Conceptual Fluid Questions

CHAPTER 9: Differential Analysis of Fluid Flow
9 – 1 The divergence theorem is … where G is a vector, V is a volume, and A is the surface area
that encloses and defines the volume. Express the divergence theorem in words?
ANSWER: For vector G, the volume integral of the divergence of G over volume V is equal to
the surface integral of the normal component of G taken over the surface A that encloses the
volume. The divergence theorem is also called Gauss’s theorem.
9 – 2 Explain the fundamental differences between a flow domain and a control volume?
ANSWER: A control volume is used in an integral, control volume solution. It is a volume over
which all mass flow rates, forces, etc. are specified over the entire control surface of the control
volume. In a control volume analysis, we do not know or care about details inside the control
volume. Rather, we solve for gross features of the flow such as net force acting on a body. A
flow domain, on the other hand, is also a volume, but is used in a differential analysis.
Differential equations of motion are solved everywhere inside the flow domain, and we are
interested in all the details inside the flow domain. Note that we also need to specify what is
happening at the boundaries of a flow domain – these are called boundary conditions.
9 – 3 What does it mean when we say that two or more differential equations are coupled?
ANSWER: A set of coupled differential equations simply means that the equations are
dependent on each other and must be solved together rather than separately. For example, the
equations of motion for fluid flow involve velocity variables in both the conservation of mass
equation and the momentum equation. To solve for these variables, we must solve the coupled
set of differential equations together. In some very simple fluid flow problems, the equations
become uncoupled, and are easier to solve.
9 – 4 For a three-dimensional, unsteady, incompressible flow field in which temperature

variations are insignificant, how many unknowns are there? List the equations required?
ANSWER: There are four unknowns (velocity components u, v, w, and pressure P) and thus we
need to solve four equations: one from conservation of mass which is a scalar equation and three
from Newton’s second law which is a vector equation. These equations are also coupled in
general. We solve for the same for variables in the steady case. Whether the proplem is steady or
not, this does not change the number of variables we solve for.
9 – 5 For an unsteady, compressible flow field that is two-dimensional in the x-y plane and in
which temperature and density variations are significant, how many unknowns are there? List the
equations required to solve for these unknowns. (Note: Assume other flow properties like
viscosity, thermal conductivity, etc., can be treated as constants.)
ANSWER: There are five unknowns (velocity components u and v, density , T, and P) and thus
we need to solve five equations: one from conservation of mass which is a scalar equation, two
from Newton’s second law which is a vector equation, one from the energy equation which is a
scalar equation and one from an equation of state (e.g., ideal gas law) which is a scalar equation.
These equations are also coupled in general.
9 – 21 In this chapter we derive the continuity equation in two ways: by using the divergence
theorem and by summing mass flow rates through each face of an infinitesimal control volume.
Explain why the former is so much less involved than the latter?
ANSWER: In the derivation using the divergence theorem, we begin with the control volume
form of conservation of mass, and simply apply the divergence theorem. The control volume
form was already derived in Chap. 5, so we begin the derivation in this chapter with an
established conservation of mass equation. On the other hand, the alternative derivation is from
“scratch” and therefore requires much more algebra.
Discussion The bottom line is that the divergence theorem enables us to quickly convert the
control volume form of the conservation law into the differential form.
9 – 22 If a flow field is compressible, what can we say about the material derivative of density?
What about if the flow field is incompressible?
ANSWER: If the flow field is compressible, we expect that as a fluid particle (a material
element) moves around in the flow, its density changes. Thus, the material derivative of density
(the rate of change of density following a fluid particle) is non-zero for compressible flow.
However, if the flow field is incompressible, the density remains constant. As a fluid particle
moves around in the flow, the material derivative of density must be zero for incompressible
flow (no change in density following the fluid particle).
Discussion The material derivative of any property is the rate of change of that property
following a fluid particle
9 – 25 Equation for volumetric strain rate, (1/v)(Dv/Dt) = Divergence . V … Write this as a word
equation and discuss what happens to the volume of a fluid element as it moves around in a
compressible fluid flow field.
ANSWER: Here is a word equation: “The time rate of change of volume of a fluid particle per
unit volume is equal to the divergence of the velocity field.” As a fluid particle moves around in
a compressible flow, it can distort, rotate, and get larger or smaller. Thus, the volume of the fluid
element can change with time; this is represented by the left-hand side of the equation. The right-
hand side is identically zero for an incompressible flow, but it is not zero for a compressible
flow. Thus, we can think of the volumetric strain rate as a measure of compressibility of a fluid
flow. Volumetric strain rate is a kinematic property as discussed in Chap. 4. Nevertheless, it is
shown here to be related to the continuity equation (conservation of mass).
9 – 39 Consider two-dimensional flow in the xy-plane. What is the significance of the difference
in value of stream function Psi from one streamline to another?
ANSWER: The difference in the value of Psi from one streamline to another is equal to the
volume flow rate per unit width between the two streamlines. This fact about the stream function
can be used to calculate the volume flow rate in certain applications.
9 – 40 In CFD lingo, the stream function is often called a non-primitive variable, while velocity
and pressure are called primitive variables. Why do you suppose this is the case?
ANSWER: The natural physical variables in a fluid flow problem are the velocity components
and the pressure. [If the flow is compressible, density and temperature are also natural physical
variables.] These variables can be considered “primitive” because we do not change them in any
way – we simply solve for them directly. Stream function, on the other hand, is a contrived or
derived variable. The stream function is not primitive in the sense that it is not one of the original
physical variables in the problem. Vorticity is another example of a non-primitive variable. In
fact, some 2-D CFD codes use stream function and vorticity as non-primitive variables.
9 – 41 What restrictions or conditions are imposed on stream function Psi so that it exactly
satisfies the two-dimensional incompressible continuity equation by definition? Why are these
restrictions necessary?
ANSWER: Stream function Psi must be a smooth function of x and y (or r and Theta). These
restrictions are necessary so that the second derivatives of Psi with respect to both variables are
equal regardless of the order of differentiation. In this case, the 2-D incompressible continuity
equation is satisfied exactly by the definition of Psi. If the stream function were not smooth,
there would be sudden discontinuities in the velocity field as well – a physical impossibility that
would violate conservation of mass.
9 – 42 What is significant about curves of constant stream function? Explain why the stream
function is useful in fluid mechanics?
ANSWER: Curves of constant stream function represent streamlines of a flow. A stream

