Turbulence Primer

A Turbulence Primer
UGM 2000
Company Confidential
Copyright 2000 Fluent Inc. All rights reserved.

1
Outline
Background
Scales
Eliminating the small scales
Reynolds Averaging
Near Wall Turbulence
Filtered Equations
Examples
Comparison with Experiments and DNS
UGM 2000
Turbulence Models
Near Wall Treatments

2
What is Turbulence?
Unsteady, irregular (aperiodic) motion in which
transported quantities (mass, momentum,
scalar species) fluctuate in time and space
Fluid properties exhibit random variations
statistical averaging results in accountable,

turbulence related transport mechanisms.
Contains a wide range of eddy sizes (scales)

typical identifiable swirling patterns
large eddies carry small eddies.
UGM 2000

3
Two Examples of Turbulence

Homogeneous, decaying, grid-generated turbulence
Turbulent boundary layer on a flat plate
UGM 2000

4
Energy Cascade
Larger, higher-energy eddies, transfer energy to
smaller eddies via vortex stretching.
Larger eddies derive energy from mean flow.
Large eddy size and velocity on order of mean flow.
Smallest eddies convert kinetic energy into thermal

energy via viscous dissipation.
Rate at which energy is dissipated is set by rate at which

they receive energy from the larger eddies at start of
cascade.
UGM 2000

5
Vortex Stretching
Existence of eddies implies vorticity (define)
Vorticity is concentrated along vortex lines or bundles.
Vortex lines/bundles become distorted from the
induced velocities of the larger eddies.
As end points of a vortex line randomly move apart
vortex line increases in length but decreases in diameter

vorticity increases because angular momentum is nearly conserved
Most of the vorticity is contained within the smallest eddies.
Turbulence is a highly 3D phenomenon.
UGM 2000

6
Smallest Scales of Turbulence

Smallest eddy (Kolmogorov) scales:
large eddy energy supply rate ~ small eddy energy

dissipation rate e = -dk/dt.
k (u2+v2+w2) is (specific) turbulent kinetic energy [l2 / t2].

e is dissipation rate of k [l2 / t3].
Motion at smallest scales dependent upon dissipation rate,

e, and kinematic viscosity, n [l2 / t].
From dimensional analysis:
h = (n3 / e)1/4; t = (n / e)1/2; v = (ne)1/4
UGM 2000

7
Small scales vs. Large scales

Largest eddy scales:
Assume l is characteristic of larger eddy size.
Dimensional analysis is sufficient to estimate order of large
eddy supply rate of k as k / tturnover.
tturnover is a time scale associated with the larger eddies
the order of tturnover can be estimated as l / k1/2
Since e ~ k / tturnover, e ~ k3/2 / l or l ~ k3/2 / e

Comparing l with h,
l
l (k 3 / 2 / l )1/ 4
3/ 4
= 3
Re
T
( / )1/ 4
3/ 4
l
>> 1
where ReT = k1/2*l / n (turbulence Reynolds number)

UGM 2000

8
Implication of Scales
Consider a mesh fine enough to resolve smallest
eddies and large enough to capture mean flow
features.
Example: 2D channel flow
h
Ncells~(4l / h)3
or
l
l
9/4
( / )
Ncells ~ (3Ret)
H
l
where
Ret = utH / 2n
ReH = 30,800 Ret = 800 Ncells = 4x107 !
3
UGM 2000
1/ 4

9
Direct Numerical Simulation

DNS is the solution of the time-dependent NavierStokes equations without recourse to modeling.
U i
U i
p
=

+Uk
+
xk
xi xk
t
Numerical time step size required, Dt ~ t
For 2D channel example

ReH = 30,800
Number of time steps ~ 48,000
U i
x j
t 2 D Channel
0.003H
Re u
DNS is not suitable for practical industrial CFD
DNS is feasible only for simple geometries and low turbulent

Reynolds numbers
DNS is a useful research tool
UGM 2000

10
Removing the Small Scales

Two methods can be used to eliminate need to
resolve small scales:
Reynolds Averaging
Transport equations for mean flow quantities are solved.

All scales of turbulence are modeled.
Transient solution Dt is set by global unsteadiness.
Filtering (LES)
Transport equations for resolvable scales.

Resolves larger eddies; models smaller ones.
Inherently unsteady, Dt set by small eddies.
Both methods introduce additional terms that

must be modeled for closure.
UGM 2000

11
Prediction Methods
h = l/Re3/4
UGM 2000

12
RANS Modeling - Velocity Decomposition

Consider a point in the given flow field:
u
u'i
Ui
G
G
G
ui ( x , t ) = U i ( x , t ) + ui( x , t )
UGM 2000
ui
time

13
RANS Modeling - Ensemble Averaging

Ensemble (Phase) average;
G
1
U i ( x , t ) = lim
N N
u
n =1
(n )
i
G
(x, t )
Applicable to nonstationary flows such as periodic or

quasi-periodic flows involving deterministic structures.
UGM 2000

14
RANS Modeling - Time Averaging

For steady flows, long-time averaging can be
used.
u
t +T
G
G
1 0
U i ( x ) = lim ui ( x , t ) dt
T T
t0
G
G
G
ui ( x , t ) = U i ( x ) + ui( x , t )
Often referred to as Reynolds average in the

literature
UGM 2000

15
Deriving RANS Equations

Substitute mean and fluctuating velocities in
instantaneous Navier-Stokes equations and
average:
(U i + ui )
(U i + ui )
( p + p) (U i + ui )
=

+
+ (U k + uk )
x j
t
x
x
x
k
i
j
Some averaging rules:
Given f = F + f and y = Y + y
; 0; = + ; = 0; 0, etc.
Mass-weighted (Favre) averaging used for

compressible flows.
UGM 2000

16
RANS Equations
Reynolds Averaged Navier-Stokes equations:
U
U i
p
=
+
i + U k
xk
xi x j
t
U i ui u j
+
x j
x j
(prime notation dropped)
New equations are identical to original except :

