Introduction to CFD -Theory - 2

Embed Size (px)

Citation preview

  • 8/13/2019 Introduction to CFD -Theory - 2

    1/58

    I CFD

    E: .@..

    &

    , 15 2011

    C F D

    CFD ,

    , ,

    .

    +

    CFD :

    S A ;

    N ;

    A

    H

    T CFD . M

    H CFD .

  • 8/13/2019 Introduction to CFD -Theory - 2

    2/58

    F

    CFD :

    S : A , O ; H ; D ; F , , ; P, , ; C ; S; C; H ; C

    C

    C

  • 8/13/2019 Introduction to CFD -Theory - 2

    3/58

  • 8/13/2019 Introduction to CFD -Theory - 2

    4/58

    R

    LAGRANGIAN APPROACH:

    P (,,)

    SUBSTANTIAL DERIVATIVE ( )

    I ()

    R

    EULERIAN APPROACH:

  • 8/13/2019 Introduction to CFD -Theory - 2

    5/58

    M

    SURFACE FORCES:

    BODY FORCES: , , ,

    (

    )

    M

  • 8/13/2019 Introduction to CFD -Theory - 2

    6/58

    E

    Conservation equation for internal energy

    Conservation equation for mechanical energy

    E

    F

    =

    F

    = , =

  • 8/13/2019 Introduction to CFD -Theory - 2

    7/58

    E

    5 EQUATIONS

    C (1)

    M (3) E (1)

    11 UNKNOWNS

    2 TD (, , , T

    = (, T) = (, T)

    = RT = T

    V (3)

    V (6)

    Liquids and gases flowing at low speeds behave as incompressiblefluids: without density variation there is no linkage between the energy

    equation and the mass and momentum conservation. The flow fieldcan be considered mass and momentum conservation only

    V

    I N

    F ()

    S

    F

    F

  • 8/13/2019 Introduction to CFD -Theory - 2

    8/58

    NS

    NS

    sMx

  • 8/13/2019 Introduction to CFD -Theory - 2

    9/58

    NS

    M

    E

    NS

    C

    MATHEMATHICALLY CLOSED PROBLEM!

  • 8/13/2019 Introduction to CFD -Theory - 2

    10/58

    G

    A :

    =1,,,, (T,)

    D

    TRANSPORT EQUATIONFOR PROPERTY DIFFERENTIAL FORM

    dV

    An

    V

    S

    G

    I 3D (CV)

    G

    INTEGRAL FORM

    Rate of change of thetotal amount of fluid

    property in the controlvolume

    Net rate ofdecrease of fluidproperty of the

    fluid element due toconvection

    Net rate ofincrease of fluidproperty of the

    fluid elementdue to diffusion

    Net rate of increaseof fluid property as

    result of sources

  • 8/13/2019 Introduction to CFD -Theory - 2

    11/58

    G

    I

    F

    S M

  • 8/13/2019 Introduction to CFD -Theory - 2

    12/58

    C

    :

    F () .

    DISCRETIZATION METHOD

    , .

    T /

    .

    C :

    N :

    T

    (.. , , ) . T

    . E , )

    A ; D

    I , .

  • 8/13/2019 Introduction to CFD -Theory - 2

    13/58

    C

    CONVERGENCECRITERIA

    SOLUTION METHODLinearization and iterative

    techniques.

    FINITE APPROXIMATIONApproximations

    for the derivatives at thegrid points for FD

    Approximating surface andvolume integrals for FV

    NUMERICAL GRID

    Structured,unstructured, etc

    REFERENCE SYSTEMCartesian, cylindrical coord.etc.

    Fixed or moving, etc

    DISCRETIZATIONMETHOD

    Differential eq. Algebraic eq.(FD, FE, FV)

    MATHEMATICALMODEL

    Set of partial differentialor integro-differential

    equations and BCs

    NUMERICAL SOLVER

    T

    T (, .).

    S () T ( )

    ( 2D) ( 3D) , .. (, , ).

    E 4 2D 6 3D

    T ,

    . S .

