow equations, to retain the desirable
 ACCEPTED MANUSCRIPT
moving-mesh features of the ST-VMS when we have spin-
ning solid surfaces, such as a turbine rotor. The mesh cov-
ering the spinning surface spins with it, retaining the high- using higher-order basis functions in space.
resolution representation of the boundary layers. The SI
between the spinning mesh and the rest of the mesh accu-
rately connects the two sides of the flow field. The starting and demonstrated in <16; 19; 20>, higher-order NURBS
point in the development of the ST-SI was the ALE-VMS
version for “sliding interfaces” <41; 42>. In the ST-SI, in-
terface terms similar to those in the ALE-VMS version are
added to the ST-VMS formulation to account for the com-
patibility conditions for the velocity and stress. An ST-SI
version where the SI is between fluid and solid domains
with weakly-enforced Dirichlet boundary conditions for
the fluid was also presented in <17>. The SI in this case is
a “fluid–solid SI” rather than a standard “fluid–fluid SI.”
The ST-SI method introduced in <18> for the coupled
incompressible-flow and thermal-transport equations re-
tain the high-resolution representation of the thermo-fluid
boundary layers near spinning solid surfaces. These ST-
SI methods have been successfully applied to aerodynamic
analysis of vertical-axis wind turbines <17>, thermo-fluid
analysis of disk brakes <18>, flow-driven string dynamics
in turbomachinery <99>, flow analysis of turbocharger
turbines <10; 11; 12>, flow around tires with road con-
tact and deformation <100; 101>, aerodynamic analysis
US NURBS in time also enable, with the “ST-C” method, ex-
AN
of ram-air parachutes <102>, and heart valve flow analysis
<96; 97; 94>.
In another version of the ST-SI presented in <17>,
the SI is between a thin porous structure and the fluid on
M
its two sides. This enables dealing with the fabric poros-
ity in a fashion consistent with how the standard fluid–
fluid SIs are dealt with and how the Dirichlet conditions
are enforced weakly with fluid–solid SIs. Furthermore,
ED
this version enables handling thin structures that have
T-junctions. This method has been successfully used in
incompressible-flow aerodynamic analysis of ram-air parachuteschinery <99>, and flow analysis of turbocharger turbines
with fabric porosity <102>. The compressible-flow ST-SI
PT
methods were introduced in <103>, including the version
where the SI is between a thin porous structure and the
fluid on its two sides. Compressible-flow porosity mod-
CE
els were also introduced in <103>. These, together with
the compressible-flow ST SUPG method <112>, extended
the ST computational analysis range to compressible-flow
aerodynamics of parachutes with fabric and geometric porosi- at a desirable level for good accuracy. It has been utilized
AC
ties. That enabled successful ST computational flow anal-
ysis of the Orion spacecraft drogue parachute in the compressible-
flow regime <103>. The computations were in the context
of finite element discretization.
1.4. ST-IGA and STNMUM
The ST-IGA was introduced in <13>. It is the inte-
gration of the ST framework with isogeometric discretiza-
tion. First computations with the ST-VMS and ST-IGA
were reported in <13> in a 2D context, with IGA ba-
sis functions in space for flow past an airfoil, and in both
space and time for the advection equation. The stabil-
ity and accuracy analysis given <13> for the advection
4
equation showed that using higher-order basis functions
in time would be essential in getting full benefit out of
In the early stages of the ST-IGA, the emphasis was on
IGA basis functions in time. As pointed out in <13; 14>
basis functions in time provide a more accurate represen-
tation of the motion of the solid surfaces and a mesh mo-
tion consistent with that. They also provide more effi-
ciency in temporal representation of the motion and de-
formation of the volume meshes, and better efficiency in
remeshing. That is how the STNMUM was introduced and
T
demonstrated in <16; 19; 20>. The name “STNMUM”
was given in <21>. The STNMUM has a wide scope
IP
that includes spinning solid surfaces. With the spinning
motion represented by quadratic NURBS basis functions
in time, and with sufficient number of temporal patches
CR
for a full rotation, the circular paths are represented ex-
actly, and a “secondary mapping” <13; 16; 14; 34> en-
ables also specifying a constant angular velocity for in-
variant speeds along the paths. The ST framework and
tracting a continuous representation from the computed
data and, in large-scale computations, efficient data com-
pression <113; 15; 18; 99>. The STNMUM and desir-
able features of the ST-IGA with IGA basis functions in
time have been demonstrated in many 3D computations.
The classes of problems solved are flapping-wing aerody-
namics for an actual locust <16; 19; 34; 86>, bioinspired
MAVs <20; 87; 84; 85> and wing-clapping <88; 89>,
separation aerodynamics of spacecraft <78>, aerodynam-
ics of horizontal-axis <21; 84; 85; 8> and vertical-axis
<17> wind-turbines, thermo-fluid analysis of ground ve-
hicles and their tires <15>, thermo-fluid analysis of disk
brakes <18>, flow-driven string dynamics in turboma-
<10; 11; 12>.
The ST-IGA with IGA basis functions in space pro-
vides more accurate representation of the geometry and
increased accuracy in the flow solution. Because it accom-
plishes that with less number of control points, and conse-
quently with larger effective element sizes, it enables using
larger time-step sizes while keeping the Courant number
in ST computational flow analysis of turbocharger tur-
bines <10; 11; 12>, ram-air parachutes <102>, tires with
road contact and deformation <101>, and heart valves
<96; 97; 94>.
In the flow analysis presented here, the ST-IGA enables
more accurate representation of the turbine and manifold
geometries, increased accuracy in the flow solution, and
using larger time-step sizes. The STNMUM enables exact
representation of the mesh rotation.
1.5. General-purpose NURBS mesh generation method
To make the ST-IGA use, and in a wider context the
IGA use, even more practical in computational flow anal-
 ACCEPTED MANUSCRIPT
ysis with complex geometries, NURBS volume mesh gen-
eration needs to be easier and more automated. To that
end, a general-purpose NURBS mesh generation method
was introduced in <11>. The method is based on multi-
block-structured mesh generation with existing techniques,
projection of that mesh to a NURBS mesh made of patches
that correspond to the blocks, and recovery of the original
model surfaces. The recovery of the original surfaces is to
the extent they are suitable for accurate and robust fluid
mechanics computations. The method is expected to re-
tain the refinement distribution and element quality of the
multi-block-structured mesh that we start with. Because
there are ample good techniques and software for gener-
ating multi-block-structured meshes, the method makes
general-purpose mesh generation relatively easy. Mesh-
quality performance studies for 2D and 3D meshes, includ-
ing those for complex models, were presented in <12>.
A test computation for a turbocharger turbine and ex-
haust manifold was also presented in <12>. The perfor-
mance studies and test computation demonstrated that
the general-purpose NURBS mesh generation method makes
the IGA use in fluid mechanics computations even more
practical.
The general-purpose NURBS mesh generation method
is used also in the turbocharger turbine and exhaust man-
US
AN
ifold flow analysis presented here.
1.6. ST-SI-IGA
An SI also provides mesh generation flexibility in a
M
general context by accurately connecting the two sides of
the solution computed over nonmatching meshes. This
type of mesh generation flexibility is especially valuable
ED
in complex-geometry flow computations with isogeometric
discretization. The integration of the ST-SI and ST-IGA
allows, without loss of accuracy, C −1 continuity between
NURBS patches and thus removes the matching require-
PT
ment between the patches. This feature was used in the
heart valve flow analysis, for the purpose of independent
meshing in the inlet and outlet regions of the computa-
tional domain. It is also used in the turbocharger turbine
CE
and exhaust manifold flow analysis, for the purpose of in-
dependent meshing in the manifold and volute regions of
the computational domain.
AC
1.7. Stabilization parameters and element length defini-
tions
In the ST-SUPS and ST-VMS, and in stabilized meth-
ods in general, an embedded stabilization parameter, known
as “τ ,” plays a significant role. This parameter involves
a measure of the local length scale (also known as “ele-
ment length”) and other parameters such as the element
Reynolds and Courant numbers. Various element lengths
and τ s were proposed, starting with those in <114; 25>
and <106; 107; 108>, followed by the the ones introduced
in <104; 105>. In many cases, the “element length” was
seen as an advection length scale. The set of τ s introduced
5
in <106; 107; 108> in conjunction with (SUPG)82 is now
called “τ82 .” The τ definition introduced in <105>, which
is for the advective limit and is now called “τSUGN1 ” (and
the corresponding element length is now called “hUGN ”),
automatically yields lower values for higher-order finite ele-
ment basis functions (see <115; 116>). Later, other τ def-
initions that are applicable to higher-order elements were
proposed in <117> in the context of advective-diffusive
systems. The τ used in <110> with (SUPG)82 was a
slightly modified version of τ82 . Subsequent minor modi-
fications of τ82 took into account the interaction between
the DC and (SUPG)82 terms in a fashion similar to how
T
it was done in <105> for the advection–diffusion–reaction
equation. Until 2004, all these slightly modified versions of
IP
τ82 were always used with the same DC parameter, which
was introduced in the 1991 ASME paper <110> and is
now called “δ91 .” This DC parameter was derived from
CR
the one given in <109> for the entropy variables.
Calculating the τ s based on the element-level matri-
ces and vectors was introduced in <118> in the context
of the advection–diffusion equation and the Navier–Stokes
equations of incompressible flows. These definitions are ex-
pressed in terms of the ratios of the norms of the matrices
or vectors. They automatically take into account the lo-
cal length scales, advection field and the element Reynolds
number. The definitions based on the element-level vectors
were shown <118; 119> to address the difficulties reported
at small time-step sizes. A second element length scale,
based on the solution gradient and called “hRGN ,” was in-
troduced in 2001 <120; 23>. Recognizing this as a diffu-
sion length scale, a new stabilization parameter for the dif-
fusive limit, “τSUGN3 ,” was introduced in <23; 121>, to be
used together with τSUGN1 and “τSUGN2 ,” the parameters
for the advective and transient limits. For the stabilized
ST methods, “τSUGN12 ,” representing both the advective
and transient limits, was also introduced in <23>.
New ways of calculating the τ and DC parameter to
be used with (SUPG)82 were introduced in 2004 <122;
121; 123>. The new τ s, now categorized under the la-
bel “τ04 ,” have a matrix structure for viscous flows and
reduce to a scalar for inviscid flows. The new DC pa-
rameters were of two types: one defined in a style the
Discontinuity-Capturing Directional Dissipation (DCDD)
<23; 123; 124> parameter was defined, and one that is now
called “YZβ” DC parameter. The YZβ DC parameter is
residual based, and it is simpler than δ91 . It has options
for smoother or sharper computed shocks. A number of
2D and 3D test computations with YZβ DC were reported
in <125; 126; 127>. These computations showed that in
addition to being simpler than δ91 , the YZβ DC parameter
was superior in accuracy. The computations reported in
<125; 126; 127> were based on the compressible-flow ST
SUPG.
Some new options for the stabilization parameters used
with the SUPS and VMS methods were proposed in <24;
83; 16; 21; 15>. These include a fourth τ component,
“τSUGN4 ” <15>, which was introduced for the VMS method,
 ACCEPTED MANUSCRIPT

