유체역학 강의노트
유체역학 강의노트
유체역학 강의노트
in
Chemical Engineering
Eun-Suok Oh
2022
This classnote has been written based on the textbooks for fluid mechanics, Introduction to Chem-
ical Engineering Fluid Mechanics, W. M. Deen, 2016 and Fundamentals of Momentum, Heat, and
Mass Transfer, J. R. Welty, C. E. Wicks, R. E. Wilson and G. L. Rorrer, 5th ed., 2009 to teach
students in University of Ulsan. Many of paragraphs, figures, and tables of those are inserted without
editing. Some of figures and examples are also from the other textbooks, Transport Phenomena, R. B.
Bird, W. E. Stewart, and E. N. Lightfoot, 2nd ed., 2002 and Unit Operations of Chemical Engineer-
ing, W. L. McCabe, J. C. Smith, and P. Harriott, 7th ed., 2010. Most of examples in this note are
extracted from these textbooks. It is written only for undergraduate students who are taking
my Fluid Mechanics course. Commercial use of this classnote is strictly prohibited in
any forms. Any types of transfer to others are not allowed due to the copyright of the
books.
3
G03312
Fluid Mechanics
Books: Textbook
· W. M. Deen, Introduction to Chemical Engineering Fluid Mechanics, Cambridge Univer-
sity Press (2016): reference ChD . and SectD .
Additional References
Grading
· Class Attendance (10%): Attendance at class is compulsory. In the case that students won’t
make it for some reasons, let me know in advance.
· Quiz (30%): On-line Quizzes will be given in the beginning of the first class of every week.
· Midterm Examination (30%): The midterm exam is scheduled at the midpoint of the course
and will cover material up to that point.
· Final Examination (30%): The final exam is scheduled at the last day of classes and will
cover the material after the midterm exam.
Office Hours
No regular time is assigned for office hours. I would be happy to meet any of you anytime
I am free.
Comments
The aim of this class is for undergraduate students to understand basic knowledge on the
engineering mechanics for fluid. This class begins by the vector and tensor analysis and
introduces basic conservation laws such as mass, momentum, and energy conservation. Note
that I WON’T follow the order in the main textbooks during the classes. Rather, I will
make the order more systematically in the viewpoint of sophomores. In addition, some of
stuff used in this class will be from the above references as well as internet data sources.
4
유체역학을 수강하면
1 흐르는 유체에서 질량이 어떻게 보존되는 이해할 수 있다.
2 뉴튼의 2법칙은 단순히 힘은 질량 곱하기 가속도 뿐만 아니라, 이로 부터 물체에 주어진 힘이
운동량을 어떻게 변화시키는지 알 수 있다.
3 힘은 물체 안에 작용하는 body force (중력 등)와 표면에 작용하는 surface force (압력, 마찰력
등)로 구성 됨을 알 수 있다.
4 점도의 물리적 의미와 변할 수 있는 값임을, 또한 점도가 물체의 운동량을 전달하는 주 원인
임을 알 수 있다.
5 역학적에너지가 항상 보존되는 것만은 아니라는 것을 알 수 있다.
6 모든 점에서 압력에너지+운동에너지+퍼텐셜에너지 합은 항상 일정 (베르누이식)을 알 수
있다.
7 산업현장에서 사용하는 펌프의 사양을 이해할 수 있다.
8 유체가 파이프나 판 사이에서 압력 또는 중력에 의하여 흐를 때 속도가 포물선 형태가 됨을 알
수 있다.
9 유체가 파이프 내 또는 유동층 반응기에서 흘러갈 때 마찰에 의하여 떨어지는 단위길이당
압력강하를 알 수 있다.
10 공기중에 기포 또는 방울은 일정한 속도로 떨어짐을 알 수 있다.
Contents
3 Conservation of Mass 33
3.1 Mass Balance for a Steady-State Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2 Equation of Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.2.1 The Differential Mass Balance for an Incompressible Flow . . . . . . . . . . . . 39
5
6 CONTENTS
9
10 CHAPTER 1. PRELIMINARY FOR FLUID MECHANICS
This field was well-established by Dr. Bird’s group in University of Wisconsin and developed more
by Dr. Slattery’s group at Texas A&M University.
Stewart
Bird
Lightfoot
Vector
2
(1, 2, 2)
a a
2a
2 y
1
-a
(1, 2, 3)
u
ez 3
ey
y
ex
1
2
In the 3-dimensional vector space, any vectors can be expressed by the combination of the three
independent vectors ex , ey , ez , which are called Basis Vectors.
u = ux ex + uy ey + uz ez (1.1)
The scalars ux , uy , uz are called the components of ex , ey , ez . The magnitude of u is given by
q
|u| = ux 2 + uy 2 + uz 2 (1.2)
12 CHAPTER 1. PRELIMINARY FOR FLUID MECHANICS
Generally, we choose the basis vectors whose magnitude is 1 (unit vector) and they are perpendicular
each other (orthonormal vector).
In addition, the basis vectors in 3-D space could be expressed as (er , eθ , ez ) or (er , eθ , eϕ )
z z ez z
ez er
eθ
eφ
ex ey uθ
uz er u
uz u ur
u eθ
y ur y y
ux uθ uφ
uy
x x x
Figure 1.2: Basis vectors in (a) Rectangular coordinate system, (b) Cylindrical coordinate system,
and (c) Spherical coordinate system
u · v = (ux ex + uy ey + uz ez ) · (vx ex + vy ey + vz ez )
(1.6)
= (ux vx ) + (uy vy ) + (uz vz )
where n is a unit vector perpendicular to both u and v and pointing in the direction that a right-
handed screw will move if turned from u toward v:
u × v = (ux ex + uy ey + uz ez ) × (vx ex + vy ey + vz ez )
= (uy vz − uz vy )ex + (uz vx − ux vz )ey + (ux vy − uy vx )ez
(1.10)
ex ey ez
= ux uy uz
vx vy vz
F F
A A
Figure 1.4: A force at a point exerted by an external force depends on both the direction of the
external force and the direction of the surface.
pressure
viscous force
viscous
force pressure force + viscous force
unit area = stress
(tensor property)
- Generally, the 2nd order tensor is just called a tensor (magnitude, 2 directions). A vector has
magnitude and one direction.
or in a matrix form,
Axx Axy Axz 0 1 0
A = Ayx Ayy Ayz , ex. A = ex ey = 0 0 0 (1.13)
Azx Azy Azz 0 0 0
I·u=u (1.14)
In a matrix form and an index notation
1 0 0
I = 0 1 0 = ex ex + ey ey + ez ez (1.15)
0 0 1
Symmetry
A tensor A is symmetric if
1 2 3
A = AT ⇔ Aij = Aji , ⃝ 4 5 (1.16)
⃝ ⃝ 6
6 independent components
∂ ∂ ∂
∇= ex + ey + ez (1.17)
∂x ∂y ∂z
It cannot stand alone and must operate on a scalar, vector, or tensor function, ∇a, ∇ · a. Note that
some of books use different symbols for
x = x1 , y = x2 , z = x3 , ex = e1 = i, ey = e2 = j, ez = e3 = z (1.18)
1.4. VECTOR DIFFERENTIAL OPERATOR 15
z z z
y y y
e xe x e xe y e xe z
x x x
z z z
y y y
e ye x e ye y e ye z
x x x
z z z
y y y
e ze x e ze y e ze z
x x x
The gradient of s at a point is a vector pointing in the direction of the maximum rate of change. The
maximum rate of change of s at that point is given by the magnitude of the gradient vector ∇s.
∂s ∂s ∂s
∇s = grad s ≡ ex + ey + ez (1.19)
∂x ∂y ∂z
Ex. 1. If s = x2 + 2xy + y + 2z 2
Answer:
∂v ∂v ∂v
∇v ≡ ex + ey + ez
∂x ∂y ∂z
∂ ∂v ∂v
= (vx ex + vy ey + vz ez )ex + ey + ez
∂x ∂y ∂z
∂ ∂ ∂ ∂v ∂v
= (vx ex ) + (vy ey ) + (vz ez ) ex + ey + ez
∂x ∂x ∂x ∂y ∂z
∂vx ∂vy ∂vz ∂v ∂v
= ex + ey + ez ex + ey + ez
∂x ∂x ∂x ∂y ∂z (1.20)
∂vx ∂vy ∂vz ∂v ∂v
= ex ex + ey ex + ez ex + ey + ez
∂x ∂x ∂x ∂y ∂z
∂vx ∂vx ∂vx
∂x ∂y ∂z
∂v
y ∂vy ∂vy
=
∂x ∂y ∂z
∂vz ∂vz ∂vz
∂x ∂y ∂z
∂v ∂v ∂v
∇ · v = div v ≡ · ex + · ey + · ez
∂x ∂y ∂z
∂ ∂v ∂v
= (vx ex + vy ey + vz ez ) · ex + · ey + · ez
∂x ∂y ∂z
∂ ∂ ∂ ∂v ∂v
= (vx ex ) + (vy ey ) + (vz ez ) · ex + · ey + · ez
∂x ∂x ∂x ∂y ∂z
∂vx ∂vy ∂vz ∂v ∂v
= ex + ey + ez · ex + · ey + · ez (1.21)
∂x ∂x ∂x ∂y ∂z
∂vx ∂vy ∂vz ∂v ∂v
= ex · ex + ey · ex + ez · ex + · ey + · ez
∂x ∂x ∂x ∂y ∂z
∂vx ∂v ∂v
= + · ey + · ez
∂x ∂y ∂z
∂vx ∂vy ∂vz
= + +
∂x ∂y ∂z
∂T ∂T ∂T
∇ · T = div T ≡ · ex + · ey + · ez
∂x ∂y ∂z
∂ ∂T ∂T
= (Txx ex ex + Txy ex ey + · · · + Tzz ez ez ) · ex + · ey + · ez
∂x ∂y ∂z
∂Txx ∂Txy ∂Tzz ∂T ∂T
= ex ex · ex + ex ey · ex + · · · + ez ez · ex + · ey + · ez
∂x ∂x ∂x ∂y ∂z
∂Txx ∂Tyx ∂Tzx ∂T ∂T
= ex + ey + ez + · ey + · ez
∂x ∂x ∂x ∂y ∂z
∂Txx ∂Txy ∂Txz ∂Tyx ∂Tyy ∂Tyz ∂Tzx ∂Tzy ∂Tzz
= + + ex + + + ey + + + ez
∂x ∂y ∂z ∂x ∂y ∂z ∂x ∂y ∂z
(1.22)
x 2xyz 1
Ex. 4. If T = 2xyz y y
1 y z
∂Txx ∂Txy ∂Txz ∂Tyx ∂Tyy ∂Tyz ∂Tzx ∂Tzy ∂Tzz
∇·T= + + ex + + + ey + + + ez
∂x ∂y ∂z ∂x ∂y ∂z ∂x ∂y ∂z
= (1 + 2xz) ex + (2yz + 1) ey + 2ez
18 CHAPTER 1. PRELIMINARY FOR FLUID MECHANICS
It should be noticed that the minus sign is required when the curl of a vector is calculated by the
above method.