function is useful because by drawing curves of constant Psi, we can visualize the instantaneous
velocity field. In addition, the change in the value of Psi from one streamline to another is equal
to the volume flow rate per unit width between the two streamlines
9 – 67 Unique Conceptual Question
9 – 78 What is mechanical pressure Pm, and how is it used in an incompressible flow solution?
ANSWER: Mechanical pressure is the mean normal stress acting inwardly on a fluid element.
For an incompressible fluid, the density is constant and therefore we have no equation of state
available for calculation of the thermodynamic pressure. In fact, thermodynamic pressure cannot
even be defined for an incompressible fluid. Fluid elements and surfaces still “feel” a pressure,
however, and this pressure is the so-called mechanical pressure. When dealing with
incompressible fluid flows, pressure variable P is always interpreted as the mechanical pressure
Pm.
9 – 79 What are constitutive equations, and to which fluid mechanics equation are they applied?
ANSWER: The constitutive equations are relationships between the components of the stress
tensor and the primary unknowns of the problem, namely pressure and velocity. The constitutive
equations enable us to write the components of the stress tensor in Cauchy’s equation in terms of
the velocity field and the pressure field. Cauchy’s equation by itself is useless without the
constitutive equations, because of the too many unknowns for the number of available equations.
9 – 80 An airplane flies at constant velocity V. Discuss the velocity boundary conditions on the
air adjacent to the surface of the airplane from two frames of reference: (a) standing on the
ground, and (b) moving with the airplane. Likewise, what are the far-field velocity boundary
conditions of the air (far away from the airplane) in both frames of reference?
ANSWER: (a) From the stationary frame of reference, V = V airplane on all surfaces of the
airplane, (no-slip boundary condition). Far from the airplane the air is still (V= 0).
(b) From the reference frame moving with the airplane, V = 0 on all surfaces, (no-slip boundary
condition). Far from the airplane the air is moving towards the airplane at a speed that is opposite
the airplane’s speed (V = - V airplane). The no-slip condition requires that the fluid velocity
equal the airplane velocity everywhere on the airplane surface, regardless of the geometry of the
airplane, and regardless of the frame of reference.
9 – 81 What is the main distinction between a Newtonian fluid and a non-Newtonian fluid?
Name at least three Newtonian fluids and three non-Newtonian fluids.
ANSWER: The main distinction between a Newtonian fluid and a non-Newtonian fluid is that
for flow of a Newtonian fluid, shear stress is linearly proportional to shear strain rate, whereas
for flow of a non-Newtonian fluid, the relationship between shear stress and shear strain rate is
nonlinear. There are many examples of Newtonian fluids. Most pure, common liquids like water,
oil, gasoline, alcohol, etc. are Newtonian. Most gases also behave like Newtonian fluids. Non-
Newtonian fluids include paint, pastes and creams, polymer solutions, cake batter, slurries and
colloidal suspensions like quicksand, blood, etc. The Navier-Stokes equations apply only to
Newtonian fluids. For non-Newtonian fluids, you would need to insert nonlinear
constitutive equations into Cauchy’s equations in order to obtain a useful differential
equation for conservation of linear momentum.
9 – 82 Define or describe each type of fluid: (a) viscoelastic fluid, (b) pseudoplastic fluid, (c)
dilatant fluid, (d) Bingham plastic fluid.
ANSWER: All these fluids are examples of non-Newtonian fluids.
a) A viscoelastic fluid is a fluid that returns (either fully or partially) to its original shape
after the applied stress is released.
b) A pseudoplastic fluid is a shear thinning fluid - the more the fluid is sheared, the less
viscous it becomes.
c) A dilatant fluid is a shear thickening fluid - the more the fluid is sheared, the more
viscous it becomes.
d) A Bingham plastic fluid is an extreme type of pseudoplastic fluid that requires a finite
stress called the yield stress in order for the fluid to flow at all.
9 – 83 The general control volume form of the linear momentum equation is … Discuss the
meaning of each term in this equation. The terms are labeled for convenience. Write the equation
as a word equation?
ANSWER: Term I is the net body force acting on the control volume. Term II is the net surface
force acting on the control volume. Term III is the net rate of change of linear momentum within
the control volume. Term IV is the net rate of outflow of linear momentum through the control
surface. In words, the equation can be expressed as: “The total force acting on the control
volume is the sum of body forces and surface forces, and is equal to the rate at which momentum
changes within the control volume plus the rate at which momentum flows out of the control
volume.”
Discussion The dimensions of each term in the equation are those of momentum per time. Each
term has primary dimensions of {mLt ^ -2}.
Channel flow / Poiseuille flow  flow between two infinite parallel plates
separated by distance h, with both the top plate and bottom plate stationary.
9 – 108 List the six steps used to solve the Navier–Stokes and continuity equations for
incompressible flow with constant fluid properties.
ANSWER:
1. Lay out the problem and the geometry. Identify all relevant dimensions and parameters.
2. List all appropriate assumptions, approximations, simplifications, and boundary
conditions.
3. Write out and simplify the differential equations (continuity and the required components
of Navier- Stokes) as much as possible.
4. Solve (integrate) the differential equations. This leads to one or more constants of
integration.
5. Apply boundary conditions to obtain values for the constants of integration.
6. Verify the results by checking that the flow field meets all the specifications and
boundary conditions.
Discussion These steps are not always followed in the same order. For example, in CFD
applications the boundary conditions are applied before the equations are integrated.
9 – 109 For each part, write the official name for the differential equation, discuss its restrictions,
and describe what the equation represents physically.
ANSWER:
(a) This is the continuity equation. The form given here is valid for any fluid. It describes
conservation of mass in a fluid flow.
(b) This is Cauchy’s equation. The form given here is valid for any fluid. It describes
conservation of linear momentum in a fluid flow.
(c) This is the Navier-Stokes equation. The form given here is valid for a specific type of fluid,
namely an incompressible Newtonian fluid. The equation describes conservation of linear
momentum in a fluid flow.
9 – 110 Explain why the incompressible flow approximation and the constant temperature
approximation usually go hand in hand?
ANSWER: For an incompressible flow, the density is assumed to be constant. In addition, the
incompressible flow approximation usually implies that all fluid properties (viscosity, thermal
conductivity, etc.) are constant as well. These assumptions go hand in hand because a flow with
constant density implies a flow with little or no temperature changes and no buoyancy effects.
Since viscosity is a strong function of temperature but generally a weak function of pressure, the
fluid’s viscosity is approximately constant whenever temperature is constant. When dealing with
incompressible fluid flows, pressure variable P is interpreted as the mechanical pressure Pm, and
we don’t need an equation of state. In effect, the equation of state is replaced by the assumption
of constant density and constant temperature. Discussion Mechanical pressure Pm is determined
by the flow field, not by thermodynamics.
9 – 111 For each statement, choose whether the statement is true or false and discuss your
answer briefly. For each statement it is assumed that the proper boundary conditions and fluid
properties are known.
a) A general incompressible flow problem with constant fluid properties has four unknowns.
b) A general compressible flow problem has five unknowns.
c) For an incompressible fluid mechanics problem, the continuity equation and Cauchy’s
equation provide enough equations to match the number of unknowns.
d) For an incompressible fluid mechanics problem involving a Newtonian fluid with
constant properties, the continuity equation and the Navier–Stokes equation provide
enough equations to match the number of unknowns.
ANSWER:
a) True: The unknowns for an incompressible flow problem with constant fluid properties
are pressure and the three components of velocity. Density and viscosity are constants
and are therefore not unknowns.
b) False: The unknowns for a compressible flow problem are pressure, the three components
of velocity, and the density. However, density is a thermodynamic function of pressure
and temperature. Hence, temperature appears as an additional unknown, as does some
kind of equation of state. In summary, there are actually at least 6 unknowns (P, u, v, w,
Density, and T). We therefore need 6 equations (continuity, 3 components of Navier
Stokes, equation of state, and energy). In addition, fluid properties such as viscosity may
change as well, and we need either more equations or look-up table for these properties.
c) False: Cauchy’s equation contains additional unknowns – the components of the stress
tensor, which must be written in terms of the velocity and pressure fields through
constitutive equation.
d) True: For an incompressible flow problem involving a Newtonian fluid, there are only 4
unknowns (P, u, v, and w). We therefore need only 4 equations (continuity and 3
components of Navier-Stokes).
9 – 112 Discuss the relationship between volumetric strain rate and the continuity equation. Base
your discussion on fundamental definitions.
ANSWER: Analysis Volumetric strain rate is defined as the rate of increase of volume of a fluid
element per unit volume. In a compressible flow field, the volume of a fluid particle may
increase or decrease as it moves along in the flow, but its mass must remain constant. (This is a
fundamental statement of conservation of mass of a system, since the fluid particle can be
thought of as an infinitesimal system.) Mathematically it turns out that volumetric strain rate is
the sum of the three normal strain rates and is identically zero for incompressible flow (density
cannot change, and hence volume cannot change). The continuity equation is based on the same
fundamental principle of mass conservation. It is a differential form of the equation of
conservation of mass. Its incompressible form also shows that the sum of the three normal strain
rates must be zero. On the other hand, if the density is not constant, the sum of the three normal
strain rates is not zero, but is still equal to the volumetric strain rate, which is also non-zero.
Discussion Volumetric strain rate is derived and discussed in Chap. 4 as a kinematic property.
CHAPTER 10: Potential Flow
10–1 Discuss how nondimensionalization of the Navier– Stokes equation is helpful in obtaining
approximate solutions. Give an example.
ANSWER: When we properly nondimensionalize the Navier-Stokes equation, all the terms are
re-written in the form of some nondimensional parameter times a quantity of order unity. Thus,
we can simply compare the orders of magnitude of the nondimensional parameters to see which
terms (if any) can be ignored because they are very small compared to other terms. For example,
if the Strouhal number is much smaller than the Euler number, we can ignore the term that
contains the Strouhal number, but must retain the term that contains the Euler number.
Discussion This method works only if the characteristic scales of the problem (length, speed,
frequency, etc.) are chosen properly.
10–2 A box fan sits on the floor of a very large room (Fig. P10–2C). Label
regions of the flow field that may be approximated as static. Label regions
in which the irrotational approximation is likely to be appropriate. Label
regions where the boundary layer approximation may be appropriate.
Finally, label regions in which the full Navier– Stokes equation most likely
needs to be solved (i.e., regions where no approximation is appropriate).
ANSWER: A boundary layer grows along the floor, both

upstream and downstream of the fan. The flow upstream of the fan
is largely irrotational except very close to the floor. The air is
nearly static far upstream and far above the fan. Downstream of
the fan, the flow is most likely swirling and turbulent, and none of
the approximations are expected to be appropriate there. In other
words, the full Navier-Stokes equation must be solved in that
region.
10–3 Explain the difference between an “exact” solution of the Navier–Stokes equation (as
discussed in Chap. 9) and an approximate solution (as discussed in this chapter).
ANSWER: In an “exact” solution, we begin with the full Navier-Stokes equation. As we solve
the problem, some terms may drop out due to the specified geometry or other simplifying
assumptions in the problem. In an approximate solution, we eliminate some terms in the Navier-
Stokes equation right from the start. In other words, we begin with a reduced or simplified
approximate form of the equation.
Discussion The approximations are based on the class of flow problem and/or the region in
which such approximations are appropriate (e.g. irrotational, boundary layer, etc.).
10–4 Which nondimensional parameter in the nondimensionalized Navier–Stokes equation is

