Professional Documents
Culture Documents
u u
u
u u
+
(
c
c
+
c
c
c
c
+
|
.
|
\
|
c
c
c
c
+
c
c
=
|
|
.
|
\
|
c
c
+
c
c
+
c
c
+
c
c
2
2
2 2
2
2
2
2 1 ) ( 1
57
Cylindrical Coordinates
The Navier-Stokes equation in the u-direction is:
u
u u u
u u u u u u
u u
u
u
g
z
v v
r
v
r r
rv
r r
p
r
z
v
v
r
v v v
r
v
r
v
v
t
v
r
z
r
r
+
(
c
c
+
c
c
+
c
c
+
|
.
|
\
|
c
c
c
c
+
c
c
=
|
.
|
\
|
c
c
+ +
c
c
+
c
c
+
c
c
2
2
2 2
2
2
2 1 ) ( 1 1
58
Cylindrical Coordinates
The Navier-Stokes equation in the z-direction is:
z
z z z
z
z
z z
r
z
g
z
v v
r r
v
r
r r z
p
z
v
v
v
r
v
r
v
v
t
v
u
+
(
c
c
+
c
c
+
|
.
|
\
|
c
c
c
c
+
c
c
=
|
.
|
\
|
c
c
+
c
c
+
c
c
+
c
c
2
2
2
2
2
1 1
59
We will return to the pipe flow problem from the start of
the lecture and solve it using the Navier-Stokes
equations.
Continuity:
0
0
=
=
u
v
v
r
0
1 ) ( 1
=
c
c
+
c
c
+
c
c
z
v v
r r
rv
r
z r
u
u
0 0
60
r-direction Navier-Stokes:
0
2 1 ) ( 1
2
2
2
2
2 2
2
2
2
=
c
c
=
c
c
+
(
c
c
+
c
c
c
c
+
|
.
|
\
|
c
c
c
c
+
c
c
=
|
|
.
|
\
|
c
c
+
c
c
+
c
c
+
c
c
z
v
r
p
g
z
v v
r
v
r r
rv
r r r
p
z
v
v
r
v v
r
v
r
v
v
t
v
z
r
z r r
r
z
r r
r
r
u u
u
u u
61
u-direction Navier-Stokes:
0
2 1 ) ( 1 1
2
2
2 2
2
2
=
c
c
+
(
c
c
+
c
c
+
c
c
+
|
.
|
\
|
c
c
c
c
+
c
c
=
|
.
|
\
|
c
c
+ +
c
c
+
c
c
+
c
c
u
u u
u
u
u
u u u
u u u u u u
p
g
z
v v
r
v
r r
rv
r r
p
r
z
v
v
r
v v v
r
v
r
v
v
t
v
r
z
r
r
62
z-direction Navier-Stokes
z
z
z
z z z
z
z
z z
r
z
g
r
v
r
r r z
p
g
z
v v
r r
v
r
r r z
p
z
v
v
v
r
v
r
v
v
t
v
u
+
|
.
|
\
|
c
c
c
c
=
c
c
+
(
c
c
+
c
c
+
|
.
|
\
|
c
c
c
c
+
c
c
=
|
.
|
\
|
c
c
+
c
c
+
c
c
+
c
c
2
2
2
2
2
1 1
63
Integrate:
2
2
1
1
1
2
4
0 0
2
2
c
r
g
L
p p
v
r at finite
r
v
keep to c
r
v
r
c r
g
L
p p
r
v
r c
r
g
L
p p
g
r
v
r
r r z
p
z
o L
z
z
z
z
o L
z
z
o L
z
z
+
|
.
|
\
|
=
=
c
c
=
c
c
= +
|
.
|
\
|
c
c
= +
|
.
|
\
|
+
|
.
|
\
|
c
c
c
c
=
c
c
64
Viscous Flows: Exact Solutions/Pipe Flow
Solving the equations with the no slip
conditions applied at r = R (the walls of the
pipe).
Parabolic Velocity Profile
65
Viscous Flows: Exact Solutions/Pipe Flow
The volumetric flow rate:
The mean velocity:
Pressure drop per length of pipe:
The maximum velocity:
Non-Dimensional velocity profile:
For Laminar Flow:
Substituing Q,
66
Conservation of Energy
The energy equation is developed similar to the momentum equation for an
infinitesimal control volume.
67
Conservation of Energy
The energy equation is developed similar to the momentum equation for an
infinitesimal control volume.
