CFD Slayt Friburg

Acknowledgement
This slide set is based on:

Computational Fluid Dynamics
An Introduction n John D. Anderson, Jr., “Computational Fluid Dynamics –
The Basics with Applications,” McGraw-Hill, Inc.,
New York, ISBN 0-07-001685-2
Simulation in Computer Graphics
University of Freiburg
WS 04/05
University of Freiburg - Institute of Computer Science - Computer Graphics Laboratory
Motivation Outline
n introduction
n pre-requisites
n governing equations
n continuity equation
n momentum equation
n summary
n solution techniques
D. Enright, S. Marschner, R. Fedkiw, n Lax-Wendroff
“Animation and Rendering of
n MacCormack
Complex Water Surfaces,”
Siggraph 2002, ACM TOG, n numerical aspects
vol. 21, pp. 736-744, 2002. n recent research in graphics
University of Freiburg - Institute of Computer Science - Computer Graphics Laboratory University of Freiburg - Institute of Computer Science - Computer Graphics Laboratory
•1
Introduction Numerical Solution - Overview
n physical aspects of fluid flow are governed by

three principles
n mass is conserved System of
Governing Equation Approx.
n force = mass × acceleration (Newton’s second law) Discretization Algebraic
Equations Solver Solution
n energy is conserved (not considered in this lecture) Equations
n principles can be described with integral equations or
partial differential equations
Navier-Stokes Finite Discrete Lax-Wendroff
n in CFD, these governing equations are replaced by
Euler Difference Nodal MacCormack
discretized algebraic forms Values
n algebraic forms provide quantities at discrete points
in time and space, no closed-form analytical solution
Continuous Quantities Problem
n x, y, z – 3D coordinate system
n t – time
n ρ ( x,y,z,t ) – density
n v ( x,y,z,t ) – velocity
n v ( x,y,z,t ) = (u ( x,y,z,t ), v ( x,y,z,t ), w ( x,y,z,t ) )T
v (x,y,z,t ) v (x,y,z,t+∆t )
ρ (x,y,z,t ) ρ (x,y,z,t+∆t )
•2
Outline Substantial Derivative of ρ
n introduction n infinitesimally small fluid element moving with the flow

n pre-requisites
n governing equations n ( x1,y1,z1 ) – position at time t1
n continuity equation n ( x2,y2,z2 ) – position at time t2
n momentum equation
n summary
n v1(x1,y1,z1,t1) = (u (x1,y1,z1,t1 ),v (x1,y1,z1,t1 ),w (x1,y1,z1,t1 ))T
n v2(x2,y2,z2,t2) = (u (x2,y2,z2,t2 ),v (x2,y2,z2,t2 ),w (x2,y2,z2,t2 ))T
n Lax-Wendroff
n MacCormack
n numerical aspects n ρ1 ( x1,y1,z1,t1 )
n recent research in graphics n ρ2 ( x2,y2,z2,t2 )
Substantial Derivative of ρ Substantial Derivative
n Taylor series at point 1 n substantial derivative of ρ

 ∂ρ   ∂ρ   ∂ρ   ∂ρ  Dρ ∂ρ ∂ρ ∂ρ ∂ρ
ρ2 = ρ1 +   ( x 2 − x1 ) +   ( y 2 − y 1 ) +   (z 2 − z 1 ) +   (t2 − t1 ) ≡u +v +w +
 ∂x  1  ∂y 1  ∂ z 1  ∂ t 1 Dt ∂x ∂y ∂z ∂t
→
n substantial derivative = local derivative + convective derivative
ρ 2 − ρ 1  ∂ ρ  x 2 − x1  ∂ ρ  y 2 − y1  ∂ ρ  z 2 − z 1  ∂ ρ 
=  +   +  +  T
t2 − t1  ∂ x  1 t 2 − t1  ∂ y 1 t 2 − t1  ∂ z  1 t 2 − t1  ∂ t  1 D ∂  ∂ ∂ ∂ 
≡ + (v ⋅∇) ∇ =  , , 
t2 → t1 Dt ∂t  ∂x ∂y ∂z 
ρ 2 − ρ 1  ∂ρ   ∂ρ   ∂ρ   ∂ρ  local derivative – time rate of change at a fixed location
=  u +   v +   w +  n
t2 − t1  ∂ x 1  ∂y 1  ∂z  1  ∂t  1
n convective derivative – time rate of change due to fluid flow
n subst. derivative = total derivative with respect to time
•3
Divergence of v Outline
n divergence of velocity v = time rate of change of the n introduction