Z
considering one of the two extra stabilization terms the + ε (wh ) : σ (uh , ph )dQ
VMS method has compared to the SUPS method. They Qn
Z
also include stabilization parameters <15> for the thermal- h h
transport part of the VMS method for the coupled incompressible-− (P ) w · h dP
n h
flow and thermal-transport equations. Z Z
h −


The stabilization and DC parameters discussed in this + q h∇ · uh dQ + (wh )+ h +
n · ρ (u )n − (u )n dΩ
section so far were all originally intended for finite element Qn Ωn

discretization, but quite often used also for isogeometric (n

X el )n Z   
τSUPS ∂wh h h
discretization. The stabilization and DC parameters intro- + ρ + u · ∇ w
Qen ρ ∂t
duced in <128> target isogeometric discretization, but are e=1

also applicable to finite element discretization. They were
+ ∇ q h · rM (uh , ph )dQ
introduced in the context of the advection–diffusion equa-

T
tion and the Navier–Stokes equations of incompressible (nel )n Z
X
flows. The parameters are based on a direction-dependent + νLSIC∇ · wh ρrC (uh )dQ

IP
element length expression. The expression is outcome of e=1 Q e
n
an easy to understand derivation. The key components of (nel )n Z
X

the derivation are mapping the direction vector from the τSUPS wh · rM (uh , ph ) · ∇ uh dQ

CR
−
physical ST element to the parent ST element, accounting e=1 Qne

for the discretization spacing along each of the parametric (nel )n Z

X 2
τSUPS

coordinates, and mapping what we have in the parent ele- − rM (uh , ph ) · ∇ wh · rM (uh , ph )dQ
ment back to the physical element. The test computations ρ

US
Qne
e=1
presented in <128> for pure-advection cases showed that
= 0, (1)
the new parameters yield good solution profiles.
Element length definitions play a significant role also where
in the ST-SI. Until now, the definitions were based on the  h 
AN
h h ∂u h h h
one introduced in <105> and its ST version <23>. rM (u , p ) = ρ + u · ∇u − f − ∇ · σ (uh , ph ),
In the flow analysis presented here, the stabilization ∂t
parameters come mostly from <128>. For the ST-SI, we (2)
h h
are introducing a new element length definition, which can rC (u ) = ∇ · u (3)
M

be seen as an extension of the one introduced in <128>.

are the residuals of the momentum equation and incom-
1.8. Computation presented pressibility constraint. Here, ρ, u, p, f , and h are the
density, velocity, pressure, body force, and the traction
ED

The model we actually compute with consists of the

specified at the boundary. The stress tensor is defined as
exhaust manifold, turbocharger turbine and the exhaust
σ (u, p) = −pI + 2µεε(u), where I is the identity tensor,
gas purifier. The gas purifier has about 700 narrow chan-
µ = ρν is the viscosity,ν is the kinematic viscosity, and
nels. Because of the sponsor restrictions, we will not be
able to provide any additional information on that com- ε (u) = (∇ ∇u) + (∇ ∇u)T /2 is the strain-rate tensor. The
PT

ponent or show a picture of it. For the presentation of test functions associated with the velocity and pressure are
the material here, we will see it just as a component that w and q. A superscript “h” indicates that the function is
makes the turbine outflow conditions more realistic. We coming from a finite-dimensional space. The symbol Qn
CE

compute the flow for a full intake/exhaust cycle.
1.9. Outline of the remaining sections
In Section 2 we provide the ST-VMS and ST-SI to-
AC
represents the ST slice between time levels n and n + 1,
(Pn )h is the part of the lateral boundary of that slice as-
sociated with the traction boundary condition h, and Ωn
is the spatial domain at time level n. The superscript “e”

gether with the stabilization parameters and ST-SI ele- is the ST element counter, and nel is the number of ST
ment length definitions. The flow analysis is presented elements. The functions are discontinuous in time at each
Section 3, and the concluding remarks in Section 4. In time level, and the superscripts “−” and “+” indicate the
the Appendix, we provide some additional details related values of the functions just below and just above the time
to the stabilization parameters and element length defini- level.
tions.
Remark 1. The ST-SUPS can be obtained from the ST-
VMS by dropping the eighth and ninth integrations.
2. ST-VMS and ST-SI
There are various ways of defining the stabilization pa-
We first describe, mostly from <17>, the ST-VMS: rameters τSUPS and νLSIC . Here, τSUPS is mostly from
Z  h  <128>:
h ∂u h h h
− 21
w ·ρ + u · ∇u − f dQ −2 −2 −2
Qn ∂t τSUPS = τSUGN12 + τSUGN3 + τSUGN4 . (4)
6
 ACCEPTED MANUSCRIPT

Z
The first component is given as 1  
 


− h
ρwA · FAh − FAh  uhA − FAh − FAh  uhB dP
   (Pn )SI 2
1 1 Z
−2
τSUGN12 = : GST , (5) h h

1 h

u u + nB · wB + nA · wA pB + phA dP
(Pn )SI 2
Z
ST
where G is the element metric tensor in the ST frame- h h




− wB − wA · n̂B · µ ε (uhB ) + ε (uhA ) dP
work (see Appendix A.2). The second component is de- (Pn )SI
fined as Z
h

h




− n̂B · µ ε wB + ε wA · uhB − uhA dP
−1 (Pn )SI
τSUGN3 = νrr : G, (6) Z
µC h h



where r is the solution direction: + wB − wA · uhB − uhA dP, (11)
(Pn )SI h

T
∇ kuk where
r= , (7)
∇ kukk
k∇

IP
FBh = nB · uhB − vB
h
, (12)
and G is the element metric tensor (see Appendix A.1). h h h


FA = nA · uA − vA , (13)
The third component, originating from <15>, is defined

CR
 −1 −1
as hB + h−1 A
h= , (14)
−1 2
τSUGN4 = ∇ uh F
(8)
− 21
hB = 2 (nB nB : G) (for Side B), (15)
or
−1
τSUGN4 = ε (uh ) F .

Here k · kF represents the Frobenius norm. The stabiliza-

US
(9)
hA = 2 (nA nA : G)

n̂B =
nB − nA
knB − nA k
.
− 21
(for Side A), (16)

(17)
AN
tion parameter νLSIC is from <21>: Here, (Pn )SI is the SI in the ST domain, n is the unit
normal vector, v is the mesh velocity, and C is a nondi-
h2LSIC mensional penalty constant. We note that the expressions
νLSIC = , (10)
τSUPS given by Eqs. (14)–(16) are being introduced here. At
M

the same time we note that the element lengths given by

where hLSIC is set equal to the minimum element length Eqs. (15) and (16) are straightforward extensions of the
hMIN (see Appendix A.1). For more ways of calculating one in <128>.
the stabilization parameters in flow computations, see <129; A number of remarks were provided in <17> to ex-
ED

111; 123; 125; 126; 127; 130; 131; 124; 132; 133; 134; 135;
136; 137; 119; 138; 139; 1; 140; 2; 141; 23; 24; 21; 15; 17>.
The expression for rM (uh , ph ) includes second deriva-
tives of the velocity. For linear basis functions these terms
PT
plain the added terms and to comment on related inter-
pretations. We refer the reader interested in such details
to <17>.
On solid surfaces where we prefer to have weakly-imposed

vanish, and for bilinear and trilinear basis functions they
are grossly underrepresented. This means that rM (uh , ph )
does not explicitly depend on the Reynolds number. When
we use quadratic or higher-order basis functions, on the
CE
Dirichlet conditions for the fluid, we use the ST-SI version
where the SI is between the fluid and solid domains. That
version was obtained in <17> by starting with the terms

added to Side B and replacing the Side A velocity with

other hand, the term is nonzero and therefore explicit de- the velocity gh coming from the solid domain. Then the
pendence of the residual on the Reynolds number is taken terms added to Eq. (1) to represent the weakly-imposed
into account. Dirichlet conditions become
AC