(Heat and Work)
Internal
Kinetic
Potential
(Time rate of change
following the particle)
Differentiate:
To get the L.H.S:
Now for the R.H.S., define the fluid properties of Heat and Work:
Heat Conduction into the element, Fouriers Law
Heat per Unit Area
Heat:
68
Conservation of Energy
Now, we do a control volume analysis on our control element:
Heat Flow into the left x-face of the element
Heat Flow out of the right x-face of the element
The above can be written for all six faces of the cube with the net result
between the in and out:
The net heat flow is transferred to the element, neglecting production terms
Heat: Heat Conduction into the element, Fouriers Law
Heat per Unit Area
69
Conservation of Energy
Work: Work is done on the element per unit area.
on the left x-face
on the right x-face
We can do the same for the other faces, and the net rate of work done is:
In condensed form:
70
Conservation of Energy
We can rewrite the equation using and identity:
We note, then, that from the momentum equation:
Now, the rate of change of work is the following:
Kinetic Potential
=
Now, when we substitute work and heat back into the governing equation:
We note potential and kinetic energy portions cancelled on each side!
71
Conservation of Energy
Now, we can split the stress tensor into pressure and viscous terms:
Using continuity, we can rewrite the pressure term:
Now, rewriting the Conservation of Energy:
Noting, the definition of fluid enthalpy:
And, defining the dissipation function:
This term always takes energy from the flow!
72
Conservation of Energy
Writing out the terms of Viscous Dissipation for a Newtonian Fluid:
Now, with the substitutions, the energy equation take the following form:
We note,
73
Conservation of Energy
Now, lets assume the flow is incompressible:
Enthalpy:
Then,
If the flow velocity is low relative to Heat Transfer then terms of order
U disappear.
| is the thermal expansion coefficient, for a
perfect gas the second term goes to zero!
If, we assume constant thermal conductivity:
Heat Convection Equation
74
Summary of Governing Equations
Mass:
Momentum:
Energy:
Most General forms of the Equations:
Only Assumptions:
(1) The fluid is a continuum
(2) the particles are essentially in thermodynamics equilibrium
(3) Only body forces are gravity
(4) The Heat conduction follows Fouriers Law
(5) There are no internal heat sources.
75
Summary of Governing Equations
Some general comments on the general form of the governing equations:
1. They are a coupled system of non-linear partial differential equations
you must solve energy, continuity, and linear momentum
simultaneously. No closed form solution exists!
2. For Newtonian flow, the shear and normal stresses can be written in
terms of the velocity gradients introducing no new unknowns.
3. There appear to be five equations and nine unknowns in the system of
equations: , , k, p, u, v, w, h, and T.
4. However, we note the following:
5. Now, we have five unknowns and five equations
) , ( ), , (
) , ( ), , (
T p k k T p h h
T p T p
= =
= =
76
Summary of Governing Equations
In, general in Fluid Mechanics/CFD we often work with a simplified form of the
equations known as the Navier-Stokes Equations:
Additional Assumptions:
(1) The fluid is Newtonian
(2) Incompressible
(3) Constant properties (k, , )
where,
uncoupled equations: The fluid flow
can be solved independent of the Heat
Transfer
77
Summary of Governing Equations
Some general comments on the Navier-Stokes governing equations:
1. They are non-linear partial differential equations which are uncoupled
in energy, and linear momentum. We can solve linear momentum and
continuity equations separately for the flow field without knowledge of
the Temperature field (4 Equations, 4 unknowns, u, v, w, p).
2. For Newtonian flow, the shear and normal stresses can be written in
terms of the velocity gradients introducing no new unknowns.
3. There appear to be five equations and 5 unknowns in the system of
equations: p, u, v, w, and T.
4. If the convective term disappears we have a linear solution.
5. If the convective term remains we have a non-linear solution.
6. The Energy equations relies on the solution of the flow field for its
solution.
Viscous Flow Equations
78
Summary of Governing Equations
Summary of the Euler form of the governing equations: Inviscid Flow Equations
Linear Momentum:
Continuity and Energy are the same as for Navier-Stokes Equations
Some general remarks:
(1) The system of equations have five unknowns and five equations (same as
Navier-Stokes)
(2) Flow is Inviscid (frictionless), Pressure is the only normal stress, and
there are no shear stresses.
(3) A specialized case of inviscid flow is irrotational flow.
(4) The energy and momentum equations are also uncoupled in this set of
equations.
79
Physical Boundary Conditions
Types of Boundary Conditions:
Fluid/Gas-Solid Interface
Fluid-Fluid Interface
Gas-Fluid Interface
80
Physical Boundary Conditions
No Slip Condition:
At the fluid-boundary interface the velocities must be equal. If the boundary
is stationary, then u, v, w = 0.