volume δV of a moving fluid element per unit volume n pre-requisites
∂u ∂v ∂w 1 D (δ V ) n governing equations
∇⋅v = + + = continuity equation
∂x ∂y ∂z δV Dt n
n momentum equation
n summary
n Lax-Wendroff
n MacCormack
n numerical aspects
n recent research in graphics
Continuity Equation Continuity Equation
n mass is conserved
n infinitesimally small fluid element moving with the flow
Dρ
n fixed mass δm, variable volume δV, variable density ρ + ρ∇ ⋅ v = 0
δm = ρ ⋅ δV Dt
n time rate of change of the mass of the moving fluid substantial derivative – divergence of velocity -
element is zero time rate of change of time rate of change of volume
D δm D (ρ ⋅ δ V ) Dρ D δV density of a moving of a moving fluid element per
= = δV +ρ =0 fluid element volume
Dt Dt Dt Dt
Dρ 1 D δV Dρ
+ρ =0 → + ρ∇ ⋅ v = 0
Dt δ V Dt Dt
•4
Continuity Equation Continuity Equation
n non-conservation form (considers moving element) n motivation for conservation form

Dρ n infinitesimally small element at a fixed location
+ ρ∇ ⋅ v = 0 n mass flux through element
Dt
n difference of mass inflow and outflow = net mass flow
n manipulation n net mass flow = time rate of mass increase
Dρ ∂ρ ∂ρ
+ ρ∇ ⋅ v = + v ⋅ ∇ ρ + ρ∇ ⋅ v = + ∇ ⋅ (ρ v )
Dt ∂t ∂t
∂ρ
n conservation form (considers element at fixed location) + ∇ ⋅ (ρ v ) = 0
∂t
∂ρ
+ ∇ ⋅ (ρ v ) = 0
∂t time rate of mass net mass flow
increase per volume per volume
Outline Momentum Equation
n introduction n consider a moving fluid element

n pre-requisites n physical principle: F = m⋅a (Newton’s second law)
n governing equations n two sources of forces
n continuity equation n body forces
n momentum equation n gravity
n summary n surface forces

n solution techniques n based on pressure distribution on the surface
n based on shear and normal stress distribution on the surface

n Lax-Wendroff
due to deformation of the fluid element friction
n MacCormack
n numerical aspects fluid pressure normal shear velocity
n recent research in graphics element stress stress
gravity velocity
•5
Body Force Pressure Force
n f = body force per unit mass n consider the x component