In describing the ST-SI (see <17>), we use the labels

“Side A” and “Side B” to represent the two sides of the SI. Z Z
In the ST-SI version of the formulation given by Eq. (1), − h
qB nB · uhB dP − h
ρwB · FBh uhB dP
there are added boundary terms corresponding to the SI. (Pn )SI (Pn )SI
Z
The boundary terms for the two sides are first added sep- h
arately, using test functions wA h h
and qA and wB h h
and qB . + qB nB · gh dP
(Pn )SI
Then, putting together the terms added to each side, the Z
1  

complete set of terms added becomes + h
ρwB · FBh + FBh  uhB
(Pn )SI 2
Z
h h

1 h
 


− qB nB − qA nA · uB − uhA dP + FBh − FBh  gh dP
(Pn )SI 2
Z Z
h 1
 
 


− ρwB · FBh − FBh  uhB − FBh − FBh  uhA dP + n− w− : σ − dP
(Pn )SI 2 (Pn )SI

7
 ACCEPTED MANUSCRIPT
Z
h


− wB · nB · σ hB dP
(Pn )SI
Z
h



− nB · 2µεε wB · uhB − gh dP
(Pn )SI
Z
µC h

+ wB · uhB − gh dP.
(Pn )SI hB
3. Computation
3.1. Problem setup
The model we use is for a four-cylinder engine and is
shown in Figure 1. The model we actually compute with
M
ED
Figure 1: Turbocharger turbine and exhaust manifold.
consists of the exhaust manifold, turbocharger turbine and
PT
the exhaust gas purifier, which has about 700 narrow chan-
nels. Because of the sponsor restrictions, we are unable to
provide any additional information on the gas purifier or
CE
show a picture of it. In Figure 1, we only show the manifold
and turbine. For the presentation of the material here, we
will see the purifier just as a component that makes the
turbine outflow conditions more realistic. In that sense,
AC
we can see the manifold as a component that makes the
inflow conditions more realistic. The rotor diameter is 30
mm and the rotor speed is 30,000 rpm, which translates to
a turbine rotation period of TTR = 2.0×10−3 s. The engine
speed is 2,000 rpm, which translates to an intake/exhaust
cycle of T = 6.0×10−2 s. The gas density and kinematic
viscosity are 0.9 kg/m3 and 2.8×10−5 m2 /s.
3.2. Mesh
Figure 2 shows the quadratic NURBS control mesh
for the turbine and manifold. The mesh was generated
with the general-purpose NURBS mesh generation method
<11; 12>. Table 1 shows the number of control points and
(18)
T
IP
CR
Figure 2: Quadratic NURBS control mesh for the turbine and
US manifold.
Table 1: Number of control points (nc) and elements (ne) in different
parts of the mesh.
AN
Part nc ne
Turbine and manifold 274,500 170,137
Exhaust gas purifier 161,365 49,426
Total 435,865 219,563
elements in different parts of the mesh. Figure 3 shows
the five SIs of the mesh. Two of the SIs have an actual
slip. The other three are just for mesh generation purpose
and connect nonmatching meshes. They are used for in-
dependent meshing in the manifold and volute regions of
the computational domain. The STNMUM enables exact
representation of the mesh rotation.
3.3. Computational conditions
We compute the flow for a full intake/exhaust cycle.
The flow rate at each manifold inlet is shown in Figure
4, with the color code for the inlets shown in Figure 5.
In temporal representation of the mesh rotation, we again
use quadratic NURBS basis functions. There are 90 time
steps per rotation, which is equivalent to a time-step size of
2.22×10−5 s. The number of nonlinear iterations per time
step is 4, with 500, 500, 600 and 800 GMRES iterations
for the first, second, third and fourth nonlinear iterations,
respectively. The first two nonlinear iterations are based
on the ST-SUPS, and the last two the ST-VMS. We use the
stabilization parameters given by Eqs. (4)–(8) and (10),
and in the scaling (see Appendix A.2), Dθ = 1 and D = I.
In the ST-SI, we use C = 8.
Remark 2. The computational settings and mesh resolu-
tions were determined based on the stability and accuracy
8
 ACCEPTED MANUSCRIPT

0.05

Volumetric flow rate (m3 /s)

0.04

0.03

0.02

0.01

0.00
0 1/4 2/4 3/4 1

T
t/T
atCylinder 1 atCylinder 2

IP
atCylinder 3 atCylinder 4
0.05

CR
Volumetric flow rate (m3 /s)
0.04

Figure 3: Five SIs of the mesh. The red SIs have an actual slip, 0.03
and the blue SIs are just for mesh generation purpose and connect
nonmatching meshes.

evaluations in the test computations we conducted prior

US 0.02
AN
0.01
to the computations we report here. The time-step size
based on 90 steps per turbine rotation is very large com- 0.00
pared to what is typical in computation of rotating ma- 0 1/4 2/4 3/4 1
chineries. We are able to use larger time-step sizes while t/T
M

keeping the Courant number at a desirable level for good

accuracy. That is because the ST-IGA provides higher ac- Figure 4: Volumetric flow rate at each inlet (top) and in total (bot-
curacy (in both geometry representation and flow solution) tom). The color code for the inlets is shown in Figure 5.
with less number of control points, and consequently with
ED

larger effective element sizes. where P (t) is the instantaneous power extracted from the
turbine. Figures 9–13 show the turbine efficiency measures
3.4. Results ηIB (0, T, t), ηIB (0, T /4, t), ηIB (T /4, T /2, t), ηIB (T /2, 3T /4, t)
Figure 6 shows the velocity magnitude, and Figure 7
PT

and ηIB (3T /4, T, t).

shows the isosurfaces corresponding to a positive value of We see in Figures 6 and 7 that immediately after the
the second invariant of the velocity gradient tensor, colored peak flow rate is reached, and that is when we have a
2
by the total pressure, p + 12 ρ kuk . strong deceleration, the vortex structure breaks down into
CE

In reporting the turbine efficiency, we first define two smaller structures. We also see that small vortices appear
quantities, one instantaneous and the other interval-based: in the manifold where the flow rate is small. Efficiency is
Z   strongly related to the total volumetric flow rate (shown
1 2
K(t) = (−n · u) ρkuk + p dΓ in the lower part of Figure 4), and is also influenced by
AC

ΓINF 2
Z the history of the flow rate. Since the vortex breakdown is
+ (−n · u) pdΓ, (19) a key source of the energy loss, reducing the deceleration
ΓOUTF rate would play a key role in increasing the efficiency.

Z t2
1 4. Concluding remarks
KIB (t1 , t2 ) = K(t)dt. (20)
t2 − t1 t1
We have addressed the computational challenges en-
The inflow and outflow boundaries ΓINF and ΓOUTF are countered in turbocharger turbine and exhaust manifold
shown in Figure 8. Now we define the interval-based effi- flow analysis and presented results for a configuration that
ciency as includes the exhaust gas purifier. The challenges include
unsteady flow through a complex geometry with multiple
P (t)
ηIB (t1 , t2 , t) = , (21) inlets, the need for high-resolution flow representation near
KIB (t1 , t2 )
9
 ACCEPTED MANUSCRIPT
Figure 5: Color code for the inlets. Each inlet color corresponds to
the color used in Figure 4.
the rotor surface, high Reynolds numbers, and multiscale
flow behavior. One of the main sources of the flow un-
steadiness is the intake/exhaust cycle. The time scale of
the intake/exhaust cycle is much larger than that of the
turbine because of high turbine rotation speeds, requir-
ing long-duration computations in the turbine time scale.
M
This is an additional computational challenge that we have
overcome.
In the flow analysis presented, the core computational Appendix A. Element metric tensor
ED
method was the ST-VMS, and the other key methods were
the ST-IGA, ST-SI, STNMUM, and a general-purpose NURBS
mesh generation method for complex geometries. The ST
framework, in a general context, provides higher-order ac- Section 2 in calculation of the stabilization parameters and
PT
curacy. The VMS feature of the ST-VMS addressed the
computational challenges associated with the multiscale
nature of the unsteady flow in the manifold and turbine,
and the moving-mesh feature of the ST framework enabled
CE
high-resolution computation near the rotor surface. The
ST-SI enabled moving-mesh computation of the spinning
rotor. The mesh covering the rotor spins with it, and the
SI between the spinning mesh and the rest of the mesh
AC
accurately connects the two sides of the solution. The
ST-IGA enabled more accurate representation of the tur-
bine and manifold geometries and increased accuracy in
the flow solution. The STNMUM enabled exact represen-
tation of the mesh rotation. The general-purpose NURBS
mesh generation method made it somewhat easier to deal
with the complex geometries we have here. An SI also pro- as
vides mesh generation flexibility in a general context by
accurately connecting the two sides of the solution com-
puted over nonmatching meshes. That enabled us to use
nonmatching NURBS meshes. Stabilization parameters
and element length definitions play a significant role in
the ST-VMS and ST-SI. For the ST-VMS, we used the
stabilization parameters introduced recently, and for the
ST-SI, the element length definition we introduced in this
article.
Including the exhaust gas purifier in the computation
made the turbine outflow conditions more realistic, just
like how including the manifold made the inflow condi-
tions more realistic. We computed the flow for a full in-
take/exhaust cycle. The computation showed that the
methods we used here are very effective in this class of
challenging flow analyses.
T
Acknowledgement
IP
This work was supported in part by Grant-in-Aid for
Challenging Exploratory Research 16K13779 from Japan
Society for the Promotion of Science; Grant-in-Aid for Sci-
CR
entific Research (S) 26220002 from the Ministry of Edu-
cation, Culture, Sports, Science and Technology of Japan
(MEXT); Council for Science, Technology and Innovation
(CSTI), Cross-Ministerial Strategic Innovation Promotion
US Program (SIP), “Innovative Combustion Technology” (Fund-
ing agency: JST); and Rice–Waseda research agreement.
This work was also supported (third author) in part by
ARO Grant W911NF-17-1-0046 and Top Global Univer-
AN
sity Project of Waseda University. The authors acknowl-
edge the Texas Advanced Computing Center (TACC) at
The University of Texas at Austin for providing HPC re-
sources that have contributed to the research results re-
ported within this paper.
Here provide from <128> the element metric tensor
in space and in the ST framework. These are used in
element lengths.
Appendix A.1. Element metric tensor in space
Components of the Jacobian matrix Q are written as
∂xi
Qij = , (A.1)
∂ξj
where ξj is the parametric coordinate in jth direction. We
first scale it with a matrix D to take into account the
polynomial order or other factors such as the dimensions
of the element domain in the parametric space:
Q̂ = QD−1 . (A.2)
With this vector, we define the element length (see <128>)
−1
hRQD = 2 (rr : G) 2 . (A.3)
where
G = Q̂−T Q̂−1 . (A.4)
10
 ACCEPTED MANUSCRIPT