Answer:
∇ × v = (1 − 2xy) ex + (2yz) ez
∂2 ∂2 ∂2
∴ ∆ ≡ ∇2 = 2 + 2 + (1.26)
∂x ∂y ∂z 2
Ex. 6. If s = x2 + 2xy + y + 2z 2
Answer:
∂2s ∂2s ∂2s
∆s = + + =6
∂x2 ∂y 2 ∂z 2
z z z
ez er
ez eθ eφ
ey θ
z er
r eθ
ex y y
y r
θ φ
x x x
1/2
x = r cos θ r = (x 2 + y 2 ) r 2 = x2 + y 2
{ y = r sin θ
z=z
{ θ = arctan (y/x)
z=z
or
{ tan θ = y/x
z=z
1/2
x = r sin θ cos φ r = (x 2 + y 2 +z 2 )
{ y = r sin θ sin φ
z = r cos θ
{ 1/2
θ = arctan ((x2 + y2 ) /z)
φ = arctan (y/x)
Figure 1.6: (a) Rectangular coordinate system, (b) Cylindrical coordinate system, and (c) Spherical
coordinate system
20 CHAPTER 1. PRELIMINARY FOR FLUID MECHANICS
and we have
∂ ∂ ∂
∇= ex + ey + ez
∂x ∂y ∂z
(1.28)
∂2 ∂2 ∂2
∆ = ∇2 = 2 + 2 +
∂x ∂y ∂z 2
For s = s(x, y, z) and v = vx ex + vy ey + vz ez , we have
∂s ∂s ∂s
∇s = ex + ey + ez
∂x ∂y ∂z
∂v ∂v ∂v ∂vx ∂vy ∂vz
∇·v = · ex + · ey + · ez = + +
∂x ∂y ∂z ∂x ∂y ∂z
2 2 2
∂ v ∂ v ∂ v
∇2 v = + + 2
∂x2 ∂y 2 ∂z
2 2
∂ 2 vx
2
∂ 2 vy ∂ 2 vy
2
∂ 2 vz ∂ 2 vz
∂ vx ∂ vx ∂ vy ∂ vz
= + + ex + + + ey + + + ez
∂x2 ∂y 2 ∂z 2 ∂x2 ∂y 2 ∂z 2 ∂x2 ∂y 2 ∂z 2
(1.29)
and we can express ∇ and ∆ in terms of the cylindrical coordinate system, (r, θ, z) and (er , eθ , ez ).
The relations between the Rectangular Cartesian and Cylindrical Coordinates are
p 2 2
2
r = x +y
2
r =x + y2
x = r cos θ
y
y
y = r sin θ or θ = arctan ⇔ tan θ = (1.31)
x x
z=z
z=z z =z
Let’s express the first terms in equation (1.29) in terms of (r, θ, z) and (er , eθ , ez ). For here s =
s(r, θ, z) where r(x, y, z), θ(x, y, z), z(x, y, z)
∂s sin θ ∂s
sx = sr rx + sθ θx + sz zx = cos θ + − (1.32)
∂r r ∂θ
and
∂r ∂θ sin θ ∂z
= cos θ, =− , =0 (1.34)
∂x y,z ∂x y,z
r ∂x y,z
In addition,
ex = cos θ er − sin θ eθ + (0) ez (1.35)
1 For example, s(u, v, w) = 2u + v + 3w where u = x + y, v = 2y + x, w = z
∂s substitution, s = 2(x + y) = (2y + x) + 3z = 3x + 4y + 3z, sx = 3
sx = =
∂x y,z chain rule, sx = su ux + sv vx + sw wx = 3
1.5. VECTOR DIFFERENTIAL OPERATORS IN OTHER COORDINATES (APP. A,B) 21
y
ey
eθ
θ er
θ
ex
p (x, y, z) = p (r, θ, z)
θ
x
Thus
sin2 θ ∂s
∂s ∂s sin θ cos θ ∂s ∂s
ex = cos2 θ − er + − sin θ cos θ + eθ (1.36)
∂x ∂r r ∂θ ∂r r ∂θ
Finally in the cylindrical coordinates we have
∂s 1 ∂s ∂s
∇s = er + eθ + ez (1.37)
∂r r ∂θ ∂z
Therefore we can say in this cylindrical coordinate system,
∂ 1 ∂ ∂
∇= er + eθ + ez (1.38)
∂r r ∂θ ∂z
Using this, we can express ∇ · v in terms of the cylindrical coordinates as follows:
∂v 1 ∂v ∂v
∇·v = · er + · eθ + · ez (1.39)
∂r r ∂θ ∂z
where v = v(r, θ, z) Since v = vr er + vθ eθ + vz ez , we have
∂v ∂ ∂vr ∂vθ ∂vz
= (vr er + vθ eθ + vz ez ) = er + eθ + ez
∂r ∂r ∂r ∂r ∂r
∂v ∂ ∂vr ∂er ∂vθ ∂eθ ∂vz
= (vr er + vθ eθ + vz ez ) = e r + vr + eθ + vθ + ez
∂θ ∂θ ∂θ ∂θ ∂θ ∂θ ∂θ (1.40)
∂vr ∂vθ ∂vz
= er + vr eθ + eθ − vθ er + ez
∂θ ∂θ ∂θ
∂v ∂ ∂vr ∂vθ ∂vz
= (vr er + vθ eθ + vz ez ) = er + eθ + ez
∂z ∂z ∂z ∂z ∂z
and finally
1 ∂ 1 ∂vθ ∂vz
∇·v = (rvr ) + + (1.41)
r ∂r r ∂θ ∂z
Similarly
∂2 ∂2 ∂2
2 ∂ ∂ ∂ ∂ ∂ ∂
∆=∇ = 2 + 2 + 2 = + +
∂x ∂y ∂z ∂x ∂x ∂y ∂y ∂z ∂z
∂ ∂ ∂r ∂ ∂ ∂θ ∂ ∂ ∂z
= + +
∂r ∂x ∂x ∂θ ∂x ∂x ∂z ∂x ∂x
∂ ∂ ∂r ∂ ∂ ∂θ ∂ ∂ ∂z
+ + + (1.42)
∂r ∂y ∂y ∂θ ∂y ∂y ∂z ∂y ∂y
∂ ∂ ∂r ∂ ∂ ∂θ ∂ ∂ ∂z
+ + +
∂r ∂z ∂z ∂θ ∂z ∂z ∂z ∂z ∂z
1 ∂2 ∂2
1 ∂ ∂
= r + 2 2+ 2
r ∂r ∂r r ∂θ ∂z
22 CHAPTER 1. PRELIMINARY FOR FLUID MECHANICS
and thus
1 ∂ 2 vr ∂ 2 vr
2 ∂ 1 ∂ 2 ∂vθ
∇ v= (rvr ) + 2 − 2 + er
∂r r ∂r r ∂θ2 r ∂θ ∂z 2
1 ∂ 2 vθ ∂ 2 vθ
∂ 1 ∂ 2 ∂vr
+ (rvθ ) + 2 + 2 + eθ (1.43)
∂r r ∂r r ∂θ2 r ∂θ ∂z 2
1 ∂ 2 vz ∂ 2 vz
1 ∂ ∂vz
+ r + 2 + ez
r ∂r ∂r r ∂θ2 ∂z 2
0 ≤ r ≤ +∞, 0 ≤ θ ≤ π, 0 ≤ ϕ ≤ 2π (1.44)
and we can express ∇ and ∆ in terms of the spherical coordinate system, (r, θ, ϕ) and (er , eθ , eϕ ).
The relations between the Rectangular Cartesian and Spherical Coordinates are
p
r = x2 + y 2 + z 2 !
x = r sin θ cos ϕ
p
x2 + y 2
y = r sin θ sin ϕ or θ = arctan (1.45)
z
z = r cos θ
y
ϕ = arctan
x
and thus
ex = (sin θ cos ϕ) er + (cos θ cos ϕ) eθ + (− sin ϕ) eϕ
ey = (sin θ sin ϕ) er + (cos θ sin ϕ) eθ + (cos ϕ) eϕ (1.46)
ez = (cos θ) er + (− sin θ) eθ + (0) eϕ
∂ 1 ∂ 1 ∂
∇= er + eθ + eϕ
∂r r ∂θ r sin θ ∂ϕ
(1.47)
∂2
2 1 ∂ 2 ∂ 1 ∂ ∂ 1
∆=∇ = 2 r + 2 sin θ + 2 2
r ∂r ∂r r sin θ ∂θ ∂θ r sin θ ∂ϕ2
In this coordinate systems,
∂s 1 ∂s 1 ∂s
∇s = er + eθ + eϕ
∂r r ∂θ r sin θ ∂ϕ
1 ∂ 1 ∂ 1 ∂vϕ
r 2 vr +
∇·v = 2 (vθ sin θ) +
r ∂r r sin θ ∂θ r sin θ ∂ϕ
2
∂ 2 vr
2 1 ∂ 2
1 ∂ ∂v r 1
∇ v = 2 2 r vr + 2 sin θ + 2 2 er
r ∂r r sin θ ∂θ ∂θ r sin θ ∂ϕ2
∂ 2 vθ
1 ∂ 2 ∂vθ 1 ∂ 1 ∂ 1 2 ∂vr 2 cot θ ∂vϕ
+ 2 r + 2 (vθ sin θ) + 2 2 + 2 − 2 eθ
r ∂r ∂r r ∂θ sin θ ∂θ r sin θ ∂ϕ2 r ∂θ r sin θ ∂ϕ
∂ 2 vϕ
1 ∂ ∂vϕ 1 ∂ 1 ∂ 1 2 ∂vr 2 cot θ ∂vθ
+ 2 r2 + 2 (vϕ sin θ) + 2 2 + + eϕ
r ∂r ∂r r ∂θ sin θ ∂θ r sin θ ∂ϕ2 r2 sin θ ∂ϕ r2 sin θ ∂ϕ
(1.48)
Chapter 2
- changes its shape (mechanically, deforms continuously when subjected to shear forces)
- more mechanically, its shear stress is proportional to its rate of shear strain. On the contrary,
the shear stress of a solid is proportional to the shear strain of the solid.
2.1.2 Mechanics
What is Mechanics?
Equilibrium
Force Matter
w/wo
Deformation Motion
- statics + dynamics
- expressed in terms of scalar, vector, and tensor such as force (vector property), deformation
(tensor property)
Molecular &
Macro Micro Nano
Quantum
-mechanics -mechanics -mechanics
-mechanics
10 0 10 -3 10 -6 10 -9 10 -12
23
24 CHAPTER 2. BASIC CONCEPTS OF FLUID MECHANICS
Control surface
Material
Body (or system)
particles
Moti
on (t
)
Control volume
Boundary
Solid Fluid
Figure 2.1: Material particles, body, control volume, control surface, and motion
at t=t0
y y y y
at t=t1
x x x x
Uniform translation Rigid-body rotation Dilatation - Simple shear -
without deformation without deformation shape-preserving volume-preserving
Figure 2.2: Two-dimensional motion of fluids; uniform translation, rigid-body motion, pure dilatation,
simple shear motions.
include spreading butter on bread or applying sunscreen. In each case one level of the liquid (butter on the knife) is
moving relative to the adjacent layer (butter on the bread). Shear rate is defined as the measure of the extent or rate
of relative motion between adjacent layers of a moving liquid = Velocity / Distance. Using the same example, we can
raise the shear rate by either raising the velocity of the knife or by placing the knife closer to the bread.
2.3. TIME DERIVATIVES: PARTIAL, TOTAL, MATERIAL (SUBSTANTIAL) 25
h(x) here.
26 CHAPTER 2. BASIC CONCEPTS OF FLUID MECHANICS
Let c is the concentration of fish in a river. Because fish swim around, the fish concentration c will
be in general a function of position (x, y, z) and time t, c(x, y, z, t) and the total differentiation in c
(t와 x, y, z가 동시에 변화할 때 c의 총변화량) is
∂c ∂c ∂c ∂c
dc = dt + dx + dy + dz (2.6)
∂t x,y,z ∂x y,z,t ∂y x,z,t ∂z x,y,t
| {z } | {z } | {z } | {z }
the change in c the change in c the change in c the change in c
due to an infinitesimal due to an infinitesimal due to an infinitesimal due to an infinitesimal
change in t only change in x only change in y only change in z only
That is, the total change in c is the sum of the changes in c according to infinitesimal changes in t, x,
y, and z, respectively.
c(t, x, y, z) : the concentration of fish in a location
v
u
v
u
where u denotes the velocity of the observer which is the same as the boat. The x, y, and z components
of the velocity correspond to ux = dx/dt, uy = dy/dt, and uz = dz/dt, respectively. In fact, ∇c · u/|u|
is a directional derivative of c in the direction of s = u/|u|, which is the rate of change of c in the
direction given by a unit vector dc/ds. Here s is a magnitude in the direction of the unit vector
s = u/|u|.