n gravity: f = δm⋅g / δm = g n pressure force dz
→ body force on fluid element acts orthogonal
p dy dz ( p + ∂p/ ∂x dx ) dy dz
to surface into
ρ f (dx dy dz )
the fluid element dx
dy
dy
fluid
element n net pressure force in x direction
  ∂p  ∂p
dz  p −  p + ∂ dx   dy dz = − dx dy dz
  x  ∂x
dx
Stress Stress
n normal stress – related to the time rate of change of volume n normal stress is related to pressure orthogonal to surface
n shear stress – related to the time rate of change n shear stress is related to friction parallel to surface
of the shearing deformation n friction and pressure are related to the velocity gradient
n τjk – stress in k direction applied to a surface perpendicular n friction (shear stress) models viscosity (viscous flow)
to the j axis n in contrast to inviscid flow, where friction is not considered
n τxx – normal stress in x direction
friction
n τzx , τyx – shear stresses in x directions
pressure normal shear velocity
τyx
stress stress
y
velocity
τxx
x
•6
Surface Forces in x Momentum Equation in x
(τyx + ∂ τyx / ∂y dy) dx dz
dz
m a x = Fx
p dy dz ( p + ∂p/ ∂x dx ) dy dz
Du  ∂p ∂τ xx ∂τ yx ∂τ zx 
τxx dy dz (τxx + ∂ τxx / ∂x dx) dy dz ρ dx dy dz = − + + +  dx dy dz + ρ f x dx dy dz
dx Dt  ∂x ∂x ∂y ∂z 
dy
τyx dx dz τzx not illustrated mass acceleration surface force body force
  ∂p   ∂τ   time rate of change pressure gravity
Fx =  p −  p + dx  dy dz + τ xx + xx dx  − τ xx  dy dz
  ∂ x   ∂x   of velocity of a normal stress
 ∂τ    ∂τ  moving fluid element shear stress

+ τ yx + yx dy  − τ yx  dx dz + τ zx + zx dz  − τ zx  dy dz
 ∂y    ∂z  
Du ∂p ∂τ ∂τ yx ∂τ zx
ρ = − + xx + + + ρ fx
 ∂p ∂τ xx ∂τ yx ∂τ zx  Dt ∂x ∂x ∂y ∂z
Fx =  − + + +  dx dy dz
 ∂x ∂x ∂y ∂z 
Momentum Equation Momentum Equation in x
n viscous flow, non-conversation form n obtaining the conservation form of Navier-Stokes

n Navier-Stokes equations Du ∂u
ρ =ρ + ρ v ⋅ ∇u
∂τ yx ∂τ zx ∂t
Du ∂p ∂τ Dt
ρ = − + xx + + + ρ fx
Dt ∂x ∂x ∂y ∂z ∂( ρu) ∂u ∂ρ ∂u ∂ρ ∂(ρ u)
=ρ +u ρ =u −
∂t ∂t ∂t ∂t ∂t ∂t
Dv ∂p ∂τ xy ∂τ yy ∂τ zy ∇ ⋅ (ρ u v ) = u ∇ ⋅ ( ρ v) + ( ρ v )∇ u (ρ v ) ∇u = ∇ ⋅ ( ρ u v) − u ∇ ⋅ ( ρ v)
ρ =− + + + + ρ fy
Dt ∂y ∂x ∂y ∂z
Du ∂( ρu)  ∂ρ  Du ∂ ( ρu )
ρ = − u  + ∇ ⋅ (ρ v) + ∇ ⋅ (ρ u v ) ρ = + ∇ ⋅ (ρ u v )
Dw ∂p ∂τ ∂τ yz ∂τ zz Dt ∂t  ∂t  Dt ∂t
ρ = − + xz + + + ρ fz continuity equation
Dt ∂z ∂x ∂y ∂z
•7
Momentum Equation Newtonian Fluids
n Navier-Stokes equations in conservation form n Newton: shear stress in a fluid is proportional

to velocity gradients
∂(ρ u ) ∂p ∂τ ∂τ yx ∂τ zx most fluids can be assumed to be newtonian,
+ ∇ ⋅ ( ρ u v ) = − + xx +
n
+ + ρ fx
∂t ∂x ∂x ∂y ∂z but blood is a popular non-newtonian fluid
n Stokes:
∂(ρ v ) ∂p ∂τ xy ∂τ yy ∂τ zy
+ ∇ ⋅ (ρ v v) = − + + + + ρ fy τ xx = λ (∇ ⋅ v ) + 2 µ
∂u
τ yy = λ (∇ ⋅ v ) + 2 µ
∂v
τ zz = λ (∇ ⋅ v ) + 2µ
∂w
∂t ∂y ∂x ∂y ∂z ∂x ∂y ∂z
∂(ρ w ) ∂p ∂τ ∂τ yz ∂τ zz  ∂v ∂u   ∂u ∂w   ∂w ∂v 
τ xy = τ yx = µ  +  τ xz = τ zx = µ  + τ yz = τ zy = µ  + 
+ ∇ ⋅ (ρ w v ) = − + xz + + + ρ fz  ∂x ∂y 