Remark 3. From this derivation, what we get with D = I References

has been used in many methods of calculating the stabi- [1] A. Corsini, F. Rispoli, A.G. Sheard, and T.E. Tezduyar, “Com-
lization parameters (see, for example, <34>). In those putational analysis of noise reduction devices in axial fans
methods, a scaling factor taking the polynomial order into with stabilized finite element formulations”, Computational
account is applied to the element length, and here we do the Mechanics, 50 (2012) 695–705, doi: 10.1007/s00466-012-0789-
4.
scaling in the parametric space, for each of the parametric [2] A. Corsini, F. Rispoli, A.G. Sheard, K. Takizawa, T.E. Tez-
directions. duyar, and P. Venturini, “A variational multiscale method for
particle-cloud tracking in turbomachinery flows”, Computa-
Sweeping over all the directions represented by r, we ob- tional Mechanics, 54 (2014) 1191–1202, doi: 10.1007/s00466-
tain the minimum and maximum element lengths: 014-1050-0.
  [3] L. Cardillo, A. Corsini, G. Delibra, F. Rispoli, and T.E. Tez-
−1 duyar, “Flow analysis of a wave-energy air turbine with the
hMIN ≡ 2 min (rr : G) 2 , (A.5) SUPG/PSPG method and DCDD”, in Y. Bazilevs and K. Tak-
r
 

T
−1 izawa, editors, Advances in Computational Fluid–Structure In-
hMAX ≡ 2 max (rr : G) 2 . (A.6) teraction and Flow Simulation: New Methods and Challenging
r
Computations, Modeling and Simulation in Science, Engineer-

IP
They are equivalent to ing and Technology, 39–53, Springer, 2016, ISBN 978-3-319-
40825-5.
 − 12 [4] L. Cardillo, A. Corsini, G. Delibra, F. Rispoli, and T.E. Tez-

CR
hMIN = 2 max (rr : G) , (A.7) duyar, “Flow analysis of a wave-energy air turbine with the
r SUPG/PSPG stabilization and Discontinuity-Capturing Di-
− 12 rectional Dissipation”, Computers & Fluids, 141 (2016) 184–
= 2 (λmax (G)) , (A.8)
190, doi: 10.1016/j.compfluid.2016.07.011.
[5] A. Castorrini, A. Corsini, F. Rispoli, P. Venturini, K. Tak-
and

hMAX = 2 min (rr : G)
r

= 2 (λmin (G))
− 12

− 12
,

,
US
(A.9)

(A.10)
izawa, and T.E. Tezduyar, “SUPG/PSPG computational anal-
ysis of rain erosion in wind-turbine blades”, in Y. Bazilevs
and K. Takizawa, editors, Advances in Computational Fluid–
Structure Interaction and Flow Simulation: New Methods
and Challenging Computations, Modeling and Simulation in
AN
Science, Engineering and Technology, 77–96, Springer, 2016,
ISBN 978-3-319-40825-5.
where λmax and λmin are the maximum and minimum [6] A. Castorrini, A. Corsini, F. Rispoli, P. Venturini, K. Tak-
eigenvalues of the argument matrix. izawa, and T.E. Tezduyar, “Computational analysis of
wind-turbine blade rain erosion”, Computers & Fluids,
Remark 4. In the implementation, we take measures to 141 (2016) 175–183, doi: 10.1016/j.compfluid.2016.08.013.
M

keep the calculated element length between hMIN and hMAX .
Appendix A.2. Element metric tensor in the ST frame-
ED
[7] Y. Bazilevs, M.-C. Hsu, I. Akkerman, S. Wright, K. Takizawa,
B. Henicke, T. Spielman, and T.E. Tezduyar, “3D simulation
of wind turbine rotors at full scale. Part I: Geometry model-
ing and aerodynamics”, International Journal for Numerical

work Methods in Fluids, 65 (2011) 207–235, doi: 10.1002/fld.2400.

The ST Jacobian matrix is [8] Y. Bazilevs, K. Takizawa, T.E. Tezduyar, M.-C. Hsu, N. Kos-
tov, and S. McIntyre, “Aerodynamic and FSI analysis of wind
 ∂t ∂t  turbines with the ALE-VMS and ST-VMS methods”, Archives
ST ∂θ ∂ξξ
Q = ∂x (A.11) of Computational Methods in Engineering, 21 (2014) 359–398,
Q
PT

∂θ doi: 10.1007/s11831-014-9119-7.
 ∂t ∂t  [9] A. Korobenko, Y. Bazilevsi, K. Takizawa, and T.E. Tezduyar,
= ∂θ∂t ∂ξξ , (A.12) “Recent advances in ALE-VMS and ST-VMS computational
v ∂θ Q aerodynamic and FSI analysis of wind turbines”, to appear in
a special volume to be published by Springer, 2018.
CE

where θ is the parametric coordinate in time, and the mesh [10] K. Takizawa, T.E. Tezduyar, Y. Otoguro, T. Terahara,
velocity v is T. Kuraishi, and H. Hattori, “Turbocharger flow compu-
tations with the Space–Time Isogeometric Analysis (ST-

∂x  IGA)”, Computers & Fluids, 142 (2017) 15–20, doi:
v= . (A.13) 10.1016/j.compfluid.2016.02.021.
∂t ξ
AC

[11] Y. Otoguro, K. Takizawa, and T.E. Tezduyar, “Space–

time VMS computational flow analysis with isogeometric dis-
The ST scaling matrix is given as cretization and a general-purpose NURBS mesh generation
  method”, Computers & Fluids, 158 (2017) 189–200, doi:
ST Dθ 0T 10.1016/j.compfluid.2017.04.017.
D = , (A.14) [12] Y. Otoguro, K. Takizawa, and T.E. Tezduyar, “A general-
0 D
purpose NURBS mesh generation method for complex ge-
and the scaling becomes ometries”, to appear in a special volume to be published by
Springer, 2018.

−1 [13] K. Takizawa and T.E. Tezduyar, “Multiscale space–time fluid–
Q̂ST = QST DST . (A.15) structure interaction techniques”, Computational Mechanics,
48 (2011) 247–267, doi: 10.1007/s00466-011-0571-z.
The ST metric tensor is defined as [14] K. Takizawa and T.E. Tezduyar, “Space–time fluid–structure
 −T  −1 interaction methods”, Mathematical Models and Methods
GST = Q̂ST Q̂ST . (A.16) in Applied Sciences, 22 (supp02) (2012) 1230001, doi:
10.1142/S0218202512300013.