The transported variables, U, , etc., now represent the
mean flow quantities
Additional terms appear: Rij = ui u j
Rij are called the Reynolds Stresses
Effectively a stress, apparent when rewritten

x j
These are the terms to be modeled.
UGM 2000
U i
ui u j
x

17
Turbulence Modeling Approaches

Boussinesq approach
isotropic
relies on dimensional analysis
Reynolds stress transport models

no assumption of isotropy
contains more physics
most complex and computationally expensive
UGM 2000

18
The Boussinesq Approach

Relates the Reynolds stresses to the mean flow by
a turbulent (eddy) viscosity, t
2 U
2
Rij = uiu j = 2t Sij t k ij kij ;
3 xk
3
1 Ui U j
Sij =
+
2 x j xi
Relation is drawn from analogy with molecular transport of

momentum.
t xy = uv = 2Sij
Assumptions valid at molecular level, not necessarily valid

at macroscopic level
t is a scalar (Rij aligned with strain-rate tensor, Sij)
Taylor series expansion valid if lmfp|d2U/dy2| << |dU/dy|

Average time between collisions lmfp / vth << |dU/dy|-1
UGM 2000

19
Modeling t
Oh well, focus attention on modeling t anyways.
Basic approach made through dimensional
arguments
Units of t = t/ are [m2/s]
Typically one needs 2 out of the 3 scales:
velocity - length - time
Models classified in terms of number of transport

equations solved, e.g.,
zero-equation
one-equation
two-equation
UGM 2000

20
Zero Equation Model

Prandtl mixing length
model:
1
2
1
v th lmfp
2
t = vmixlmix
The mixing length model:
assumes that vmix is proportional to lmix& strain rate: vmix lmix 2Sij Sij
requires that lmix be prescribed
lmix
2 Sij Sij ;
1 U i U j
+
Sij =
xi
2 x j
Relation is drawn from same analogy with molecular

transport of momentum:
=
t = l
2
mix
must be calibrated for each problem
Very crude approach, but economical
Not suitable for general purpose CFD though can be useful

where a very crude estimate of turbulence is required.
UGM 2000

21
Other Zero Equation Models

Mixing length observed to behave differently in flows
near solid boundaries than in free shear flows.
Modifications made to the Prandtl mixing length model to

account for near wall flows.
Van Driest / Clauser / Klebanoff

Reduce mixing length in viscous sublayer (inner boundary layer) with
damping factor to effect reduced mixing.
Define appropriate mixing length in velocity defect (outer boundary) layer.
Account for intermittency dependency.
Cebeci-Smith and Baldwin-Lomax
Accounts for all of above adjustments in two layer models.
Mixing length models typ. fail for separating flows.
Large eddies persist in the mean flow and cannot be modeled

from local properties alone.
UGM 2000

22
One-Equation Models
Traditionally, one-equation models were based on
transport equation for k (turbulent kinetic energy) to
calculate velocity scale, v = k1/2.
Circumvents assumed relationship between v and turbulence
length scale (mixing).
Use of transport equation allows history effects to be
accounted for.
Length scale still specified algebraically based on the

mean flow
very dependent on problem type
approach not suited to general purpose CFD
UGM 2000

23
Turbulence Kinetic Energy Equation

Exact k equation derived from sum of products of
Navier-Stokes equations with fluctuating velocities.
(Trace of the Reynolds Stress transport equations)
U i
k
+U j
= Rij
+
t
x j
x j
x j
unsteady &
convective
k 1
u i ui u j p ' u j
x j 2
molecular
production
dissipation diffusion
where =
UGM 2000
ui ui
xk xk
turbulent
transport
pressure
diffusion
(incompressible form)

24
Modeled Equation for k

The production, dissipation, turbulent transport, and
pressure diffusion terms must be modeled.
Rij in production term is calculated from Boussinesq formula.
Turbulent transport and pressure diffusion:
k
1
ui ui u j + p ' u j = t
k x j
2
Using mt/sk assumes k

can be transported by
turbulence as can U.
e = CDk3/2/l from dimensional arguments

mt = CDrk2/ e (recall mt rk1/2l)
CD, sk, and l are model parameters to be specified
Necessity to specify l limits usefulness of this model.
Advanced one equation models are complete
solves for eddy viscosity

UGM 2000

25
Spalart-Allmaras Model Equations

modified turbulent viscosity
~ ~ 1
= cb1S +
Dt
~ x j
D~
t ~f v1 ,
f v1 =
~
~
S S + 2 2 fv2 ,
d
3
3
+ cv1
3
f v2 = 1
1 + f v1
1/ 6
1+ 6
f w = g 6 cw63 ,
g + cw3
2
2
~
~
~
~
cw1 f w
( + )
+ cb 2
x j
d
x j
damping functions
g = r + cw 2 r r ,
6
~
r ~ 2 2
S d
distance from wall
Wall boundary condition :~ = 0

UGM 2000

26
Spalart-Allmaras Production Term

Default definition uses rotation rate tensor only:
S
2 ijij ;
1 U i U j
ij =
2 x j
xi
Alternative formulation also uses strain rate

tensor:
S ij + C prod min(0, Sij - ij )
reduces turbulent viscosity for vortical flows
more correctly accounts for the effects of rotation
UGM 2000

27
Spalart-Allmaras Model
Spalart-Allmaras model developed for unstructured
codes in aerospace industry
Increasingly popular for turbomachinery applications
Low-Re formulation by default
Economical and accurate for:
can be integrated through log layer and viscous sublayer to wall

Fluents implementation can also use law-of-the-wall
wall-bounded flows
flows with mild separation and recirculation
Weak for:
massively separated flows

free shear flows
simple decaying turbulence
UGM 2000

28
Two-Equation Models
Two transport equations are solved, giving two
independent scales for calculating t
Virtually all use the transport equation for the turbulent
kinetic energy, k
Several transport variables have been proposed, based
on dimensional arguments, and used for second
equation.
Kolmogorov, w:
mt rk / w, l k1/2 / w, e k / w
w is specific dissipation rate

defined in terms of large eddy scales that define supply rate of k.
Chou, e:
Rotta, l:
mt rk2 / e, l k3/2 / e
mt rk1/2l, e k3/2 / l
Boussinesq relation still used for Reynolds Stresses.