 ∂z ∂x   ∂y ∂z 
∂t ∂z ∂x ∂y ∂z
n µ is the molecular viscosity coefficient
n λ is the second viscosity coefficient with λ = -2/3 ⋅ µ
Perfect Gas Momentum Equation in x
intermolecular forces are negligible

∂(ρ u ) ∂τ yx ∂τ zx
n
∂p ∂τ
R – specific gas constant + ∇ ⋅ (ρ u v ) = − + xx + + + ρ fx
∂t ∂x ∂x ∂y ∂z
n
n T – temperature
n termal equation of state p = ρ R T n now,
p can be expressed using density and
τ can be expressed using velocity gradients,
fx is user-defined (e. g. gravity)
•8
Governing Equations –
Outline
Viscous Compressible Flow
n introduction n thus, four equations and four unknowns ρ, u, v, w
pre-requisites ∂ρ
+ ∇ ⋅ (ρ v ) = 0
n
continuity equation
n governing equations ∂t
n continuity equation
∂(ρ u ) ∂p ∂τ ∂τ yx ∂τ zx
n momentum equation
+ ∇ ⋅ ( ρ u v ) = − + xx + + + ρ fx
∂t ∂x ∂x ∂y ∂z
momentum equation
n summary
∂ (ρ v ) ∂p ∂τ xy ∂τ yy ∂τ zy
n Lax-Wendroff + ∇ ⋅ (ρ v v) = − + + + + ρ fy
n MacCormack ∂t ∂y ∂x ∂y ∂z
n numerical aspects
∂(ρ w ) ∂p ∂τ ∂τ yz ∂τ zz
n recent research in graphics + ∇ ⋅ (ρ w v ) = − + xz + + + ρ fz
∂t ∂z ∂x ∂y ∂z
Governing Equations – Comments

Inviscid Compressible Flow on the Governing Equations
n Euler equations n coupled system of nonlinear partial differential equations
∂ρ normal and shear stress terms are functions of the
+ ∇ ⋅ (ρ v ) = 0
n
continuity equation
∂t velocity gradients
n pressure is a function of the density
∂(ρ u ) ∂p
+ ∇ ⋅ (ρ u v ) = − + ρ f x n only momentum equations are Navier-Stokes equations,
∂t ∂x however the name is commonly used for the full set
momentum equation
of governing equations (plus energy equation)

∂(ρ v ) ∂p
+ ∇ ⋅ (ρ v v ) = − + ρ f y
∂t ∂y
∂(ρ w ) ∂p
+ ∇ ⋅ (ρ w v ) = − + ρ f z
∂t ∂z
•9
Governing Equations in
Generic Form
Conservation Form
n can be obtained directly from a control volume fixed in n viscous flow
space  ρ   ρu   ?v   ?w   0 
         
n consider flux of mass, momentum (and energy) into and ρu
U= 
F=
ρ u u + p − τ xx   ? u v − t yx 
G =

H =
? u w − t zx 
J =
 ? fx 
   ? v w + p − t zy  ? f 
out of the volume 
ρv
 
ρ v u − τ xy
 
? v v + p − t yy
    y
 ρ w  ρ w u −τ   ? wv − t   ? ww + p − t  ? f 
n have a divergence term which involves the flux of    xz   yz   zz   z
mass (ρv), momentum in x, y, z (ρu v, ρv v, ρw v)
n inviscid flow
n can be expressed in a generic form
 ρ   ρu   ?v   ?w   0 
         