eliminated by use of modified pressure instead of actual pressure? Explain.
ANSWER: Modified pressure effectively combines the effects of actual pressure and gravity. In
the nondimensionalized Navier-Stokes equation in terms of modified pressure, the Froude
number disappears. The reason Froude number is eliminated is because the gravity term is
eliminated from the equation.
Discussion Keep in mind that we can employ modified pressure only for flows without free
surface effects.
10–5 What criteria can you use to determine whether an approximation of the Navier–Stokes
equation is appropriate or not? Explain.
ANSWER: We determine if an approximation is appropriate by comparing the orders of

magnitude of the various terms in the equations of motion. If the neglected terms are negligibly
small compared to other terms, then the approximation is appropriate. If not, then it is not
appropriate to neglect those terms.
Discussion It is important that the proper scales be used for the nondimensionalization of the
equation. Otherwise, the order of magnitude analysis may be incorrect.
10–6 In the nondimensionalized incompressible Navier– Stokes equation (Eq. 10–6), there are
four nondimensional parameters. Name each one, explain its physical significance (e.g., the ratio
of pressure forces to viscous forces), and discuss what it means physically when the parameter is
very small or very large.
ANSWER: The four parameters are discussed individually below:
1. Strouhal number: St is the ratio of some characteristic flow time to some period of oscillation.
If St << 1, the oscillation period is very large compared to the characteristic flow time, and the
problem is quasi-steady; the unsteady term in the Navier-Stokes equation may be ignored. If St
>> 1, the oscillation period is very short compared to the characteristic flow time, and the
unsteadiness dominates the problem; the unsteady term must remain.
2. Euler number: Eu is the ratio of a characteristic pressure difference to a characteristic pressure

due to fluid inertia. If Eu << 1, pressure gradients are very small compared to inertial pressure,
and the pressure term can be neglected in the Navier-Stokes equation. If Eu >> 1, the pressure
term is very large compared to the inertial term, and must remain in the equation.
3. Froude number: Fr is the ratio of inertial forces to gravitational forces. Note that Fr appears in
the denominator of the nondimensionalized Navier-Stokes equation. If Fr << 1, gravitational
forces are very large compared to inertial forces, and the gravity term must remain in the Navier-
Stokes equation. If Fr >> 1, gravitational forces are negligible compared to inertial forces, and
the gravity term in the Navier-Stokes equation can be ignored.
4. Reynolds number: Re is the ratio of inertial forces to viscous forces. Note that Re appears in
the denominator of the nondimensionalized Navier-Stokes equation. If Re << 1, viscous forces
are very large compared to inertial forces, and the viscous term must remain. (In fact, it may
dominate the other terms, as in creeping flow). If Re >> 1, viscous forces are negligible
compared to inertial forces, and the viscous term in the Navier-Stokes equation can be ignored.
Note that this applies only to regions outside of boundary layers, because the characteristic
length scale for a boundary layer is generally much smaller than that for the overall flow.
Discussion You must keep in mind that the approximations discussed here are appropriate only
in certain regions of the flow field. In other regions of the same flow field, different
approximations may apply.
10–7 What is the most important criterion for use of the modified pressure P' rather than the
thermodynamic pressure P in a solution of the Navier–Stokes equation?
ANSWER: Modified pressure can be used only when there are no free surface effects in the
problem.
Discussion: Modified pressure is simply a combination of thermodynamic pressure and
hydrostatic pressure. It turns out that if there are no free surface effects, the hydrostatic pressure
component is independent of the flow pressure component, and these two can be separated.
10–8 What is the most significant danger associated with an approximate solution of the Navier–
Stokes equation? Give an example that is different than the ones given in this chapter.
ANSWER: The danger of an approximate solution of the Navier-Stokes equation is this: If the
approximation is not appropriate to begin with, our solution will be incorrect – even if we
perform all the mathematics correctly. There are many examples. For instance, we may assume
that a boundary layer exists in a region of flow. However, if the Reynolds number is not large
enough, the boundary layer is too thick and the boundary layer approximations break down.
Another example is that we may assume a fluid statics region, when in reality there are swirling
eddies in that region. The unsteady motion of the eddies makes the problem unsteady and
dynamic – the approximation of fluid statics would be inappropriate.
Discussion When you make an approximation and solve the problem, it is best to go back and
verify that the approximation is appropriate.
10–16 Discuss why fluid density has negligible influence on the aerodynamic drag on a particle
moving in the creeping flow regime.
ANSWER: It turns out that fluid density drops out of the creeping flow equations, since the
terms that contain ρ in the Navier-Stokes equation are negligibly small compared to the pressure
and viscous terms (which do not contain ρ). Another way to think about this is: In creeping flow,
there is no fluid inertia, and since inertia is associated with fluid mass (density), density cannot
contribute to the aerodynamic drag on a particle moving in creeping flow. In creeping flow, there
is a balance between pressure forces and viscous forces, neither of which depend on fluid
density.
Discussion Density does have an indirect influence on creeping flow drag. Namely, ρ is needed
in the Reynolds number calculation, and Re determines whether the flow is in the creeping flow
regime or not.
10–17 Write a one-word description of each of the

five terms in the incompressible Navier–Stokes
equation, When the creeping flow approximation is
made, only two of the five terms remain. Which
two terms remain, and why is this significant?
ANSWER: The terms in the equation are identified as follows:
I  Unsteady – II  Inertial – III  Pressure – IV  Gravity – V  Viscous term
When the creeping flow approximation is made, only terms III (pressure) and V (viscous)
remain. The other three terms are very small compared to these two and can be ignored. The
significance is that all unsteady and inertial effects (terms I and II) have disappeared, as has
gravity. We are left with a flow in which pressure forces and viscous forces must balance.
Another significant result is that density has disappeared from the creeping flow equation, as
discussed in the text.
Discussion There are other acceptable one-word descriptions of some of the terms in the
equation. For example, the inertial term can also be called the convective term, the
advective term, or the acceleration term.
10–38 What is the main difference between the steady, incompressible Bernoulli equation for
irrotational regions of flow, and the steady incompressible Bernoulli equation for rotational but
inviscid regions of flow?
ANSWER: The Bernoulli equation itself is identical in these two cases, but the “constant” for
the case of rotational but inviscid regions of flow is constant only along streamlines of the flow,
not everywhere. For irrotational regions of flow, the same Bernoulli constant holds everywhere.
Discussion: A simple example is that of solid body rotation, which is rotational but inviscid. In
this flow, the Bernoulli “constant” changes from one streamline to another.
10–39 In what way is the Euler equation an approximation of the Navier–Stokes equation?
Where in a flow field is the Euler equation an appropriate approximation?
ANSWER: The Euler equation is simply the Navier-Stokes equation with the viscous term
neglected; it is therefore an inviscid approximation of the Navier-Stokes equation. The Euler
equation is appropriate in high Reynolds number regions of the flow where net viscous forces are
negligible, far away from walls and wakes.
Discussion: The Euler equation is not appropriate very close to solid walls, since frictional forces
are always present there. Note that the same Euler equation is appropriate in an irrotational
region of flow as well.
10–49 What is D’Alembert’s paradox? Why is it a paradox?