UGM 2000

29
Standard k- Model Equations

k-transport equation
Dk
=
Dt x j
t
+
+ C1 t S C2
x j k
turbulent viscosity
UGM 2000
inverse time scale
coefficients
k , , Ci , C 2
S = 2 Sij Sij
production dissipation
e-transport equation
=
Dt x j
k
2
+ t S ;
x j
Empirical constants determined from benchmark

experiments of simple flows using air and water.
t = C
k2

30
Closure Coefficients
Simple flows render simpler equations
from models.
Coefficients can be isolated and compared
with experiment.
e.g.,
Uniform flow past grid
convection and dissipation terms
2
U
= ; U
= C2
x
x
k
UGM 2000
Homogeneous Shear Flow

Near-Wall (Log layer) Flow

31
Standard k- Model
High-Reynolds number model
(i.e., must be modified for the near-wall region)
The term standard refers to the choice of

coefficients
Sometimes additional terms are included
production due to buoyancy
unstable stratification (gT >0) supports k production.
dilatation dissipation due to compressibility
added dissipation term, prevents overprediction of

spreading rate in compressible flows.
Dk
=
Dt x j
UGM 2000
k
t
k
2
t
i
x
Prt xi
RT
j
Buoyancy production
Dilatation dissipation

32
Standard k- Model Pros & Cons

Strengths:
robust
economical
reasonable accuracy for a wide range of flows
Weaknesses:
overly diffusive for many situations
flows involving strong streamline curvature, swirl, rotation,

separating flows, low-Re flows.
cannot predict round jet spreading rate
Variants of the k- model have been developed

to address its deficiencies
RNG and Realizable
UGM 2000

33
RNG k- Model Equations

Derived using renormalization group theory
scale-elimination technique applied to Navier-Stokes

equations (sensitizes equations to specific flow regimes)
k equation is similar to standard k-e model

Additional strain rate term in e equation
most significant difference between standard and RNG k-e models
Analytical formula for turbulent Prandtl numbers

Differential-viscosity relation for low Reynolds numbers
Boussinesq model used by default
-transport equation
D

2
*
=
+ C1 t S C2 where
( eff )
Dt x j
x j k
eff = + t
UGM 2000
C2* = C2
=S
C 3 1
0
+
1 + 3
0 , are coefficients
34
RNG k- Model Pros &Cons

For large strain rates:
where > 0, is augmented, and therefore k and t

are reduced
Buoyancy and compressibility terms can be

included.
Improved performance over std. k- model for
rapidly strained flows
flows with streamline curvature
Still suffers from the inherent limitations of an

isotropic eddy-viscosity model
UGM 2000

35
Realizable k- Model: Motivation

Ensures positivity of normal stresses
u 2 0
Ensures Schwarzs inequality of shear stresses
(u u )
u 2 u 2
k equation is same; new formulation for mt and e

Cm is variable
e equation is based on a transport equation for
the mean-square vorticity fluctuation
UGM 2000

36
Realizable k- Model: Cm
C is not a constant, but varies as a function of
mean velocity field and turbulence (0.09 in loglayer Sk/ = 3.3, 0.05 in shear layer of Sk/ = 6)
C along
bottom-wall
C contours for 2D backward-facing step

UGM 2000

37
Realizable k- Model: Realizability

How can normal stresses become negative?
Std. k-e Boussinesq viscosity relation:
k 2 U i U j 2
- k ij
ui u j = C
+
x j xi 3
Normal component:
k 2 U
2
u = k 2C
x
3
2
Normal stress will be negative if:

1
k U
>
3.7
x 3C
UGM 2000

38
Realizable k- Model Equations

=
Dt x j
+ C1 S C2
k +
x j

C1 = max 0.43,
, = Sk / , C2 = 1.0
+ 5
turbulent viscosity
t = C
where
UGM 2000
k2
C =
1
A0 + As
U *k
1
A0 = 4.04, As = 6 cos , = cos 1 6W
3
S ij S ji S ki ~
*
U S ij S ij + ij ij , W =
~ , S = S ij S ij
S

39
Realizable k- Model Pros & Cons

Performance generally exceeds the standard k model
Buoyancy and compressibilty terms can be
included.
Good for complex flows with large strain rates
recirculation, rotation, separation, strong p
Resolves the round-jet/plane jet anomaly
predicts the speading rate for round and plane jets
Still suffers from the inherent limitations of an

isotropic eddy-viscosity model
UGM 2000

40
Why use ? A k-L Model Case Study

Other scales can be used
Consider using L, the length scale
mt rk1/2L, e k3/2 / L
L varies linearly near the wall (unlike ~ 1/y)
L scales simply with distance
easy to resolve
Easy to use with 2-layer wall models

No production term
From the chain rule for differentiation:

1 DL 3 k 1/ 2 Dk k 3 / 2 D
=
2
C L Dt 2 Dt Dt
We can use this to transform the k- model
UGM 2000

41
k-L Model from the Std. k- Model

L-transport equation
1.5 - 1.44 ~ 0
t L
x
j

t 2 L L 3 1
1 t L k
+

L x j x j 2 k x j k k x j
DL 3
L
3
= C1 t S 2 C L C2 k 1/ 2 +
Dt 2
x j
k
2
3
9
+
2 k 4
L k k 1
1 t k L
t 2
+

k x j x j k k x j x j
Note additional cross-derivative terms

Can we just neglect them?
UGM 2000

42
Smiths k-L Model

k-L model of Smith (1994) (high-Re form)
k-transport equation
k 3/ 2
Dk
= t S CL
+
Dt
L
x j
L-transport equation
t k
k x
j
k L
DL
= C1 k 1/ 2 + C2 t
+
Dt
k x j x j x j
coefficients
t L
x
j

C1 , C2 , C L , C , k , L
turbulent viscosity
t = C 1 / 4 k 1 / 2 L
UGM 2000

43
k-L Model Pros & Cons

k-L model is quite successful for compressible wallbounded shear flows
But
Model performs poorly for free shear flows

jets (round and plane)
far wakes
mixing layers
radial jets
Sensitive to inlet and far-field boundary values of L

The k- model does not have this problem!
k- is robust to inlet and far-field values of

UGM 2000

44
k- Model of Wilcox
Wilcoxs k- model is a popular alternative to k-e
w~e/k
mt rk / w
Initially found to be quite sensitive to inlet and far-field

boundary values of w
Can be used in near-wall region without modification
Latest version contains several refinements:
reduced sensitivity to boundary conditions
modifcation for the round-jet/plane-jet anomaly
compressibility effects
low-Re (transitional) effects
UGM 2000

45
Faults in the Boussinesq Assumption

Boussinesq: Rij = 2mtSij
Is simple linear relationship sufficient?
Rij is strongly dependent on flow conditions and history

Rij changes at rates not entirely related to mean flow processes
Rij is not strictly aligned with Sij for flows with:
sudden changes in mean strain rate

extra rates of strain (e.g., rapid dilatation, strong streamline
curvature)
rotating fluids
stress-induced secondary flows
Modifications to two-equation models cannot be

generalized for arbitrary flows.
UGM 2000

46
Reynolds Stress Models

Starting point is the exact transport equations for
the transport of Reynolds stresses, Rij.
six transport equations in 3d
Equations are obtained by Reynolds-averaging

the product of the exact momentum equations
and a fluctuating velocity.
uiNS (u j ) + u j NS (ui ) = 0
The resulting equations contain several terms

that must be modeled.
UGM 2000

47
Reynolds Stress Transport Equations

Reynolds Stress
Transport Eqns.
DR ij
Dt
= Pij + ij ij +
J ijk
x k
Generation
U j
Ui
+ u j uk
Pij uiuk
x
k
k
(computed)
Pressure-Strain
Redistribution
ui u j
ij p
+
x
j xi
(modeled)
Dissipation
u i u j
ij 2
x k x k
(related to e)
Turbulent
Diffusion
J ijk pui jk + pu j ik + ui u j u k

(modeled)
Pressure/velocity
fluctuations
Turbulent
transport
(equations written for incompressible flow w/o body forces)
UGM 2000

48
Dissipation Modeling
Dissipation rate is predominantly associated
with small scale eddy motions.
Large scale eddies affected by mean shear.
Vortex stretching process breaks eddies down into
continually smaller scales
The directional bias imprinted on turbulence by mean flow is

gradually lost.
Small scale eddies assumed to be locally isotropic.
2
ij = ij
3
e is calculated with its own (or related) transport equation.
Compressibility and near-wall anisotropy effects can be
accounted for.
UGM 2000

49
Turbulent Transport and

Pressure Diffusion
Most closure models combine the pressure diffusion
with the triple products and use a simple gradient
diffusion hypothesis.
k
u u
C s u k ul i j

xl
C k 2 ui u j
Or even a simpler model =
xk xk
( )
p'
ui u j u k + ( kj ui + ik u j )
ui u j =
xk
xk
xk
Overall performance of models for these terms is

generally inconsistent based on isolated comparisons to
measured triple products.
DNS data indicate that above p terms are negligible.
UGM 2000

50
Pressure-Strain Modeling
Pressure-strain term of same order as production
Pressure-strain term acts to drive turbulence p ui + u j
x x
towards an isotropic state by redistributing the ij
i
j
Reynolds stresses
Decomposed into parts ij = ij,1 + ij,2 + ij, w
mean
Slow Part
u j ui u j
1 2 p
= u i
+
xi xi
xi x j xi
Rapid Part
u U j
1 2 p
= 2 i
xi xi
x j xi
gradient
Model of Launder, Reece & Rodi (1978)
U j
U l
U i 2
ij = c1bij + c 2 u i u l
+ u j ul
3 u l u m x ij
x
x
m
l
l
where bij
ui u j k ij
k
UGM 2000

51
Pressure-Strain Modeling Options

Wall-reflection effect
contains explicit distance from wall
damps the normal stresses perpendicular to wall
enhances stresses parallel to wall
SSG (Speziale, Sarkar and Gatski) Pressure

Strain Model
Expands the basic LRR model to include non-linear
(quadratic) terms
Superior performance demonstrated for some basic
shear flows
plane strain, rotating plane shear, axisymmetric

expansion/contraction
UGM 2000

52
Characteristics of RSM
Effects of curvature, swirl, and rotation are directly
accounted for in the transport equations for the
Reynolds stresses.
When anisotropy of turbulence significantly affects

the mean flow, consider RSM.
More cpu resources (vs. k- models) is needed
50-60% more cpu time per iteration and 15-20%

additional memory
Strong coupling between Reynolds stresses and

the mean flow
number of iterations required for convergence may

increase.
UGM 2000

53
Heat Transfer
The Reynolds averaging process produces an
additional term in the energy equation: ui
Analogous to the Reynolds stresses, this is termed

the turbulent heat flux
It is possible to model a transport equation for the heat

flux, but this is not common practice
Instead, a turbulent thermal diffusivity is defined
proportional to the turbulent viscosity
The constant of proportionality is called the turbulent Prandtl
number
Generally assumed that Prt ~ 0.85-0.9
Applicable to other scalar transport equations.