1
2
1 3 5
3 fish at time t + Δt
4
fish at III
time t 2
I 5
Figure 2.4: Fish at time t and t + ∆t in a square. The body moves in the velocity u due to the boat
velocity u.
The next limiting expression is the net rate of fish efflux across the control surface during
the time interval ∆t since region I and II correspond to the control surface as ∆t → 0.
cII |t+∆t − cI |t
lim = ċout − ċin (2.11)
∆t→0 ∆t
Therefore, we can see
u · ∇c = ċout − ċin (2.12)
and
dc ∂c ∂c
= + u · ∇c = + ċout − ċin (2.13)
dt ∂t ∂t
Acceleration
the time derivative of velocity:
dv
a≡ (2.15)
dt
∂c
=0 (2.16)
∂t
If the flow at a point varies with time, the flow termed unsteady.
coming from classical mechanics. The law has to be modified to comply with the laws of quantum mechanics and
special relativity under the principle of mass-energy equivalence, which states that energy and mass form one conserved
quantity. For very energetic systems the conservation of mass-only is shown not to hold, as is the case in nuclear
reactions and particle-antiparticle annihilation in particle physics.
2.4. FUNDAMENTAL PHYSICS LAWS 29
Figure 2.5: Conservation of Mass, Momentum, and Energy (Bird et al., Fig. 7.0-1). Macroscopic flow
system with fluid entering at plane 1 and leaving at plane 2.
F m v
The Second Law: Law of Acceleration, Conservation of Linear Momentum (운동량 보존법
칙)
1. Force acting on a body(고체 뿐만 아니라 유체도 포함) is the cause of its moving.
2. Force changes the direction of motion or the velocity of a body. It is possible for a body to keep
moving without force, but not possible to change the state of the motion.
3. Net force is equal to acceleration produced in unit mass
X
F = ma (2.21)
4. The net force acting on the body is also equal to the time rate of change of linear momentum
of a body.
X dv
F = ma = m
dt (2.22)
d(mv) dP
= = ∵ m=constant by conservation of mass
dt dt
where the linear momentum, P = mv, indicates the tendency of an object moving in a certain
direction to keep going at the same speed in the same direction. See Fig. 2.8.
Figure 2.7: 운동량보존. x 방향의 반대로 힘이 주어져서 노즐에서 나온 물의 x 방향의 운동량이 감소하
였고, y 방향으로 힘이 주어지지 않았기 때문에 전체 y 방향의 운동량은 0을 유지함.
2 m/s
500 kg 500 kg
5000 N 6000 N
5000 N 5000 N
Scale (5000N) Scale (6000 N)
Etot W
F m, E v
1. In any process, the total energy of the universe (system + surrounding) remains the same.
dEtot
=0 (2.24)
dt
2. The time rate of change of energy of a body (closed system) is equal to the rate of energy
transmission to the body from its surroundings plus the rate of work done on the body by its
surrounding.
dE
= Q̇ − Ẇ (2.25)
dt
Summary
32 CHAPTER 2. BASIC CONCEPTS OF FLUID MECHANICS
유체 및 역학
기체 & 액체
평형
힘
고체의 반대이다 변형
운동
응력 µ 변형율
Scalar, Vector, Tensor
(고체, 응력 µ 변형)
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
뉴튼의 운동법칙
!" 질량 ! !
∑ " = $% = $ = $& = ' , '(운동량) = $&
!# 보존 !# !#
• 노즐에서 나온 물의 운동량(속
도가 12m/s에서 0으로 변함)을
변화시키기 위해서는 힘이 필
요함.
• 여기서 힘의 방향은 – 방향임
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
Chapter 3
Conservation of Mass
The conservation of mass can be expressed within either a body or a control volume.
1. (Solid mechanics) The time rate of change of mass in a body or a system is zero.
dm
dt
motion (v)
dm
=0 (3.1)
dt
min
дm
дt mout
CS CV
∂m
= ṁin − ṁout (3.2)
∂t
|{z}
|{z} | {z }
mass flow rate mass flow rate
rate of mass accumulation
into control volume out of control volume
within control volume
Here m is the mass in the control volume which is a fixed region in space as mentioned in Sect.
2.2. Mass can be expressed as
When the density is not constant in the control volume, mas can be obtained from
Z
m = ρ dV (3.4)
33
34 CHAPTER 3. CONSERVATION OF MASS
where v⊥ is the velocity component perpendicular to surface A. Generally, ρv⊥ is called mass
flux defined as the amount of mass flowing through a unit cross-sectional area per unit time
[g/cm2 ·s] and Av⊥ is called the volume flow rate.
When the velocity v⊥ is not constant on the surface A, the mass flow rate becomes
Z
ṁ = ρv⊥ dA (3.6)
n1
n2
v1y v2y
y
ρ1 v1 A1 v2 ρ2
A2
mass per v1x v2x
x
unit volume
Figure 3.1 shows inclined surfaces through which some of mass enters or leaves. The mass flow
rate entering through surface A1 is
For clear understanding, let’s consider a body composed of three fishes as shown in Fig. 3.2.
Their masses are 1, 2, and 3 kg, respectively.
Ex. 7. Find the amount of salt (mS ) in the tank at any time t. Initially, the amount of salt in the
tank was 100 kg out of 1000 kg brine.
3.1. MASS BALANCE FOR A STEADY-STATE FLOW 35
Ex. 8. Find the exiting water’s average velocity. Assume that the density is the same everywhere.
Answer: First, let control volume be the total fluid in the channel. Mass flow rate entering into the
control volume is
ṁin = ρvin Ain = 8ρ (3.12)
Mass flow rate exiting from the control volume is
ṁout = ρvout Aout = ρvout (3.13)
since ṁin = ṁout , vout = 8 m/s. However, it is the velocity perpendicular to surface Aout because
the velocity in (3.5) should be perpendicular to the area. Thus the outer velocity becomes
v2 = vout / cos 60 = 16 m/s (3.14)
36 CHAPTER 3. CONSERVATION OF MASS
1m
2m Aout
1m
2 m/s
60o
2m
Figure 3.4: Steady one dimensional flow into and out of a control volume.
(x, y, z)
∂m
= ṁin − ṁout (3.16)
∂t
• Mass in the differential element shown in Fig. 3.5 is
∂
(ρ∆x∆y∆z) (3.18)
∂t
• Mass flow rate entering into and leaving from the element
3.2. EQUATION OF CONTINUITY 37
The mass flow rate entering the differential element through the surface at x is
(ρvx )x ∆y∆z (3.19)
and the mass flow rate leaving the differential element through the surface at x + ∆x is
(ρvx )x+∆x ∆y∆z (3.20)
Thus the net mass flow rate out of the differential element in the x direction is written as
[(ρvx )x+∆x − (ρvx )x ] ∆y∆z (3.21)
Similarly we can get the net mass flow rates in the y and z directions.
[(ρvy )y+∆y − (ρvy )y ] ∆x∆z, and [(ρvz )z+∆z − (ρvz )z ] ∆x∆y (3.22)
Using the mass balance (3.15) and (3.18) through (3.22), we finally have
∂
(ρ∆x∆y∆z) + [(ρvx )x+∆x − (ρvx )x ] ∆y∆z
∂t
+ [(ρvy )y+∆y − (ρvy )y ] ∆x∆z + [(ρvz )z+∆z − (ρvz )z ] ∆x∆y = 0 (3.23)
Since the volume does not change with time, we divide (3.23) by ∆x∆y∆z and take the limit as
∆x → 0, ∆y → 0, and ∆z → 0
∂ρ ∂ ∂ ∂
+ (ρvx ) + (ρvy ) + (ρvz ) = 0
∂t ∂x ∂y ∂z
∂ρ
=⇒ = −∇ · (ρv)
∂t
|{z} | {z }
net influx of mass per unit length in all direction
rate of mass per volume accumulation at a point
=한 점에서 각 방향으로 단위길이당 순유입 mass flux
=한 점에서 시간에 대한 단위부피당 질량(=밀도)의 변화율
(3.24)
This is called the differential mass balance or the equation of continuity which is the mass balance
required at a point.
With the help of the definition of divergence (1.21)
∂(ρvx ) ∂(ρvy ) ∂(ρvz ) ∂ρ ∂ρ ∂ρ ∂vx ∂vy ∂vz
∇ · (ρv) = + + = vx + vy + vz +ρ + +
∂x ∂y ∂z ∂x ∂y ∂z ∂x ∂y ∂z (3.25)
= v · ∇ρ + ρ(∇ · v)
The equation of continuity can be also expressed in terms of the material derivative (2.13)
Dρ
+ ρ(∇ · v) = 0 (3.26)
Dt
As shown in Fig. 3.6, the Dρ/Dt is the time rate of change of density in a fluid body composed of
the material particles when the fluid particles are in a motion. Thus,
m', V'
m, V
motion (v)
Dρ/Dt
Body
Figure 3.6: Physical meaning of ∇ · v. Here V and V ′ are very small such as a point.
D
Dρ Dt m/V DV Volumetric increase
∇·v =− ρ=− = V = (3.27)
Dt m/V Dt Original volume
The physical meaning of divergence is well explained in movie clips (https://www.khanacademy.org
/math/multivariable-calculus/multivariable-derivatives/divergence-grant-videos/v/divergence-intuition-
part-1?modal=1). You can clearly see the concept of the divergence of velocity through Fig. 3.7
38 CHAPTER 3. CONSERVATION OF MASS
v=xex-yey
v=xex
(2,0) (2,0)
Figure 3.7: The divergences of the points which are composed of material particles. The point (2, 0)
is an example.
∇·v > 0
∇·v < 0
∇·v = 0
A
point A: ∇·v = 0
Figure 3.8: Divergences of fluid flows. The above three flows have the same sign at any points whereas
the forth flow has a positive divergence of v for upper region and negative for the lower region,
respectively. (see Khan Academy)
3.2. EQUATION OF CONTINUITY 39
no volume
Incompressible Fluid Incompressible flow (isochoric motion)
change
(ρ = constant) ∇·v = 0
Fluid
almost no
Incompressible flow ( ρ ≈ constant
vol. change
Compressible Fluid ∇·v ≈ 0 during the flow)
dependence of volume
ρ on pressure Compressible flow ( ρ ≠ constant
change ∇·v ≠ 0 during the flow)
According to the coordinates used, equation (3.29) could be expressed in terms of velocity com-
ponents as
Ex. 9. You can easily see that a fluid is not incompressible when its velocity profile is v = xex
because ∇ · v = 1.
1 The density at any point in a fluid is defined as
ρ = lim ∆m/∆V (3.28)
∆V →0
2 The compressibility is a measure of the relative volume change of a fluid or solid as a response to a pressure (or
Ex. 10. Filtration in a hollow fiber. A common type of membrane filtration unit consists of many
hollow fibers arranged in parallel. The wall of each fiber has fine pores, and a pressure difference causes
outward flow of the solvent, thereby concentrating macromolecules or particles that are retained inside
the fiber. Figure 3.10 shows a single fiber of radius R and length L. Predicting the performance of
the device requires knowledge of v inside the fiber. The cylindrical velocity components of interest are
vz (r, z) and vr (r, z). The flow is assumed to be axisymmetric (independent of θ) and without swirl
(vθ = 0). At any position along the fiber the mean axial velocity is u(z) and the radial velocity at
the wall is vw (z) = vr (R, z). The mean velocities at the inlet and outlet are u0 and uL , respectively.