    I

  • 8/13/2019 Introduction to CFD -Theory - 2

    14/58

    T

    B

    T

    U

    FVM FEM

    A , / 2D, /

    3D.

    .

    N . T ,

    .

    S .

  • 8/13/2019 Introduction to CFD -Theory - 2

    15/58

    T

    H

    T

  • 8/13/2019 Introduction to CFD -Theory - 2

    16/58

    D

    T CFD :

    F V (80%).

    F E (15%).

    O (5%)F ()

    S .

    B .

    V .

    L / B.

    P :

    3 :M ,

    ;

    D ,

    ;

    I ( ), .

  • 8/13/2019 Introduction to CFD -Theory - 2

    17/58

    D : F D

    M

    O (E 18 )

    C .

    S

    A ,

    T .

    O , .

    ADV:

    FDM .

    I

    DIS:

    C .

    R

    D : F VM

    C

    T (CV), CV. A CV

    .

    I CV (CV) .

    S .

    O CV, .

    ADV: , .

    DIS: 3D. T

    FV : ,, .

  • 8/13/2019 Introduction to CFD -Theory - 2

    18/58

    F V M

    F V M

    N

    A

    A

    I

    U

    L

    Q

    H

    I

    A

    GENERIC CONSERVATION EQUATIONIN INTEGRAL FORM

  • 8/13/2019 Introduction to CFD -Theory - 2

    19/58

    N

    T

    (CV) .

    nodes centered in CVsnodes centered in CVsnodes centered in CVsnodes centered in CVs CV faces centered between nodesCV faces centered between nodesCV faces centered between nodesCV faces centered between nodes

    Nodal value mean over CV More accurate CDS approximation

    A

    =

    =

  • 8/13/2019 Introduction to CFD -Theory - 2

    20/58

    A

    T , S. T , (CV

    ) . T :

    1.

    ;

    2. (CV

    ) ()

    A

    A 1:

    T INTERPOLATION

    ( ).

    T (CV )

    INTERPOLATION ( ).

    T , (CV )

    INTERPOLATION ( ,

    ).

    2 order accuracy

    4 order accuracy

    2 order accuracy

  • 8/13/2019 Introduction to CFD -Theory - 2

    21/58

    I

    A 2: T (CV ).

    , , ,

    CV .

    INTERPOLATION

    U I (UDS)

    L I (CDS)

    Q U I (QUICK)

    HO S

    O

    T

    B

    ,

    CV

    CV.

    T

    CV . W ,

    A . T

    CV;

    ,

    ,

    .

  • 8/13/2019 Introduction to CFD -Theory - 2

    22/58

    S L ES

    S

    D FDM FVM ,

    G

    LU

    TDMA

    G

    LU

    TDMA

    I , , .

    I , ( CFD )

  • 8/13/2019 Introduction to CFD -Theory - 2

    23/58

    D : G

    T A21 0

    1 A21/ A11 2 (1 A21/ A11)

    E

    U

    N = 3/3

    B = 2

    /2,

    P

    FORWARDFORWARDFORWARDFORWARD

    ELIMINATIONELIMINATIONELIMINATIONELIMINATION

    BACK SUBSTITUTIONBACK SUBSTITUTIONBACK SUBSTITUTIONBACK SUBSTITUTIONHIGH COSTHIGH COSTHIGH COSTHIGH COST

    I

    A .

    T

    I T

    M (P )

    Correction or updateCorrection or updateCorrection or updateCorrection or update(approximation of the iteration(approximation of the iteration(approximation of the iteration(approximation of the iteration

    error)error)error)error)

  • 8/13/2019 Introduction to CFD -Theory - 2

    24/58

    C C I E

    I .

    T

    ; U,

    .

    I

    norm

    S

  • 8/13/2019 Introduction to CFD -Theory - 2

    25/58

    U

    T

    T

    : ( ) ,

    ( )

    T , (" )

    M

    (ODE)

    M I V P ODE: T L M

    TL M

    ODE

    .

    1= + , 2 = + ,. . .

