Numerical Methods in Thermo-Fluid Dynamics I

Simplification of Conservation Eqs. up to

Boundary Layer Equations

Dr.-Ing. Manuel Münsch

Friedrich-Alexander Universität Erlangen-Nürnberg
Lehrstuhl für Strömungsmechanik

Erlangen
 Erlangen

Schedule: Lectures
No. Week(s) Practical Lecture Exercise
Monday: 10:00-11:30 Monday: 14:15-15:45 Wednesday: 10:15-11:45
Room: online Room: online Room: online
1. 02.11.20-06.11.20 Matlab Tutorial I Introduction and Navier-Stokes Eq. No Exercise
2. 09.11.20-13.11.20 Matlab Tutorial II; Computer Exercise 1 Simplification of NSE up to Boundary Layer Eq. Hand out Exercise 1
3. 16.11.20-20.11.10 Work on Computer Ex. 1 Initial Value Problems: Time Stepping (ODEs) Work on Exercise 1
4. 23.11.20-27.11.20 Work on Computer Ex. 1 Finite Difference Method 1 Presentation Ex. 1 &
Hand out Ex. 2
5. 30.11.20-04.12.20 Work on Computer Ex. 1 Finite Difference Method 2 Work on Ex. 2
6. 07.12.20-11.12.20 Work on Computer Ex. 1 Consistency, Stability and Convergence Presentation Ex. 2 &
Hand out Ex. 3 + 4
7. 14.12.20-18.12.20 Submit & Present Computer Ex. 1 Finite Volume Method 1 Work on Ex. 3 + 4
Computer Ex. 2
8. 21.12.20-23.12.20 Work on Computer Ex. 2 Finite Volume Method 2 Work on Ex. 3 + 4
24.12.20-06.01.21 Holiday Holiday Holiday
07.01.21-08.01.21 Presentation Ex. 3 + 4 &
Hand out Ex. 5
9. 11.01.21-15.01.21 Submit & Present Computer Ex. 2 Numerical solution of the Navier-Stokes-Equations 1 Work on Ex. 5
Computer Ex. 3
10. 18.01.21-22.01.21 Work on Computer Ex. 3 Numerical solution of the Navier-Stokes-Equations 2 Work on Ex. 5
11. 25.01.21-29.01.21 Work on Computer Ex. 3 Numerical solution of the Navier-Stokes-Equations 3 Presentation Ex. 5 &
Hand out Ex. 6
12. 01.02.21-05.02.21 Work on Computer Ex. 3 Numerical solution of the Navier-Stokes-Equations 4 Work on Ex. 6
Hand out Ex.6
13. 08.02.21-12.02.21 Submit & Present Computer Ex. 3 Boundary Conditions Presentation Ex. 6

Dr.-Ing. Manuel Münsch | FAU | 2

 Erlangen

Outline

• PART 1: Repetition
• Navier-Stokes Equations

• PART 2: Special Types of Conservation Equations

• Conservation Equations for incompressible fluids
• Non-dimensional Conservation Equations
• Stokes Equations
• Euler Equations
• Boundary Layer Equations

Dr.-Ing. Manuel Münsch | FAU | 3

PART 1: Repetition

Erlangen
 Erlangen

Conservation Equations
System of Conservation Equations
Mass:
∂ρ ∂(ρui )
+ =0 (1)
∂t ∂ xi
Momentum:
∂(ρui ) ∂(ρui uj ) ∂ p ∂τij
+ =− + + ρgi (2)
∂t ∂ xj ∂ xi ∂ xj
For a Newtonian Fluid:
 
∂ ui ∂ uj 2 ∂ uk
τij = µ + − µδij (3)
∂ xj ∂ xi 3 ∂ xk
Energy: ideal gas with p = ρRT
   
∂(ρT ) ∂(ρui T ) ∂ ∂T ∂ ui ∂ uj
cv + = λ −p − τij (4)
∂t ∂ xi ∂ xi ∂ xi ∂ xi ∂ xi

Dr.-Ing. Manuel Münsch | FAU | 5

 Erlangen

Conservation Equations
System of Conservation Equations
Remarks:
• most general form of the conservation equations
• non-linear due to the second term on the LHS of equation (2)
• no analytic solution available
• numerical solution required (How?-> we will see later)
• thus, some useful simplifications have to be discussed