UGM 2000

54
Importance of Near-Wall Turbulence

Walls are main source of vorticity and turbulence.
Accurate near-wall modeling is important for most
engineering applications.
Successful prediction of frictional drag for external flows,
or pressure drop for internal flows, depends on fidelity of
local wall shear predictions.
Pressure drag for bluff bodies is dependent upon extent
of separation.
Thermal performance of heat exchangers, etc., are
determined by wall heat transfer whose prediction
depends upon near-wall effects.
UGM 2000

55
Near-Wall Modeling Issues

k-e and RSM models are valid in the turbulent core
region and through the log layer.
Some of the modeled terms in these equations are

based on isotropic behavior.
Isotropic diffusion (mt/s)

Isotropic dissipation
Pressure-strain redistribution
Some model parameters based on experiments of isotropic
turbulence.
Near-wall flows are anisotropic due to presence of walls.
Special near-wall treatments are necessary since

equations cannot be integrated down to wall.
UGM 2000

56
Flow Behavior in Near-Wall Region

U/ut
Velocity profile exhibits layer structure

identified from dimensional analysis
Inner layer
k production and dissipation are

nearly equal in overlap layer
viscous forces rule, U = f(r, tw, m, y)
Outer layer
dependent upon mean flow
turbulent equilibrium
Dissipation >> production in

sublayer region
Overlap layer
log-law applies
UGM 2000

57
Near-Wall Modeling Approaches

Wall Function
Approach
Two-Layer Zonal/
Low-Re Approaches
The viscosity affected region is not

resolved.
The near wall region is resolved

down to the wall.
Wall-adjacent cell centroid assumed

to be in log layer (y+ = 30).
Cell center information bridged by
wall function.
High Re turbulence models can be
used throughout.
UGM 2000
Wall-adjacent cell assumed to be in

sublayer (y+ = 1).
Low-Re models are valid throughout.
Two-Layer Zonal Model
Simpler turbulence model used in

viscosity affected region.
58
Wall Functions
Wall functions consist of wall laws for mean velocity and

temperature and formulas for turbulent quantities.
Universal Wall Laws
Formulas for k and e
Viscous sublayer
dimensional analysis (r, t, m, y)
U = utf(y+)
(tw/r)1/2
ut =
y+ = yut/n
k = ut2/Cm1/2
e = ut3/ky
Precludes transport of
turbulence in log layer
U = Ue - utg(h)
Local turbulent equilibrium
Clauser defect layer
k and e functions of ut only
h = y/D
Overlap layer
large scale variance (n /ut<< D)
utf(y+) = Ue - utg(h)
ut2f(y+)/n = -utg(h)/D or ( x y/ut )
y+f(y+)/n = -hg(h)
U
1
u + = ln( y + ) + C ; u + =
UGM 2000

59
Standard Wall Functions

Fluent uses Launder-Spalding Wall Functions
U = U(r, t, m, y, k )
Introduces additional velocity scale for general application

y* < yv*
U = y
C 1/ 4k 1P/ 2 yP
UP C 1/ 4k 1P/ 2
for
where U
y
y* > yv*
U = 1 ln Ey
w /
Estimating k and e at wall-adjacent cells
Generally, k is obtained from solution of k transport

equation
Cell center is immersed in log layer
Local equilibrium (production = dissipation) prevails
kn = 0 at surface
e also estimated from local equilibrium assumptions
e = Cm3/4k3/2/ky
Wall
less reliable when cell intrudes viscous
UGM functions
2000
sublayer.

60
Limitations of Standard
Wall Functions
Wall functions become less reliable when:
local equilibrium assumption fails
Transpiration through wall

strong body forces
highly 3D flow
rapidly changing fluid properties near wall
low-Re flows are pervasive throughout model

small gaps are present
UGM 2000

61
Non-equilibrium Wall Functions

Log-law is sensitized to
pressure gradient for better
prediction of adverse pressure
gradient flows and separation.
Relaxed local equilibrium
assumptions for TKE in wallneighboring cells.
~
U C1/ 4 k1/ 2 1 C1/ 4 k1/ 2 y
= ln E

w /
where
Rij, k, e are estimated in each

region and used to determine
average e and production of k.
y
y
y
~
dp
y
v
v
v
1
U =U
ln +
+
2 dx k 1/ 2 yv k 1/ 2
UGM 2000

62
Two-Layer Zonal Model

Approach is to divide flow domain into two
regions
Viscosity affected near-wall region
Fully turbulent core region
Use different turbulence models for each region.