    + =

    + :

  • 8/13/2019 Introduction to CFD -Theory - 2

    26/58

    M I V P

    ODE: T L M

    EXPLICIT OR FORWARDEXPLICIT OR FORWARDEXPLICIT OR FORWARDEXPLICIT OR FORWARDEULEREULEREULEREULER

    IMPLICIT OF BACKWARDIMPLICIT OF BACKWARDIMPLICIT OF BACKWARDIMPLICIT OF BACKWARDEULEREULEREULEREULER

    MIDPOINT RULEMIDPOINT RULEMIDPOINT RULEMIDPOINT RULE

    TRAPEZOID RULETRAPEZOID RULETRAPEZOID RULETRAPEZOID RULE

    implicit methods

    M I V P ODE:PC M

    .

    .

    ,

    .

    T E :

    *+1,

    Second orderaccurate!

  • 8/13/2019 Introduction to CFD -Theory - 2

    27/58

    S N S

    NS

    Steady 2D flowSteady 2D flowSteady 2D flowSteady 2D flow

  • 8/13/2019 Introduction to CFD -Theory - 2

    28/58

  • 8/13/2019 Introduction to CFD -Theory - 2

    29/58

    S

    E (, , T, .)

    E

    I :

    I

    Storage locations for v velocities

    Storage locations for u velocities

    NON ZERO!NON ZERO!NON ZERO!NON ZERO!

    BACKWARD STAGGERED GRID

    D

    T (, J) :

    T : (1,J), (+1,J), (,J1),(,J+1)

    J (, QUICK, .)

    (F) (D) J

    E. (1, J)

    ( )

  • 8/13/2019 Introduction to CFD -Theory - 2

    30/58

    D

    G ,

    I

    I

    , ,

    P : SIMPLE

    S I M P L E (SIMPLE)

    P S (1972)

    M

    A *

    T

    = *+

    = *+ = *+

    T .

    M :

  • 8/13/2019 Introduction to CFD -Theory - 2

    31/58

    SIMPLE

    T :

    NEGLECTED to simplify the equstions:MAIN APPORXIMATION OF SIMPLE!

    EQUATION FOR PRESSURE CORRECTION

    SIMPLE

    EQUATION FOR PRESSURE CORRECTION

    I,J

  • 8/13/2019 Introduction to CFD -Theory - 2

    32/58

    SIMPLE

    *, *, *, *

    STEP 1

    STEP 2

    STEP 3

    STEP 4

    ( )

    u*, v*

    p

    , ,*+,

    p, u, v, *

    ?

    * =

    * =

    * =

    * = ,

    NO

    YES

    64

    T

    Large StructureSmall Structure

  • 8/13/2019 Introduction to CFD -Theory - 2

    33/58

    65

    T

    I () U () (R

    )

    ( ) ( ) ( ) iiiii uUtxutxutxu ',,, ' +=+= ( ) 0,' =txui

    Point velocity measurement in a turbulent flow

    66

    D

    T = /2

    T

    P D F: *

    M

    H

    *)**(**)( dPROBdP +

  • 8/13/2019 Introduction to CFD -Theory - 2

    34/58

    67

    T

    4

    13

    =

    k

    2

    1

    4

    1

    4

    3

    L

    k

    LL

    k

    L

    k

    ReReu

    uRe

    L===

    3 :

    K :

    R (>

    ,

    .

    2

    1

    =

    k

    ( )41

    =ku

    Scale di Kolmogorov

    Length

    Time

    Velocity

    .

    68

    E

    = 2/.

    T

    T

    ( =2/).