11
 ACCEPTED MANUSCRIPT
[15] K. Takizawa, T.E. Tezduyar, and T. Kuraishi, “Multiscale ST
methods for thermo-fluid analysis of a ground vehicle and its
tires”, Mathematical Models and Methods in Applied Sciences,
25 (2015) 2227–2255, doi: 10.1142/S0218202515400072.
[16] K. Takizawa, B. Henicke, A. Puntel, T. Spielman, and
T.E. Tezduyar, “Space–time computational techniques for the
aerodynamics of flapping wings”, Journal of Applied Mechan-
ics, 79 (2012) 010903, doi: 10.1115/1.4005073.
[17] K. Takizawa, T.E. Tezduyar, H. Mochizuki, H. Hattori,
S. Mei, L. Pan, and K. Montel, “Space–time VMS method for
flow computations with slip interfaces (ST-SI)”, Mathemati-
cal Models and Methods in Applied Sciences, 25 (2015) 2377–
2406, doi: 10.1142/S0218202515400126.
[18] K. Takizawa, T.E. Tezduyar, T. Kuraishi, S. Tabata, and
H. Takagi, “Computational thermo-fluid analysis of a disk
brake”, Computational Mechanics, 57 (2016) 965–977, doi:
10.1007/s00466-016-1272-4.
[19] K. Takizawa, B. Henicke, A. Puntel, N. Kostov, and T.E. Tez-
duyar, “Space–time techniques for computational aerodynam-
ics modeling of flapping wings of an actual locust”, Compu-
tational Mechanics, 50 (2012) 743–760, doi: 10.1007/s00466-
012-0759-x.
[20] K. Takizawa, N. Kostov, A. Puntel, B. Henicke, and T.E. Tez-
duyar, “Space–time computational analysis of bio-inspired
flapping-wing aerodynamics of a micro aerial vehicle”, Compu-
tational Mechanics, 50 (2012) 761–778, doi: 10.1007/s00466-
012-0758-y.
[21] K. Takizawa, T.E. Tezduyar, S. McIntyre, N. Kostov, R. Kole-
sar, and C. Habluetzel, “Space–time VMS computation of
wind-turbine rotor and tower aerodynamics”, Computational
Mechanics, 53 (2014) 1–15, doi: 10.1007/s00466-013-0888-x.
US
AN
[22] T.E. Tezduyar, “Stabilized finite element formulations for in-
compressible flow computations”, Advances in Applied Me-
chanics, 28 (1992) 1–44, doi: 10.1016/S0065-2156(08)70153-4.
[23] T.E. Tezduyar, “Computation of moving boundaries and in-
terfaces and stabilization parameters”, International Journal
M
for Numerical Methods in Fluids, 43 (2003) 555–575, doi:
10.1002/fld.505.
[24] T.E. Tezduyar and S. Sathe, “Modeling of fluid–structure in-
teractions with the space–time finite elements: Solution tech-
niques”, International Journal for Numerical Methods in Flu-
ED
ids, 54 (2007) 855–900, doi: 10.1002/fld.1430.
[25] A.N. Brooks and T.J.R. Hughes, “Streamline upwind/Petrov-
Galerkin formulations for convection dominated flows with
particular emphasis on the incompressible Navier-Stokes equa-
tions”, Computer Methods in Applied Mechanics and Engi-
PT
neering, 32 (1982) 199–259.
[26] T.J.R. Hughes, “Multiscale phenomena: Green’s functions, the
Dirichlet-to-Neumann formulation, subgrid scale models, bub-
bles, and the origins of stabilized methods”, Computer Meth-
ods in Applied Mechanics and Engineering, 127 (1995) 387–
CE
401.
[27] T.J.R. Hughes, A.A. Oberai, and L. Mazzei, “Large eddy sim-
ulation of turbulent channel flows by the variational multiscale
method”, Physics of Fluids, 13 (2001) 1784–1799.
AC
[28] Y. Bazilevs, V.M. Calo, J.A. Cottrell, T.J.R. Hughes, A. Reali,
and G. Scovazzi, “Variational multiscale residual-based tur-
bulence modeling for large eddy simulation of incompressible
flows”, Computer Methods in Applied Mechanics and Engi-
neering, 197 (2007) 173–201.
[29] Y. Bazilevs and I. Akkerman, “Large eddy simulation of tur-
bulent Taylor–Couette flow using isogeometric analysis and
the residual–based variational multiscale method”, Journal of
Computational Physics, 229 (2010) 3402–3414.
[30] T.J.R. Hughes, W.K. Liu, and T.K. Zimmermann,
“Lagrangian–Eulerian finite element formulation for in-
compressible viscous flows”, Computer Methods in Applied
Mechanics and Engineering, 29 (1981) 329–349.
[31] Y. Bazilevs, V.M. Calo, T.J.R. Hughes, and Y. Zhang, “Iso-
geometric fluid–structure interaction: theory, algorithms, and
computations”, Computational Mechanics, 43 (2008) 3–37.
12
[32] K. Takizawa, Y. Bazilevs, and T.E. Tezduyar, “Space–time
and ALE-VMS techniques for patient-specific cardiovascu-
lar fluid–structure interaction modeling”, Archives of Com-
putational Methods in Engineering, 19 (2012) 171–225, doi:
10.1007/s11831-012-9071-3.
[33] Y. Bazilevs, M.-C. Hsu, K. Takizawa, and T.E. Tez-
duyar, “ALE-VMS and ST-VMS methods for computer
modeling of wind-turbine rotor aerodynamics and fluid–
structure interaction”, Mathematical Models and Methods
in Applied Sciences, 22 (supp02) (2012) 1230002, doi:
10.1142/S0218202512300025.
[34] Y. Bazilevs, K. Takizawa, and T.E. Tezduyar, Computational
Fluid–Structure Interaction: Methods and Applications. Wi-
ley, February 2013, ISBN 978-0470978771.
[35] Y. Bazilevs, K. Takizawa, and T.E. Tezduyar, “Challenges
T
and directions in computational fluid–structure interaction”,
Mathematical Models and Methods in Applied Sciences,
IP
23 (2013) 215–221, doi: 10.1142/S0218202513400010.
[36] Y. Bazilevs, K. Takizawa, and T.E. Tezduyar, “New direc-
tions and challenging computations in fluid dynamics modeling
with stabilized and multiscale methods”, Mathematical Models
CR
and Methods in Applied Sciences, 25 (2015) 2217–2226, doi:
10.1142/S0218202515020029.
[37] V. Kalro and T.E. Tezduyar, “A parallel 3D computational
method for fluid–structure interactions in parachute systems”,
Computer Methods in Applied Mechanics and Engineering,
190 (2000) 321–332, doi: 10.1016/S0045-7825(00)00204-8.
[38] Y. Bazilevs and T.J.R. Hughes, “Weak imposition of Dirich-
let boundary conditions in fluid mechanics”, Computers and
Fluids, 36 (2007) 12–26.
[39] Y. Bazilevs, C. Michler, V.M. Calo, and T.J.R. Hughes, “Iso-
geometric variational multiscale modeling of wall-bounded tur-
bulent flows with weakly enforced boundary conditions on un-
stretched meshes”, Computer Methods in Applied Mechanics
and Engineering, 199 (2010) 780–790.
[40] M.-C. Hsu, I. Akkerman, and Y. Bazilevs, “Wind turbine
aerodynamics using ALE-VMS: Validation and role of weakly
enforced boundary conditions”, Computational Mechanics,
50 (2012) 499–511.
[41] Y. Bazilevs and T.J.R. Hughes, “NURBS-based isogeometric
analysis for the computation of flows about rotating compo-
nents”, Computational Mechanics, 43 (2008) 143–150.
[42] M.-C. Hsu and Y. Bazilevs, “Fluid–structure interaction mod-
eling of wind turbines: simulating the full machine”, Compu-
tational Mechanics, 50 (2012) 821–833.
[43] M.E. Moghadam, Y. Bazilevs, T.-Y. Hsia, I.E. Vignon-
Clementel, A.L. Marsden, and M. of Congenital Hearts Al-
liance (MOCHA), “A comparison of outlet boundary treat-
ments for prevention of backflow divergence with rele-
vance to blood flow simulations”, Computational Mechanics,
48 (2011) 277–291, doi: 10.1007/s00466-011-0599-0.
[44] Y. Bazilevs, M.-C. Hsu, J. Kiendl, R. Wüchner, and K.-
U. Bletzinger, “3D simulation of wind turbine rotors at full
scale. Part II: Fluid–structure interaction modeling with com-
posite blades”, International Journal for Numerical Methods
in Fluids, 65 (2011) 236–253.
[45] M.-C. Hsu, I. Akkerman, and Y. Bazilevs, “High-performance
computing of wind turbine aerodynamics using isogeometric
analysis”, Computers and Fluids, 49 (2011) 93–100.
[46] Y. Bazilevs, M.-C. Hsu, and M.A. Scott, “Isogeometric fluid–
structure interaction analysis with emphasis on non-matching
discretizations, and with application to wind turbines”, Com-
puter Methods in Applied Mechanics and Engineering, 249-
252 (2012) 28–41.
[47] M.-C. Hsu, I. Akkerman, and Y. Bazilevs, “Finite element sim-
ulation of wind turbine aerodynamics: Validation study using
NREL Phase VI experiment”, Wind Energy, 17 (2014) 461–
481.
[48] A. Korobenko, M.-C. Hsu, I. Akkerman, J. Tippmann, and
Y. Bazilevs, “Structural mechanics modeling and FSI simula-
tion of wind turbines”, Mathematical Models and Methods in
 ACCEPTED MANUSCRIPT
Applied Sciences, 23 (2013) 249–272.
[49] Y. Bazilevs, A. Korobenko, X. Deng, and J. Yan, “Novel struc-
tural modeling and mesh moving techniques for advanced FSI
simulation of wind turbines”, International Journal for Nu-
merical Methods in Engineering, 102 (2015) 766–783, doi:
10.1002/nme.4738.
[50] A. Korobenko, M.-C. Hsu, I. Akkerman, and Y. Bazilevs,
“Aerodynamic simulation of vertical-axis wind turbines”,
Journal of Applied Mechanics, 81 (2013) 021011, doi:
10.1115/1.4024415.
[51] Y. Bazilevs, A. Korobenko, X. Deng, J. Yan, M. Kinzel,
and J.O. Dabiri, “FSI modeling of vertical-axis wind tur-
bines”, Journal of Applied Mechanics, 81 (2014) 081006, doi:
10.1115/1.4027466.
[52] J. Yan, A. Korobenko, X. Deng, and Y. Bazilevs, “Compu-
tational free-surface fluid–structure interaction with applica-
tion to floating offshore wind turbines”, Computers and Fluids,
141 (2016) 155–174, doi: 10.1016/j.compfluid.2016.03.008.
[53] Y. Bazilevs, A. Korobenko, J. Yan, A. Pal, S.M.I. Gohari,
and S. Sarkar, “ALE–VMS formulation for stratified turbulent
incompressible flows with applications”, Mathematical Models
and Methods in Applied Sciences, 25 (2015) 2349–2375, doi:
10.1142/S0218202515400114.
[54] Y. Bazilevs, A. Korobenko, X. Deng, and J. Yan, “FSI mod-
eling for fatigue-damage prediction in full-scale wind-turbine
blades”, Journal of Applied Mechanics, 83 (6) (2016) 061010.
[55] Y. Bazilevs, V.M. Calo, Y. Zhang, and T.J.R. Hughes,
“Isogeometric fluid–structure interaction analysis with appli-
cations to arterial blood flow”, Computational Mechanics,
38 (2006) 310–322.
[56] Y. Bazilevs, J.R. Gohean, T.J.R. Hughes, R.D. Moser, and
US
AN
Y. Zhang, “Patient-specific isogeometric fluid–structure in-
teraction analysis of thoracic aortic blood flow due to im-
plantation of the Jarvik 2000 left ventricular assist device”,
Computer Methods in Applied Mechanics and Engineering,
198 (2009) 3534–3550.
M
[57] Y. Bazilevs, M.-C. Hsu, D. Benson, S. Sankaran, and A. Mars-
den, “Computational fluid–structure interaction: Methods and
application to a total cavopulmonary connection”, Computa-
tional Mechanics, 45 (2009) 77–89.
[58] Y. Bazilevs, M.-C. Hsu, Y. Zhang, W. Wang, X. Liang,
ED
T. Kvamsdal, R. Brekken, and J. Isaksen, “A fully-coupled
fluid–structure interaction simulation of cerebral aneurysms”,
Computational Mechanics, 46 (2010) 3–16.
[59] Y. Bazilevs, M.-C. Hsu, Y. Zhang, W. Wang, T. Kvamsdal,
S. Hentschel, and J. Isaksen, “Computational fluid–structure
PT
interaction: Methods and application to cerebral aneurysms”,