Dr.-Ing. Manuel Münsch | FAU | 6

PART 2: Special Types of Conservation Equations

Erlangen
 Erlangen

Incompressible Newtonian Fluid

Assumption: ρ = const . and ν , λ, cv independent from
temperature changes
Mass:
∂ ui
=0 (5)
∂ xi
Momentum:  
∂ ui ∂(ui uj ) ∂ p ∂τij
ρ + =− + (6)
∂t ∂ xj ∂ xi ∂ xj
with:  
∂ ui ∂ uj
τij = µ + (7)
∂ xj ∂ xi
and:
p := p + ρgi xi (8)
• if ρ = const . and gi = const . gravitation term ρgi is just leading to
additional hydrostatic pressure
• the gravitational force does not change the dynamics
Dr.-Ing. Manuel Münsch | FAU | 8
 Erlangen

Incompressible Newtonian Fluid

Assumption: ρ = const . and ν , λ, cv independent from
temperature changes
Energy: ideal gas with p = ρRT
   
∂ T ∂(ui T ) ∂ ∂T ∂ uj
cv ρ + = λ − τij (9)
∂t ∂ xi ∂ xi ∂ xi ∂ xi
• term p ∂∂uxi = 0 due to equation (5)
i

Dr.-Ing. Manuel Münsch | FAU | 9

 Erlangen

Incompressible, isothermal Newtonian Fluid

Assumption: ρ = const . and T = const .
-> Energy equation can be neglected
-> Thus, only mass and momentum conservation are considered
Mass:
∂ ui
=0 (10)
∂ xi
Momentum:  
∂ ui ∂(ui uj ) ∂ p ∂τij
ρ + =− + (11)
∂t ∂ xj ∂ xi ∂ xj
with:  
∂ ui ∂ uj
τij = µ + (12)
∂ xj ∂ xi
and:
p := p + ρgi xi (13)

Dr.-Ing. Manuel Münsch | FAU | 10

 Erlangen

Incompressible, isothermal Newtonian Fluid

Component-wise notation:
Mass:
∂u ∂v ∂w
+ + =0 (14)
∂x ∂y ∂z
Momentum:

2 2 2
   
∂u ∂u ∂u ∂u ∂p ∂ u ∂ u ∂ u
ρ +u +v +w =− +µ 2
+ 2+ 2
∂t ∂x ∂y ∂z ∂x ∂x ∂y ∂z
   2 2 2

∂v ∂v ∂v ∂v ∂p ∂ v ∂ v ∂ v
ρ +u +v +w =− +µ 2
+ 2+ 2 (15)
∂t ∂x ∂y ∂z ∂y ∂x ∂y ∂z
 2
∂ w ∂ 2w ∂ 2w
  
∂w ∂w ∂w ∂w ∂p
ρ +u +v +w =− +µ + + 2
∂t ∂x ∂y ∂z ∂z ∂x 2 ∂y 2 ∂z

Dr.-Ing. Manuel Münsch | FAU | 11

 Erlangen

Dimensionless Governing Differential Equations

Assume the relevant equations to describe the fluid flow connected to some
arbitrary fluid mechanics problem are given by the conservation equations
(mass and momentum) for a compressible fluid (Newtonian):
• Mass:
∂ρ ∂(ρui )
+ =0 (16)
∂t ∂ xi
• Momentum:
∂(ρuj ) ∂(ρui uj ) ∂ p ∂τij
+ =− + + ρgj (17)
∂t ∂ xi ∂ xj ∂ xi
• for a Newtonian Fluid:
 
∂ uj ∂ ui 2 ∂ uk
τij = µ + − µδij (18)
∂ xi ∂ xj 3 ∂ xk
• in addition, initial and boundary conditions are assumed to be known
• the variables of the equations / the problem are: xi , t, uj , ρ, gi , µ, p