ρu  ρ uu + p  ?u v   ?u w   ? fx 
∂U ∂F ∂G ∂H U= F = = H = J =
+ + + =J ρv  ρ vu  ?v w + p 
G  ?v v + p ? f 
         y
∂t ∂x ∂y ∂z  ρ w  ρ wu   ?w w + p ? f 
     ?w v     z
Motivation for Different Forms

Generic Form
of the Governing Equations
n U – solution vector (to be solved for) n in many real-world problems, discontinuous changes in
n F, G, H – flux vectors the flow-field variables ρ, v occur (shocks, shock waves)
n J – source vector (external forces, energy changes) n problem: differential form of the governing equations
∂U ∂F ∂G ∂H
assumes differentiable (continuous) flow properties
=J− + + simple 1-D shock wave: ρ1 u1 = ρ2 u2
∂t ∂x ∂y ∂z n
n dependent flow field variables can be solved progressively ρ u ρu

in steps of time (time-marching )
n spatial derivatives are know from previous time steps
x x x
n numbers can be obtained for density ρ and flux variables
ρu, ρv, ρw n conservation form solves for ρu, sees no discontinuity
n positive effect on numerical accuracy and stability
•10
Outline Lax-Wendroff Technique
n introduction n flow-field variables ρ, u, v, w are known at each discrete

n pre-requisites spatial position ( x, y, z ) at time t
n governing equations n Lax-Wendroff computes all information at time t +∆t
n continuity equation n second-order accuracy in space and time
n momentum equation
n summary n explicit, finite-difference method
n particularly suited to marching solutions
n Lax-Wendroff
n marching suited to hyperbolic, parabolic PDEs
n MacCormack
n numerical aspects n unsteady (time-dependent), compressible flow
n recent research in graphics is governed by hyperbolic PDEs
Lax-Wendroff
Taylor Series Expansions
Example
n 2D, inviscid flow, no body force, in non-conservation form n ρti,j – density at position (i, j ) and time t
n ρti,j , u ti,j , v ti,j are known
 (∆ t )2
t
 ∂ρ   ∂2ρ
t
∂ρ  ∂u ∂ρ ∂v ∂ρ 
= −  ρ +u +ρ +v  ρ it,+j ∆ t = ρ it, j +   ∆ t +   + ...
∂t ∂ ∂ ∂ ∂ y   ∂t i , j  ∂t
2
 x x y  i, j 2
∂u  ∂u ∂u R T ∂ρ 
 (∆ t )
t
 ∂u   ∂ 2u
t 2
= −  u +v + 
∂t  ∂x ∂y ρ ∂x 
u it ,+j ∆ t = u it, j +   ∆ t +  2  + ...
 ∂t i , j  ∂t i , j 2
∂v  ∂v ∂v R T ∂ρ 
= −  u +v +   ∂ 2 v  (∆ t )2
t
ρ ∂ y   ∂v 
t
∂t  ∂ x ∂y v it,+j ∆ t = v it, j +   ∆ t +  2  + ...
 ∂ t i , j  ∂t  i, j 2
•11
Taylor Expansion - Density Taylor Expansion - Density
 (∆ t )
t
 ∂ρ   ∂ 2ρ
t 2
ρ it,+j ∆ t = ρ it, j +   ∆ t +  2  + ... n ∂2ρ/∂t 2 can be obtained by differentiating the governing