ANSWER: D’Alembert’s paradox states that with the irrotational flow approximation, the
aerodynamic drag force on any non-lifting body of any shape immersed in a uniform stream is
zero. It is a paradox because we know from experience that bodies in a flow field have non-zero
aerodynamic drag. Irrotational flow over a non-lifting immersed body has neither pressure drag
nor viscous drag. In a real flow, both of these drag components are present.
10–50 Consider the flow field produced by a hair dryer (Fig. P10–50C).
Identify regions of this flow field that can be approximated as irrotational,
and those for which the irrotational flow approximation would not be
appropriate (rotational flow regions).
ANSWER: Flow in the air far away from the hair dryer and its jet is
certainly irrotational. As the air approaches the inlet, it is irrotational except very close to the
surface of the hair dryer. Flow in the jet is rotational, but flow outside of the jet can be
approximated as irrotational.
Discussion Flow near solid walls is nearly always rotational because of the viscous rotational
boundary layer that grows there. There are sharp velocity gradients in a jet, so the vorticity
cannot be zero in that region, and the flow must be rotational in the jet as well.
10–51 In an irrotational region of flow, the velocity field can be calculated without need of the
momentum equation by solving the Laplace equation for velocity potential function ∅, and then
solving for the components of ⃗V from the definition of ∅, namely, ⃗ V =⃗∇ ∅. Discuss the role of
the momentum equation in an irrotational region of flow.
ANSWER: Although it is true that the momentum equation is not required in order to solve for
the velocity field, it is required in order to solve for the pressure field. In particular, the Navier-
Stokes equation reduces to the Bernoulli equation in an irrotational region of flow.
Discussion Mathematically, it turns out that in an irrotational flow field the continuity equation is
uncoupled from the momentum equation, meaning that we can solve continuity for ∅ by itself,
without need of the momentum equation. But, the momentum equation cannot be solved by itself
10–52 A subtle point, often missed by students of fluid mechanics (and even their professors!), is
that an inviscid region of flow is not the same as an irrotational (potential) region of flow.
Discuss the differences and similarities between these two approximations. Give an example of
each.
ANSWER: The two approximations are similar in that in both cases, the viscous terms in the
Navier-Stokes equation drop out, leaving the Euler equation. Also, in both cases the Bernoulli
equation results form integration of the Euler equation. However, these two approximations
differ significantly from each other. When making the inviscid flow approximation, we assume
that the viscous terms are negligibly small. A good example, as discussed in this chapter, is solid
body rotation. In this case, although the fluid itself is viscous, all effects of viscosity are gone,
and the flow field can be considered “inviscid” (although it is rotational). On the other hand, the
irrotational approximation is made when the vorticity (a measure of rotationality of fluid
particles) is negligibly small. In this case, viscosity still acts on fluid particles – it shears them
and distorts them, yet the net rate of rotation of fluid particles is zero. In other words, in an
irrotational region of flow, the net viscous force on a fluid particle is zero, but viscous stresses on
the fluid particle are certainly not zero. Examples of irrotational, but viscous flows include any
irrotational flow field with curved streamlines, such as a line vortex, a doublet, irrotational flow
over a circular cylinder, etc. Freestream flow is both inviscid and irrotational since fluid particles
do not shear or distort or rotate, and viscosity does not enter into the picture.
Discussion In either case, the viscous terms in the Navier-Stokes equation disappear, but
for different reasons. In the inviscid flow approximation, the viscous terms disappear
because we neglect viscosity. In the irrotational flow approximation, the viscous terms
disappear because they cancel each other out due to the fact that the vorticity (hence the
rate of rotation) of fluid particles is negligibly small.
10–53 What flow property determines whether a region of flow is rotational or irrotational?
ANSWER: The vorticity determines whether a region of flow is rotational or irrotational.

Specifically, if the vorticity is zero (or negligibly small), the flow is approximated as irrotational,
but if the vorticity is not negligibly small, the flow is rotational.
Discussion Another acceptable answer is the rate of rotation vector or the angular velocity vector
of a fluid particle.
10–54 Write the Bernoulli equation, and discuss how it differs between an inviscid, rotational
region of flow and a viscous, irrotational region of flow. Which case is more restrictive (in
regards to the Bernoulli equation)?
P V2
ANSWER: Steady incompressible Bernoulli equation is the same in both cases: + + gz=C
ρ 2
However, in an inviscid, rotational region of flow, Eq. 1 is applicable only along a streamline.
The Bernoulli “constant” C is constant along any particular streamline, but may change from
streamline to streamline. In a viscous, irrotational region of flow, however, the Bernoulli
constant is constant everywhere, even across streamlines. Thus, the inviscid, rotational region of
flow has more restrictions on the use of the Bernoulli equation. In either case, the viscous terms
in the Navier-Stokes equation disappear, but for different reasons.
10–72 We usually think of boundary layers as occurring along solid walls. However, there are
other flow situations in which the boundary layer approximation is also appropriate. Name three
such flows, and explain why the boundary layer approximation is appropriate.
ANSWER: The boundary layer approximation is appropriate for the three basic types of shear
layers: wakes, jets, and mixing layers. These flows have a predominant flow direction, and for
high Reynolds numbers, the shear layer is very thin, causing the viscous terms to be much
smaller than the inertial terms, just as in the case of a boundary layer along a wall.
Discussion For the wake and the mixing layer, there is an irrotational outer flow in the
streamwise direction, but for the jet, the flow outside the jet is nearly stagnant.
10–73 For each statement, choose whether the statement is true or

false and discuss your answer briefly. These statements concern a
laminar boundary layer on a flat plate (Fig. P10–73C).
A. At a given x-location, if the Reynolds number were to

increase, the boundary layer thickness would also increase.
B. As outer flow velocity increases, so does the boundary
layer thickness.
C. As the fluid viscosity increases, so does the boundary layer thickness.
D. As the fluid density increases, so does the boundary layer thickness.
ANSWER:
A. False: If the Reynolds number at a given x location were to increase, all else being equal,
viscous forces would decrease in magnitude relative to inertial forces, rendering the
boundary layer thinner.
B. False: Actually, as V increases, so does Re, and the boundary layer thickness decreases
with increasing Reynolds number.
C. True: Since μ appears in the denominator of the Reynolds number, Re decreases as μ
increases, causing the boundary layer thickness to increase.
D. False: Since ρ appears in the numerator of the Reynolds number, Re increases as ρ
increases, causing the boundary layer thickness to decrease.
10–74 In this chapter, we make a statement that the boundary layer approximation “bridges the
gap” between the Euler equation and the Navier–Stokes equation. Explain.
ANSWER: The Euler equation neglects the viscous terms compared to the inertial terms. For
external flow around a body, this is a reasonable approximation over the majority of the flow
field, except very close to the body, where viscous effects dominate. The Navier-Stokes
equation, on the other hand, includes both viscous and inertial terms, but is much more difficult
to solve. The boundary layer equations bridge the gap between these two: we solve the simpler
Euler equation away from walls, and then fit in a thin boundary layer to account for the no-slip
condition at walls.
10–75 A laminar boundary layer growing along a flat plate is sketched in Fig. P10–75C. Several
velocity profiles and the boundary layer thickness d(x) are also shown. Sketch several
streamlines in this flow field. Is the curve representing d(x) a streamline?
ANSWER: Five streamlines are sketched in Fig. 1. In order to satisfy conservation of mass, the
streamlines must cross the curve δ ( x). Thus, δ ( x) cannot itself be a streamline of the flow.
Discussion As the boundary layer grows in thickness, streamlines diverge slowly away from the
wall (and become farther apart from each other) in order to conserve mass. However, the upward
displacement of the streamlines is not as fast as the growth of δ ( x).
10–76 What is a trip wire, and what is its purpose?
ANSWER: A trip wire is a rod or wire stretched normal to the streamwise direction along the
wall. Its purpose is to create a large disturbance in the laminar boundary layer that causes the
boundary layer to “trip” to turbulence much more quickly than it would otherwise. Dimples on a
golf ball serve the same purpose.
10-77 Discuss the implication of an inflection point in a boundary layer profile. Specifically,
does the existence of an inflection point infer a favorable or adverse pressure gradient? Explain.
ANSWER: As sketched in Fig. 10-124, the existence of an inflection point in the boundary layer
profile indicates an adverse or unfavorable pressure gradient. The reason for this is due to the
fact that the second derivative of the velocity profile u(y) at the wall is directly proportional to
the pressure gradient (Eq. 10-86). In an adverse pressure gradient field, dP/dx is positive, and
thus, ∂2 u/ ∂2 y = 0 is also positive. However, since ∂2 u/ ∂2 y must be negative as u approaches
U(x) at the edge of the boundary layer, there has to be an inflection point (∂2 u/ ∂2 y = 0)
somewhere in the boundary layer.
Discussion If the adverse pressure gradient is large enough, the boundary layer separates off the
wall, leading to reverse flow near the wall.
10–78 Compare flow separation for a laminar versus turbulent boundary layer. Specifically,
which case is more resistant to flow separation? Why? Based on your answer, explain why golf
balls have dimples.
ANSWER: Turbulent boundary layers are more “full” than are laminar boundary layers.
Because of this, a turbulent boundary layer is much less likely to separate compared to a laminar
boundary layer under the same adverse pressure gradient. A smooth golf ball, for example,
would maintain a laminar boundary layer on its surface, and the boundary layer would separate
fairly easily, leading to large aerodynamic drag. Golf balls have dimples (a type of surface
roughness) in order to create an early transition to a turbulent boundary layer. Flow still separates
from the golf ball surface, but much farther downstream in the boundary layer, resulting in
significantly reduced aerodynamic drag.
Discussion Turbulent boundary layers have more skin friction drag than do laminar boundary
layers, but this effect is less significant than the pressure drag caused by flow separation. Thus, a
rough golf ball (at appropriate Reynolds numbers) ends up with less overall drag, compared to a
smooth golf ball at the same conditions.
10–79 In your own words, summarize the five steps of the boundary layer procedure.
ANSWER: We list the five steps below, with a description of each:
Step 1 Solve for the outer flow, ignoring the boundary layer (assuming that the region of flow
outside the boundary layer is approximately inviscid and/or irrotational). Transform coordinates
as necessary to obtain U(x).
Step 2 Assume a thin boundary layer – so thin in fact that it does not affect the outer flow
solution of Step 1.
Step 3 Solve the boundary layer equations. For this step we use the no-slip boundary condition at
the wall, u = v = 0 at y = 0, the known outer flow condition at the edge of the boundary layer, u
 U(x) as y  , and some known starting profile, u = ustarting(y) at x = xstarting.
Step 4 Calculate quantities of interest in the flow field. For example, once the boundary layer
equations have been solved (Step 3), we can calculate (x), shear stress along the wall, total
skin friction drag, etc.
Step 5 Verify that the boundary layer approximations are appropriate. In other words, verify that
the boundary layer is indeed thin – otherwise the approximation is not justified.
10–80 In your own words, list at least three “red flags” to look out for when performing laminar
boundary layer calculations.
ANSWER: We list four below. (Students are asked to list at least three.)
- The boundary layer approximation breaks down if Reynolds number is not large enough. For
example, /L ~ 0.01 (1%) for ReL = 10,000.
- The assumption of zero pressure gradient in the y direction breaks down if wall curvature is of
similar magnitude as . In such cases, centripetal acceleration effects due to streamline
curvature cannot be ignored. Physically, the boundary layer is not “thin” enough for the
approximation to be appropriate when  is not << R.
- When Reynolds number is too high, the boundary layer does not remain laminar. The boundary
layer approximation itself may still be appropriate, but the laminar boundary layer equations are
not valid if the flow is transitional or fully turbulent. The laminar boundary layer on a smooth
flat plate under clean flow conditions begins to transition towards turbulence at Rex  1  105.
In practical engineering applications, walls may not be smooth and there may be vibrations,
noise, and fluctuations in the freestream flow above the wall, all of which contribute to an even
earlier start of the transition process.
- If flow separation occurs, the boundary layer approximation is no longer appropriate in the
separated flow region. The main reason for this is that a separated flow region contains reverse
flow, and the parabolic nature of the boundary layer equations is lost.
10–81 Two definitions of displacement thickness are given in this chapter. Write both definitions
in your own words. For the laminar boundary layer growing on a flat plate, which is larger—
boundary layer thickness d or displacement thickness d*? Discuss.
ANSWER: The two definitions of displacement thickness are:

* Displacement thickness is the distance that a streamline just outside of the boundary layer is
deflected away from the wall due to the effect of the boundary layer.
* Displacement thickness is the imaginary increase in thickness of the wall, as seen by the outer
flow, due to the effect of the growing boundary layer.
For a laminar boundary layer,  is larger than *.  is defined by the overall thickness of the
boundary layer, whereas * is an integrated thickness across the boundary layer that averages
the mass deficit across the boundary layer. Therefore, it is not surprising that * is less than .
10–82 Explain the difference between a favorable and an adverse pressure gradient in a boundary
layer. In which case does the pressure increase downstream? Why?
ANSWER: When the pressure decreases downstream, the boundary layer is said to experience
to a favorable pressure gradient. When the pressure increases downstream, the boundary layer is
subjected to an adverse pressure gradient. The term “favorable” is used because the boundary
layer is unlikely to separate off the wall. On the other hand, “adverse” or “unfavorable” indicates
that the boundary layer is more likely to separate off the wall.
Discussion A favorable pressure gradient occurs typically at the front of a body, whereas an
adverse pressure gradient occurs typically at the back portion of a body.
10–107 For each statement, choose whether the statement is true or false, and discuss
A. The velocity potential function can be defined for threedimensional flows.

B. The vorticity must be zero in order for the stream function to be defined.
C. The vorticity must be zero in order for the velocity potential function to be defined.
D. The stream function can be defined only for two-dimensional flow fields.
ANSWER:
A. True: We do not have to make the 2-D approximation in order to define the velocity
potential function –  can be defined for any flow if the vorticity is zero.
B. False: The stream function is definable for any two-dimensional flow field, regardless of
the value of vorticity.
C. True: The velocity potential function is valid only for irrotational flow regions where the
vorticity is zero.
D. True: The stream function is defined from the continuity equation, and is valid only for
two-dimensional flows. Note that some researchers have defined three-dimensional forms
of the stream function, but these are beyond the scope of the present introductory text
book.
CHAPTER 15: CFD
15 – 1 A CFD code is used to solve a two-dimensional (x and y), incompressible, laminar flow
without free surfaces. The fluid is Newtonian. Appropriate boundary conditions are used. List the
variables (unknowns) in the problem, and list the corresponding equations to be solved by the
computer?
Answer: There are only three unknowns in this problem, u, v, and P (orP’). Thus, we require
three equations: continuity, x momentum (or x component of Navier-Stokes), and y momentum
(or y component of Navier-Stokes). These equations, when combined with the appropriate
boundary conditions, are sufficient to solve the problem.
The actual equations to be solved by the computer are discretized versions of the differential
equations.
15 – 2 Write a brief (a few sentences) definition and description of each of the following, and
provide example(s) if helpful: (a) computational domain, (b) mesh, (c) transport equation, (d)
coupled equations?
Answer:
(a) A computational domain is a region in space (either 2-D or 3-D) in which the numerical
equations of fluid flow are solved by CFD. The computational domain is bounded by edges (2D)
or faces (3D) on which boundary conditions are applied.
(b) A mesh is generated by dividing the computational domain into tiny cells. The numerical
equations are then solved in each cell of the mesh. A mesh is also called a grid.
(c) A transport equation is a differential equation representing how some property is transported
through a flow field. The transport equations of fluid mechanics are conservation equations. For
example, the continuity equation is a differential equation representing the transport of mass, and
also conservation of mass. The Navier-Stokes equation is a differential equation representing the
transport of linear momentum, and also conservation of linear momentum.
(d) Equations are said to be coupled when at least one of the variables (unknowns) appears in
more than one equation. In other words, the equations cannot be solved alone, but must be solved
simultaneously with each other. This is the case with fluid mechanics since each component of
velocity, for example, appears in the continuity equation and in all three components of the
Navier-Stokes equation.
15 – 3 What is the difference between a node and an interval and how are they related to cells? In
Fig. P15–3C, how many nodes and how many intervals are on each edge?
Answer: Nodes are points along an edge of a computational domain that represent the vertices of
cells. In other words, they are the points where corners of the cells meet. Intervals, on the other
hand, are short line segments between nodes. Intervals represent the small edges of cells
themselves. In Fig. P15-3 there are 6 nodes and 5 intervals on the top and bottom edges. There
are 5 nodes and 4 intervals on the left and right edges.
15 – 4 For the two-dimensional computational domain of Fig. P15–3C, with the given node
distribution, sketch a simple structured grid using four-sided cells and sketch a simple
unstructured grid using three-sided cells. How many cells are in each? Discuss.
Answer:
We construct the two grids in the figure: (a) structured, and (b) unstructured. There are 5 * 4 =
20 cells in the structured grid, and there are 36 cells in the unstructured grid (depends on each
student solution).
15 – 5 For the two-dimensional computational domain of Fig. P15–3C, with the given node
distribution, sketch a simple structured grid using four-sided cells and sketch a simple
unstructured polyhedral grid using at least one of each: 3-sided, 4-sided, and 5-sided cells. Try to
avoid large skewness. Compare the cell count for each case and discuss your results.
Answer: We construct the two grids in the figure: (a) structured, and (b) unstructured polyhedral.
We show two other options in (c) and (d). There are many possible answers for the polyhedral
mesh, depending on how large you want your cells to be. There are 5 * 4 = 20 cells in the
structured grid. There are 22 cells in polyhedral grid (b). There are some cells with 3 sides, 4
sides, and 5 sides, as required. Compared to the triangular mesh with 36 cells, we have reduced
the cell count considerably. In (c) and (d), there are 21 cells and 18 cells respectively. In case (d)
we have reduced the cell count below that of even the structured grid. In that case, 3 of the cells
have 6 sides each. The cell reduction is particularly useful in large 3-D problems where CPU
time and computer memory are important limitations.
15 – 6 Summarize the eight steps involved in a typical CFD analysis of a steady, laminar
flow field?
Answer:
1. Specify a computational domain and generate a grid.

