You are on page 1of 20

Getting started: CFD notation

PDE of p-th order scalar unknowns vector unknowns Nabla operator


u u u u u f u, x, t, x , . . . , , , , . . . , =0 xn t x1 x2 tp 1
2 p

u = u(x, t), v = v(x, t),

x Rn , t R, v Rm ,

n = 1, 2, 3

m = 1, 2, . . . v = (v x , v y , v z )

= i x + j y + k z T

x = (x, y, z ),

u =

i u x

j u y +

k u z + j
y

u u u x , y , z

gradient
z

v =

vx x

vy y

vz z

divergence =
2u x2

i
x

k
z

v = det

vx vy vz

vz y vx z vy x

vy z vz x vx y

curl
x

u = (u) = 2 u =

2u y 2

2u z 2

Laplacian

Tensorial quantities in uid dynamics


Velocity gradient
vx x vx y vx z vy x vy y vy z vz x vz y vz z

v = [ v x , v y , v z ] =

1111111111111111 0000000000000000 0000000000000000 1111111111111111

1111111111111111 0000000000000000 0000000000000000 1111111111111111

Remark. The trace (sum of diagonal elements) of v equals v. Deformation rate tensor (symmetric part of v) 1 D (v ) = ( v + v T ) = 2
vx x 1 2 vx y

1 2

vy x vz x

1 2

vx z

1 2

vy x + vy y vy z +

vx y

1 2 1 2

vz x vz y

+ +

vx z vy z

vz y

vz z

Spin tensor

S (v ) = v D (v )

(skew-symmetric part of v)

Vector multiplication rules


Scalar product of two vectors a, b R3 , b1 T a b = a b = [a 1 a2 a3 ] b 2 = a1 b 1 + a2 b 2 + a3 b 3 R b3 v u = vx u u u + vy + vz x y z convective derivative

Example.

Dyadic product of two vectors a1 b 1 a1 a b = abT = a2 [b 1 b 2 b 3 ] = a2 b 1 a3 b 1 a3 a1 b 2 a2 b 2 a3 b 2 a1 b 3 33 R a2 b 3 a3 b 3

a, b R ,

Elementary tensor calculus


1. 2. T = {tij }, T = {tij } R33 , R

3.

4.

5.

2 T 1 + T 2 = {t1 T 1 , T 2 R33 , a R3 ij + tij }, t t t 3 11 12 13 a T = [a1 , a2 , a3 ] t21 t22 t23 = ai [t i 1 , t i 2 , t i 3 ] i=1 t31 t32 t33 i-th row t a t t t 3 1j 11 12 13 1 T a = t21 t22 t23 a2 = t2j aj (j -th column) j =1 t3j a3 t31 t32 t33 2 2 1 1 t2 t1 3 11 t12 t13 11 t12 t13 1 1 1 2 2 2 2 1 2 t1 T T = t21 t22 t23 t21 t22 t23 = ik tkj k=1 2 2 1 1 t2 t1 31 t32 t33 31 t32 t33 3 3 2 t1 ik tik

6.

T : T = tr (T (T ) ) =

2 T

i=1 k=1

Divergence theorem of Gau


\. Let R3 and n be the outward unit normal to the boundary = Then

f dx =

f n ds

for any dierentiable function f (x)

Example. A sphere: = {x R3 : ||x|| < 1}, where ||x|| = xx= so that x2 + y 2 + z 2

= {x R3 : ||x|| = 1}

is the Euclidean norm of x and n =


x ||x||

Consider f (x) = x volume integral:

f 3 in

on

f dx = 3 f n ds =

d x = 3 | | = 3

4 3 1 = 4 3 ds = 4

surface integral:

xx ds = ||x||

||x|| ds =

Governing equations of uid dynamics


Physical principles 1. Mass is conserved 2. Newtons second law 3. Energy is conserved Mathematical equations continuity equation momentum equations energy equation

It is important to understand the meaning and signicance of each equation in order to develop a good numerical method and properly interpret the results

Description of uid motion Eulerian Lagrangian monitor the ow characteristics in a xed control volume
i

z v

k j

track individual uid particles as they move through the ow eld


x

Description of uid motion


Trajectory of a uid particle
z v

x = x(x0 , t) x = x(x0 , y0 , z0 , t) y = y (x0 , y0 , z0 , t) z = z (x0 , y0 , z0 , t)

k i j

dx = vx (x, y, z, t), dt dy = vy (x, y, z, t), dt dz = vz (x, y, z, t), dt

x|t0 = x0 y |t0 = y0 z |t0 = z0

Denition. A streamline is a curve which is tangent to the velocity vector v = (vx , vy , vz ) at every point. It is given by the relation
y

dx dy dz = = vx vy vz

v y(x)

dy dx
x

vy vx

Streamlines can be visualized by injecting tracer particles into the ow eld.

Flow models and reference frames


Eulerian
S

Lagrangian
S

xed CV of a nite size

moving CV of a nite size

dS dV

dS dV

xed innitesimal CV

moving innitesimal CV

Good news: all ow models lead to the same equations

dierential

integral

Eulerian vs. Lagrangian viewpoint


d Denition. Substantial time derivative dt is the rate of change for a moving is the rate of change at a xed point. uid particle. Local time derivative t

Let u = u(x, t), where x = x(x0 , t). The chain rule yields du u u dx u dy u dz u = + + + = + v u dt t x dt y dt z dt t substantial derivative = local derivative + convective derivative Reynolds transport theorem d dt u(x, t) dV =
Vt V Vt

u(x, t) dV + t +

S St

u(x, t)v n dS

rate of change in a moving volume

rate of change in a xed volume

convective transfer through the surface

Derivation of the governing equations


Modeling philosophy 1. Choose a physical principle conservation of mass conservation of momentum conservation of energy 2. Apply it to a suitable ow model Eulerian/Lagrangian approach for a nite/innitesimal CV 3. Extract integral relations or PDEs which embody the physical principle
Generic conservation law Z Z Z u dV + f n dS = q dV t V S V
S n

f = v u d u
f

V dS

ux function

Divergence theorem yields Z Z Z u dV + f dV = q dV V V V t Partial dierential equation u +f =q t in V