One-equation (k) near-wall model for the viscosity
affected near-wall region.
High-Re k-e or RSM models for turbulent core region.
Wall functions, and their limitations, are

avoided.
UGM 2000

63
Two-Layer Zones
The two regions are
demarcated on a cellby-cell basis:
Rey > 200
turbulent core region
Rey < 200
viscosity affected region
Rey = rk1/2y/m
y is shortest distance to
nearest wall
zoning is dynamic and
solution adaptive
UGM 2000

64
Two-Layer Zones
In the turbulent core region, the selected high-Re
turbulence model is used.
In the viscosity-affected region, a one-equation
model is used.
k equation is same as high-Re model.
Length scale used in evaluation of mt is not from e.
mt = rCmk1/2lm
lm = cly(1-exp(-Rey/Am)
cl = kCm3/4
Dissipation rate, e, is calculated algebraically
e = k3/2/le
le = cly(1-exp(-Rey/Ae)
UGM 2000

65
Damping-Function Low-Re Models

Std. k- model modified by damping functions:
f , f1 , f 2
=
Dt x j
+ ( f1C1 t S f 2 C2 )
x j k
turbulent viscosity
t = f C
k2
k and equations solved on fine mesh (required)

right to the wall
UGM 2000

66
Typical Damping Functions

Damping functions written in terms of
Reynolds numbers:
k 2
Re t =
ky
( / )1 / 4 y
; Re y =
; Re =
e.g., Abids model:
f = tanh(0.008 Re y ) 1 + 4 Ret
3 / 4
f1 = 1
2
Ret 2
Re y
1 exp
f 2 = 1 exp
9
12
36
UGM 2000

67
Low-Re k- Models
Several full low-Re k- models now available
Lam Bremhorst
Launder-Sharma
Abid
Chang et al.
Abe-Kondo-Nagano
Enables modeling of low-Re effects including

transitional flows
Implementations are problem specific
Features are not visible in GUI
Access from TUI
UGM 2000

68
V2F low-Re k-e Model

Durbin (1990) suggests that wall normal
fluctuations, v 2, are responsible for near-wall
transport
v 2 behaves quite differently to u 2 and w2
attenuation of v 2 is a kinematic effect
2
damping of u is a dynamic effect
Model t ~ v 2T instead of t ~ kT
Requires two additional transport equations:
equation for wall-normal fluctuations, v 2
equation for an elliptic relaxation function, f
UGM 2000

69
V2F k-e Model Equations

Dt x j
t
+
v2 -transport equation
Dv 2
=
x j
Dt
t
+
k
+ (C1 t S 2 C2 )
x j T
v 2
2
+ k f v
k
x j
relaxation equation
2
2
2
2
(
)
S
f
C
v
N
v
2
f L2
= 1 + C2 t +
k
kT
x j x j T 3 k
scales
k3
k
3
2
2
T = max ,6 ; L = C L max 2 , C
UGM 2000

70
V2F k- Model Pros and Cons

Very promising results for a wide range of flow
and heat transfer test cases
at least as good as the best of the damping function

approaches in most test cases
Still an isotropic eddy-viscosity model
Can be extended for RSM
Needs 2 additional equations, so requires more

memory and CPU than damping functions
UGM 2000

71
Large Eddy Simulation (LES)

Recall: Two methods can be used to eliminate
the need to resolve small scales.
Reynolds Averaging Approach
Periodic and quasi-periodic unsteady flows
Filtering (LES)
Transport equations are filtered such that only larger eddies

need be resolved.
Difficult to model large eddies since they are
Smaller eddies are modeled.
anisotropic
subject to history effects
dependent upon flow configuration, boundary conditions, etc.
Typically isotropic and so more amenable to modeling.
Random unsteadiness can be resolved.
UGM 2000

72
Filtering
In finite-volume schemes, the cell size in a mesh can
ui
determine the filter width.
u
ui
e.g., in 1-D,
ui ( x + ) ui ( x ) d 1 x +
=
u
(
)
d
i
2
dx 2 x
xi
D
more or less information is filtered as D is varied.
In general,
1
G
ui ( x , t )
V
G G G
V ui , t; x d , V
where V is the volume of cell
where the subgrid scale (SGS) velocity, ui = ui ui

Resolvable-scale filtered velocity
UGM 2000

73
LES - Governing Equations

The governing equations for LES are obtained by
filtering (space-averaging) Navier-Stokes equations:
ui
=0
xi
ui (ui u j )
1 p
ui ij
=
+
+
xi x j x j x j
t
x j
ij ui u j ui u j
The SGS stresses consist of terms that must be

modeled:
ui u j = ui u j + ui u j + uiu j + uiu j
UGM 2000

74
SGS Stress Modeling

The subgrid-scale stress is modeled by;
1
ij kk ij 2 Sij ;
3
1 ui u j
Sij =
+
2 x j xi
The subgrid-scale eddy viscosity nD is modeled by:
Smagorinskys subgrid-scale model

(CS )2 2Sij Sij
Smagorinsky constant varied from flow to flow
RNG-based subgrid-scale model

1/ 3
eff
s 2 eff
, S (C RNG )2 2 S ij S ij
+ 1 + H
C
3
UGM 2000

75
Grid and Time Step Size

Requires 3-D
transient
modeling
Spatial and
log E
temporal
resolution of
scales in inertial
subrange
Not easily
defined
l@
(10nk/e)1/2
t @ l/U
UGM 2000
Kolmogorov scale
l >> lc >> >>
Large
eddy
scale
Cell
size
Energy
containing
eddies
Taylor
scale
Inertial
subrange
Dissipating
eddies
-5/3
Kl 1 / l
K c 1 / lc
K d 1 / Re 3 / 4 / l
76
LES Example - Dump Combustor

A 3-D model of a lean premixed combustor
studied by Gould (1987) at Purdue University
Non-reacting (cold) flow was simulated with a
170K cell hexahedral mesh using second-order
temporal and spatial discretization schemes.
UGM 2000
Iso-surface of instantaneous
vorticity magnitude colored by
velocity angle

77
LES Examples - Dump Combustor

Simulation
done for
Mean axial velocity

prediction at x/h = 5;
Re d = 105 (Re 150 )
Computed
using RNGbased subgridscale model
Mean axial velocity at x/h = 5

UGM 2000

78
LES Examples - Dump Combustor

RMS velocities predictions at x/h = 10;
UGM 2000

79
Fluent 6: Turbulence Models

Customizability enhancement to turbulence
modeling via UDFs
BCs for turbulence equations
source terms (e.g., production and dissipation
turbulent viscosity, turbulent Prandtl/Schmidt
numbers
k- model (Wilcox, SST)