Biomechanics and Modeling in Mechanobiology, 9 (2010) 481–
498.
[60] M.-C. Hsu and Y. Bazilevs, “Blood vessel tissue prestress mod-
eling for vascular fluid–structure interaction simulations”, Fi-
CE
nite Elements in Analysis and Design, 47 (2011) 593–599.
[61] C.C. Long, A.L. Marsden, and Y. Bazilevs, “Fluid–structure
interaction simulation of pulsatile ventricular assist de-
vices”, Computational Mechanics, 52 (2013) 971–981, doi:
AC
10.1007/s00466-013-0858-3.
[62] C.C. Long, M. Esmaily-Moghadam, A.L. Marsden, and
Y. Bazilevs, “Computation of residence time in the simula-
tion of pulsatile ventricular assist devices”, Computational Me-
chanics, 54 (2014) 911–919, doi: 10.1007/s00466-013-0931-y.
[63] C.C. Long, A.L. Marsden, and Y. Bazilevs, “Shape op-
timization of pulsatile ventricular assist devices using FSI
to minimize thrombotic risk”, Computational Mechanics,
54 (2014) 921–932, doi: 10.1007/s00466-013-0967-z.
[64] M.-C. Hsu, D. Kamensky, Y. Bazilevs, M.S. Sacks, and
T.J.R. Hughes, “Fluid–structure interaction analysis of bio-
prosthetic heart valves: significance of arterial wall deforma-
tion”, Computational Mechanics, 54 (2014) 1055–1071, doi:
10.1007/s00466-014-1059-4.
[65] M.-C. Hsu, D. Kamensky, F. Xu, J. Kiendl, C. Wang,
M.C.H. Wu, J. Mineroff, A. Reali, Y. Bazilevs, and M.S. Sacks,
13
“Dynamic and fluid–structure interaction simulations of bio-
prosthetic heart valves using parametric design with T-splines
and Fung-type material models”, Computational Mechanics,
55 (2015) 1211–1225, doi: 10.1007/s00466-015-1166-x.
[66] D. Kamensky, M.-C. Hsu, D. Schillinger, J.A. Evans, A. Ag-
garwal, Y. Bazilevs, M.S. Sacks, and T.J.R. Hughes, “An
immersogeometric variational framework for fluid-structure
interaction: Application to bioprosthetic heart valves”,
Computer Methods in Applied Mechanics and Engineering,
284 (2015) 1005–1053.
[67] I. Akkerman, Y. Bazilevs, D.J. Benson, M.W. Farthing, and
C.E. Kees, “Free-surface flow and fluid–object interaction
modeling with emphasis on ship hydrodynamics”, Journal of
Applied Mechanics, 79 (2012) 010905.
[68] I. Akkerman, J. Dunaway, J. Kvandal, J. Spinks, and
T
Y. Bazilevs, “Toward free-surface modeling of planing vessels:
simulation of the Fridsma hull using ALE-VMS”, Computa-
IP
tional Mechanics, 50 (2012) 719–727.
[69] C. Wang, M.C.H. Wu, F. Xu, M.-C. Hsu, and Y. Bazilevs,
“Modeling of a hydraulic arresting gear using fluid–structure
interaction and isogeometric analysis”, Computers and Fluids,
CR
142 (2017) 3–14, doi: 10.1016/j.compfluid.2015.12.004.
[70] M.C.H. Wu, D. Kamensky, C. Wang, A.J. Herrema, F. Xu,
M.S. Pigazzini, A. Verma, A.L. Marsden, Y. Bazilevs, and M.-
C. Hsu, “Optimizing fluid–structure interaction systems with
immersogeometric analysis and surrogate modeling: Applica-
[71]
tion to a hydraulic arresting gear”, Computer Methods in Ap-
plied Mechanics and Engineering, (2017), Published online.
doi:10.1016/j.cma.2016.09.032.
J. Yan, X. Deng, A. Korobenko, and Y. Bazilevs, “Free-surface
flow modeling and simulation of horizontal-axis tidal-stream
turbines”, Computers and Fluids, 158 (2017) 157–166, doi:
10.1016/j.compfluid.2016.06.016.
[72] B. Augier, J. Yan, A. Korobenko, J. Czarnowski, G. Ket-
terman, and Y. Bazilevs, “Experimental and numerical FSI
study of compliant hydrofoils”, Computational Mechanics,
55 (2015) 1079–1090, doi: 10.1007/s00466-014-1090-5.
[73] J. Yan, B. Augier, A. Korobenko, J. Czarnowski, G. Ket-
terman, and Y. Bazilevs, “FSI modeling of a propulsion
system based on compliant hydrofoils in a tandem config-
uration”, Computers and Fluids, 141 (2016) 201–211, doi:
10.1016/j.compfluid.2015.07.013.
[74] K. Takizawa and T.E. Tezduyar, “Computational methods
for parachute fluid–structure interactions”, Archives of Com-
putational Methods in Engineering, 19 (2012) 125–169, doi:
10.1007/s11831-012-9070-4.
[75] K. Takizawa, M. Fritze, D. Montes, T. Spielman, and
T.E. Tezduyar, “Fluid–structure interaction modeling of ring-
sail parachutes with disreefing and modified geometric poros-
ity”, Computational Mechanics, 50 (2012) 835–854, doi:
10.1007/s00466-012-0761-3.
[76] K. Takizawa, T.E. Tezduyar, J. Boben, N. Kostov, C. Boswell,
and A. Buscher, “Fluid–structure interaction modeling of clus-
ters of spacecraft parachutes with modified geometric poros-
ity”, Computational Mechanics, 52 (2013) 1351–1364, doi:
10.1007/s00466-013-0880-5.
[77] K. Takizawa, T.E. Tezduyar, C. Boswell, Y. Tsutsui, and
K. Montel, “Special methods for aerodynamic-moment calcula-
tions from parachute FSI modeling”, Computational Mechan-
ics, 55 (2015) 1059–1069, doi: 10.1007/s00466-014-1074-5.
[78] K. Takizawa, D. Montes, M. Fritze, S. McIntyre, J. Boben,
and T.E. Tezduyar, “Methods for FSI modeling of space-
craft parachute dynamics and cover separation”, Mathematical
Models and Methods in Applied Sciences, 23 (2013) 307–338,
doi: 10.1142/S0218202513400058.
[79] K. Takizawa, T.E. Tezduyar, C. Boswell, R. Kolesar, and
K. Montel, “FSI modeling of the reefed stages and disreefing of
the Orion spacecraft parachutes”, Computational Mechanics,
54 (2014) 1203–1220, doi: 10.1007/s00466-014-1052-y.
[80] K. Takizawa, T.E. Tezduyar, R. Kolesar, C. Boswell, T. Kanai,
and K. Montel, “Multiscale methods for gore curvature calcula-
 ACCEPTED MANUSCRIPT
tions from FSI modeling of spacecraft parachutes”, Computa-
tional Mechanics, 54 (2014) 1461–1476, doi: 10.1007/s00466-
014-1069-2.
[81] K. Takizawa, T.E. Tezduyar, and R. Kolesar, “FSI modeling of
the Orion spacecraft drogue parachutes”, Computational Me-
chanics, 55 (2015) 1167–1179, doi: 10.1007/s00466-014-1108-
z.
[82] K. Takizawa, B. Henicke, T.E. Tezduyar, M.-C. Hsu,
and Y. Bazilevs, “Stabilized space–time computation of
wind-turbine rotor aerodynamics”, Computational Mechanics,
48 (2011) 333–344, doi: 10.1007/s00466-011-0589-2.
[83] K. Takizawa, B. Henicke, D. Montes, T.E. Tezduyar, M.-
C. Hsu, and Y. Bazilevs, “Numerical-performance studies for
the stabilized space–time computation of wind-turbine rotor
aerodynamics”, Computational Mechanics, 48 (2011) 647–657,
doi: 10.1007/s00466-011-0614-5.
[84] K. Takizawa, Y. Bazilevs, T.E. Tezduyar, M.-C. Hsu,
O. Øiseth, K.M. Mathisen, N. Kostov, and S. McIntyre, “En-
gineering analysis and design with ALE-VMS and space–time
methods”, Archives of Computational Methods in Engineer-
ing, 21 (2014) 481–508, doi: 10.1007/s11831-014-9113-0.
[85] K. Takizawa, “Computational engineering analysis with the
new-generation space–time methods”, Computational Me-
chanics, 54 (2014) 193–211, doi: 10.1007/s00466-014-0999-z.
[86] K. Takizawa, B. Henicke, A. Puntel, N. Kostov, and T.E. Tez-
duyar, “Computer modeling techniques for flapping-wing aero-
[87]
dynamics of a locust”, Computers & Fluids, 85 (2013) 125–
134, doi: 10.1016/j.compfluid.2012.11.008.
K. Takizawa, T.E. Tezduyar, and N. Kostov, “Sequentially-
coupled space–time FSI analysis of bio-inspired flapping-
wing aerodynamics of an MAV”, Computational Mechanics,
US
AN
54 (2014) 213–233, doi: 10.1007/s00466-014-0980-x.
[88] K. Takizawa, T.E. Tezduyar, A. Buscher, and S. Asada,
“Space–time interface-tracking with topology change (ST-
TC)”, Computational Mechanics, 54 (2014) 955–971, doi:
10.1007/s00466-013-0935-7.
M
[89] K. Takizawa, T.E. Tezduyar, and A. Buscher, “Space–
time computational analysis of MAV flapping-wing aero-
dynamics with wing clapping”, Computational Mechanics,
55 (2015) 1131–1141, doi: 10.1007/s00466-014-1095-0.
[90] K. Takizawa, Y. Bazilevs, T.E. Tezduyar, C.C. Long,
ED
A.L. Marsden, and K. Schjodt, “ST and ALE-VMS meth-
ods for patient-specific cardiovascular fluid mechanics model-
ing”, Mathematical Models and Methods in Applied Sciences,
24 (2014) 2437–2486, doi: 10.1142/S0218202514500250.
[91] K. Takizawa, K. Schjodt, A. Puntel, N. Kostov, and T.E. Tez-
PT
duyar, “Patient-specific computer modeling of blood flow in
cerebral arteries with aneurysm and stent”, Computational
Mechanics, 50 (2012) 675–686, doi: 10.1007/s00466-012-0760-
4.
[92] K. Takizawa, K. Schjodt, A. Puntel, N. Kostov, and
CE
T.E. Tezduyar, “Patient-specific computational analysis of
the influence of a stent on the unsteady flow in cerebral
aneurysms”, Computational Mechanics, 51 (2013) 1061–1073,
doi: 10.1007/s00466-012-0790-y.
AC
[93] H. Suito, K. Takizawa, V.Q.H. Huynh, D. Sze, and T. Ueda,
“FSI analysis of the blood flow and geometrical charac-
teristics in the thoracic aorta”, Computational Mechanics,
54 (2014) 1035–1045, doi: 10.1007/s00466-014-1017-1.
[94] K. Takizawa, T.E. Tezduyar, H. Uchikawa, T. Terahara,
T. Sasaki, K. Shiozaki, A. Yoshida, K. Komiya, and G. In-
oue, “Aorta flow analysis and heart valve flow and structure
analysis”, to appear in a special volume to be published by
Springer, 2018.
[95] K. Takizawa, T.E. Tezduyar, A. Buscher, and S. Asada,
“Space–time fluid mechanics computation of heart valve mod-
els”, Computational Mechanics, 54 (2014) 973–986, doi:
10.1007/s00466-014-1046-9.
[96] K. Takizawa, T.E. Tezduyar, T. Terahara, and T. Sasaki,
“Heart valve flow computation with the Space–Time Slip In-
terface Topology Change (ST-SI-TC) method and Isogeomet-
14
ric Analysis (IGA)”, in P. Wriggers and T. Lenarz, editors,
Biomedical Technology: Modeling, Experiments and Simula-
tion, Lecture Notes in Applied and Computational Mechanics,
77–99, Springer, 2018, ISBN 978-3-319-59547-4.
[97] K. Takizawa, T.E. Tezduyar, T. Terahara, and T. Sasaki,
“Heart valve flow computation with the integrated Space–Time
VMS, Slip Interface, Topology Change and Isogeometric Dis-
cretization methods”, Computers & Fluids, 158 (2017) 176–
188, doi: 10.1016/j.compfluid.2016.11.012.
[98] K. Takizawa, D. Montes, S. McIntyre, and T.E. Tezdu-
yar, “Space–time VMS methods for modeling of incompress-
ible flows at high Reynolds numbers”, Mathematical Models
and Methods in Applied Sciences, 23 (2013) 223–248, doi:
10.1142/s0218202513400022.
[99] K. Takizawa, T.E. Tezduyar, and H. Hattori, “Computa-
T
tional analysis of flow-driven string dynamics in turboma-
chinery”, Computers & Fluids, 142 (2017) 109–117, doi:
IP
10.1016/j.compfluid.2016.02.019.
[100] K. Takizawa, T.E. Tezduyar, S. Asada, and T. Kuraishi,
“Space–time method for flow computations with slip inter-
faces and topology changes (ST-SI-TC)”, Computers & Fluids,
CR
141 (2016) 124–134, doi: 10.1016/j.compfluid.2016.05.006.
[101] T. Kuraishi, K. Takizawa, and T.E. Tezduyar, “Space–time