Derivation of the continuity equation


Physical principle: conservation of mass d dm = dt dt dV =
Vt V Vt

dV + t

S St

v n dS = 0

accumulation of mass inside CV = net inux through the surface Divergence theorem yields + (v) dV = 0 t Continuity equation + (v) = 0 t

Lagrangian representation (v) = v + v Incompressible ows:


d dt

d + v = 0 dt

=v =0

(constant density)

Conservation of momentum
Physical principle:
dS n h g

f = ma

(Newtons second law) f = g dV + h dS, g h where h=n

total force body forces surface forces Stress tensor

dV

gravitational, electromagnetic,. . . pressure + viscous stress momentum ux

= pI +

For a newtonian uid viscous stress is proportional to velocity gradients: = ( v)I + 2D (v), Normal stress: stretching
xx = yy = zz =
x v + 2 v x vy v + 2 y z v + 2 v z

where D (v) =

1 ( v + v T ), 2

2 3

Shear stress: deformation


xy = yx =

xz = zx yz = zy

` x vz + = v x z vy z + z = v y

vy x

vx y

Derivation of the momentum equations


Newtons law for a moving volume d dt v dV
Vt

=
V Vt

(v) dV + t g dV +

S St

(v v) n dS

=
V Vt

S St

n dS

Transformation of surface integrals (v) + (v v) dV = t [ + g] dV, = pI +

Momentum equations

(v) + (v v) = p + + g t + (v) t
continuity equation

(v) v + (v v) = + v v + v t t
substantial derivative

dv dt

Conservation of energy
Physical principle:
dV dS n h g

e = s + w e s w

(rst law of thermodynamics)

accumulation of internal energy heat transmitted to the uid particle rate of work done by external forces Fouriers law of heat conduction fq = T the heat ux is proportional to the local temperature gradient total force velocity = pI +

Heating: s = q dV fq dS q fq T internal heat sources diusive heat transfer absolute temperature thermal conductivity

Work done per unit time =

w = f v = g v dV + v ( n) dS,

Derivation of the energy equation


Total energy per unit mass: e
|v| 2
2

E =e+

|v|2 2

specic internal energy due to random molecular motion specic kinetic energy due to translational motion

Integral conservation law for a moving volume d dt E dV


Vt

=
V Vt

(E ) dV + t q dV +

S St

E v n dS

accumulation heating work done

=
V Vt

S St

T n dS v ( n) dS

+
V Vt

g v dV +

S St

Transformation of surface integrals (E ) + (E v) dV = t [ (T ) + q + ( v) + g v] dV,

where ( v) = (pv) + ( v) = (pv) + v ( ) + v :

Dierent forms of the energy equation


Total energy equation (E ) + (E v) = (T ) + q (pv) + v ( ) + v : + g v t (E ) E + (E v) = + v E + E + (v) t t t
substantial derivative continuity equation

dE dt

Momentum equations

dv = p + + g dt

(Lagrangian form)

de dv (e) dE = +v = + (ev) + v [p + + g] dt dt dt t

Internal energy equation (e) + (ev) = (T ) + q p v + v : t

Summary of the governing equations


1. Continuity equation / conservation of mass + (v) = 0 t 2. Momentum equations / Newtons second law (v) + (v v) = p + + g t 3. Energy equation / rst law of thermodynamics (E ) + (E v) = (T ) + q (pv) + v ( ) + v : + g v t |v |2 E =e+ , 2 (e) + (ev) = (T ) + q p v + v : t

This PDE system is referred to as the compressible Navier-Stokes equations

Conservation form of the governing equations


Generic conservation law for a scalar quantity u + f = q, t where f = f (u, x, t) is the ux function

Conservative variables, uxes and sources v U = v v + pI v , F = (E + p)v T v E Navier-Stokes equations in divergence form U +F=Q t U R5 ,

Q=

0 g (q + g v)

F R35 ,

Q R5

representing all equations in the same generic form simplies the programming it suces to develop discretization techniques for the generic conservation law

Constitutive relations
Variables: , v, e, p, , T Equations: continuity, momentum, energy

The number of unknowns exceeds the number of equations. 1. Newtonian stress tensor = ( v)I + 2D (v), 2. Thermodynamic relations, e.g. p = RT e = cv T ideal gas law caloric equation of state R cv specic gas constant specic heat at constant volume D (v ) = 1 ( v + v T ), 2 2 3

Now the system is closed: it contains ve PDEs for ve independent variables , v, e and algebraic formulae for the computation of p, and T . It remains to specify appropriate initial and boundary conditions.

Initial and boundary conditions


Initial conditions |t=0 = 0 (x), v|t=0 = v0 (x), e|t=0 = e0 (x)

in

Boundary conditions Inlet in = {x : v n < 0} = in , v = vin , e = ein

Let = in w out

prescribed density, energy and velocity

Solid wall v=0 T = Tw


T n

w = {x : v n = 0}
no-slip condition given temperature or
f

Outlet

v s = vs

v n = vn

out = {x : v n > 0}
or or

p + n n = 0 s n=0
vanishing stress

q =

prescribed heat ux

prescribed velocity

The problem is well-posed if the solution exists, is unique and depends continuously on IC and BC. Insucient or incorrect IC/BC may lead to wrong results (if any).

You might also like