Illustrate the three levels of detail at which conservation of mass can be applied: all liquid inside the
fiber, the dotted region, a point in the fibre.
one of the dimensions is differential, what is to be constructed is a kind of shell balance. Here a shell
indicates a differential element of the flow. Equating volume inflow and outflow for this region gives
Dividing by ∆z and letting ∆z → 0 provides a differential equation that governs the mean velocity,
du 2vw
=− , u(0) = u0 (3.36)
dz R
If vw (z) is known, this can be solved for u(z). Thus, this approach provides the mean velocity at
intermediate locations along the fiber. The simplest possibility is a constant wall velocity, for which
2vw z
u(z) = u0 − , (constant vw ) (3.37)
R
A point: The most detailed information on v is contained in (3.29). With ρ constant and vθ = 0, the
cylindrical continuity equation reduces to
1 ∂ ∂vz
(rvr ) + =0 (3.38)
r ∂r ∂z
Equations (3.34) and (3.36) each could have been obtained also from a more detailed form of conser-
vation of mass. Equation (3.34) could have been found by integrating (3.36) over the channel length.
Likewise, Eq. (3.36) could have been derived by integrating (3.38) over the cross-section, as follows.
The differential area for the circular cross-section is dS = rdrdθ. Integrating each term in (3.38) gives
Z 2π Z R r=R
1 ∂
(rvr ) rdrdθ = 2π(rvr ) = 2πRvw
0 0 r ∂r r=0
Z 2π Z R Z R (3.39)
∂vz d du
rdrdθ = 2π vz r dr = πR2
0 0 ∂z dz 0 dz
and assembling these results leads to (3.36). Although either approach gives the same result, a shell
balance such as (3.35) is generally the simpler way to determine how a mean velocity varies along a
conduit.
Of the three ways to express conservation of mass, only the continuity equation permits evaluation of
an unknown velocity component. If vz (r, z) is given as
2
r
vz (r, z) = 2u(z) 1 − (3.40)
R
r3
2vw (z) C
vr (r, z) = r− + (3.42)
R 2R2 r
Summary
42 CHAPTER 3. CONSERVATION OF MASS
질량보존의 법칙
!"
= 0, 물질입자로 이루어진 body내 질량은 시간에 대하여 변하지 않는다.
!#
!"
= #̇ $% − #̇ &'# , 고정된 공간인 control volume내 시간에 대한 질량 변화는
!#
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
연속방정식 및 비압축성 흐름
• 연속방정식은 한점에서 만족해야 할 질량보존법칙이다.
%&
+ ∇ * &+ = 0
%'
- 비압축성 유체는 압력에 따라 밀도가 변하지 않는 유체이다. 따라서 속도는
발산하지 않는다. ∇ * + = 0
- 압축성 유체 또한 많은 경우 속도가 발산하지 않는 흐름을 보인다. ∇ * + = 0.
따라서 흐름 동안 밀도가 일정하게 유지된다.
• 속도의 발산은 유체가 흘러갈 때 부피의 변화 크기를 나타낸다.
- ∇ * + > 0, 들어오는 속도에 비하여 나가는 속도가 커 부피가 커지는 흐름
- ∇ * + = 0, 들어오는 속도와 나가는 속도가 동일하여 부피증가가 없는
isochoric 흐름
43
44 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
• Friction force (Wikipedia) is a surface force resisting the relative motion of solid surfaces,
fluid layers, and material elements sliding against each other.2
– Fluid friction (viscous force): describes the friction between layers of a viscous fluid
that are moving relative to each other. Layers of a fluid can slide past each other, and
the viscous force resists the sliding. The internal resistance to flow is named viscosity.
– Skin friction: describes the surface force resisting the motion of a fluid across the
surface of a body. Fluids do not slide along surfaces, the so-called no slip condition.
This causes the skin friction.
Figure 4.1: 유체에 작용하는 힘은 체적에 작용하는 body force (Fb ), 표면에 작용하는 surface force
(Fs ), 계면위 임의의 line 또는 contour에 작용하는 surface tension force(Fγ )로 이루어짐. Surface force
는 수직방향의 pressure과 수평 및 수직 방향의 friction의 합으로 표현됨.
3. Interfacial (surface) tension force is a tensile tangential force acting on an imaginary line
or contour within a interface between two different phases.
Ex. 11. Determine all forces acting throughout the volume and on the surface of a stationary fluid
in the container shown in Fig. 4.2
FP=압력힘
Patm FP1
FP4 FP2
ρ(밀도) h Fb=중력
y
S(밑면적) x
FP3
Figure 4.2: Forces acting throughout the volume and on the surface of a stationary fluid
2 Besides the following two frictions, there are other types of friction. Dry friction is a surface force that opposes
the relative lateral motion of two solid surfaces in contact. Internal friction is the force resisting motion between the
elements making up a solid material while it undergoes deformation.
4.1. FORCE ACTING ON A FLUID 45
X
F = Fb + FP 1 + FP 2 + FP 3 + FP 4
(4.4)
= −ρShgey − Patm Sey − Pside hlex + (Patm + ρgh)Sey + Pside hlex = 0
P
즉, 외부에서 유체에 작용하는 모든 힘의 합은 0이다. (유체가 정지해 있기 때문에 a = 0 =⇒ F = 0)
Ex. 12. Using Fig. 4.3, show that the horizontal pressure force on the inclined surface is its projected
area times the pressure averaged over that projection.
Air
P = Patm
ρ FP1
h
x FP2
유체가 정지해 있으므로 수평힘 및 수직힘 각각의 합이 0이 되어야 한다. 따라서 기울어진 면에 작용하는
수평성분의 압력힘은 왼쪽 수직면에 작용하는 수평성분의 압력힘과 크기가 같아야 한다. 왼쪽면에 작용
하는 수평성분 압력힘 FP 1 은 평균압력 (Pside = Patm + ρgh/2)과 면적 hl의 곱이므로, 기울어진면에
작용하는 힘 FP 2 의 수평성분 또한
Ex. 13. Show that the friction force acting on the upper surface of a moving fluid is equal to that
on the lower surface of the fluid (Fig. 4.4). Here the fluid moves due to the constant velocity of the
upper plate v.
Fb : 중력 = 유체의 무게 = mg = ρAw hg
FP 1 : 공기가 왼쪽면의 유체에 가하는 압력에 의한 힘 = Patm A
FP 2 : 공기가 오른쪽면의 유체에 가하는 압력에 의한 힘 = Patm A
FP 3 : 아랫판이 아랫면의 유체에 가하는 압력에 의한 힘 = (Pm + ρgh)Aw
FP 4 : 윗판이 윗면의 유체에 가하는 압력에 의한 힘 = Pm Aw
FP 5 : 공기가 앞면의 유체에 가하는 압력에 의한 힘 = Patm As
FP 6 : 공기가 뒷면의 유체에 가하는 압력에 의한 힘 = Patm As
Fτ 1 : 움직이는 윗판이 유체에 가하는 마찰력(no slip) = τ1 Aw
* 윗면의 유체가 움직이는 윗판에 가하는 마찰력은 크기는 같고 방향은 반대방향임 = −Fτ 1
Fτ 2 : 아랫판이 아랫면의 유체에 가하는 마찰력(no slip) = τ2 Aw
*흐르는 유체가 아랫판에 가하는 마찰력은 크기는 같고 방향은 반대방향임 = −Fτ 2
X
F = Fb + FP 1 + FP 2 + FP 3 + FP 4 + FP 5 + FP 6 + Fτ 1 + Fτ 2
= −ρhgAw ey + Patm Aex − Patm Aex + (Pm + ρgh)Aw ey − Pm Aw ey
(4.6)
− Patm As ez + Patm As ez + τ1 Aw ex − τ2 Aw ex = 0
=⇒ τ1 = τ2
P
유체가 일정속도로 움직이므로 운동량의 변화가 없어 a = 0 =⇒ F = 0.
46 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
Aw(판면적) v
FP4=압력
x 압력=FP6
z Pm
Fτ1=마찰력
FP2=압력
압력=FP1 Fb=중력
압력=FP5 Fτ2=마찰력
FP3=압력
Figure 4.4: Forces acting throughout the volume and on the surface of a moving fluid between two
plates
Suppose an air-water interface shown in Fig. 4.5. Each molecule in water bulk is pulled equally
in every direction by neighboring water molecules, resulting in zero net force. On the other hand,
the interface molecules experience air-water adhesive forces with the above air molecules as well as
water-water cohesive forces with the below water molecules. Because the air-water adhesive force is
not as strong as the water-water cohesive force, the surface molecules feel a net force toward the
bulk of water. This inward force makes the interface to be tense, ultimately forcing liquid surface
to contract to the minimal area. In this viewpoint, the interfacial (or surface) tension3 (γ) is a
tensile force per unit length (N/m) that acts on an imaginary line or contour within a
interface. The direction of the surface tension force is tangent to the interface and away from the line
of contour. The forces (γL) are tensile, meaning that each pulls on the imaginary line. If γ is constant
and the interface is planar, the surface tension force will balance and the surface tension will not
be evident. However, at an interface with curvature, the surface tension generally creates a pressure
difference between the phases. For example, the pressure inside a small drop or bubble exceeds that
in the surrounding fluid. (See Ex. 14)
Two examples on the effects of surface tension are also shown in Fig. 4.5. The above shows water
adhering to the faucet gaining mass until it is stretched to a point where the surface tension can no
longer keep the drop linked to the faucet. It then separates and surface tension forms the drop into a
sphere. If a stream of water was running from the faucet, the stream would break up into drops during
its fall. Gravity stretches the stream, then surface tension pinches it into spheres. Water striders use
3 In general, the interfacial tension is called surface tension in the case that the interface is composed of gas-liquid
Air
Interface
γL
γL
Bulk
Liquid
Figure 4.5: Surface tension of the interface between air and water. A force of magnitude γL pulls away
from each side of the imaginary cut shown by the dashed line. Two effects of surface tension are also
shown; the formation of a drop from faucet and the flotation of water strider (Wikipedia).
surface tension to walk on the surface of a pond in the following way. The nonwettability of the water
strider’s leg means there is no attraction between molecules of the leg and molecules of the water,
so when the leg pushes down on the water, the surface tension of the water only tries to recover
its flatness from its deformation due to the leg. This behavior of the water pushes the water strider
upward so it can stand on the surface of the water as long as its mass is small enough that the water
can support it. The surface of the water behaves like an elastic film (Wikipedia).
Table 4.1 shows values of γ for several liquids in contact with air. The surface tension for water is
about three times that for benzene, ethanol, or other common organic solvents. Mercury is unusual in
that γ is about 20 times that for the organics. Surface tensions decrease with increasing temperature,
as emplified by the entries for water and mercury.
Besides, we can see the surface tension at work when we fill a glass. Even if the water is at the rim
of the glass, we can add just a few more drops so that the water is slightly taller than the rim. When
surfactants are added to water, the surfactants gather at the surface of water because they are lighter
than water. The intermolecular force between water molecules becomes weaker due to the existence
of surfactant molecules between them. The surfactants actually reduce the surface tension of water
by a factor of three or more. Thus it is possible to make thin soap bubble.
Sometimes, surface tension is viewed as energy per unit area (J/m2 ). A molecule in a bulk phase
is in a lower state of energy due to balancing with neighbor molecules, whereas the surface molecules
are missing neighbors and thus have higher energy. For the liquid to minimize its energy state, the
number of higher energy surface molecules should be minimized. Additionally, in order to have new
surface molecules by creating new interface, some of energy should be required. With this viewpoint,
the surface tension is called the interfacial energy (or surface energy only in the case of gas-liquid
or liquid-liquid interface), which is an energy per unit area (J/m2 ) required to increase the
interfacial area in thermodynamics. It can be thermodynamically expressed as γ = ∂G ∂A T,P,n .
More details on the surface tension is described with mathematical approach in ”Interfacial Trans-
48 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
Ex. 14. Young-Laplace equation. Determine the pressure difference across the surface of a sta-
tionary spherical bubble of radius R in terms of the surface tension and bubble size. It is assumed
that the internal and external pressures are nearly uniform. The limitation of that assumption will
be examined later.