Enhanced wall function option
Aimed at removing y+ constraint
Accounts for the effects of pressure gradient,
transpiration, and compressibility
UGM 2000

80
Summary
CFD codes solve the Reynolds average NavierStokes equations
Closure problem requires Reynolds stresses to be
modeled
Boussinesq models:
zero equation
one-equation
works quite well for some classes of flows
two-equation (e.g. k-)
seldom used for complex problems
work-horse of industrial CFD
Reynolds stress models (7 equations in 3d)
needed for complex, anisotropic flows
UGM 2000

81
Summary
Near-wall modeling options
wall functions
two-layer
do not resolve viscous sublayer and buffer

still can give surprisingly good results
resolves sublayer, so computationally expensive
not a true low-Re formulation
low-Re models
resolves sublayer, so computationally expensive

mostly use ad hoc damping functions
V2F is first low-Re model with potential for being universal
UGM 2000

82
Turbulent Flow Examples
UGM 2000

83
2D Backstep
Driver & Seegmiller (1985)

UGM 2000
ReH = 37,400

84
2D Backstep: Skin Friction Distribution
X/H
151x91 mesh with 10 nodes less than y+=10

UGM 2000

85
Low-Re Backstep
Re = 5,100
Comparison with DNS data of Le and Moin (1994)
Comparison of std. k- + 2-layer, Yang-Shih lowRe model and V2F low-Re model
UGM 2000

86
Low-Re Backstep
Cp
Cfx
Pressure coefficient and x-component of skin friction

2-layer model less accurate than V2F and Yang-Shih
UGM 2000

87
Low-Re Backstep
X/h = 1
X/h = 3
X
Velocity
Velocity
X/h = 7
X/h = 5
X
Velocity
UGM 2000
X
Velocity

88
Low-Re Backstep
X/h = 3
Y
Velocity
Velocity
X/h = 1
X/h = 5
X/h = 7
Y
Velocity
UGM 2000
Y
Velocity

89
Low-Re Backstep
Contours of Rey < 200
e and nt prescribed algebraically for 2-layer model in

region where Rey < 200
For low Re, much of the flow is in this region
2-layer model is not always a good substitute for a lowRe model
UGM 2000

90
2D U-Bend
Comparison with exp. data of Monson et al. (1990)
UGM 2000

91
Streamwise Velocity Comparisons
r*
r*
= 180
=0
U/Uref
U/Uref
r*
= 90
UGM 2000
U/Uref

92
Pressure Coefficients
Cp
Cp
Inner
Wall
Outer
Wall
S/H
UGM 2000
S/H

93
Stream Function Contours
Spalart-Allmaras
std k-
RNG k-
RSM
UGM 2000

94
Lessons from 2-D U-Bend

Only the RSM correctly predicts the effects of
streamline curvature
Std. k-e does not predict any separation
RNG k-e predicts slight separation
Both RSM and Spalart-Allmaras predict
significant separation
UGM 2000

95
Turbulent Vortex Breakdown
Comparison with experimental data of

Sarpkaya (1999)
2D axisymmetric calculation
Simulation courtesy of R. Spall, Utah State
University
UGM 2000

96
Comparisons of Axial Velocity Profiles
x/r0 = 8.3
x/r0 = 5
Axial
Velocity
Axial
Velocity
r/r0
r/r0
UGM 2000

97
Comparisons of Swirl Velocity Profiles
x/r0 = 5
Swirl
Velocity
x/r0 = 8.3
Swirl
Velocity
r/r0
r/r0
UGM 2000

98
Lessons from Turbulent Vortex

Breakdown
k-e model cannot predict vortex breakdown
in high strain rates, turbulent kinetic energy
increases and increases turbulent viscosity
RNG k-e model is better (additional strain-rate term,
and an ad hoc swirl correction, reduce the turbulent
viscosity) but not acceptable
RSM results show significant improvement for

this and many other swirling flow cases
UGM 2000

99
Axisymmetric Afterbody
Compared to experiment of Huang et al.

Spalart-Allmaras, std k-, RNG k-, realizable k-
and RSM
Run on both coarse and fine quad meshes
wall functions on coarse meshes
2-layer formulation on fine mesh with k- and RSM
low-Re formulations investigated
UGM 2000

100
Spalart-Allmaras
model (fine mesh)
Std. k- model +
2-layer (fine mesh)
No separation
on afterbody
UGM 2000

101
Pressure coeff. on coarse mesh (y+ ~ 40) using wall functions
Model
Separates?
Std k-
n
RNG k-
n
Real. k-
y
RSM
y
S-A
y
Cp
UGM 2000

102
Pressure coeff. on fine mesh (y+ ~ 0.5) using two-layer model
Model
Separates?
Std k-
n
RNG k-
n
Real. k-
y
RSM
y
S-A
y
Cp
UGM 2000

103
Spalart-Allmaras gives consistent results on both
meshes
Separation not predicted by std. k- on either mesh
RSM separates on both meshes
Cp on body somewhat overpredicted on coarse mesh
Wall reflection term, or quadratic pressure-strain term,

necessary to obtain coarse mesh separation
Subtle separation illustrates effect of near-wall

treatment
Realizable k- has smaller separation bubble on fine mesh
Difficult to get grid-independent solutions using wall

functions --- would a low-Re formulation work?
UGM 2000

104
Pressure coeff. on fine mesh using Low-Re k-e models
Model
Separates?
V2F
y
Abid
n
Launder-Sharma
n
Yang-Shih
n
Abe-Kondo-Nagano n
Chang-Hsieh-Chen n
Cp
UGM 2000
Position (m)

105
Low-Re models using damping functions do
not predict the separation
Durbins V2F (4-equation) model predicts
separation
UGM 2000