computational analysis of tire aerodynamics with actual ge-
ometry, road contact and tire deformation”, to appear in a
special volume to be published by Springer, 2018.
[102] K. Takizawa, T.E. Tezduyar, and T. Terahara, “Ram-
air parachute structural and fluid mechanics computa-
tions with the space–time isogeometric analysis (ST-
IGA)”, Computers & Fluids, 141 (2016) 191–200, doi:
10.1016/j.compfluid.2016.05.027.
[103] K. Takizawa, T.E. Tezduyar, and T. Kanai, “Porosity mod-
els and computational methods for compressible-flow aerody-
namics of parachutes with geometric porosity”, Mathematical
Models and Methods in Applied Sciences, 27 (2017) 771–806,
doi: 10.1142/S0218202517500166.
[104] T.J.R. Hughes, M. Mallet, and A. Mizukami, “A new finite
element formulation for computational fluid dynamics: II. Be-
yond SUPG”, Computer Methods in Applied Mechanics and
Engineering, 54 (1986) 341–355.
[105] T.E. Tezduyar and Y.J. Park, “Discontinuity capturing fi-
nite element formulations for nonlinear convection-diffusion-
reaction equations”, Computer Methods in Applied Mechan-
ics and Engineering, 59 (1986) 307–325, doi: 10.1016/0045-
7825(86)90003-4.
[106] T.E. Tezduyar and T.J.R. Hughes, “Development
of time-accurate finite element techniques for first-
order hyperbolic systems with particular empha-
sis on the compressible Euler equations”, NASA
Technical Report NASA-CR-204772, NASA, 1982,
http://www.researchgate.net/publication/24313718/.
[107] T.E. Tezduyar and T.J.R. Hughes, “Finite element formula-
tions for convection dominated flows with particular emphasis
on the compressible Euler equations”, in Proceedings of AIAA
21st Aerospace Sciences Meeting, AIAA Paper 83-0125, Reno,
Nevada, (1983), doi: 10.2514/6.1983-125.
[108] T.J.R. Hughes and T.E. Tezduyar, “Finite element methods
for first-order hyperbolic systems with particular emphasis on
the compressible Euler equations”, Computer Methods in Ap-
plied Mechanics and Engineering, 45 (1984) 217–284, doi:
10.1016/0045-7825(84)90157-9.
[109] T.J.R. Hughes, L.P. Franca, and M. Mallet, “A new finite ele-
ment formulation for computational fluid dynamics: VI. Con-
vergence analysis of the generalized SUPG formulation for lin-
ear time-dependent multi-dimensional advective-diffusive sys-
tems”, Computer Methods in Applied Mechanics and Engi-
neering, 63 (1987) 97–112.
[110] G.J. Le Beau and T.E. Tezduyar, “Finite element computa-
tion of compressible flows with the SUPG formulation”, in Ad-
vances in Finite Element Analysis in Fluid Dynamics, FED-
Vol.123, ASME, New York, (1991) 21–27.
 ACCEPTED MANUSCRIPT
[111] G.J. Le Beau, S.E. Ray, S.K. Aliabadi, and T.E. Tezdu-
yar, “SUPG finite element computation of compressible flows
with the entropy and conservation variables formulations”,
Computer Methods in Applied Mechanics and Engineering,
104 (1993) 397–422, doi: 10.1016/0045-7825(93)90033-T.
[112] T.E. Tezduyar, S.K. Aliabadi, M. Behr, and S. Mittal, “Mas-
sively parallel finite element simulation of compressible and
incompressible flows”, Computer Methods in Applied Mechan-
ics and Engineering, 119 (1994) 157–177, doi: 10.1016/0045-
7825(94)00082-4.
[113] K. Takizawa and T.E. Tezduyar, “Space–time computa-
tion techniques with continuous representation in time
(ST-C)”, Computational Mechanics, 53 (2014) 91–99, doi:
10.1007/s00466-013-0895-y.
[114] T.J.R. Hughes and A.N. Brooks, “A multi-dimensional upwind
scheme with no crosswind diffusion”, in T.J.R. Hughes, edi-
tor, Finite Element Methods for Convection Dominated Flows,
AMD-Vol.34, 19–35, ASME, New York, 1979.
[115] J.E. Akin, T. Tezduyar, M. Ungor, and S. Mittal, “Stabiliza-
tion parameters and Smagorinsky turbulence model”, Journal
of Applied Mechanics, 70 (2003) 2–9, doi: 10.1115/1.1526569.
[116] J.E. Akin and T.E. Tezduyar, “Calculation of the advec-
tive limit of the SUPG stabilization parameter for linear
and higher-order elements”, Computer Methods in Applied
Mechanics and Engineering, 193 (2004) 1909–1922, doi:
10.1016/j.cma.2003.12.050.
[117] L.P. Franca, S.L. Frey, and T.J.R. Hughes, “Stabilized fi-
model”, Computer Methods in Applied Mechanics and Engi-
neering, 95 (1992) 253–276.
[118] T.E. Tezduyar and Y. Osawa, “Finite element stabilization
US
nite element methods: I. Application to the advective-diffusive
AN
parameters computed from element matrices and vectors”,
Computer Methods in Applied Mechanics and Engineering,
190 (2000) 411–430, doi: 10.1016/S0045-7825(00)00211-5.
[119] M.-C. Hsu, Y. Bazilevs, V.M. Calo, T.E. Tezduyar, and
T.J.R. Hughes, “Improving stability of stabilized and multi-
M
scale formulations in flow simulations at small time steps”,
Computer Methods in Applied Mechanics and Engineering,
199 (2010) 828–840, doi: 10.1016/j.cma.2009.06.019.
[120] T.E. Tezduyar, “Adaptive determination of the finite element
stabilization parameters”, in Proceedings of the ECCOMAS
ED
Computational Fluid Dynamics Conference 2001 (CD-ROM),
Swansea, Wales, United Kingdom, (2001).
[121] T.E. Tezduyar, “Finite element methods for fluid dynam-
ics with moving boundaries and interfaces”, in E. Stein,
R.D. Borst, and T.J.R. Hughes, editors, Encyclopedia of Com-
PT
putational Mechanics, Volume 3: Fluids, Chapter 17, Wiley,
2004, ISBN 978-0-470-84699-5.
[122] T.E. Tezduyar, “Determination of the stabilization and shock-
capturing parameters in SUPG formulation of compressible
flows”, in Proceedings of the European Congress on Computa-
CE
tional Methods in Applied Sciences and Engineering, ECCO-
MAS 2004 (CD-ROM), Jyvaskyla, Finland, (2004).
[123] T.E. Tezduyar, “Finite elements in fluids: Stabi-
lized formulations and moving boundaries and inter-
AC
faces”, Computers & Fluids, 36 (2007) 191–206, doi:
10.1016/j.compfluid.2005.02.011.
[124] F. Rispoli, A. Corsini, and T.E. Tezduyar, “Finite ele-
ment computation of turbulent flows with the discontinuity-
capturing directional dissipation (DCDD)”, Computers & Flu-
ids, 36 (2007) 121–126, doi: 10.1016/j.compfluid.2005.07.004.
[125] T.E. Tezduyar and M. Senga, “Stabilization and shock-
capturing parameters in SUPG formulation of com-
pressible flows”, Computer Methods in Applied Me-
chanics and Engineering, 195 (2006) 1621–1632, doi:
10.1016/j.cma.2005.05.032.
[126] T.E. Tezduyar and M. Senga, “SUPG finite element com-
putation of inviscid supersonic flows with YZβ shock-
capturing”, Computers & Fluids, 36 (2007) 147–159, doi:
10.1016/j.compfluid.2005.07.009.
[127] T.E. Tezduyar, M. Senga, and D. Vicker, “Computation of
15
inviscid supersonic flows around cylinders and spheres with the
SUPG formulation and YZβ shock-capturing”, Computational
Mechanics, 38 (2006) 469–481, doi: 10.1007/s00466-005-0025-
6.
[128] K. Takizawa, T.E. Tezduyar, and Y. Otoguro, “Stabilization
and discontinuity-capturing parameters for space–time flow
computations with finite element and isogeometric discretiza-
tions”, Computational Mechanics, published online, DOI:
10.1007/s00466-018-1557-x, April 2018, doi: 10.1007/s00466-
018-1557-x.
[129] T.E. Tezduyar and D.K. Ganjoo, “Petrov-Galerkin formula-
tions with weighting functions dependent upon spatial and
temporal discretization: Applications to transient convection-
diffusion problems”, Computer Methods in Applied Mechan-
ics and Engineering, 59 (1986) 49–71, doi: 10.1016/0045-
T
7825(86)90023-X.
[130] T.E. Tezduyar and S. Sathe, “Enhanced-discretization selec-
IP
tive stabilization procedure (EDSSP)”, Computational Me-
chanics, 38 (2006) 456–468, doi: 10.1007/s00466-006-0056-7.
[131] A. Corsini, F. Rispoli, A. Santoriello, and T.E. Tezduyar, “Im-
proved discontinuity-capturing finite element techniques for re-
CR
action effects in turbulence computation”, Computational Me-
chanics, 38 (2006) 356–364, doi: 10.1007/s00466-006-0045-x.
[132] T.E. Tezduyar, S. Ramakrishnan, and S. Sathe, “Stabi-
lized formulations for incompressible flows with thermal cou-
pling”, International Journal for Numerical Methods in Flu-
[133]
ids, 57 (2008) 1189–1209, doi: 10.1002/fld.1743.
F. Rispoli, R. Saavedra, A. Corsini, and T.E. Tezduyar, “Com-
putation of inviscid compressible flows with the V-SGS sta-
bilization and YZβ shock-capturing”, International Journal
for Numerical Methods in Fluids, 54 (2007) 695–706, doi:
10.1002/fld.1447.
[134] Y. Bazilevs, V.M. Calo, T.E. Tezduyar, and T.J.R. Hughes,
“YZβ discontinuity-capturing for advection-dominated pro-
cesses with application to arterial drug delivery”, International
Journal for Numerical Methods in Fluids, 54 (2007) 593–608,
doi: 10.1002/fld.1484.
[135] A. Corsini, C. Menichini, F. Rispoli, A. Santoriello, and
T.E. Tezduyar, “A multiscale finite element formulation with
discontinuity capturing for turbulence models with dom-
inant reactionlike terms”, Journal of Applied Mechanics,
76 (2009) 021211, doi: 10.1115/1.3062967.
[136] F. Rispoli, R. Saavedra, F. Menichini, and T.E. Tezduyar,
“Computation of inviscid supersonic flows around cylinders
and spheres with the V-SGS stabilization and YZβ shock-
capturing”, Journal of Applied Mechanics, 76 (2009) 021209,
doi: 10.1115/1.3057496.
[137] A. Corsini, C. Iossa, F. Rispoli, and T.E. Tezduyar, “A DRD
finite element formulation for computing turbulent reacting
flows in gas turbine combustors”, Computational Mechanics,
46 (2010) 159–167, doi: 10.1007/s00466-009-0441-0.
[138] A. Corsini, F. Rispoli, and T.E. Tezduyar, “Stabilized finite
element computation of NOx emission in aero-engine combus-
tors”, International Journal for Numerical Methods in Fluids,
65 (2011) 254–270, doi: 10.1002/fld.2451.
[139] A. Corsini, F. Rispoli, and T.E. Tezduyar, “Computer mod-
eling of wave-energy air turbines with the SUPG/PSPG for-
mulation and discontinuity-capturing technique”, Journal of
Applied Mechanics, 79 (2012) 010910, doi: 10.1115/1.4005060.
[140] P.A. Kler, L.D. Dalcin, R.R. Paz, and T.E. Tezduyar, “SUPG
and discontinuity-capturing methods for coupled fluid mechan-
ics and electrochemical transport problems”, Computational
Mechanics, 51 (2013) 171–185, doi: 10.1007/s00466-012-0712-
z.
[141] F. Rispoli, G. Delibra, P. Venturini, A. Corsini, R. Saavedra,
and T.E. Tezduyar, “Particle tracking and particle–shock in-
teraction in compressible-flow computations with the V-SGS
stabilization and YZβ shock-capturing”, Computational Me-
chanics, 55 (2015) 1201–1209, doi: 10.1007/s00466-015-1160-
3.
 ACCEPTED MANUSCRIPT