Dr.-Ing. Manuel Münsch | FAU | 12

 Erlangen

Dimensionless Governing Differential Equations

Derivation:
• to derive dimensionless numbers characteristic reference values are
introduced for each of the upper variables
• the characteristic reference values depend on the flow field, i.e., the
problem definition
• here we choose: Lc , tc , uc = u∞, ρc , gc , µc , ∆pc
• from this the following dimensionless variables are obtained:
xi gj
xi∗ = => xi = Lc · xi∗ gj∗ = => gj = gc · gj∗
Lc gc
∗ t ∗ ∗ µ
t = => t = tc · t µ = => µ = µc · µ∗
tc µc
uj ρ
uj∗ = => uj = uc · uj∗ ρ∗ = => ρ = ρc · ρ∗
uc ρc
∗ p
p = => p = ∆pc · p∗ (19)
∆pc

Dr.-Ing. Manuel Münsch | FAU | 13

 Erlangen

Dimensionless Governing Differential Equations

Mass Conservation:
• inserting the dimensionless variables into the mass conserv. equations will
lead to the no-dimensional mass conserv. equation:
ρc ∂ρ∗ ρc uc ∂ρ∗ui∗
∗
+ ∗ =0 (20)
tc ∂ t Lc ∂ xi
• dividing by ρc uc /Lc to make eq. non-dimensional:
Lc ∂ρ∗ ∂(ρ∗ui∗)
∗
+ ∗ =0 (21)
tc uc
|{z} ∂t ∂ xi
Strouhal number
• finally:
∂ρ∗ ∂(ρ∗ui∗)
Sr ∗ + =0 (22)
∂t ∂ xi∗

Dr.-Ing. Manuel Münsch | FAU | 14

 Erlangen

Dimensionless Governing Differential Equations

Momentum Conservation:
• inserting the dimensionless variables into the momentum conserv.
equations will lead to the no-dimensional momentum conserv. equation:
ρc uc ∂(ρ∗ui∗) ρc uc2 ∂(ρ∗ui∗uj∗)
∗
+ ∗ =
tc ∂t Lc ∂x
" i ! #
∗ ∗ ∗ ∗
∆pc ∂ p µc uc ∂ ∗
∂ uj ∂ ui 2 ∗ ∂ uk
− ∗ + 2 ∗ µ ∗ + ∗ − µ δij ∗ + ρc gc ρ∗gj∗
Lc ∂ xj Lc ∂ xi ∂ xi ∂ xj 3 ∂ xk
(23)
• dividing by ρc uc2/Lc to make eq. non-dimensional is leading finally to:
∂(ρ∗uj∗) ∂(ρ∗ui∗uj∗)
Sr ∗
+ ∗ =
∂t ∂ xi " ! #
∗ ∗ ∗ ∗
∂p 1 ∂ ∗
∂ uj ∂ ui 2 ∗ ∂ uk 1 ∗ ∗
− Eu ∗ + µ + − µ δij ∗ + 2 ρ gj (24)
∂ xj Re ∂ xi∗ ∂ xi∗ ∂ xj∗ 3 ∂ xk Fr

Dr.-Ing. Manuel Münsch | FAU | 15

 Erlangen

Dimensionless Governing Differential Equations

Characteristic Numbers:
• Strouhal number:
Lc ( utcc ) local accel.
Sr = = u2 = (25)
tc uc c spatial accel.
Lc
• Reynolds number:
ρc uc2
Lc ρc uc Lc inertial force
Re = µc uc = = (26)
L2c
µc viscous force
• Euler number:
∆pc static pressure
Eu = = (27)
ρc uc2 dynamics pressure
• Froude number:
ρc uc2
inertial force
Lc
Fr = = (28)
ρc gc gravitational force

Dr.-Ing. Manuel Münsch | FAU | 16

 Erlangen

Dimensionless Governing Differential Equations

Remarks
• Flows can be characterized basing on the Re-Number.
• flows at very low Re numbers: Re → 0 -> Stokes Equations
• flows at very high Re numbers: Re → ∞ -> Euler Equations
• flows at very high but finite Re numbers: Re >> 1 -> Boundary Layer
Equations

Dr.-Ing. Manuel Münsch | FAU | 17

 Erlangen

Simplified Conservation Equations

Stokes Equations (incompressible)
Flows at very low Re numbers: Re → 0
• in so called creeping flows viscous forces are much larger than inertia
• creeping flows occur at very low velocities uc , high viscosity µc or very
small dimensions Lc
-> non linear inertia term in momentum conservation is neglected