106
Turbulent Heat Transfer

Over a Blunt Plate
Ota & Kan

UGM 2000
151x75 quad mesh


107
Blunt Plate
The standard k- model gives spuriously large
turbulent kinetic energy on the front face,
underpredicting the size of the recirculation.
Reynolds-Stress model
(exact)
Standard k- model
Contours of TKE production

UGM 2000

108
Blunt Plate
Predicted separation bubble
Standard k-
Realizable k-
Experimentally observed
reattachment point is at x/d = 4.7
UGM 2000

109
Heat Transfer Over a Blunt Plate
noneq
noneq
UGM 2000

110
HSVA Ship Model

Measured in an wind
tunnel with a double-body
model
ReL = 5 x 106
70x41x56 hex mesh was
used.
UGM 2000
Computed using different

turbulence models with
the standard wall functions
Girthwise friction-velocity
at x/L = 0.945

111
HSVA Ship Model

U-contours at x/L = 0.974
SKE
RKE
Measured
RSM
RNG
UGM 2000

112
Flow in a Rotating Channel

Represents flows through
Flow configuration;
rotating internal passages
(e.g. turbomachinery
Johnston et al. (1972)
applications)
Rotation affects mean axial
momentum equation through
turbulent stresses.
ReH = 11,500
Rotation makes mean axial
velocity asymmetrical.
Ro = 0.21
Computations are carried
out using SKE, RNG, RKE
and RSM models are with
the standard wall functions.
UGM 2000

113
Flow in a Rotating Channel

Predicted axial velocity profiles (ReH = 11.500, Ro = 0.21)
UGM 2000

114
NACA 0012 Airfoil

Transonic flow (Ma =
0.8, Rec = 9x106) with
boundary-layer
separation
Cp predictions by standard wall functions
a = 2.26 degrees
Computed using
different one-and twoequation turbulence
models (S-A, SKE,
RNG, RKE) with
standard wall functions
UGM 2000

115
NACA 0012 Airfoil

Cp predictions by RNG
Predictions of B.L.
model with nonseparation
equilibrium wall functions
RNG + Standard wall functions
RNG + Non-eq. wall functions
Spalart-Allmaras
UGM 2000

116
Flow in a Cyclone
0.1 m
40,000 cell hexahedral

mesh
High-order upwind
0.12 m
scheme was used.
Uin = 20 m/s Computed using SKE,
RNG, RKE and RSM
models with the
standard wall functions
Represents highly
swirling flows (Wmax =
1.8 Uin)
0.97 m
0.2 m
UGM 2000

117
Flow in a Cyclone
Tangential velocity profile predictions at 0.41 m
below the vortex finder
UGM 2000

118

Turbulence Primer

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Turbulence Primer

Uploaded by

Copyright:

Available Formats

A Turbulence Primer

Copyright 2000 Fluent Inc. All rights reserved.

Comparison with Experiments and DNS

Copyright 2000 Fluent Inc. All rights reserved.

statistical averaging results in accountable,

Contains a wide range of eddy sizes (scales)

Copyright 2000 Fluent Inc. All rights reserved.

Two Examples of Turbulence

Turbulent boundary layer on a flat plate

Copyright 2000 Fluent Inc. All rights reserved.

Smallest eddies convert kinetic energy into thermal

Rate at which energy is dissipated is set by rate at which

Copyright 2000 Fluent Inc. All rights reserved.

As end points of a vortex line randomly move apart

vortex line increases in length but decreases in diameter

Most of the vorticity is contained within the smallest eddies.

Turbulence is a highly 3D phenomenon.

Copyright 2000 Fluent Inc. All rights reserved.

Smallest Scales of Turbulence

large eddy energy supply rate ~ small eddy energy

k (u2+v2+w2) is (specific) turbulent kinetic energy [l2 / t2].

Motion at smallest scales dependent upon dissipation rate,

h = (n3 / e)1/4; t = (n / e)1/2; v = (ne)1/4

Copyright 2000 Fluent Inc. All rights reserved.

Small scales vs. Large scales

the order of tturnover can be estimated as l / k1/2

Since e ~ k / tturnover, e ~ k3/2 / l or l ~ k3/2 / e

where ReT = k1/2*l / n (turbulence Reynolds number)

Copyright 2000 Fluent Inc. All rights reserved.

Copyright 2000 Fluent Inc. All rights reserved.

Direct Numerical Simulation

Numerical time step size required, Dt ~ t

For 2D channel example

DNS is not suitable for practical industrial CFD

DNS is feasible only for simple geometries and low turbulent

Copyright 2000 Fluent Inc. All rights reserved.

Removing the Small Scales

Transport equations for mean flow quantities are solved.

Transport equations for resolvable scales.

Both methods introduce additional terms that

Copyright 2000 Fluent Inc. All rights reserved.

Copyright 2000 Fluent Inc. All rights reserved.

RANS Modeling - Velocity Decomposition

Copyright 2000 Fluent Inc. All rights reserved.

RANS Modeling - Ensemble Averaging

Applicable to nonstationary flows such as periodic or

Copyright 2000 Fluent Inc. All rights reserved.

RANS Modeling - Time Averaging

Often referred to as Reynolds average in the

Copyright 2000 Fluent Inc. All rights reserved.

Deriving RANS Equations

Some averaging rules:

Mass-weighted (Favre) averaging used for

Copyright 2000 Fluent Inc. All rights reserved.

(prime notation dropped)

New equations are identical to original except :

Rij are called the Reynolds Stresses

Effectively a stress, apparent when rewritten

These are the terms to be modeled.

Copyright 2000 Fluent Inc. All rights reserved.

Turbulence Modeling Approaches

Reynolds stress transport models

Copyright 2000 Fluent Inc. All rights reserved.