Answer: A convenient control volume is a hemispherical shell of infinitesimal thickness that encloses
half of the gas-liquid interface, as illustrated in Fig. 4.6. With this choice, the surface tension force
that the surroundings exert on the control volume is downward in the diagram, which is 2πRγ. The
pressure force is simply calculated by the projected-area concept. The uniform pressures each act on
a projected area πR2 . Gravity need not be considered because the mass in the thin control volume is
negligible. The net force acting downward is then
Liquid
Pout
Pin R
γ Gas
Figure 4.6: Pressure and surface tension acting on a spherical bubble. The control volume is the
hemispherical shell indicated by the dashed curves.
The assumption of uniform pressures requires the bubble (or droplets) to be small. More precisely,
the pressure variation within each fluid must be negligible relative to ∆P . In the case of liquid, the
gravity-induced pressure variation over a height 2R of liquid is 2ρL gR. Accordingly, the Young-Laplace
equation requires that
Ex. 15. Capillary rise. Capillary action (sometimes, capillarity or capillary motion) is the ability
of a liquid to flow in narrow spaces without the assistance of, or even in opposition to, external
forces like gravity. A consequence of the pressure difference resulting from the surface tension is the
phenomenon of capillary action.4 A common apparatus used to demonstrate the first phenomenon is
the capillary tube. When the lower end of a vertical glass tube is placed in water, a concave meniscus
forms (capillary rise). On the other hand, a convex meniscus forms when it is placed in mercury
4 This effect is related to how well a liquid wets a solid boundary. The indicator for wetting or nonwetting is the
contact angle α. A nonwetting case is associated with α > 90◦ , such as mercury in contact with a clean glass tube,
α ≈ 130◦ . For a wetting case α < 90◦ .
4.2. INTERFACIAL TENSION FORCE (SURFACE TENSION, INTERFACIAL ENERGY) 49
(capillary depression) due to its strong intermolecular forces. This is caused by the difference between
the liquid cohesion and the liquid adhesion with the solid tube.
In the case of water, the adhesion between water molecules and the glass tube is stronger than the
cohesion between water molecules at the surface, resulting in the formation of thin film of water on
the tube wall. Then, the surface tension pulls water molecules at the center to minimize the interfacial
area. The tube suck up water to a certain height where the weight of the sucked water reaches to the
adhesion, as shown in Fig. 4.7.
γ α
CV1
Air 1 R γ
P = P0 CV2
2 α
Rm
H
α
α
Water
Figure 4.7: Rise of water inside a small, wettable glass tube. 1 and 2 indicates the tops of control
volumes CV1 and CV2, respectively.
Show how H depends on the tube radius (R), surface tension (γ), and contact angle (α).
Answer: The analysis can be done using force balances on either of two control volumes; CV1 and
CV2. The difference between them is that the top of CV1 is on the air side of the meniscus and that
of CV2 is on the liquid side.
• Method 1: With CV1 the control surface cuts through the air-water interface. The surroundings
therefore exert a surface tension force on the liquid, the upward component of which is 2πRγ cos α.
The gravitational force on the liquid of CV1 acts in the downward direction, which is ρgπR2 H
under the assumption that H ≫ R. There is no net pressure force, because the top and bottom
of CV1 are the same pressure (P0 ) and have the same projected arer (πR2 ). Requiring that the
upward forces in this static system sum to zero, we find that
2πRγ cos α − ρgπR2 H = 0 (4.11)
Solving for H gives
2γ cos α
H= (4.12)
ρgR
Thus, the smaller the tube radius and the smaller the contact angle, the higher the liquid will
rise.
• Method 2: With CV2 the control surface does not cut through the air-water interface, and
therefore there is no surface tension force from the surroundings. However, there is a net pressure
force, because surface tension within the curved surface lowers the pressure at the top of the
liquid column. If Bo ≪ 1, the gas-water interface will resemble part of a spherical bubble and
the Young-Laplace equation will apply. From Eq. (4.9), the pressure on the liquid side is then
P0 − 2γ/Rm . Accordingly, the vertical force balance for CV2 is
2 2γ cos α
P0 πR − P0 − πR2 − ρgπR2 H = 0 (4.13)
R
This will give the same result as Eq. (4.12).
It is equally valid to attribute capillary rise to (1) surface tension pulling upward on the liquid, or (2)
surface tension creating a partial vacuum at the top of the liquid column.
50 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
As shown in Fig. 4.8(a), the linear momentum represents the effect of mass and velocity simultaneously.
By the Newton’s second law of motion, the linear momentum should be conserved if the external forces
are not given. The linear momentum flow rate becomes
m
v
r mv
+
mv mv
r
where v⊥ is the velocity component perpendicular to surface A. When the velocity v⊥ is not constant
on the surface A, the linear momentum flow rate becomes
Z
Ṗ = ρv⊥ v dA (4.16)
Now consider an object rotating with an equal speed shown in Fig. 4.8(b). In this case, the linear
momentum mv is continuously changed due to the change of direction of motion. However the vector
quantity
H ≡ r × P = r × mv = Iω (4.17)
keeps constant during the rotational motion. H is called the angular momentum (moment of momen-
tum) which measures how much the linear momentum is rotating around a certain point called the
origin. Similar to the linear momentum expressed as linear inertia(m) and linear velocity(v), the angu-
lar(rotational or moment of) momentum can be written as rotational inertia times rotational(angular)
velocity since
I = m|r|2 (4.18)
and
r× v
v=ω × r → ω= (4.19)
|r|2
The angular momentum should be conserved if the external rotational forces(moment) are not
given.5
5 피겨스케이팅 선수가 빙판위에서 회전운동을 할때는 팔을 몸에 바짝 붙어서 회전을 한다. 회전할 때 외부에서 토크(돌림 힘)
가 주어지지 않기 때문에 각운동량 H = Iω은 보존된다. 따라서 rotational inertia(회전관성)을 I = mr2 을 줄여서 각속도가
크게 하여야 하므로 팔이 몸쪽에 붙도록, 즉 r 적어지도록 하여야 한다. 반면에 회전을 멈추고자 할때는 손을 펼쳐서 회전관성을
최대한 크게 해야 각속도가 줄어들게 된다.
4.4. LINEAR MOMENTUM BALANCE FOR A STEADY FLOW 51
It means that
X X X
(Ṗout,x − Ṗin,x )ex + (Ṗout,y − Ṗin,y )ey + (Ṗout,z − Ṗin,z )ez = Fx ex + Fy ey + Fz ez (4.21)
and thus
X X X
Ṗout,x − Ṗin,x = Fx , Ṗout,y − Ṗin,y = Fy , Ṗout,z − Ṗin,z = Fz (4.22)
Ex. 16. A jet of water exits a nozzle and strikes a vertical plane surface as shown in Fig. 4.9.
Determine the force required to hold the plate stationary.
A1 A2
Figure 4.9: A fluid jet striking a vertical plate. 유체로만 CV를 정하면 x 방향의 반대로 720 N 힘이 plate
오른쪽에 가해지고 이로 인하여 plate 왼쪽면이 유체에 720 N힘을 가해서 노즐에서 나온 물의 x 방향의
운동량이 720 N 만큼 감소하였고, y 방향으로 힘이 주어지지 않았기 때문에 전체 y 방향의 운동량은 0을
유지함.
where F is the external (surface) force acting on the control surface, which is given from outside.
Therefore
F = ρAj vj 2 = 1000 kg/m3 (0.005 m2 )(12 m/s)2 = 720 N (4.24)
The control volume shown in Fig. 4.9 for which the above solution was obtained represents only
possible choice. Another can be chosen, even though wrong choice of the control volume may not give
the solution.
Ex. 17. Water flows steadily through the 90◦ reducing elbow shown in Fig. 4.10. At the inlet to the
elbow, the absolute pressure is 220 kPa and the cross-sectional area is 0.01 m2 . At the outlet, the
cross-sectional area is 0.0025 m2 and the velocity is 16 m/s. The elbow discharge to the atmosphere.
Determine the force required to hold the elbow in place, Rx and Ry . Here the density of incompressible
water is 999 kg/m3 . Neglect weight of elbow and water in elbow.
52 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
As2
As1
y P1 1 W Rx
R1 Ry 2
R2
x
P2
Figure 4.10: Flow in a reducing elbow. Here we choose the dot line as a control volume. The second
figure shows all forces acting on the control volume. 파이프 내 유체만을 CV로 정하면 다음과 같은
해석이 가능함. 입구에서 유체의 x방향 운동량과 gauge 압력힘이 1.35 kN 이었는데 출구에서는 0이 되
었으므로 −x 방향으로 1.35 kN 힘이 파이프의 안쪽면에서 유체로 가해졌음. 이와 같은 힘은 외부에서
파이프의 바깥쪽면으로 동일한 힘(크기 및 방향)이 주어졌기 때문임. 입구에서 y방향 운동량 0 이었는데
출구에서는 -639 N이 되었으므로 −y 방향으로 639 N 힘이 가해졌음. 이 힘에는 중력도 포함되어 있음.
예를 들어 유체의 중력이 10 N 작용한다면 파이프 안쪽면에서 유체에 가해진 −y 방향 총힘은 629 N임.
where the gauge pressure at position 1 is P1g = P1 − Patm . The Rx and Ry are the external
(body) forces acting on the control volume from outside (more exactly, the force given to the
elbow in order to hold it in place).
The linear momentum balance in the y direction becomes
X
Ṗout,y = ρA2 v2 (−v2 ) = Fy = Ry + Patm As2 − Patm As2 − W
(4.27)
=⇒ Ry = −ρA2 v2 2 + W
Thus, 1.35 kN and 639 N are the forces given to the elbow in the negative x and y directions in order
to hold the elbow.
Ex. 18. In the steady-state flow system of incompressible water shown in Fig. 4.11, the total drag on
the object (흐르는 유체가 고체에 미치는 흐름방향 힘) is measured to be 100,000 N/m of length normal
to the direction of flow, and frictional force at the walls are neglected, find the pressure difference
between inlet and outlet sections
4.4. LINEAR MOMENTUM BALANCE FOR A STEADY FLOW 53
2m 4m
5 m/s
P3
τw
P1 τo P2
τw
P4
Figure 4.11: A two-dimensional object is placed in a 4-ft-wide water tunnel.
Answer: The conservation of mass of a fluid in the steady-state incompressible flow says that at
Z
ṁin = ṁout (Qin = Qout ) =⇒ v1 A1 = vout dA
Z 1 Z 2
=⇒ 4v1 = 2 v2 y dy + v2 dy (4.29)
0 1
4 20
=⇒ v2 = v1 = m/s
3 3
The linear momentum balance in the x direction is
X
Ṗout,x − Ṗin,x = Fx = P1 A − P2 A − τo Ao − 2τw Aw (4.30)
Ex. 19. (Control volume is moving with a constant velocity) Consider the steam locomotive tender
schematically illustrated in Fig. 4.12, which obtains water from a trough by means of a scoop. Find
the force on the train due to the water. Here the force due to pressure and friction are to be neglected.
Answer: The logical choice for a control volume in this case is the water-tank/scoop combination.
Our control volume will be selected as the interior of the tank and scoop. As the train is moving with
a uniform velocity, there are two possible choices of coordinate systems. We may select a coordinate
system either fixed in space or moving with the velocity of the train v0 . The control volume is inertial,
since it is not accelerating.
• Case 1: Stationary coordinate system but moving control volume (Fig. 4.13).