Stokes Equations (dimensional):

∂ ui
=0 (29)
∂ xi  
∂p ∂ ∂ ui ∂ uj
=µ + (30)
∂ xi ∂ xj ∂ xj ∂ xi

Dr.-Ing. Manuel Münsch | FAU | 18

 Erlangen

Simplified Conservation Equations

Euler Equations (compressible)
Flows at very high Re numbers: Re → ∞
• this types of flows occur at high velocities uc , low viscosity µc or large
dimensions Lc
-> in a good approximation viscous forces can be neglected compared to
inertia and pressure (just valid far away from any walls)

Euler Equations (dimensional):

∂ρ ∂(ρui )
+ =0 (31)
 ∂ t ∂ x i 
∂(ρui ) ∂(ρui uj ) ∂p
+ =− + ρgi (32)
∂t ∂ xj ∂ xi
 
∂(ρT ) ∂(ρui T ) ∂ ui
cv + = −p (33)
∂t ∂ xi ∂ xi

Dr.-Ing. Manuel Münsch | FAU | 19

 Erlangen

Simplified Conservation Equations

Euler Equations
• External flow: inviscid -> Euler equations
• at the boundary layer viscous forces dominate

(Source: Vorlesungsskript Numerische Strömungsmechanik, M. Breuer, LSTM, 2006)

with the conditions: u (x , 0) = 0 and u (y , ∞) = u∞

Dr.-Ing. Manuel Münsch | FAU | 20

 Erlangen

Simplified Conservation Equations

Boundary Layer Equations

Flows at very high but finite Re numbers: Re >> 1

• viscous forces are most dominant in the boundary layer
• fluid is sticking to the wall due to no-slip boundary condition (Stokes)
• at the edge of the boundary layer the fluid reaches the undisturbed flow
velocity u∞
• splitting the flow field into a boundary layer flow and a non-viscous outer
flow simplifies the underlying conservation equations

-> Boundary Layer Equations

• derivation starts from the 2D, steady, incompressible & isothermal

conservation eqs.

Dr.-Ing. Manuel Münsch | FAU | 21

 Erlangen

Boundary Layer Equations

Derivation:
Starting point: 2D, steady, incompressible &
isothermal conservation eqs.

(Source: Vorlesungsskript Numerische Strömungsmechanik, M. Breuer,

Mass: LSTM, 2006)
∂u ∂v
+ =0 (34)
∂x ∂y
Momentum:
2 2
 
∂u ∂u 1 ∂p ∂ u ∂ u
u +v =− +ν 2
+ 2 (35)
∂x ∂y ρ ∂x ∂x ∂y
 2
∂ v ∂ 2v

∂v ∂v 1 ∂p
u +v =− +ν + (36)
∂x ∂y ρ ∂y ∂x 2 ∂y 2

Dr.-Ing. Manuel Münsch | FAU | 22

 Erlangen

Boundary Layer Equations

Derivation: Assumptions
• flat plate parallel to the flow
• thin boundary layer: induced normal vel. components much smaller than
parallel component u >> v
• rate of change across the boundary layer much grater than rate of change
in flow direction:
∂ ∂ ∂2 ∂2
∂x << ,
∂y ∂x 2 << ∂y 2
∂p
• no streamwise pressure gradient: ∂ x =0

Dr.-Ing. Manuel Münsch | FAU | 23

 Erlangen

Boundary Layer Equations

Derivation: Assumptions
• from the upper assumptions the conservation eq. simplify to:
Mass:
∂u ∂v
+ =0 (37)
∂x ∂y
Momentum:
∂u ∂u ∂ 2u
u +v =ν 2 (38)
∂x ∂y ∂y
1 ∂p
0=− (39)
ρ ∂y
-> since ∂∂ yp = 0, the pressure normal to the plate is const. (under the upper
assumptions)

Dr.-Ing. Manuel Münsch | FAU | 24

 Erlangen

Boundary Layer Equations

Blasius (flat-plate) boundary layer (1908):

∂u ∂v
+ =0 (40)
∂x ∂y
∂u ∂u ∂ 2u
u +v =ν 2 (41)
∂x ∂y ∂y
Boundary Conditions:
•y = 0: u = v = 0
• y − > ∞: u − > u∞

