Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Lec3 Intro Turb

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Notes

DEN403/DENM010 Computational Fluid Dynamics


Part 3: Introduction to turbulence Dr. Jens-Dominik Muller School of Engineering and Materials Science, Queen Mary, University of London
j.mueller@qmul.ac.uk Room: Eng 122 ofce hours: any reasonable time c Jens-Dominik Muller, 2011

1 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Organisation of the lectures on turbulence


1. Introduction
Motivating examples, description of turbulence The Reynolds number The Kolmogorov cascade

Notes

2. Reynolds-Averaged Navier-Stokes
Averaging the Navier-Stokes equations Reynolds stresses, closure Modelling the Reynolds stresses

3. Using RANS
The near-wall structure of turbulent boundary layers Mesh spacing requirements, wall functions Limits of applicability

4. Alternative approaches
DNS, LES, DES

2 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Outline of this part

Notes

Examples of turbulent ow Description of turbulence The Reynolds number The Kolmogorov Cascade Summary

3 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Outline of this part

Notes

Examples of turbulent ow Description of turbulence The Reynolds number The Kolmogorov Cascade Summary

4 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

On Turbulence
Benoit Mandelbrot: The techniques I developed for studying turbulence, like weather, also apply to the stock market. Werner Heisenberg: When I meet God, I am going to ask him two questions: Why relativity? And why turbulence? I really believe he will have an answer for the rst.
(Source: Great-Quotes.com, Wikipedia)

Notes

5 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

An artists view of turbulence

Notes

Leonardo da Vinci
6 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Turbulence on a global scale

Notes

Flow around Selkirk island


(Source: NOAA) 7 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Reynolds experiment

Notes

(Source: Reynolds, 1883)

Turbulence in smooth pipes typically occurs above Re = 2000.

8 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Effect of Reynolds number

Notes

Re = 15, 000
(Source: van Dyke: Album of uid motion)

Re = 30, 000

9 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Laminar and turbulent ow II

Notes

10 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Turbulent combustion I

Notes

Turbulent mixing downstream of a swirler


(Source: CERFACS) 11 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Turbulent combustion II

Notes

Turbulent combustor
(Source: CERFACS) 12 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Turbulent combustion III

Notes

Ignition simulation in an annular combustor


(Source: CERFACS) 13 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Outline of this part

Notes

Examples of turbulent ow Description of turbulence The Reynolds number The Kolmogorov Cascade Summary

14 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Characteristics of turbulence
Turbulence is inherently unsteady and 3-dimensional. Turbulence is dominated by chaotic - but not random

Notes

motion of swirling structures, the eddies.


There is a cascade of eddies, largest eddies determined

e.g. by the geometry.


Largest scales take their energy from mean ow. Larger eddies break up, passing their energy to smaller

scales.
Smallest scales dissipate their energy into heat. Is always dissipative, i.e. increases mixing, disorder. The Reynolds number will play a major role. Nearly all relevant industrial ows are turbulent!

15 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

How does turbulence arise?


A at plate boundary layer
starts out laminar transitions from laminar

Notes

to turbulence after some running length


remains turbulent

downstream of transition
transition modelling is

very complex: in CFD typically full turbulence is assumed.

16 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Laminar vs. turbulent boundary layers


The laminar b.l. prole

Notes

has a lower velocity gradient u near the y wall, hence a lower wall shear stress
The turb. b.l. prole has

a higher velocity near the wall, hence is more resistant to separation


The turb. b.l. has more

mixing, hence heat transfer or surface reactions are enhanced.


17 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Outline of this part

Notes

Examples of turbulent ow Description of turbulence The Reynolds number The Kolmogorov Cascade Summary

18 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

The Reynolds number in the momentum equations


The momentum equ. incompr. Navier-Stokes equations in vector notation: u + u u = p + 2 u, t where u is the vector of velocities. The unit of the equation, as stated above, is force per volume: F /V = m a/V = a . Dividing the equation by this factor of this dimension, u 2 /D, which is equivalent to normalising the variables by u 1 D u = , p = p 2, , =D = U U t U t makes the equ. nondimensional: u +u t u = p + 1 Re
2

Notes

u,

Note: for Re 0 the effect of the viscous term vanishes, but the no-slip condition at the wall may remain!

19 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Reynolds number and Turbulence


Re = = momentum of the ow inertial forces = viscous forces viscous stress u2 u2 ul ul = u = = l u y

Notes

When Re 1 the ow is very viscous (creeping ow). As Re the ow becomes less dominated by viscosity,

and boundary-layers conned to small region near surfaces.


The Reynolds number depends on the choice of

length-scale!
Choosing an overall length-scale, e.g. aerofoil chord

length, provides only analysis of overall effects.


20 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Reynolds number based on length


Most simply: base Reynolds number on the length of the

Notes

body L,
but a boundary-layer grows with distance ( x 0.5 for a

laminar boundary-layer) L and are inter-related.