All the velocities must be measured relative to the stationary coordinate system. The x compo-
nent of the momentum balance is
∂Px X ∂
= Ṗin,x − Ṗout,x + Fx =⇒ (mvx ) = ṁin (0) + Fx (4.35)
∂t ∂t
Here Fx the total force exerted on the fluid by the train and scoop. The entering velocity of the
water is zero relative to the stationary coordinate. Also the velocity of the control volume in the
x direction vx is
vx = v0 (4.36)
∂m
= ṁin − ṁout = ρhv0 (4.37)
∂t | {z }
0
Finally, the force acting on the train by the fluid R can be also expressed as
Ex. 20. (Control volume is moving with an acceleration) A small rocket, with an initial mass of 400
kg, is to be launched vertically. Upon ignition the rocket consumes fuel at the rate of 5 kg/s and ejects
gas at atmospheric pressure with a speed of 3500 m/s relative to the rocket. Determine the initial
acceleration of the rocket and the rocket speed after 10 sec, if air resistance is neglected.
ve-v(t) ve
ve=3500 m/s vs rocket x
Stationary Moving
me=5 kg/s coordinate coordinate
Since the mass m and the velocity vy of the control volume are changed with time, it becomes
∂m ∂v
v +m = −ṁe (−ve + v) − mg (4.43)
∂t ∂t
The conservation of mass says
∂m
= ṁin −ṁout = −ṁe =⇒ m = M0 − ṁe t (4.44)
∂t |{z}
0
6 가속되는 좌표계 속에 놓여 있는 물체는 좌표계가 가속되는 방향과 반대 방향으로 힘을 받는 것처럼 느낀다. 이 힘을 관성력
이라고 한다. 가속도 a로 가속되는 좌표계 속에 있는 질량이 m인 물체는 f=ma 만큼의 힘을 가속도의 반대 방향으로 받는 것처럼
느낀다. [네이버 지식백과] 관성력 [慣性力] (Basic 고교생을 위한 물리 용어사전, 2002. 4. 15., 신근섭). 가속도가 중력과 반대
방향으로 작용하는 물체에 작용하는 body force는 중력방향으로 질량×(중력가속도+가속도)로 작용하며, 만약 가속도가 중력과
같은 방향이면 질량×(중력가속도-가속도)의 body force가 중력방향으로 작용한다. 예를 들어 윗 방향으로 2 m/s2 로 가속하고
있는 승강기에 타고 있는 100kg의 질량은 아랫방향으로 약 1200N 의 힘을 가한다. 아인슈타인의 일반 상대성이론에 따르면 가속에
이한 관성력과 질량에 의한 중력은 구분이 안됨.
4.5. STRESS : SURFACE FORCE PER UNIT AREA AT A POINT 57
ΔFs t
tp+tτn
tτt
ΔA
ΔA→0
Figure 4.16: Surface force on an element of fluid. The stress vector is a surface force per unit area
acting at a point on the surface.
The stress vector defined as (4.51) does not include the direction of the surface on which the
surface forces act. As depicted in Fig. 4.17, however, a force at a point exerted by an external surface
force depends on the direction of the surface as well as the direction of the external force. Thus it
is necessary to add information on the direction of the surface to the stress vector. This could be
achieved by introducing the stress tensor (or just stress) T (σ in our textbook p. 132) as
∆F
∆F ∆F
∆A ∆A ∆A
Figure 4.17: A force at a point exerted by an external force depends on both the direction of the
external force and the direction of the surface.
t = TT · n (4.52)
Here n is the unit normal vector to the surface, which gives the information on the direction of the
surface. The unit normal vectors are evaluated generally as
∇G
n= (4.53)
|∇G|
where G(x, y, z) = 0 defines the location of the surface.
In order to know physical meaning of stress T, we expand (4.52) in terms of components:
tx ex + ty ey + tz ez ≡ (Txx ex ex + Tyx ex ey + Tzx ex ez + Txy ey ex + Tyy ey ey + Tzy ey ez
(4.54)
+ Txz ez ex + Tyz ez ey + Tzz ez ez ) · (nx ex + ny ey + nz ez )
and thus
tx = Txx nx + Tyx ny + Tzx nz
ty = Txy nx + Tyy ny + Tzy nz (4.55)
tz = Txz nx + Tyz ny + Tzz nz
For reference, suppose that a stress vector acting on the surface shown in Fig. 4.18. Since n = ex ,
we have from (4.55)
Txx = tx > 0, and Txy = ty > 0 (4.56)
In other words, Txx is the magnitude of the surface force in the x direction per unit surface area of
which the unit normal vector is ex and Txy is the magnitude of the surface force in the y direction
per unit surface area of which the unit normal vector is ex .
More generally,
∆Fj i − the direction of n
Tij = lim ,
∆Ai →0 ∆Ai j − the direction of t (or ∆F)
(4.57)
Tij > 0 if both of n and t are either in positive or in negative directions.
Tij < 0 if either of n and t is in the positive direction.
58 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
n
y t
tyey
x txex
Figure 4.18: The stress vector acting on a point. (한점에 작용하고 있는 단위면적당 표면힘)
Figure 4.19 shows positive stress components acting at a point on the surfaces of a differential control
volume. When n and t are in the same direction regardless of their signs, it is call the normal stress
Tyy
y Tyz Tyx
Txy
Tzy
Txx
Tzx
Txz
Tzz
z
Figure 4.19: Positive stress components acting at a point.
4.5.2 Viscosity
As written in (4.60), the viscous portion of the stress has both normal and shear components of the
stress.
τxx τxy τxz
τ = τyx τyy τyz (4.63)
τzx τzy τzz
4.5. STRESS : SURFACE FORCE PER UNIT AREA AT A POINT 59
As previously mentioned in the surface force acting on fluid surface, when a shear force (viscous
force) is applied to the layers of a fluid, one of the fluid layer slides against the other layer and thus
forms a velocity gradient, more exactly shear rate (전단률)7 (Fig. 4.20). Similar to the voltage-current
system8 , the resistance of a fluid to the shear force (or sliding) is called viscosity defined as
For a given shear force, thicker viscosity(=larger resistance) leads to smaller velocity gradient.
F/A F/A
Δv
Fluid Δh
Tar and honey (approximately 10,000 cP) as highly viscous; air and water are low viscous. In
Table 4.2, the viscosities of water and air are shown with respect to temperature.
20 ◦ C 60 ◦ C 100 ◦ C
Water (cP = mPa·s) 1 0.46 0.28
Air (cP) 0.018 0.02 0.021
A simple theory for noninteracting gas molecules says that the viscosities of the gases are
(M T )1/2
µ = 0.00267 (4.65)
σ2
where µ is in cP, M is molecular weight, T is absolute temperature, and σ is the molecular diameter
in Å. The viscosities of liquids are much greater than those of gas at the same temperature. A good
approximation for the viscosities of liquid below the normal boiling point is
B
ln µ = A + (4.66)
T
As the temperature increases, the viscosity of gas increases due to increase in the molecular inter-
actions. However, the viscosity of liquid is highly dependent of cohesion between molecules and thus
decreases as the temperature increases. In Fig. 4.21, the viscosities of some gases and liquids are shown
with respect to temperature.
Sometimes, we use the other terminology of viscosity, the kinematic9 viscosity defined as
m2
µ
ν≡ (4.67)
ρ s
It is like the diffusion coefficient in mass transfer. As the diffusion coefficient increases, the amount of
mass transfer increases. The rate of momentum transfer can be determined by the kinematic viscosity
(ν), not viscosity (µ). More details are in Sect. 5.1.
7 Shear rate is defined as the flow velocity gradient in the direction perpendicular to the flow direction. γ̇ = ∇v+∇vT
forces.
60 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
Bingham
F/A plastic Real plastic
Shear Shear
stress thinning
τ Newtonian
F/A F/A
Yield μ Shear
Δv
stress thickening
Fluid Δh
Figure 4.22: Shear stress versus shear rate for newtonian and non-newtonian fluids.
For the non-Newtonian fluids, experimental data show nonlinear relationships between the shear
force and the velocity gradient, as shown in Fig. 4.22. In order words, their viscosities vary with
the shear force (or the velocity gradient because an increase in the shear force increases the velocity
4.5. STRESS : SURFACE FORCE PER UNIT AREA AT A POINT 61
∞
(
for τ < τ0
µ= τ0 (4.69)
µ0 + for τ ≥ τ0
dv/dh
A material that obeys Eq. (4.69) is called a Bingham fluid or Bingham plastic. The infinite viscosity
for τ < τ0 means simply that the material behaves as a rigid solid when subjected to insufficient
shear. If dv/dh greatly exceeds τ0 /µ0 , the fluid becomes Newtonian with µ = µ0 . House paint and
various foods behave much like this.
In summary, the shear stress for a fluid, which can be described by (4.68) and (4.69), can be
expressed as n−1
dv dv
τ (= F/A) = τ0 + K = τ0 + K γ̇ n (4.70)
dh dh
| {z }
µ
• Newtonian (τ0 = 0 and n = 1): Most of gases and liquids with molecular weights less than 103 ,
and suspensions of non-aggregating particles
• Shear-thinning (τ0 = 0 and n < 1): lava, blood, polymer melt (see Oobleck2008.wmv at Yotube)
• Shear-thickening (τ0 = 0 and n > 1): cornstarch + water mixture
• Bingham plastic (τ0 ̸= 0 and n ≤ 1): at low stress it is like a rigid body, at high stress it flows
like a viscous fluid. ex, slurry, toothpastes, ketchup, paint
Ex. 21. The functional relationship between stress and shear rate of a fluid is experimentally de-
termined as shown below. Plot shear stress vs. shear rate and finally determine a specific rheological
model for this fluid.
Answer: Bingham plastic, τ = 14.5 + 0.2γ̇
Shear stress (N/m2 ) Shear rate (1/s) Shear stress (N/m2 ) Shear rate (1/s)
14.54 0.2 16.52 10.1
14.56 0.3 18.52 20.1
14.6 0.5 22.52 40.1
14.66 0.8 26.52 60.1
14.72 1.1 30.52 80.1
14.92 2.1 46.52 160.1
15.32 4.1 66.52 260.1
15.52 5.1 86.52 360.1
62 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
Ex. 22. Some of non-Newtonian fluids do not follow the power-law given in (4.68). Other models
are available for a shear-thinning fluid such as polymer melt.
µ − µ∞ 1
Cross model : = , Sisko model : µ = µ∞ + K γ̇ n−1 (4.71)
µ0 − µ∞ 1 + (λγ̇)m
Viscosity of a polymer solution is measured over a wide range of shear rate as shown below. Plot shear
stress versus shear rate. Use Cross, Power law, and Sisko viscosity models to fit the viscosity-shear
rate relationship.
Viscosity (Pa·s) Shear rate (1/s) Viscosity (Pa·s) Shear rate (1/s)
0.0000012 60000 0.07 80
0.000004 70000 0.2 30
0.000008 30000 0.8 9
0.00002 28000 2.3 3.5
0.00012 11000 8 1
0.00031 7000 20 0.75
0.0008 3500 100 0.2
0.0013 1300 170 0.105
0.0036 900 800 0.11
0.008 350 1600 0.095
0.02 180 7000 0.105
65000 − 0.1
Cross : µ = 0.1 + , Power-law : µ = 8γ̇ −0.84 Sisko : µ = 0.1 + 8γ̇ −0.84 (4.72)
1 + (50000γ̇)0.84
τ = µ ∇v + ∇vT
(4.74)
4.6. EQUATION OF MOTION 63
It is clear that the viscous stress tensor is symmetric. The components of viscous stress for an incom-
pressible Newtonian fluid becomes
∂vx ∂vx ∂vy ∂vx ∂vz
2µ µ + µ +
∂x ∂y ∂x ∂z ∂x
τxx τxy τxz
τyx
∂vx ∂vy ∂vy ∂vy ∂vz
τyy τyz = µ
+ 2µ µ + (4.75)
τzx τzy τzz ∂y ∂x ∂y ∂z ∂y
∂vx ∂vz ∂vy ∂vz ∂vz
µ + µ + 2µ
∂z ∂x ∂z ∂y ∂z
(x, y, z)
The linear momentum balance(or the conservation of linear momentum) within the differential
64 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
∂P X
= Ṗin − Ṗout + F (4.79)
∂t
• Linear momentum in the differential element shown in Fig. 4.23 is
and the time rate of change of linear momentum in the differential element is
∂
(ρv∆x∆y∆z) (4.81)
∂t
and the linear momentum rate leaving the differential element through the surface at x + ∆x is
Thus the net linear momentum rate out of the differential element in the x direction is written
as
[(ρvvx )x+∆x − (ρvvx )x ] ∆y∆z (4.84)
Similarly we can get the net linear momentum rates in the y and z directions:
[(ρvvy )y+∆y − (ρvvy )y ] ∆x∆z, and [(ρvvz )z+∆z − (ρvvz )z ] ∆x∆y (4.85)
Finally, the net rate of linear momentum from the differential element becomes
[(ρvvx )x+∆x − (ρvvx )x ] ∆y∆z + [(ρvvy )y+∆y − (ρvvy )y ] ∆x∆z + [(ρvvz )z+∆z − (ρvvz )z ] ∆x∆y
(4.86)
ρb∆x∆y∆z (4.87)
where b is the body force per mass and thus ρb is the body force per unit volume. In case
there is the gravitational force only as a body force, b = g.