    T

    E()

    Energy containing rangeInertial subrangeDissipation range

    Transfer of energy tosmaller scales

    k lDI lEI L

    Production PDissipation

    =

    =3

    1

    2'

    2

    1

    i

    iuk

  • 8/13/2019 Introduction to CFD -Theory - 2

    35/58

    69

    F :

    L (+

  • 8/13/2019 Introduction to CFD -Theory - 2

    36/58

    71

    D N S (DNS)

    ,:

    4/9

    3

    L

    k

    celle ReL

    N =

    2/1

    L

    k

    Lt ReN =

    411Re /tcelle NN timeCPU

    Integral scale

    Kolmogorov scales

    72

    D N S (DNS)

    ADVANTAGES

    H ;

    N

    P

    DISADVANTAGES

    O R

    A 3D;

    R BC

    Particle-laden jet,Longmire and

    Eaton(JFM, 1992)

    Boundariesof a DNSdomain

  • 8/13/2019 Introduction to CFD -Theory - 2

    37/58

    73

    R N S E

    (RANS)

    RANSRANSRANSRANS

    74

    R N S E(RANS)

    A 3 3 :

    N

    T

    REYNOLDS STRESSESREYNOLDS STRESSESREYNOLDS STRESSESREYNOLDS STRESSES

  • 8/13/2019 Introduction to CFD -Theory - 2

    38/58

  • 8/13/2019 Introduction to CFD -Theory - 2

    39/58

    77

    RANS:

    : R

    Navier- Stokes equation

    RANS

    Closure problem:Reynolds Stresses?

    Problem:eddy viscosity?

    Reynolds decomposition

    EDDY VISCOSITY MODELS(Boussinesq hypothesis)

    Mixing length modelk- modelk- model

    DIRECT MODELS

    Reynolds stressmodels

    Anisotropic turbulence

    Algebraic stressmodels

    78

    RANS:

    K

    S 2D

    V

    E

    PRANDTLS MIXINGPRANDTLS MIXINGPRANDTLS MIXINGPRANDTLS MIXINGLENGTH MODELLENGTH MODELLENGTH MODELLENGTH MODEL

  • 8/13/2019 Introduction to CFD -Theory - 2

    40/58

    79

    RANS:

    T

    T ; R,

    Velocity scaleVelocity scaleVelocity scaleVelocity scale Length scaleLength scaleLength scaleLength scale

    80

    RANS:

    P P L S (1974)

  • 8/13/2019 Introduction to CFD -Theory - 2

    41/58

    81

    RANS:

    ADVANTAGES

    S

    W

    S

    W

    DISADVANTAGES

    P

    P

    P

    P R

    Problems with RANS:-Low Re number flows

    Inaccuracy of log law

    -Rapidly changing flowsIn 2 eq. Models the Reynoldsstresses are proportional to thedeformation rate but thisholds only when productionand dissipation of k are inbalance

    -Stress anisotropy-Adverce pressure gradientsand recirculation regions-Extra strains (curvature,rotation)

    82

    RANS: M SST

    M (1992) : ,

    .

    H

    S

    T ( =

    I

    R =1, ,1=2, ,2=1.17, 2=0.44, 2=0.083, *=0.09

    B

    C1 C2 , (FC= 1 )

    L

    ( ) ( )kk

    ij

    j

    iijij

    t

    xx

    k

    x

    UEEgraddivdiv

    t

    +

    +

    +=+

    2,

    2

    22

    1,

    23

    22U

    Extra term!Extra term!Extra term!Extra term!

    ( ) 21 1 CFCFC CC +=

  • 8/13/2019 Introduction to CFD -Theory - 2

    42/58

    83

    RANS: R (RSM)

    6 R

    P P

    D D

    D

    T R

    T

    1 .

    Lots of constants!Lots of constants!Lots of constants!Lots of constants!

    84

    RANS: R (RSM)

    BC

    I: R

    O/ :

    F : R= 0 = 0

    : R L R

    ADVANTAGES

    P

    A R

    W , ,

    DISADVANTAGES

    V (7 PDE)

    N

    P

    C

    0= nRij 0= n0= nRij

    0= n

  • 8/13/2019 Introduction to CFD -Theory - 2

    43/58

    85

    L E S (LES)

    Turbulent flow

    Large energetic eddies are resolved

    Small universal eddies are modelled

    FILTERFILTERFILTERFILTER to separate small and large eddies

    LES

    RANS

    DNS

    86

    LES:

    S N S

    ( ) ( ) ( ) 321 ''',, dxdxdxtGt x',x'x,x +

    +

    +

    =

    Filtered functionFiltered functionFiltered functionFiltered function Unfiltered functionUnfiltered functionUnfiltered functionUnfiltered function

    Filter cutoff widthFilter cutoff widthFilter cutoff widthFilter cutoff width

    Spatial filteringSpatial filteringSpatial filteringSpatial filtering

  • 8/13/2019 Introduction to CFD -Theory - 2

    44/58

    87

    LES:

    T

    G

    S

    ( )

    >

  • 8/13/2019 Introduction to CFD -Theory - 2

    45/58

    89

    LES: SGS

    ( )( )

    ( ) jijijijiji

    jijijijijijijjiijijiij

    uuuuuuuuuu

    uuuuuuuuuuuuuuuuuuuu

    ''''

    ''''''

    +++=

    =+++=++==

    Leonard stressesEffects at theresolved scale

    cross-stressesInteractions between

    SGS ans resolvededdies

    LES Reynoldsstresses

    Convectivemomentum transferdue to interactoins

    between SGS eddiesjijiij uuuuL = jijiij uuuuC '' +=

    jiij uuR ''=

    THE SGS STRESSES MUST BE MODELLED!THE SGS STRESSES MUST BE MODELLED!THE SGS STRESSES MUST BE MODELLED!THE SGS STRESSES MUST BE MODELLED!

    90

    LES: SL SGS

    S (1963)

    , B

    T SGS R

    P K (2000)

    SGS

    ijii

    i

    j

    j

    iSGSijiiijSGSij R

    x

    u

    x

    uRER

    3

    12

    3

    12 +

    =+=

    ijii

    i

    j

    j

    iSGSijiiijSGSij

    x

    u

    x

    uE

    3

    12

    3

    12 +

    =+= SGS turbulence modelSGS turbulence modelSGS turbulence modelSGS turbulence model

    ( ) ( )

    ===

    i

    j

    j

    iijijijSGSSGSSGS

    x

    u

    x

    uEEECEC

    2

    122

  • 8/13/2019 Introduction to CFD -Theory - 2

    46/58

    91

    L E S

    Filtered with width = 8DFiltered with width = 4D

    DNS

    FLOW OVER A FLAT PLATEFLOW OVER A FLAT PLATEFLOW OVER A FLAT PLATEFLOW OVER A FLAT PLATE

    M

  • 8/13/2019 Introduction to CFD -Theory - 2

    47/58

    2D

    rpc

    cvDnf

    2

    11

    ==

    =

    =

    c

    pp

    p

    D

    18

    2

    =

    N

    N

    N N

    > 5 D 1/3

  • 8/13/2019 Introduction to CFD -Theory - 2

    48/58

    P

    dispersed flow

    .

    O , ..

    T A , ..

    T A , ..

    F A

    , .

    H , ..

    H , ..

    increasingmassorvolumefraction

    M

    W ?

    W ?

    C ?

    C

    ?

    A

    .

    From A. Bakker

    /

    L D P

    A S

    E

    E G

    V F

    S

    F

    S

    E

    R

  • 8/13/2019 Introduction to CFD -Theory - 2

    49/58

    M

    C (RANS, LES, DN)

    D : E L

    2 :

    ( )

    Eulerian/Eulerian

    k = 1 conitnuum phase, k = 2 dispersed phase

    Eulerian/Lagrangian

    Continuum phase

    Dispersed phase

    L/E

  • 8/13/2019 Introduction to CFD -Theory - 2

    50/58

    A

    I (E/E E/L) :

    I (.. )

    F E/E :

    F E/L (L) (E)

    I (, , )

    M

    E/E

    ADV:

    S

    L L/E

    S

    ( )DIS

    D

    L/E

    ADV:

    B

    ,

    D

    ( , ), ()

    ().

    CON

    H

    D

    Usually E/L for volume fractions of thedispersed phase

  • 8/13/2019 Introduction to CFD -Theory - 2

    51/58

    O : A S M

    (ASM) S

    .

    A .