T
IP
CR
US
AN
M
ED
PT
CE
AC

0 210
Figure 6: Velocity magnitude (m/s) at t/T = 0.1, 0.2, 0.3, 0.4, 0.5,
0.6, 0.7, 0.8, 0.9 and 1.0 (left to right and top to bottom).
61 263
Figure 7: Isosurfaces corresponding to a positive value of the second
invariant of the velocity gradient tensor, colored by the total pressure
(kPa), at t/T = 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1.0 (left
to right and top to bottom).

16
 ACCEPTED MANUSCRIPT

120

100

Efficiency (%)
80

60

40

20

T
0
0 1/8 1/4

IP
t/T

CR
Figure 10: Turbine efficiency ηIB (0, T /4, t). The black line shows the
average value.

120

US 100
Efficiency (%)

Figure 8: Boundaries used in calculating the turbine efficiency: ΓINF 80

(red) and ΓOUTF (blue).
AN
60

40
M

20

0
ED

1/4 3/8 2/4

120
t/T

Figure 11: Turbine efficiency ηIB (T /4, T /2, t). The black line shows
PT

100 the average value.

Efficiency (%)

80 120
CE

60 100
Efficiency (%)

40 80
AC

20 60

0 40
0 1/4 2/4 3/4 1
t/T 20

Figure 9: Turbine efficiency ηIB (0, T, t). The black line shows the 0
average value. 2/4 5/8 3/4
t/T

Figure 12: Turbine efficiency ηIB (T /2, 3T /4, t). The black line
shows the average value.

17
 ACCEPTED MANUSCRIPT

T
IP
CR
120

100
US
Efficiency (%)

80
AN
60

40
M

20

0
ED

3/4 7/8 1
t/T

Figure 13: Turbine efficiency ηIB (3T /4, T, t). The black line shows
PT

the average value.

CE
AC

18