The temperature of the fluid has to equal the temperature
of boundary at the interface.
Heat Flux in the fluid must equal the heat flux of the solid at the interface
At a solid boundary:
No Temperature Jump:
Equality of Heat Flux:
Examples:
Stationary Solid Boundary
Moving Boundary:
81
Computational Fluid Dynamics: Differential Analysis
Governing Equations:
Navier-Stokes:
Continuity:
The above equations can not be solved for most practical problems with analytical
methods so Computational Fluid Dynamics or experimental methods are
employed.
The numerical methods employed are the following:
1. Finite difference method
2. Finite element (finite volume) method
3. Boundary element method.
These methods provide a way of writing the governing equations in discrete
form that can be analyzed with a digital computer.
82
Computational Fluid Dynamics: Finite Element
These methods discretize the domain of the flow of interest (Finite Element
Method Shown):
The discrete governing equations are solved in every element. This
method often leads to 1000 to 10,000 elements with 50,000 equations
or more that are solved.
83
Computational Fluid Dynamics: Finite Difference
These methods discretize the domain of the flow of interest as well (Finite
Difference Method Shown):
Finite Difference Mesh:
Comparison between
Experiment and CFD
Analysis:
84
Computational Fluid Dynamics: Pitfalls
Numerical Solutions can diverge or exhibit unstable wiggles.
Finer grids may cause instability in the solution rather than better
results.
Large flow domains can be computationally intensive.
Turbulent flows have yet to be well described with CFD.
85
Inviscid Flow: Bernoulli Equation
Daniel Bernoulli
(1700-1782)
Earlier, we derived the Bernoulli Equation from a direct
application of Newtons Second Law applied to a fluid particle
along a streamline.
Now, we derive the equation from the Euler Equation
First assume steady state:
Select, the vertical direction as up, opposite gravity:
Use the vector identity:
Now, rewriting the Euler Equation:
Rearrange:
86
Inviscid Flow: Bernoulli Equation
Now, take the dot product with the differential length ds along a streamline:
ds and V are parrallel,
, is perpendicular to V, and thus to ds.
We note,
Now, combining the terms:
Integrate:
Then,
1) Inviscid flow
2) Steady flow
3) Incompressible flow
4) Along a streamline
87
Inviscid Flow: Irrotational Flow
Irrotational Flow: the vorticity of an irrotational flow is zero.
= 0
For a flow to be irrotational, each of the vorticity vector components must be
equal to zero.
The z-component:
The x-component lead to a similar result:
The y-component lead to a similar result:
Uniform flow will satisfy these conditions:
There are no shear forces in irrotational flow.
88
Inviscid Flow: Irrotational Flow
Example flows, where inviscid flow theory can be used:
Viscous Region
Inviscid Region
89
Inviscid Flow: Bernoulli Irrotational Flow
Recall, in the Bernoulli derivation,
However, for irrotational flow,
.
Thus, for irrotational flow, we do not have to follow a streamline.
Then,
1) Inviscid flow
2) Steady flow
3) Incompressible flow
4) Irrotational Flow
90
Potential Flow: Velocity Potential
For irrotational flow there exists a velocity potential:
Take one component of vorticity to show that the velocity potential is irrotational:
Substitute u and v components:
0
2
1
2 2
=
|
|
.
|
\
|
c c
c
c c
c
x y y x
| |
we could do this to show all vorticity components are zero.
Then, rewriting the u,v, and w components as a vector:
For an incompressible flow:
Then for incompressible irrotational flow:
And, the above equation is known as Laplaces Equation.
91
Potential Flow: Velocity Potential
Laplacian Operator in Cartesian Coordinates:
Laplacian Operator in Cylindrical Coordinates:
Where the gradient in cylindrical coordinates, the gradient operator,
Then,
May choose cylindrical
coordinates based on the
geometry of the flow problem,
i.e. pipe flow.
If a Potential Flow exists, with
appropriate boundary
conditions, the entire velocity
and pressure field can be
specified.
92
Potential Flow: Plane Potential Flows
Laplaces Equation is a Linear Partial Differential Equation, thus there are
know theories for solving these equations.
Furthermore, linear superposition of solutions is allowed:
where
and
are solutions to Laplaces equation
For simplicity, we consider 2D (planar) flows:
Cartesian:
Cylindrical:
We note that the stream functions also exist for 2D planar flows
Cartesian:
Cylindrical:
93
Potential Flow: Plane Potential Flows
For irrotational, planar flow:
Now substitute the stream function:
Then,
Laplaces Equation
For plane, irrotational flow, we use either the potential or the stream function,
which both must satisfy Laplaces equations in two dimensions.