2. Specify boundary conditions on all edges or faces.
3. Specify the type of fluid and its properties.
4. Specify numerical parameters and solution algorithms.
5. Apply initial conditions as a starting point for the iteration.
6. Iterate towards a solution.
7. After convergence, analyze the results (post processing).
8. Calculate global and integral properties as needed.
The order of some of the steps is interchangeable, particularly Steps 2 through 5.
15 – 7 Suppose you are using CFD to simulate flow through a duct in which there is a circular
cylinder as in Fig. P15–7C. The duct is long, but to save computer resources you choose a
computational domain in the vicinity of the cylinder only. Explain why the downstream edge of
the computational domain should be further from the cylinder than the upstream edge?
Answer: Flow separates over bluff bodies, generating a wake with reverse flow and eddies
downstream of the body. There are no such problems upstream. Hence it is always wise to
extend the downstream portion of the domain as far as necessary to avoid reverse flow problems
at the outlet boundary.
Discussion The same problems arise at the outlet of ducts and pipes – sometimes we need to
extend the duct to avoid reverse flow at the outlet boundary.
15 – 8 Write a brief (a few sentences) discussion about the significance of each of the following
in regards to an iterative CFD solution: (a) initial conditions, (b) residual, (c) iteration, (d)
postprocessing?
Answer: (a) In a CFD solution, we typically iterate towards a solution. In order to get started, we
make some initial conditions for all the variables (unknowns) in the problem. These initial
conditions are wrong, of course, but they are necessary as a starting point. Then we begin the
iteration process, eventually obtaining the solution.
(b) A residual is a measure of how much our variables differ from the “exact” solution. We
construct a residual by putting all the terms of a transport equation on one side, so that the terms
all add to zero if the solution is correct. As we iterate, the terms will not add up to zero, and the
remainder is called the residual. As the CFD solution iterates further, the residual should
(hopefully) decrease.
(c) Iteration is the numerical process of marching towards a final solution, beginning with initial
conditions, and progressively correcting the solution. As the iteration proceeds, the variables
converge to their final solution as the residuals decrease.
(d) Once the CFD solution has converged, post processing is performed on the solution.
Examples include plotting velocity and pressure fields, calculating global properties, generating
other flow quantities like vorticity, etc. Post processing is performed after the CFD solution has
been found and does not change the results. Post processing is generally not as CPU intensive as
the iterative process itself.
Discussion We have assumed steady flow in the above discussions.
How would it be different if the flow was unsteady?
15 – 9 Briefly discuss how each of the following is used by CFD codes to speed up the
iteration process: (a) multigridding and (b) artificial time?
Answer:
(a) With multigridding, solutions of the equations of motion are obtained on a coarse grid
first, followed by successively finer grids. This speeds up convergence because the gross
features of the flow are quickly established on the coarse grid, and then the iteration
process on the finer grid requires less time.
(b) In some CFD codes, a steady flow is treated as though it were an unsteady flow. Then,
an artificial time is used to march the solution in time. Since the solution is steady,
however, the solution approaches the steady-state solution as “time” marches on. In some
cases, this technique yields faster convergence.
Discussion There are other “tricks” to speed up the iteration process, but CFD solutions
often take a long time to converge.
15 – 10 Of the boundary conditions discussed in this chapter, list all the boundary conditions that
may be applied to the right edge of the two-dimensional computational domain sketched in Fig.
P15–10C. Why can’t the other boundary conditions be applied to this edge?
Answer: We may apply the following boundary conditions: outflow, pressure inlet, pressure
outlet, symmetry (to be discussed), velocity inlet, and wall. The curved edge cannot be an axis
because an axis must be a straight line. The edge cannot be a fan or interior because such edges
cannot be at the outer boundary of a computational domain. Finally, the edge cannot be periodic
since there is no other edge along the boundary of the computational domain that is of identical
shape (a periodic boundary must have a “partner”). The symmetry boundary condition merits
further discussion. Numerically, gradients of flow variables in the direction normal to a
symmetry boundary condition are set to zero, and there is no mathematical reason why the
curved right edge of the present computational domain cannot be set as symmetry. However, you
would be hard pressed to think of a physical situation in which a curved edge like that would be
a valid symmetry boundary condition.
Discussion Just because you can set a boundary condition and generate a CFD result does not
guarantee that the result is physically meaningful.
15 – 11 What is the standard method to test for adequate grid resolution when using CFD?
Answer: The standard method to test for adequate grid resolution is to increase the resolution (by
a factor of 2 in all directions if feasible) and repeat the simulation. If the results do not change
appreciably, the original grid is deemed adequate. If, on the other hand, there are significant
differences between the two solutions, the original grid is likely of inadequate resolution. In such
a case, an even finer grid should be tried until the grid is adequately resolved.
Discussion Keep in mind that if the boundary conditions are not specified properly, or if the
chosen turbulence model is not appropriate for the flow being simulated by CFD, no amount of
grid refinement is going to make the solution more physically correct.
15 – 12 What is the difference between a pressure inlet and a velocity inlet boundary condition?
Explain why you cannot specify both pressure and velocity at a velocity inlet boundary condition
or at a pressure inlet boundary condition?
Answer: At a pressure inlet we specify the pressure but not the velocity. At a velocity inlet we
specify the opposite – velocity but not pressure. To specify both pressure and velocity would
lead to mathematical overspecification, since pressure and velocity are coupled in the
equations of motion. When pressure is specified at a pressure inlet (or outlet), the CFD code
automatically adjusts the velocity at that boundary. In a similar manner, when velocity is
specified at a velocity inlet, the CFD code adjusts the pressure at that boundary.
Discussion Since pressure and velocity are coupled, specification of both at a boundary
would lead to inconsistencies in the equations of motion at that boundary.
Does this happen even if the specified pressure and velocity perfectly match each other?
15 – 13 An incompressible CFD code is used to simulate the flow of air through a two-
dimensional rectangular channel (Fig. P15–13C). The computational domain consists of four
blocks, as indicated. Flow enters block 4 from the upper right and exits block 1 to the left as
shown. Inlet velocity V is known and outlet pressure Pout is also known. Label the boundary
conditions that should be applied to every edge of every block of this computational domain?
15 – 14 Consider Prob. 15–13C again, except let the boundary condition on the common
edge between blocks 1 and 2 be a fan with a specified pressure rise from right to left across
the fan. Suppose an incompressible CFD code is run for both cases (with and without the
fan). All else being equal, will the pressure at the inlet increase or decrease? Why? What
will happen to the velocity at the outlet? Explain.
Answer: Since the fan helps to push air through the channel, the inlet pressure will adjust
itself so that less inlet pressure is required. In other words, the inlet pressure will decrease
when the fan is turned on. Since the inlet velocity is the same in both cases, the mass flow
rate (and volume flow rate since the flow is incompressible) must remain the same for
either case. Therefore, outlet velocity will not change.
Discussion It may seem at first glance that Vout should increase because of the fan, but in
order to conserve mass, the outlet velocity cannot change. The solution is constrained by
the specified inlet boundary condition. In a real physical experiment, there is no such
restriction. The fan would cause the inlet pressure to decrease, the inlet velocity to increase,
and the outlet velocity to increase.
15 – 15 List ten boundary conditions that are used with CFD to solve incompressible fluid
flow problems. For each one, provide a brief description and give an example of how that
boundary condition is used?
Answer:
 Axis: Used in axisymmetric flows as the axis of rotation. Example: the axis of a torpedo-
shaped body.
 Fan: An internal edge (2-D) or face (3-D) across which a sudden pressure rise is
specified. Example: an axial flow fan in a duct.
 Interior: An internal edge (2-D) or face (3-D) across which nothing special happens –
the interior boundary condition is used at the interface between two blocks. Example: all
of the multiblock problems in this chapter, which require this boundary condition at the
interface between any two blocks.
 Outflow: An outlet boundary condition in which the gradient of fluid properties is zero
normal to the outflow boundary. Outflow is typically useful far away from the object or
area of interest in a flow field. Example: the far field of flow over a body.
 Periodic: When the physical geometry has periodicity, the periodic boundary condition is
used to specify that whatever passes through one face of the periodic pair must
simultaneously enter the other face of the periodic pair. Example: in a heat exchanger
where there are several rows of tubes.
 Pressure inlet: An inflow boundary in which pressure (but not velocity) is known and
specified across the face. Example: the high pressure settling chamber of a blow-down
wind tunnel facility.
 Pressure outlet: An outflow boundary in which pressure (but not velocity) is known and
specified across the face. Example: the outlet of a pipe exposed to atmospheric pressure.
 Symmetry: A face over which the gradients of all flow variables are set to zero normal to
the face – the result is a mirror image across the symmetry plane. Fluid cannot flow
through a symmetry plane. Example: the mid-plane of flow over a circular cylinder in
which the lower half is a mirror image of the upper half.
 Velocity inlet: An inflow boundary condition in which velocity (but not pressure) is
known and specified across the face. Example: a uniform freestream inlet flow entering a
computational domain from one side.
 Wall: A boundary through which fluid cannot pass and at which the no-slip condition (or
a shear stress condition) is applied. Example: the surface of an airfoil that is being
modeled by CFD.
Discussion There are additional boundary conditions used in CFD calculations, but these
are the only ones discussed in this chapter.
15 – 16 A CFD code is used to simulate flow over a two dimensional airfoil at an angle of attack.
A portion of the computational domain near the airfoil is outlined in Fig. P15–16 (the
computational domain extends well beyond the region outlined by the dashed line). Sketch a
coarse structured grid using four-sided cells and sketch a coarse unstructured grid using three-
sided cells in the region shown. Be sure to cluster the cells where appropriate. Discuss the
advantages and disadvantages of each grid type?
Answer: Notice that the cells are clustered (more fine) near the surface of the airfoil since there
is likely to be large velocity gradients there (in the boundary layer).
The structured grid in Fig 1a is called a C-grid since it wraps around the airfoil like the letter
“C”. The main advantage of the structured grid is that we can get high resolution near the surface
with few cells. The main advantage of the unstructured grid is that it is somewhat easier to
generate when the geometry is complicated (especially for highly curved surfaces). Furthermore,
it is easier to transition between curved and straight edges with an unstructured grid. The main
disadvantage of an unstructured grid is that more cells are required for the same spatial
resolution.
15 – 17 For the airfoil of Prob. 15–16, sketch a coarse hybrid grid and explain the advantages of
such a grid?
Answer: We sketch a hybrid grid in the figure. Note that the grid is structured near the airfoil
surface, but unstructured beyond the surface. The advantage of a hybrid grid is that it combines
the advantages of both structured and unstructured grids. Near surfaces we can use a structured
grid to finely resolve the boundary layer with a minimum number of cells, and away from
surfaces we can use an unstructured grid so that we can rapidly expand the cell size. We can also
more easily blend the grid into the edges of the computational domain with an unstructured grid.
Discussion A structured grid is generally the best choice, but a hybrid grid is often a better option
than a fully unstructured grid.
15 – 18 An incompressible CFD code is used to simulate the flow of water through a two-
dimensional rectangular channel in which there is a circular cylinder (Fig. P15–18). A
timeaveraged turbulent flow solution is generated using a turbulence model. Top–bottom
symmetry about the cylinder is assumed. Flow enters from the left and exits to the right as
shown. Inlet velocity V is known, and outlet pressure Pout is also known. Generate the blocking
for a structured grid using four-sided blocks, and sketch a coarse grid using four-sided cells,
being sure to cluster cells near walls. Also be careful to avoid highly skewed cells. Label the
boundary conditions that should be applied to every edge of every block of your computational
domain. (Hint: Six to seven blocks are sufficient.)
Go back to the question answer to understand more!
15 – 19 An incompressible CFD code is used to simulate the flow of gasoline through a two-
dimensional rectangular channel in which there is a large circular settling chamber (Fig. P15–
19). Flow enters from the left and exits to the right as shown. A time-averaged turbulent flow
solution is generated using a turbulence model. Top–bottom symmetry is assumed. Inlet velocity
V is known, and outlet pressure Pout is also known. Generate the blocking for a structured grid
using four-sided blocks, and sketch a coarse grid using four-sided cells, being sure to cluster
cells near walls. Also be careful to avoid highly skewed cells. Label the boundary conditions that
should be applied to every edge of every block of your computational domain?
Go back to the question answer to understand more!
15 – 20 Redraw the structured multiblock grid of Fig. 15–12b for the case in which your CFD
code can handle only elementary blocks. Renumber all the blocks and indicate how many i- and
j-intervals are contained in each block. How many elementary blocks do you end up with? Add
up all the cells, and verify that the total number of cells does not change.
15-21 Suppose your CFD code can handle nonelementary blocks. Combine as many blocks of
Fig. 15–12b as you can. The only restriction is that in any one block, the number of i-intervals
and the number of j-intervals must be constants. Show that you can create a structured grid with
only three nonelementary blocks. Renumber all the blocks and indicate how many i- and j-
intervals are contained in each block. Add up all the cells and verify that the total number of cells
does not change.
Block 2 is called an O-grid (for obvious reasons).
15 – 22 A new heat exchanger is being designed with the goal of mixing the fluid downstream of
each stage as thoroughly as possible. Anita comes up with a design whose cross section for one
stage is sketched in Fig. P15–22. The geometry extends periodically up and down beyond the
region shown here. She uses several dozen rectangular tubes inclined at a high angle of attack to
ensure that the flow separates and mixes in the wakes. The performance of this geometry is to be
tested using two-dimensional time-averaged CFD simulations with a turbulence model, and the
results will be compared to those of competing geometries. Sketch the simplest possible
computational domain that can be used to simulate this flow. Label and indicate all boundary
conditions on your diagram. Discuss.
15 – 23 Sketch a coarse structured multiblock grid with foursided elementary blocks and
four-sided cells for the computational domain of Prob. 15–22?
Answer: We choose the computational domain of Fig. 1 of the previous problem. Since all
edges are straight, the blocking scheme can be rather simple. We sketch the blocking
topology and apply a coarse mesh in Fig. 1 for the case in which the CFD code does not
require the node distribution to be exactly the same on periodic pairs.
Unfortunately, many CFD codes require that the node distribution on periodic pairs of
edges be identical (the two edges of a periodic pair are “linked” in the grid generation
process). In such a case, the grid of Fig. 1 would not be acceptable. Furthermore, although
the edges of the blocks of Fig. 1 are not split with respect to adjacent blocks, the top edges
of Block 1 and Block 3 are split with respect to the boundary conditions (part of the edge is
periodic and part is a wall). Thus these blocks are not really elementary blocks after all.
We construct a more elaborate blocking topology in Fig. 2 to correct these problems. The
node distribution on the edges of each periodic pair are identical, at the expense of more
complexity (7 instead of 5 blocks) and more cell skewness.
15 – 24 Anita runs a CFD code using the computational domain and grid developed in Probs.
15–22 and 15–23. Unfortunately, the CFD code has a difficult time converging and Anita
realizes that there is reverse flow at the outlet (far right edge of the computational domain).
Explain why there is reverse flow, and discuss what Anita should do to correct the problem.
Answer: Reverse flow at an outlet is usually an indication that the computational domain is not
large enough. In this case the rectangular heat exchanger tubes are inclined at 35 degree, and the
flow will most certainly separate, leaving large recirculating eddies in the wakes. Anita should
extend the computational domain in the horizontal direction downstream so that the eddies have
a chance to “close” and the flow has a chance to re-develop into a flow without any reverse flow.
Discussion In most commercial CFD codes a warning will pop up on the computer monitor
whenever there is reverse flow at an outlet. This is usually an indication that the computational
domain should be enlarged.
15 – 25 As a follow-up to the heat exchanger design of Prob. 15–22, suppose Anita’s design is
chosen based on the results of a preliminary single-stage CFD analysis. Now she is asked to
simulate two stages of the heat exchanger. The second row of rectangular tubes is staggered and
inclined oppositely to that of the first row to promote mixing (Fig. P15–25). The geometry
extends periodically up and down beyond the region shown here. Sketch a computational domain
that can be used to simulate this flow. Label and indicate all boundary conditions on your
diagram. Discuss.
15 – 26 Sketch a structured multiblock grid with foursided elementary blocks for the
computational domain of Prob. 15–25. Each block is to have four-sided structured cells, but you
do not have to sketch the grid, just the block topology. Try to make all the blocks as rectangular
as possible to avoid highly skewed cells in the corners. Assume that the CFD code requires that
the node distribution on periodic pairs of edges be identical (the two edges of a periodic pair are
“linked” in the grid generation process). Also assume that the CFD code does not allow a block’s
edges to be split for application of boundary conditions.
Answer: Note that with this blocking topology we had to split the periodic “mid” boundary pair
into two edges (the tops of blocks 6 and 7 and the bottoms of blocks 9 and 10). As long as both
pairs of each segment are the same size and have the same number of nodes, this is not a
problem. In the CFD code we would have to name each periodic pair separately, however. The
block numbers are labeled. Notice that most of the blocks are nearly rectangular such that none
of the computational cells would have to be highly skewed.
15 – 27 Consider the two-dimensional wye of Fig. P15–27. Dimensions are in meters, and the
drawing is not to scale. Incompressible flow enters from the left, and splits into two parts.
Generate three coarse grids, with identical node distributions on all edges of the computational
domain: (a) structured multiblock grid, (b) unstructured triangular grid, and (c) unstructured
quadrilateral grid. Compare the number of cells in each case and comment about the quality of
the grid in each case.
Answer: The structured multi-block mesh is shown in Fig.
1a. We split the domain into four blocks for convenience,
and to achieve cells with minimal skewing. There are 1060
cells. The unstructured triangular mesh is shown in Fig. 1b.
There is only one block, and it contains 1996 cells. The
unstructured quad mesh is shown in Fig. 1c. It has 833 cells
in its one block. Comparing the three meshes, the triangular
unstructured mesh has too many cells. The unstructured
quad mesh has the least number of cells, but the clustering of
cells occurs in undesirable locations, such as at the outlets on
the right. The structured quad mesh seems to be the best
choice for this geometry – it has only about 27% more cells
than the unstructured quad mesh, but we have much more
control on the clustering of the cells. Skewness is not a
problem with any of the meshes.
15 – 28 ANSYS Solution (IMPORTANT)
15 – 29 ANSYS Solution (IMPORTANT)
15 – 44 For each statement, choose whether the statement is