Instead, base Re on distance from the L.E.:

ReL = UL/

21 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Reynolds number for a boundary layer


Reynolds number:

Notes

Re =

momentum of the ow viscous sresses

momentum of the ow: U 2 viscous sresses:

= (du/dy) (U/)
Reynolds number based on

b.l. thickness: Re = U2 U U = = (U/)

22 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Outline of this part

Notes

Examples of turbulent ow Description of turbulence The Reynolds number The Kolmogorov Cascade Summary

23 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Effect of Reynolds number on small scales

Notes

med. Re

higher Re

(Source: van Dyke: Album of uid motion) 24 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Reynolds number based on eddy diameter d


Re = = momentum of the ow inertial forces = viscous forces viscous stress u2 ud ud u2 = = u = d u y

Notes

For Re >> 1, inertial forces dominate. The ow keeps

swirling, energy is passed down to smaller scales.


When Re 1, viscous forces become equal in magnitude

to inertial forces, eddies dissipate.


There is a smallest length scale for turbulent eddies!

Smaller eddies are dissipated by viscosity.


Rotational energy d 2 , hence smaller eddies contain less

energy.
25 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

The Kolmogorov Cascade I

Notes

26 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

The Kolmogorov Cascade II


Eddy structure is fractal with higher Re, we nd

Notes

smaller scales
smaller scales have

smaller diameters, hence in a ow with the same speed lead to uctuations with higher frequencies
in turbulent ow literature,

rather than higher frequency, the term higher wavenumber k is used.

27 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Spectrum of turbulent kinetic energy

Notes

28 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Spectrum of turbulent kinetic energy


There is a peak of overall

Notes

turbulent kinetic energy E at some wavenumber k = O(L1 ), i.e. some diameter L given by the geometry.
In isotropic turbulence

energy drops at a rate of 5 k 3 with increasing wavenumber k.


There is a smallest

wavenumber/scale which increases with Re.


29 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

How small is the smallest eddy? I

Notes

The largest eddies depend on the geometry scale, e.g. b.l.

thickness or pipe diameter.


The scale of the smallest eddies, the scale at which

dissipation occurs, is independent of the scale of the largest eddies or the mean ow. At the smallest scales there is an equilibrium between
energy supplied by larger scales energy dissipated by viscosity

This is known as Kolmogorovs Universal Equilibrium

Theory

30 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

How small is the smallest eddy? II


Dene the dissipation rate per unit mass [m2 sec3 ], and use the kinematic viscosity [m2 sec1 ], using dimensional analysis we can then we can derive the

Notes

Kolmogorov microscales:
Kolmogorov length scale: Kolmogorov time scale:

= =

1/4

1/2

Using dimensional analysis we can approximate U 3 /L hence the ratio of typical length L to smallest eddy size is
L 3 1/4
3 3

= L/

= (UL/) 4 = Re 4

31 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Implications of eddy scaling for CFD


L size of smallest scales: = Re
3 4

Notes

To resolve an eddy we need at least two mesh points to

represent the velocity uctuations: x = h .


hence the number of meshpoints in one direction is
L h

= Re 4

for a three-dimensional calculation we need this many

mesh points in each direction, hence the overall number of 9 nodes N scales with Re as N = (L/h)3 = Re 4
Resolving all turbulent structures is only possible for low

Re, but is prohibitive for high Re.


DNS of a complete aircraft will require at least an exaop

(1018 ops) computer. The best performance currently is a around 500 tera ops (500 1012 ops)
32 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Approaches to deal with turbulence in CFD I


Simulation of the nest scales means: resolving these

Notes

scales with mesh points such that we can accurately model them in the conservation equations.
This approach is called Direct Navier-Stokes (DNS), but is

not affordable even in the mid-term future.


We are typically not interested in the ne scale

uctuations, in engineering we care for the long-term time-averages as they would affect the ight of an aircraft.
Hence, we could approximate the average effect of these

uctuations with an additional model that embodies our knowledge of turbulent ows.
This approach is called Reynolds-averaged Navier-Stokes

(RANS), and is the most popular approach to CFD for turbulent ows.
33 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Approaches to deal with turbulence in CFD II

Notes

34 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Outline of this part

Notes

Examples of turbulent ow Description of turbulence The Reynolds number The Kolmogorov Cascade Summary

35 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Introduction to turbulence: summary I


Turbulence is unsteady and three-dimensional. The chaotic motion of turbulent ow is fully described by

Notes

the conservation equations.


There is a cascade of eddies, largest scales determined by

geometry.
Turbulence increases skin friction, but also increases

mixing.
Nearly all ows of industrial interest are turbulent. They Reynolds number can describe turbulent effects, but

care needs to be taken to choose the correct length scale.

36 / 37

Outline

Examples

Description

Reynolds number

Kolmogorov cascade

Summary

Introduction to turbulence: summary II


There is a smallest eddy scale, the Kolmogorov scale. For

Notes

smaller scales viscosity becomes dominant and eddies are dissipated.


The ratio of smallest to largest scales is Re 4 . If we were to resolve the smallest scale in a numerical ow
3

simulation, a DNS, the required number of mesh points 9 would scale with Re 4 , an unsteady computation would require an exaop computer.
For lower computational cost, we need to model the

time-averaged effect of turbulent uctuations, the Reynolds-averaged Navier Stokes approach (RANS).
There is also an intermediate approach, the Large Eddy

Simulation (LES), where the largest scales are resolved (simulated) and the sub-grid scales are modelled.
37 / 37

Notes

Notes

You might also like