2. Surface forces acting on the surface of the differential element.
From Fig. 4.24, we can see that sum of the surface forces in the x direction is
(Txx |x+∆x − Txx |x )∆y∆z + (Tyx |y+∆y − Tyx |y )∆x∆z + (Tzx |z+∆z − Tzx |z )∆x∆y (4.88)
The surface forces in the y and z directions can be obtained in similar forms:
(Txy |x+∆x − Txy |x )∆y∆z + (Tyy |y+∆y − Tyy |y )∆x∆z + (Tzy |z+∆z − Tzy |z )∆x∆y
(4.89)
(Txz |x+∆x − Txz |x )∆y∆z + (Tyz |y+∆y − Tyz |y )∆x∆z + (Tzz |z+∆z − Tzz |z )∆x∆y
4.6. EQUATION OF MOTION 65
y Tyx y+∆y
Tzx z
∆y
Txx x
Txx x+∆x
Tzx z+∆z
Tyx y ∆z
∆x
z
Figure 4.24: Surface forces in the x direction acting on the surface of a differential element.
Using the linear momentum balance (4.78) and (4.81) through (4.89), we finally have
∂
(ρv∆x∆y∆z) + [(ρvvx )x+∆x − (ρvvx )x ] ∆y∆z
∂t X
+ [(ρvvy )y+∆y − (ρvvy )y ] ∆x∆z + [(ρvvz )z+∆z − (ρvvz )z ] ∆x∆y = F (4.90)
Since the volume does not change with time, dividing (4.90) by ∆x∆y∆z gives
This is called the differential linear momentum balance or the equation of motion that should be
applied at any differential elements, i.e., points.
Since
∂ ∂ρ ∂v ∂v
(ρv) + ∇ · (ρvv) = v +ρ + v∇ · (ρv) + ∇v · (ρv) = ρ + ∇v · v (4.93)
∂t ∂t ∂t ∂t
66 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
from the differential mass balance (3.24) and the symmetry of stres (4.58), the equation of motion
becomes
∂v
ρ + ∇v · v = ρb + ∇ · T (4.94)
∂t
It can be also expressed in terms of the material (or substantial) time derivative
Dv
ρ = ρb + ∇ · T (4.95)
Dt
from the definition of substantial time derivative (2.13).
From (4.60), the equation of motion (4.94) becomes
∂v
ρ + ∇v · v = −∇P + ∇ · τ + ρb (4.96)
∂t
because
∇ · (−P I) = −∇P · I − P (∇ · I) = −∇P (4.97)
which is known as the Euler equation for inviscid flows. This Euler equation has been found to be
useful for describing the flows of low-viscosity (µ ≈ 0) at Re≫ 1 (see Sect. 5.2.1). Of course, we
know that this equation will be inadequate in the neighborhood of solid surface, so called boundary-
layer (see Sect. 5.2.3). The subject of inviscid flow has particular application in aerodynamics, which
handles mainly the external flows having a very high Re number.
This is called the Navier-Stokes equations and is the differential form of Newton’s second law of motion
for an incompressible Newtonian fluid. In deriving (4.96), we used that
∇ · (∇v) = ∇2 v and
T
(4.100)
∇ · (∇v ) = ∇(∇ · v) = 0 ∵ ∇ · v = 0 for an incompressible flow
which is called the Stokes flow equation or the steady-state creeping flow equation.
4.7. MECHANICAL ENERGY BALANCE 67
Irrotational flow (∇ × v = 0 → ∇ · τ = 0)
Using vector calculation, we can show
1
∇v · v = ∇ (v · v) −[v × (∇ × v)] (4.102)
2 | {z }
v2
and
∇2 v = ∇(∇ · v) −∇ × (∇ × v) (4.103)
| {z }
0, incompressible
Thus, the viscous force acting on a fluid becomes
∇ · τ = −µ[∇ × (∇ × v)] (4.104)
Indeed, the curl of velocity is twice the vorticity vector, which is a measure of rotational motion (see
Appendix B). If a fluid flow is irrotational, the curl of velocity becomes zero:
∇× v=0 (4.105)
The physical meaning of curl of velocity is well explained in movie clips (https://www.khanacademy.org/
math/multivariable-calculus/multivariable-derivatives/curl-grant-videos/v/2d-curl-intuition?modal=1).
Therefore, the flow of a fluid is irrotational, the viscous effect is negligible: ∇ · τ = 0.
∇×v
Figure 4.25: Curl of fluid flows. The positive curl corresponds to counter-clockwise rotation, whereas
the negative curl corresponds to clockwise rotation. The zero curl indicates no rotation flow. (see
Khan Academy)
Overall, when the flow of a Newtonian fluid is incompressible and irrotational, the differential form
of linear momentum balance (4.99) reduces to
v2
∂v
ρ +∇ = −∇P + ρb (4.106)
∂t 2
This is further simplified for a steady-state flow to
2
ρv
∇ + P = ρb (4.107)
2
It is nothing but the differential form of linear momentum balance for the steady-state incompressible
and irrotational flow of a Newtonian fluid.
b = −∇Φ̂ (4.113)
Then
dΦ̂ ∂ Φ̂
v · Fb = v · m(−∇Φ̂) = −m(∇Φ̂ · v) = −m − (4.114)
dt ∂t
in view of (2.13). If the potential is independent of the time, ∂ Φ̂/∂t = 0 (it is true for the gravitational
field for systems; Φ̂ = gy), this becomes
dΦ̂ d
v · Fb = −m = − (mΦ̂) (4.115)
dt dt
from the conservation of mass dm/dt = 0. Finally (4.112) can be rewritten as
d 1 2 d
mv + mΦ̂ = (EK + EP ) = v · Fs , if b = −∇Φ̂ (4.116)
dt 2 dt
where EP = mΦ̂ is called the potential energy. In the gravitational field only,
b = g = −gey =⇒ Φ̂ = gy (4.117)
Let’s consider a falling object in a fluid. It is assumed that all the surface forces acting on the
object caused by the fluid are negligible, compared to the gravitational body force. Thus, (4.116)
becomes
d 1 2 d ∂
mv + mgh = (EK + EP ) = (EK + EP ) + v · ∇(EK + EP ) = 0 (4.118)
dt 2 dt ∂t
This is the conservation of mechanical energy for a falling object in an inviscid fluid.
10 a force with the property that the work done in moving a particle between two points is independent of the path
tional force is the only body force, then b = g and Φ̂ can be expressed as gy so that g = −∇Φ̂ = −gey .
4.8. EQUATION OF MECHANICAL ENERGY 69
h
y
Figure 4.26: Conservation of mechanical energy for a falling object. Here, the drag force (air resistance)
on the object is assumed to be negligible.
The directional derivative of f , the rate of change of f in the direction given by a unit vector u is
defined as
df
= u · ∇f (4.119)
dh
where h is a magnitude in the direction of u. Thus,
v d
· ∇(EK + EP ) = (EK + EP ) (4.120)
v dh
With neglecting the time derivatives, we get using the definition of the directional derivative
d 1
(EK + EP ) = 0 ⇐⇒ mv 2 + mgh = constant (4.121)
dh 2
where h is the magnitude along the direction of the velocity, the falling direction of a body:
With the assumptions that b = −∇Φ̂, neglecting the time derivatives (steady-state), and ρ constant
(incompressible), this becomes using (4.102)
ρv 2
v· ∇ + ∇P + ∇ρΦ̂ = v · [v × (∇ × v)] + v · (∇ · τ ) (4.123)
2
Using the directional derivative (4.119) and (4.104), we get
d 1 v v
P + ρv 2 + ρΦ̂ = · [v × (∇ × v)] − · {µ[∇ × (∇ × v)]} (4.124)
ds 2 v v
where s is the magnitude along the velocity direction of a particle, a streamline. It is called the
differential mechanical energy balance or the equation of mechanical energy that should be applied at
any differential elements, i.e., points on the same streamline. If a fluid flow is also irrotational and
thus no viscous effect occurs, we finally have
1
P + ρv 2 + ρΦ̂ = constant, for a steady-state incompressible irrotational flow (4.125)
2
which is called the (streamline) Bernoulli equation, a cornerstone of classical fluid dynamics, applied
at a point.12 For a steady-state incompressible irrotational(or inviscid) flow, the total
12 Daniel Bernoulli (1700-1782), Swiss mathematician, was one of the early researchers in fluid dynamics. He showed
that as the velocity of fluid increases, the pressure decreases, a statement known as the Bernoulli principle. He was a
friend of Leonhard Euler (1707-1783), who formulated the continuity and momentum equations for an inviscid fluid.
70 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
Figure 4.27: Streamline: an imaginary line in a fluid such that the tangent at any point indicates
the direction of the velocity of a particle of the fluid at that point. In a steady flow, streamlines
corresponds to path lines of particles. (유동에서 입자의 속도벡터에 접하는 선)
Take a cross product of a position vector, r, with the Newton’s 2nd laws, we have
X dv d dP d dr
r× F=r× m =r× (mv) = r × = (r × P) − × P
dt dt dt dt dt
X d dr
=⇒ r × F = (r × P) ∵ × P = v × mv = 0 (4.126)
dt dt
X dH
=⇒ M=
dt
where M is the moment (torque) and H is the angular momentum(moment of momentum), which
measures how much the linear momentum is rotating around a certain point called the origin. This is
called the conservation of angular momentum. Torque is usually used to describe a rotational force of
a shaft (the tendency of a force to rotate an object about an axis), for example a turning screw-driver
(a measure of the turning force on an object).
4.9. ANGULAR MOMENTUM BALANCE: R × 2ND LAW 71
Summary
72 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
운동량 (Momentum)
운동량: 움직이는 물체가 지니고 있는 power.
m
v
mv r´mv
mv
r
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
운동량보존 법칙
!"
= Σ#, 물질입자로 이루어진 body내 선형운동량의 시간변화율은 body에
!#
작용하는 힘들의 합과 같다.
!"
̇ − #&'#
= #$% ̇ + Σ&, 고정된 공간인 control volume내 시간에 대한 선형운동량
!#
!$
= Σ$, 물질입자로 이루어진 body내 각운동량의 시간변화율은 body에
!#
작용하는 모멘트의 합과 같다. (' = ( × *, + = ( × &)
!(
= ,̇ $% − ,̇ &'# + Σ+, 고정된 공간인 control volume내 시간에 대한 각운동량
!#
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
4.9. ANGULAR MOMENTUM BALANCE: R × 2ND LAW 73
운동량 보존법칙 예
!"