 ∂ t  i, j  ∂t  i, j 2 equation with respect to time
n ∂ρ/∂t can be replaced by spatial derivatives  ∂ 2u
∂2ρ ∂u ∂ ρ ∂ 2 ρ ∂ ρ ∂u ∂2v ∂v ∂ρ ∂ 2 ρ ∂ρ ∂v 
given in the governing equations = −  ρ + +u + +ρ + +v + 
∂t 2  ∂x ∂t ∂x ∂t ∂x ∂ t ∂x ∂t ∂y ∂t ∂y ∂t ∂ y ∂ t ∂ y ∂ t 
∂ρ  ∂u ∂ρ ∂v ∂ρ 
= −  ρ +u +ρ +v  n mixed second derivatives can be obtained by differentiating
∂t  ∂x ∂x ∂y ∂y  the governing equations with respect to a spatial variable,
using second-order central differences e. g.
n 2
∂ u 2
 ∂ u  ∂u 
2 2
∂ u
2
∂ u ∂v R T  ∂ ρ 
2
R T ∂ ρ 
= −u + +v + − 2   +
∂ x ∂t  ∂ x 2  ∂ x  ∂ x ∂ y ∂ y ∂ x ρ  ∂ x  ρ ∂ x 2 
 t u it+ 1, j − u it −1, j ρt − ρ it− 1, j − vit −1, j ρt − ρ it− 1, j  
 ∂ρ 
t
vt
  = −  ρ i , j + u it, j i + 1, j + ρ it, j i + 1, j + v it, j i + 1, j 

 ∂t  i , j  2 ∆x 2∆ x 2 ∆y 2∆ y  n brackets missing on pp. 219, 220 in Anderson’s book!
Taylor Expansion - Density Outline
n → ∂2ρ/∂t 2 can be computed using central differences for n introduction

spatial derivatives n pre-requisites
n higher-order terms are required, e. g. n governing equations
t continuity equation
∂ u − 2u + u
t t t n
2
u i +1, j i −1, j
 2  = i, j
momentum equation
(∆x )
n
 ∂ x  i, j
2
n summary
t
 ∂ 2u  ut + ut − ut − ut Lax-Wendroff
  = i+1, j +1 i −1, j−1 i −1, j +1 i +1, j −1 n
 ∂x∂y i, j 4∆x∆y n MacCormack

n numerical aspects
n ... u, v are computed in the same way n recent research in graphics
•12
MacCormack Technique MacCormack Technique
n same characteristics like Lax-Wendroff n predictor step for density

n second-order accuracy in space and time  ∂ρ 
t
 t u it +1 , j − u it−1 , j ρt − ρ it−1, j vt − v it− 1, j ρt − ρ it− 1, j 
  = −  ρ i , j + u it, j i + 1, j + ρ it, j i + 1, j + v it, j i + 1, j 

n requires only first time derivative  ∂ t i, j  2 ∆ x 2 ∆ x 2 ∆ y 2 ∆y 
n predictor – corrector  ∂ρ 
t
ρ it,+j ∆ t = ρ it, j +   ∆t
n illustrated for density  ∂ t i , j
av
 ∂ρ 
ρ it,+j ∆ t = ρ it, j +   ∆t n u, v are predicted the same way
 ∂t  i, j
 ∂u   ∂v 
t t
u it, +j ∆ t = u ti , j +   ∆t v it,+j ∆ t = vit, j +   ∆t

 ∂t  i , j  ∂ t i, j
MacCormack Technique Comments
n corrector step for density n Lax-Wendroff and MacCormack can be used for
 ∂ρ 
t + ∆t
 u t + ∆t
−u t + ∆t
ρ t + ∆t
−ρ t+∆ t
v t+∆ t
−v t + ∆t
ρ t + ∆t
−ρ t + ∆t
 unsteady flow
  = −  ρ it,+j ∆ t
i +1 , j i −1 , j
+ u it,+j ∆ t
i +1 , j i −1 , j
+ ρ it,+j ∆ t
i +1 , j i −1 , j
+ v it,+j ∆ t
i +1 , j i − 1, j