(Source: Vorlesungsskript Numerische Strömungsmechanik, M. Breuer, LSTM, 2006)

Dr.-Ing. Manuel Münsch | FAU | 25

 Erlangen

Boundary Layer Equations

Observations:
2
• the diffusion term ν ∂∂ xu2 is neglected
• the BL-equations are (first order) parabolic in x
• there is information transport only in the downstream direction:
• the flow at a position x2 does not influence what happens at x1 if x2 > x1
• this has implications on the appropriate numerical scheme
-> streamwise space-marching can be used (also interesting for FD-schemes)

Further solution steps by Blasius:

• Blasius was able to reduce the partial differential equation to an ordinary
differential equation by making use of further assumptions:
• introduction of a dimensionless similarity variable η and a stream function ψ
• similarity conditions: velocity profiles are similar along the plate

Dr.-Ing. Manuel Münsch | FAU | 26

 Erlangen

Boundary Layer Equations

Stream Function:
In two dimensions it is very useful to define a stream function ψ(x , y ), so
that:

∂ψ ∂ψ
u= and v = − (42)
∂y ∂x

Inserting these relation into the mass conservation equation:

∂u ∂v ∂ ∂ψ ∂ ∂ψ
+ = − =0 (43)
∂x ∂y ∂x ∂y ∂y ∂x

-> Mass conservation directly fulfilled!

Dr.-Ing. Manuel Münsch | FAU | 27

 Erlangen

Boundary Layer Equations

Stream Function: momentum conservation
The momentum conservation:
∂u ∂u ∂ 2u
u +v =ν 2 (44)
∂x ∂y ∂y
is written in terms of the stream function as:

∂ψ ∂ 2ψ ∂ψ ∂ 2ψ ∂ 3ψ
− =ν 3 (45)
∂y ∂y ∂x ∂x ∂y ∂y ∂y

-> this is a 3rd order nonlinear PDE, it is still quite difficult to solve, even
numerically...

Dr.-Ing. Manuel Münsch | FAU | 28

 
   
Erlangen

    !  
Boundary Layer Equations
Similarity of the velocity profile




 

Exemplary scaling relation:
 
    -/ "  /01   /02
y % '   


  !'
u (x2, y ) = u (x1, p
x2/x1
) (46)

the profile at x = 4 may be the profile at x = 1 stretched by a factor of two in

the y direction for example
Dr.-Ing. Manuel Münsch | FAU | 29
 Erlangen

Boundary Layer Equations

Similarity considerations
• the velocity profiles are similar to each other
• the velocity profiles u (y ) can be matched by using appropriate scaling
factors for u and y :
u
-> for u the undisturbed outer velocity u∞ is chosen: u∞
-> for y the√boundary layer thickness δ(x ) is chosen which is scaling with
δ(x ) ∼ x leading to a similarity variable:
y
η(x , y ) = (47)
δ(x )
• comments on δ(x ):
• δ is proportional to thickness of the fluid layer which is influenced by friction
(viscous effects)
• viscosity leads to a momentum flux normal to the wall with transport velocity uv
• uv depends on ν and δ

Dr.-Ing. Manuel Münsch | FAU | 30

 Erlangen

Boundary Layer Equations

Similarity considerations
• comments on δ(x ) con’t:
• from dimensional considerations
ν
uv ∼ (48)
δ
• δ depends on the fluid elements which are just still be captured by this
momentum transport away from the wall and the elements being convected
away from x (along the wall) without being captured
• the residence time of a fluid element with velocity u∞ over a length x is T1 = ux
∞
• for the wall normal momentum transport with uv ∼ νδ over a layer thickness δ the
2
time T2 = uδv = δν passes
• with T1 = T2 : r
xν
δ(x ) ∼ (49)
u∞

Dr.-Ing. Manuel Münsch | FAU | 31

 Erlangen

Boundary Layer Equations

Similarity considerations
• and with this: r
u∞
η(x , y ) = y (50)
νx
q
u∞
• here ν is chosen to make η non-dimensional

• the velocities u (x1, η1) and u (x2, η2) are similar if:
u (x1, η1) u (x2, η2)
= (51)
u∞ u∞
• thus, the velocity profiles are matching if uu is just a function of the
∞
similarity coordinate:
u
= F (η) (52)
u∞