Lines of constant + are streamlines:
Now, the change of from one point (x, y) to a nearby point (x + dx, y + dy):
Along lines of constant | we have d| = 0,
0
94
Potential Flow: Plane Potential Flows
Lines of constant | are called equipotential lines.
The equipotential lines are orthogonal to lines of constant +, streamlines
where they intersect.
The flow net consists of a family of streamlines and equipotential lines.
The combination of streamlines and equipotential lines are used to visualize a
graphical flow situation.
The velocity is inversely proportional
to the spacing between streamlines.
Velocity increases
along this streamline.
Velocity decreases
along this streamline.
95
Potential Flow: Uniform Flow
The simplest plane potential flow is a uniform flow in which the streamlines
are all parallel to each other.
Consider a uniform flow in the x-direction:
Integrate the two equations:
| = Ux + f(y) + C
| = f(x) + C
Matching the solution
C is an arbitrary constant, can be set to zero:
Now for the stream function solution:
Integrating the two equations similar
to above.
96
Potential Flow: Uniform Flow
For Uniform Flow in an Arbitrary direction, o:
97
Potential Flow: Source and Sink Flow
Source/Sink Flow is a purely radial flow.
Fluid is flowing radially from a line through
the origin perpendicular to the x-y plane.
Let m be the volume rate emanating from the line (per
unit length.
Then, to satisfy mass conservation:
Since the flow is purely radial:
Now, the velocity potential can be obtained:
Integrate
0
If m is positive, the flow is radially outward, source flow.
If m is negative, the flow is radially inward, sink flow.
m is the strength of the source or sink!
This potential flow does not exist at r = 0, the origin, because it is not a real flow, but can
approximate flows.
Source Flow:
98
Potential Flow: Source and Sink Flow
0
Now, obtain the stream function for the flow:
Then, integrate to obtain the solution:
The streamlines are radial lines and the equipotential
lines are concentric circles centered about the origin:
| lines
+ lines
99
Potential Flow: Vortex Flow
In vortex flow the streamlines are concentric circles, and the equipotential
lines are radial lines.
where K is a constant.
Solution:
The sign of K determines whether the flow rotates
clockwise or counterclockwise.
In this case,
,
The tangential velocity varies inversely with the distance from the origin. At the
origin it encounters a singularity becoming infinite.
| lines
+ lines
100
Potential Flow: Vortex Flow
How can a vortex flow be irrotational?
Rotation refers to the orientation of a fluid element and not the path
followed by the element.
Irrotational Flow: Free Vortex
Rotational Flow: Forced Vortex
Traveling from A to B, consider two sticks
Initially, sticks aligned, one in the flow direction, and the
other perpendicular to the flow.
As they move from A to B the perpendicular-aligned
stick rotates clockwise, while the flow-aligned stick
rotates counter clockwise.
The average angular velocities cancel each other, thus, the
flow is irrotational.
Irrotational Flow:
Velocity
increases
inward.
Velocity
increases
outward.
Rotational Flow: Rigid Body Rotation
Initially, sticks aligned, one in the flow
direction, and the other perpendicular to the
flow.
As they move from A to B they sticks move
in a rigid body motion, and thus the flow is
rotational.
i.e., water
draining from
a bathtub
i.e., a rotating
tank filled with
fluid
101
Potential Flow: Vortex Flow
A combined vortex flow is one in which there is a forced vortex at the core, and
a free vortex outside the core.
A Hurricane is
approximately a
combined vortex
Circulation is a quantity associated with vortex flow. It is defined as the line
integral of the tangential component of the velocity taken around a closed
curve in the flow field.
For irrotational flow the
circulation is generally
zero.
102
Potential Flow: Vortex Flow
However, if there are singularities in the flow, the circulation is not zero if the
closed curve includes the singularity.
For the free vortex:
The circulation is non-zero and constant for the free vortex:
The velocity potential and the stream function can be rewritten in terms of the
circulation:
An example in which the closed surface circulation will be zero:
Beaker Vortex:
103
Potential Flow: Doublet Flow
Combination of a Equal Source and Sink Pair:
Rearrange and take tangent,
Note, the following:
Substituting the above expressions,
and
Then,
If a is small, then tangent of angle is approximated by the angle:
104
Potential Flow: Doublet Flow
Now, we obtain the doublet flow by letting the source and sink approach one
another, and letting the strength increase.