    RESTRICTIONS

    A < 0.001 0.01.

    O .

    N .

    O .

    N

    N .

    O : ASM

    C

    O

    O

    M

    D

    M

    E

  • 8/13/2019 Introduction to CFD -Theory - 2

    52/58

    O : V F (VOF)

    A I .

    S

    T : J

    M

    S

    I :

    F ( )

    B

    V F (VOF)

    A (

    ).

    F , :

    = 0 ( )

    = 1 ( )

    0 < < 1

    T ()

    :

    M

    C

    kS

    xu

    t i

    kj

    k

    =+

  • 8/13/2019 Introduction to CFD -Theory - 2

    53/58

    105

    R

    106

    R

    T

    Y=

    D= (2/)

    E

    ( ) ( ) kkkkk gradYDdivYdivt

    Y

    &+=+

    U

    Reaction source:Reaction source:Reaction source:Reaction source:generation or distructiongeneration or distructiongeneration or distructiongeneration or distructionof chemical species kdueof chemical species kdueof chemical species kdueof chemical species kdue

    to chemical reactionto chemical reactionto chemical reactionto chemical reaction

    ( ) ( ) ( ) radk

    N

    k

    k

    kkh

    St

    pYgradh

    Schgraddivhdiv

    t

    h+

    +

    +=+

    =1

    11

    U

    Net rate of increaseof enthalpy due to

    diffusion alonggradients of

    enthalpy

    Net rate of increase ofenthalpy due to mass

    diffusion alonggradients of species

    concentration

    Net rate ofincrease of

    enthalpy due topressure work

    Net rate ofincrease of

    enthalpy due toradiative heat

    transfer

  • 8/13/2019 Introduction to CFD -Theory - 2

    54/58

    107

    K

    B A

    C :

    C

    D/

    S

    G

    107

    108

    F N S (FANS)

    W

    R

    F

    i

    i

    i uu

    u

    ==

    ii

    uu = ''

    iii uuu +=

    W :

    R T

    T

    M

    R

    C

    R R

    i

    k

    tk

    tik

    x

    Y

    ScuY

    =

    ''''

    it

    ti

    x

    huh

    =Pr

    ''''

    0=

    +

    i

    i

    xu

    t

    k

    i

    k

    i

    i

    ik

    i

    ikkw

    x

    J

    x

    uY

    x

    uY

    t

    Y&+

    =

    +

    ''''

    j

    i

    ij

    ji

    ij

    i

    ijjg

    xx

    p

    x

    uu

    x

    uu

    t

    u

    +

    +

    =

    +

    ''''

    ( ) radiiijihiii

    ''

    i

    ''

    i

    iQFuuJ

    xx

    uh

    x

    uh

    t

    h&+++

    =

    +

  • 8/13/2019 Introduction to CFD -Theory - 2

    55/58

    109

    T

    T :

    C

    M

    R,

    ,

    ,

    CHEMISTRY TURBULEN

    CE

    T

    ,

    ,

    LrC

    T

    S

    ulDa

    /

    / '

    ==

    110

    C

    ()

    (Y) (T

    ), (

    )

    (CMC ).

    S

    T

    L RRA

    ()

    , ,

    , .

    R

    (EDC, PASR) (

    H H

  • 8/13/2019 Introduction to CFD -Theory - 2

    56/58

    E D

    M H (19)

    I 1 F+O(1+)P

    111

    R = (R = (

    112

    E D/F R

    I 1 F+O(1+)P

    R = (R = (

  • 8/13/2019 Introduction to CFD -Theory - 2

    57/58

    113

    E D C

    FINE STRUCTURES

    H

    P S R

    F

    T*, *, *

    T, ,

    m&

    m&

    RANS CODE VARIABLES

    F

    T*, *, *

    T, ,

    0ci m&

    P

    T :

    PROGRESS VARIABLE

    =0

    =1

    MIXTURE FRACTION Z

    Z = 1

    Z=0

    114

  • 8/13/2019 Introduction to CFD -Theory - 2

    58/58

    115

    T

    T

    . ,