Dr.-Ing. Manuel Münsch | FAU | 32

 Erlangen

Boundary Layer Equations

Streamfunction
• a streamfunction is given as:
Z Z
ψ= u dy = u∞F (η)dy
r
νx
Z
= u∞F (η) dη
u∞
√
Z
= u∞ν x F (η)d η
√
= u∞ν xf (η) + C (53)
R
• here f (η) = F (η)d η is the non-dimensional stream function
• C = 0 if f (η) = 0 for y = 0 and η = 0
• Thus:
√
ψ= u∞ν x f (η) (54)

Dr.-Ing. Manuel Münsch | FAU | 33

 Erlangen

Boundary Layer Equations

Streamfunction
• from that:
∂ψ ∂ψ ∂η
u= = = u∞f 0(η)
∂y ∂η r∂y
∂ψ 1 ν u∞ 0
v =− = (η f (η) − f (η))
∂x r 2 x
∂ 2ψ u∞ 00
= u∞ f
∂y 2 νx
∂ 3ψ u∞ 000
3
= u∞ f
∂y νx
∂ 2ψ 1 u∞ 00
=− ηf
∂y ∂x 2 x
• the upper relations are introduced into the simplified conservation
equations ((40) and (41))

Dr.-Ing. Manuel Münsch | FAU | 34

 Erlangen

Boundary Layer Equations

Ordinary Differential Equation
The complex problem is now reduced to:
f f 00 + 2 f 000 = 0 (55)
with the boundary conditions:
• for η = 0: f 0 = f = 0 no-slip condition
• for η = ∞: f 0 = 1 non-viscid outer flow

Thus:
• the partial differential equations (40) and (41) are transformed to an
ordinary differential equation
• this equation can be solved numerically by an Runge-Kutta approach for
example (to be done in exercise 2)
• the solution delivers a distribution of f , f 0 and f 00

Dr.-Ing. Manuel Münsch | FAU | 35

 Erlangen


Boundary Layer Equations
     !   1   
Ordinary Differential Equation: Solution

Flow velocities:
∂ψ ∂ψ ∂η
u= = = u∞f 0(η) (56)
∂y ∂η r
∂y
∂ψ 1 ν u∞ 0
v =− = (η f (η) − f (η)) (57)
∂x 2 x
At
    
the edge of the boundary:
r
ν
v| = 0.8604u (58)

∞
  η→∞
 

      xu∞


which
is different from v |η→∞ = 0 due to dis-
placement effect (see following figure)

Dr.-Ing. Manuel Münsch | FAU | 36

 Erlangen

Boundary Layer Equations

Boundary Layer: Streamlines

Dr.-Ing. Manuel Münsch | FAU | 37

 Erlangen

Boundary Layer Equations

Blasius Solution: Characteristic Values
From the Blasius solution further characteristic quantities are derived:
• local friction coefficient:
τw 0.664 u∞x
cf = 2
= √ with Rex = (59)
0.5ρu∞ Rex ν
• drag coefficient:
1.328 u∞L
cw = √ with ReL = (60)
ReL ν
• boundary layer thickness:
δ 5.0
=√ (61)
x Rex
• displacement thickness:
δ1 1.7208
= √ (62)
x Rex
• momentum thickness:
δ2 0.664
=√ (63)
Dr.-Ing. Manuel Münsch | FAU | x Rex 38
 Erlangen

Boundary Layer Equations

Blasius Solution: Conclusion
• Blasius solution gives a good approximation for the solution of a laminar
boundary layer flow (Re < 5 · 105)
• a solution can also be derived for flows with pressure gradient (see:
Schlichting)
• Blasius solution not valid at the leading edge (singularity) since δ/L << 1
not fulfilled
• Blasius solution not valid for separated flows

Dr.-Ing. Manuel Münsch | FAU | 39

 Erlangen

Conclusion

• The Navier-Stokes equation are impossible to solve with a pencil and

paper even in very simple flows
• In any application you will need numerical simulations
• Understanding the application well, will allow you to make simplifications
and reduce simulation time
• Depending on the model equations you simulate, you will need different
numerical tools
• Next class we start with numerical methods!

Dr.-Ing. Manuel Münsch | FAU | 40