K is the strength of the doublet, and is
equal to ma/t.
is then constant.
The corresponding velocity potential then is the following:
Streamlines of a Doublet:
105
Potential Flow: Summary of Basic Flows
106
Potential Flow: Superposition of Basic Flows
Because Potential Flows are governed by linear partial differential equations,
the solutions can be combined in superposition.
Any streamline in an inviscid flow acts as solid boundary, such that there is no
flow through the boundary or streamline.
Thus, some of the basic velocity potentials or stream functions can be
combined to yield a streamline that represents a particular body shape.
The superposition representing a body can lead to describing the flow around
the body in detail.
107
Superposition of Potential Flows: Rankine Half-Body
The Rankine Half-Body is a combination of a source and a uniform flow.
Stream Function (cylindrical coordinates):
Potential Function (cylindrical coordinates):
There will be a stagnation point, somewhere along the negative x-axis where
the source and uniform flow cancel (u = t):
For the source: For the uniform flow:
Evaluate the radial velocity:
u cos U v
r
=
For u = t,
U v
r
=
Then for a stagnation point, at some r = -b, u = t:
t 2
m
v
r
= and
108
Superposition of Potential Flows: Rankine Half-Body
Now, the stagnation streamline can be defined by evaluating at r = b, and
u = t .
Now, we note that m/2 = tbU, so following this constant streamline gives
the outline of the body:
Then,
describes the half-body outline.
So, the source and uniform can be used to describe an aerodynamic body.
The other streamlines can be obtained by setting constant and plotting:
Half-Body:
109
Superposition of Potential Flows: Rankine Half-Body
The width of the half-body:
Total width then,
The magnitude of the velocity at any point in the flow:
Noting,
and
Knowing, the velocity we can now determine the pressure field using the Bernoulli
Equation:
P
o
and U are at a point far away from the body and are known.
110
Superposition of Potential Flows: Rankine Half-Body
Notes on this type of flow:
Provides useful information about the flow in the front part of streamlined body.
A practical example is a bridge pier or a strut placed in a uniform stream
In a potential flow the tangent velocity is not zero at a boundary, it slips
The flow slips due to a lack of viscosity (an approximation result).
At the boundary, the flow is not properly represented for a real flow.
Outside the boundary layer, the flow is a reasonable representation.
The pressure at the boundary is reasonably approximated with potential flow.
The boundary layer is to thin to cause much pressure variation.
111
Superposition of Potential Flows: Rankine Oval
Rankine Ovals are the combination a source, a sink and a uniform flow,
producing a closed body.
Some equations describing the flow:
The body half-length
The body half-width
Iterative
Potential and Stream Function
112
Superposition of Potential Flows: Rankine Oval
Notes on this type of flow:
Provides useful information about the flow about a streamlined body.
At the boundary, the flow is not properly represented for a real flow.
Outside the boundary layer, the flow is a reasonable representation.
The pressure at the boundary is reasonably approximated with potential flow.
Only the pressure on the front of the body is accurate though.
Pressure outside the boundary is reasonably approximated.
113
Superposition of Potential Flows: Flow Around a Circular Cylinder
Combines a uniform flow and a doublet flow:
and
Then require that the stream function is constant for r = a, where a is the
radius of the circular cylinder:
K = Ua
2
Then,
and
Then the velocity components:
114
Superposition of Potential Flows: Flow Around a Circular Cylinder
At the surface of the cylinder (r = a):
The maximum velocity occurs at the top and bottom of the cylinder,
magnitude of 2U.
115
Superposition of Potential Flows: Flow Around a Circular Cylinder
Pressure distribution on a circular cylinder found with the Bernoulli equation
Then substituting for the surface velocity:
Theoretical and experimental agree
well on the front of the cylinder.
Flow separation on the back-half in the
real flow due to viscous effects causes
differences between the theory and
experiment.
116
Superposition of Potential Flows: Flow Around a Circular Cylinder
The resultant force per unit force acting on the cylinder can be determined
by integrating the pressure over the surface (equate to lift and drag).
(Drag)
(Lift)
Substituting,
Evaluating the integrals:
Both drag and lift are predicted to be zero on fixed cylinder in a uniform flow?
Mathematically, this makes sense since the pressure distribution is
symmetric about cylinder, ahowever, in practice/experiment we see
substantial drag on a circular cylinder (dAlemberts Paradox, 1717-
1783).
Viscosity in real flows is the Culprit Again!
Jean le Rond
dAlembert
(1717-1783)