true or false, and discuss your answer briefly: (a) The
physical validity of a CFD solution always improves as the
grid is refined. (b) The x-component of the Navier–Stokes equation is an example of a transport
equation. (c) For the same number of nodes in a two-dimensional mesh, a structured grid
typically has fewer cells than an unstructured triangular grid. (d ) A time-averaged turbulent flow
CFD solution is only as good as the turbulence model used in the calculations.
Answer:
a) False: If the boundary conditions are not correct, if the computational domain is not large
enough, etc., the solution can be erroneous and nonphysical no matter how fine the grid.
b) True: Each component of the Navier-Stokes equation is a transport equation.
c) True: The four-sided cells of a 2-D structured grid require less cells than do the triangular
cells of a 2-D unstructured grid. (Note however, that some unstructured cells can be four-
sided as well as three-sided.)
d) True: Turbulence models are approximations of the physics of a turbulent flow, and
unfortunately are not universal in their application.
15 – 45 In Prob. 15–19 we take advantage of top–bottom symmetry when constructing our
computational domain and grid. Why can’t we also take advantage of the right–left
symmetry in this exercise? Repeat the discussion for the case of potential flow.
In the time-averaged CFD simulation, we are not concerned about top-bottom fluctuations
or periodicity. Thus, top-bottom symmetry can be assumed. However, fluid flows do not
have upstream-downstream symmetry in general, even if the geometry is perfectly
symmetric fore and aft. In the problem at hand for example, the flow in the channel
develops downstream. Also, the flow exiting the left channel enters the circular settling
chamber like a jet, separating at the sharp corner. At the opposite end, fluid leaves the
settling chamber and enters the duct more like an inlet flow, without significant flow
separation. We certainly cannot expect fore-aft symmetry in a flow such as this. On the
other hand, potential flow of a symmetric geometry yields a symmetric flow, so it would be
okay to cut our grid in half, invoking fore-aft symmetry.
Discussion If unsteady or oscillatory effects were important, we should not even specify
top-bottom symmetry in this kind of flow field.
15 – 46 Gerry creates the computational domain sketched in Fig. P15–46C to simulate flow
through a sudden contraction in a two-dimensional duct. He is interested in the timeaveraged
pressure drop and the minor loss coefficient created by the sudden contraction. Gerry generates a
grid and calculates the flow with a CFD code, assuming steady, turbulent, incompressible flow
(with a turbulence model). (a) Discuss one way that Gerry could improve his computational
domain and grid so that he would get the same results in approximately half the computer time.
(b) There may be a fundamental flaw in how Gerry has set up his computational domain. What is
it? Discuss what should be different about Gerry’s setup.
Answer: (a) Since Gerry is not interested in unsteady fluctuations (which may be unsymmetric),
he could eliminate half of the domain. In other words, he could assume that the axis is a plane of
symmetry between the top and bottom of the channel. Gerry’s grid would be cut in size by a
factor of two, leading to approximately half the required CPU time, but yielding virtually
identical results.
(b) The fundamental flaw is that the outflow boundary is not far enough downstream. There will
likely be flow separation at the corners of the sudden contraction. With a duct that is only about
three duct heights long, it is possible that there will be reverse flow at the outlet. Even if there is
no reverse flow, the duct is nowhere near long enough for the flow to achieve fully developed
conditions. Gerry should extend the outlet duct by many duct heights to allow the flow to
develop downstream and to avoid possible reverse flow problems.
Discussion The inlet appears to be perhaps too short as well. If Gerry specifies a fully developed
channel flow velocity profile at the inlet, his results may be okay, but again it is better to extend
the duct many duct heights beyond what Gerry has included in his computational domain.
15 – 47 Think about modern high-speed, large-memory computer systems. What feature of such
computers lends itself nicely to the solution of CFD problems using a multiblock grid with
approximately equal numbers of cells in each individual block? Discuss.
Answer: The fastest computers are multi-processor computers. In other words, the computer
system contains more than one CPU – a parallel computer. Modern parallel computers may
combine 32, 64, 128, or more CPUs or nodes, all working together. In such a situation it is
natural to let each node operate on one block. If all the nodes are identical (equal speed and equal
RAM), the system is most efficient if the blocks are of similar size.
Discussion In such a situation there must be communication between the nodes. At the interface
between blocks, for example, information must pass during the CFD iteration process.
15 – 48 What is the difference between multigridding and multiblocking? Discuss how each may
be used to speed up a CFD calculation. Can these two be applied together?
Answer: Multigridding has to do with the resolution of an established grid during CFD
calculations. With multigridding, solutions of the equations of motion are obtained on a coarse
grid first, followed by successively finer grids. This speeds up convergence because the gross
features of the flow are quickly established on the coarse grid (which takes less CPU time), and
then the iteration process on the finer grid requires less time. Multiblocking is something totally
different. It refers to the creation of two or more separate blocks or zones, each with its own grid.
The grids from all the blocks collectively create the overall grid. As discussed in the previous
problem, multiblocking can have some speed advantages if using a parallel-processing computer.
In addition, some CFD calculations would require too much RAM if the entire computational
domain were one large block. In such cases, the grid can be split into multiple blocks, and the
CFD code works on one block at a time. This requires less RAM, although information from the
dormant blocks must be stored on disk or solid state memory chips, and then swapped into and
out of the computer’s RAM. There is no reason why multigridding cannot be used on each block
separately. Thus, multigridding and multiblocking can be used together.
Discussion Although all the swapping in and out requires more CPU time and I/O time, for large
grids multiblocking can sometimes mean the difference between being able to run and not being
able to run at all.
15 – 49 Suppose you have a fairly complex geometry and a CFD code that can handle
unstructured grids with triangular cells. Your grid generation code can create an unstructured
grid very quickly. Give some reasons why it might be wiser to take the time to create a
multiblock structured grid instead. In other words, is it worth the effort? Discuss.
Answer: There are several reasons why a structured grid is “better” than an unstructured grid,
even for a case in which the CFD code can handle unstructured grids. First of all the structured
grid can be made to have better resolution with fewer cells than the unstructured grid. This is
important if computer memory and CPU time are of concern. Depending on the CFD code, the
solution may converge more rapidly with a structured grid, and the results may be more accurate.
In addition, by creating multiple blocks, we can more easily cluster cells in certain blocks and
locations where high resolution is necessary, since we have much more control over the final
grid with a structured grid.