 ∂t 
 i , j 2∆ x 2∆x 2∆ y 2∆ y  n non-conservation form
 
n conservation form
1  ∂ ρ  
t + ∆t
 ∂ρ   ∂ρ 
av t
ρ it,+j ∆ t = ρ it, j +   ∆ t = ρ i, j + 
t
 +   ∆t n viscous flow
 ∂  i, j   ∂  i , j  ∂t 
t 2 t  i , j  n inviscid flow
n higher-order accuracy required to avoid numerical
n corresponds to Heun scheme for ODEs dissipation, artificial viscosity, numerical dispersion
n other higher-order schemes, e. g. Runge-Kutta 2 or 4, n note, that viscosity is represented by second partial
could be used as well derivatives in the governing equations
•13
Outline Numerical Effects
n introduction n truncation error causes dissipation and dispersion

n pre-requisites n numerical dissipation is caused by even-order terms
n governing equations of the truncation error
n continuity equation n numerical dispersion is caused by odd-order terms
n momentum equation n leading term in the truncation error dominates the
n summary behavior
u u u
n Lax-Wendroff
n MacCormack
n numerical aspects x x x
n recent research in graphics original function dissipation dispersion
Numerical Effects Summary
n artificial viscosity compromises the accuracy, but n governing equations for compressible flow
improves the stability of the solution n continuity equation, momentum equation
n adding artificial viscosity increases the probability of n Navier-Stokes (viscous flow) and Euler (inviscid flow)
making the solution less accurate, but improves n discussion of conservation and non-conservation form
the stability n explicit time-marching techniques
n similar to iterative solution schemes for implicit integration n Lax-Wendroff
schemes, where a smaller number of iterations introduces n MacCormack
“artificial viscosity”, but improves the stability n numerical aspects
n recent research in graphics
•14
Recent Research Recent Research
M. Carlson, P. J. Mucha, G. Turk, “Rigid Fluid: Animating the T. G. Goktekin, A. W. Bargteil, J. F. O’Brien,
Interplay Between Rigid Bodies and Fluid,” Siggraph 2004. “A Method for Animating Viscoelastic Fluids,” Siggraph 2004.
•15

CFD Slayt Friburg

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

CFD Slayt Friburg

Uploaded by

Copyright:

Available Formats

Acknowledgement

This slide set is based on:

University of Freiburg - Institute of Computer Science - Computer Graphics Laboratory

n physical aspects of fluid flow are governed by

Continuous Quantities Problem

n v ( x,y,z,t ) = (u ( x,y,z,t ), v ( x,y,z,t ), w ( x,y,z,t ) )T

n introduction n infinitesimally small fluid element moving with the flow

Substantial Derivative of ρ Substantial Derivative

n Taylor series at point 1 n substantial derivative of ρ

n divergence of velocity v = time rate of change of the n introduction

Continuity Equation Continuity Equation

n non-conservation form (considers moving element) n motivation for conservation form

Outline Momentum Equation

n introduction n consider a moving fluid element

n summary n surface forces

n based on shear and normal stress distribution on the surface

n f = body force per unit mass n consider the x component

Momentum Equation Momentum Equation in x

n viscous flow, non-conversation form n obtaining the conservation form of Navier-Stokes

n Navier-Stokes equations in conservation form n Newton: shear stress in a fluid is proportional

Perfect Gas Momentum Equation in x

intermolecular forces are negligible

Governing Equations – Comments

of governing equations (plus energy equation)

Motivation for Different Forms

n dependent flow field variables can be solved progressively ρ u ρu

n introduction n flow-field variables ρ, u, v, w are known at each discrete

ρ it,+j ∆ t = ρ it, j +   ∆ t +  2  + ... n ∂2ρ/∂t 2 can be obtained by differentiating the governing

Taylor Expansion - Density Outline

n → ∂2ρ/∂t 2 can be computed using central differences for n introduction

 ∂x∂y i, j 4∆x∆y n MacCormack

n same characteristics like Lax-Wendroff n predictor step for density

u it, +j ∆ t = u ti , j +   ∆t v it,+j ∆ t = vit, j +   ∆t

MacCormack Technique Comments

n introduction n truncation error causes dissipation and dispersion

Numerical Effects Summary

You might also like