Conceptual Fluid Questions

Încărcat de

Informații document

Descriere originală:

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Conceptual Fluid Questions

Încărcat de

Drepturi de autor:

Formate disponibile

CHAPTER 9: Differential Analysis of Fluid Flow

9 – 4 For a three-dimensional, unsteady, incompressible flow field in which temperature

ANSWER: Curves of constant stream function represent streamlines of a flow. A stream

9 – 67 Unique Conceptual Question

ANSWER: All these fluids are examples of non-Newtonian fluids.

9 – 84 Unique Conceptual Question

9 – 88 Unique Conceptual Question

9 – 98 Unique Conceptual Question

9 – 100 Unique Conceptual Question

CHAPTER 10: Potential Flow

ANSWER: A boundary layer grows along the floor, both

10–4 Which nondimensional parameter in the nondimensionalized Navier–Stokes equation is

ANSWER: We determine if an approximation is appropriate by comparing the orders of

2. Euler number: Eu is the ratio of a characteristic pressure difference to a characteristic pressure

10–17 Write a one-word description of each of the

I  Unsteady – II  Inertial – III  Pressure – IV  Gravity – V  Viscous term

10–49 What is D’Alembert’s paradox? Why is it a paradox?

ANSWER: The vorticity determines whether a region of flow is rotational or irrotational.

10–73 For each statement, choose whether the statement is true or

A. At a given x-location, if the Reynolds number were to

10–76 What is a trip wire, and what is its purpose?

ANSWER: We list the five steps below, with a description of each:

ANSWER: The two definitions of displacement thickness are:

A. The velocity potential function can be defined for threedimensional flows.

CHAPTER 15: CFD

1. Specify a computational domain and generate a grid.

The order of some of the steps is interchangeable, particularly Steps 2 through 5.

Block 2 is called an O-grid (for obvious reasons).

15 – 28 ANSYS Solution (IMPORTANT)

15 – 29 ANSYS Solution (IMPORTANT)

15 – 44 For each statement, choose whether the statement is

S-ar putea să vă placă și