!#
= #̇ $% − #̇ &'# + Σ&
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
힘 (Force)
Interfacial
Body force Surface force tension force
계면에 작용
물체의 Body(체적)에 작용한 힘 물체의 표면에 작용하는 힘
하는 힘
압력 + 마찰력(점성력)
중력, 정전기력 = 응력(stress)
점성력 : 유체 층들 사이 미끄러짐
에 대항하는 힘 표면장력 (g)
m q1
압력 (P)
중력
정전기력 점성력 (t)
M q2
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
74 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
W
h(높이) P1 P2
r(밀도)
L(너비) t2
SF= 중력 + 압력 + 마찰력
P4
W = 유체의 무게에 의한 중력 =mg=rShg
FP1 = 좌측면과 접하고 있는 유체가 좌측면의 유체에 가한 압력힘 = P1 hL
FP2 = 우측면과 접하고 있는 유체가 우측면의 유체에 가한 압력힘 = P2 hL
FP3 = 용기의 윗면이 윗면의 유체에 가한 압력힘 = P3 S
FP4 = 용기의 아랫면이 아랫면의 유체에 가한 압력힘 = P4 S
Ft1 = 용기의 윗면이 윗면의 유체에 가한 마찰힘 = t1 S
Ft2 = 용기의 아랫면이 아랫면의 유체에 가한 마찰힘 = t2 S
표면장력
• 계면분자들은 벌크분자들과 달리 분자간 인력이 균형을 이
Air
루지 않아 순힘이 작용하며, 이를 계면장력힘이라 함.
Interface
응력 (Stress)
• 한점에 작용하는 단위면적당 표면의 힘 (압력+점성력)
• 응력벡터는 t로 표시하며 힘의 방향만 있음.
t = T・n
• 응력텐서 (일반적으로 응력이라 함)는 T로 표시하고
힘 및 단면의 방향 정보가 있음.
Tij i: n의 방향, j: 표면힘(t) 방향
Tii: 수직응력, Tij (i ≠ j) 전단응력
• Txy : x축에 수직인 면에 작용하는 y방향의
단위면적당 표면힘.
• 응력은 대칭임 (Tij=Tji)
• T = -P I + t (응력 = 압력 +점성력)
• 압력은 수직응력만 존재하고 점성응력은 수직응력과 전단응력 모두 존재함.
따라서 전단응력은 점성력에 의해서만 존재함.
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
점도(Viscosity)
유체의 한 층에 shear force(전단응력)를 가하면 유체의 층이 다른 층에서 미끄러지
며, 속도구배를 형성한다. 유체의 점도란 이와 같은 shear force (or sliding)에 대항
하는 저항을 일컫는다. 이는 도선에 전위차를 가하여 전류가 흐르게 하는 것과 같은
원리로서, 전위차에 대한 저항을 도선의 저항이라 칭함 (; = ∆< ⁄=).
F/A F/A
Δv
Fluid Δh
shear force 4 ⁄5 g
점도 ! = = [ ] = Poise
velocity gradient ∆7⁄∆ℎ cm : s
점도(!)
동점도 > =
!
[! Bs]
밀도(A)
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
76 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
점도와 온도
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
점도에 따른 뉴튼 및 비뉴튼 유체
tic
두 평판 사이에 유체를 두고 힘을 가하 Bingham
p las
F/A plastic al
Re Shear thinning
면서 속도기울기를 측정함.
Newtonian
F
d" Shear
$
d# thickening
y 항복
! d& 응력
유체 = $
" d'
x
d"
Ø Newtonian – 점도가 주어진 전단응력에 따라 변하지 않는 일정한 유체
d#
(공기와 물과 같은 대부분의 기체 및 액체)
Ø Bingham plastic - 일정한 힘(항복응력) 이상을 가하지 않으면 속도의 기울기가 생기지
않는 유체 (치약, 케찹, 등)
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
4.9. ANGULAR MOMENTUM BALANCE: R × 2ND LAW 77
점성 응력(Viscous stress)
F
y ! d(
& = $!" = %!" = & ∵ T = -P I + t
" d)
x
! = #(∇& + ∇&! )
운동방정식 및 비회전성 흐름
• 운동방정식은 한점에서 만족해야 할 운동량보존법칙(뉴튼 2법칙)이다.
• 뉴튼유체가 비압축성흐름을 보이면 운동방정식은 Navier-Stokes 식이 됨.
*+
) + + - .+ = −∇0 + #∇" + + )1
*,
- 어느 한점에서 시간에 대한 운동량의 변화량 (관성력)은 그 점에 작용하는
압력 + 점성력 + 중력의 합과 같다.
- Creeping flow : 속도가 매우 천천히 변하여 관성력을 무시할 수 있는 흐름
!"
(# !# =0), 이와 같은 흐름을 Stokes flow라고도 함.
- Irrotational flow: 유체입자들이 회전없이(속도의 curl이 0임) 흘러 점성력이
0이 되는 흐름
- Potential flow: 속도가 potential 함수 (+ = −∇2)로서 표현되어 비회전성 흐
름을 형성, 결국 점성력이 0 (마찰이 생기지 않는 흐름)이 됨.
- Inviscid flow: 모든 유체는 점도를 있지만 점도가 거의 없는 것처럼 흐름.
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
78 CHAPTER 4. CONSERVATION OF LINEAR MOMENTUM AND ITS DERIVATIVES
역학적에너지 방정식(베르누이 식)
• 베르누이식(equation of mechanical energy)은 속도와 운동방정식 (한점에 작
용하는 운동량보존법칙)의 dot product로 부터 유도됨.
• 베르누이식은 유체가 정상상태 비압축성 및 비점성 흐름을 보일 때 한 유선
상의 어느 점들에서도 총에너지 (압력일 + 운동에너지 + 위치포텐셜에너지)
는 항상 일정함.
#! $!" ## $#"
!! + + #! &ℎ! = !# + + ## &ℎ#
2 2
b
a
a
b
Conservation dm dP dH
body =0 ma = å F Û = åF r ´ ma = r ´ å F Û = åM
dt dt dt
F M
¶m ¶P ¶H
min mout Pin Hin
¶t ¶t Pout ¶t Hout
¶m ¶P ¶H
control volume + mout - min = 0 + Pout - Pin = å F + Hout - Hin = å M
¶t ¶t ¶t
Integral m = òòò r dV P = mv = òòò r v dV H = r ´ P = òòò r ´ r v dV
Balance CV CV CV
å F = Fb + Fs
= òòò r b dV + òò t dA å M = r ´ å F = r ´ (F + F ) b s
= òòò r ´ r b dV + òò r ´ t dA
CV CS
CV CS
t º T×n
Differential dr dv d (r ´ v)
Balance + rÑ × v = 0 r = Ñ × T + rb r = r ´ (Ñ × T) + r ´ r b
dt dt dt
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
Chapter 5
In what follows, we will mostly use the differential mass (equation of continuity) and momentum
balances (equation of motion) to analyze fluid flows. These will give some information on every point
in a control volume.
The differential mass balance is
∂ρ Dρ
+ ∇ · (ρv) = 0, or + ρ∇ · v = 0 (5.1)
∂t Dt
and for incompressible fluids or flows it reduces to
∇·v =0 (5.2)
For the incompressible flow of a Newtonian fluid, it becomes the Navier-Stokes equation:
∂v
ρ + ∇v · v = −∇P + µ∇2 v + ρb (5.4)
∂t
Momentum transfer τ1
in the vertical direction
Δy
by viscosity Δx
τ2
fluid in motion
Momentum(=mv) transfer in the
at t=0
horizontal direction by motion v
Figure 5.1: Momentum transfer of a moving flud in the left corner of the bottom.
momentum to the neighborhood of the fluid in both horizontal and vertical directions. The momentum
transfer in the horizontal direction is dependent upon the velocity of the fluid, whereas the momentum
79
80 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
transfer in the vertical direction is due to the viscosity of the fluid, represented by the cohesion force.
In other words, fluid layers can slide each other and the viscosity is the resistance for the sliding,
leading to momentum transfer from one to the other layer.
The conservation of linear momentum over a fluid layer in Fig. 5.1 is
X d ∂
F= (mv) = (mv) + (ṁv)out − (ṁv)in (5.5)
dt ∂t | {z }
=0
∂vx µ ∂ 2 vx
= (5.7)
∂t ρ ∂y 2
Therefore the time rate of change of vx at a point is dependent of µ/ρ, not µ only. That is, the rate
of momentum transfer in the vertical direction is dependent upon kinematic viscosity ν = µ/ρ. The
viscosities of water and air are 1 cP and 0.018 cP, respectively, whereas the kinematic viscosities of
water and air are 1 × 10−6 m2 /s and 15 × 10−6 m2 /s. This indicates that the momentum transfer of
air in the vertical direction is much faster than that of water.
ρDv
Re ≡ (5.8)
µ
It is a criterion for laminar and turbulent flows. The critical Re in a circular pipe is 2100. As the Re
increases, the flow transits from laminar to turbulent. When the Re is much larger than this value,
the flow becomes really turbulent.
Re < 2100
Re > 2100
Fluid
Inertia force v→
( ρv2 L2 )
Viscous force
(μvL)
Figure 5.3: Physical meaning of Reynolds number.
따라서 고체면에 가까운 얇은 유체층에는 점성력이 여전히 중요하다. 이러한 얇은 유체층을 경계층(boundary layer) 라고 부른다.
2 (4 February 1875 – 15 August 1953) was a German fluid dynamicist, physicist and aerospace scientist.
3 was a Hungarian-American mathematician, aerospace engineer, and physicist who was active primarily in the
fields of aeronautics and astronautics. He joined Ludwig Prandtl at the University of Göttingen, where he received his
doctorate in 1908.
82 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
of mass and momentum. Also, the quadratic velocity profile in the laminar boundary area was used
and determined by the boundary conditions, which are no-slip at the surface, and v = v∞ at the
boundary, and zero velocity derivative at the boundary. From this, he derived the boundary thickness
is expressed as
δ 5.5 ρv∞ x
= 1/2
where Rex = in a laminar flow (5.10)
x Rex µ
Ex. 23. Calculate the thickness of boundary layers for air and water when the fluids flow with
v∞ = 0.5 m/s over a flat pate L = 1m. ρair = 1.20 kg/m3 and µ = 1.80 × 10−5 Pa·s and ρwater = 998
kg/m3 and µ = 1.0 × 10−3 Pa·s.
Figure 5.5: Development of boundary layer flow in pipe: Fully developed flow
Experimental results and numerical simulations for laminar flow of Newtonian fluids indicate that
LE
= 0.49 + 0.11Re (5.13)
R
where LE is the distance at which the centerline velocity reaches 99% of its final value. Thus, the
entrance region will be as long as 230 radii if laminar flow is maintained up to Re = 2100.
P=P0 P=PL
r R vz
z
∆r ∆z
· r-component
∂P 1 ∂ 1 ∂τθr ∂τzr τθθ
0=− + (rτrr ) + + − + ρbr =⇒ P = ρgr sin θ + f (θ, z)
∂r r ∂r r ∂θ ∂z r |{z}
| {z } ρg sin θ
0, τrr = τθr = τθθ = 0, τzr ̸= f (z)
84 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
· θ-component
1 ∂P 1 ∂ 1 ∂τθθ ∂τzθ
0=− + 2 r2 τrθ + + + ρbθ =⇒ P = ρgr sin θ + g(r, z)
r ∂θ r ∂r r ∂θ ∂z |{z}
| {z } ρg cos θ
0, τrθ = τθθ = τzθ = 0
· z-component4
∂P 1 ∂ 1 ∂τθz ∂τzz 1 ∂ ∂P
0=− + (rτrz ) + + + ρbz =⇒ (rτrz ) = (5.16)
∂z r ∂r r ∂θ ∂z |{z} r ∂r ∂z
0
| {z }
0, τθz =τzz =0
1 ∂ ∂P ∆P r c1
(rτrz ) = = constant =⇒ τrz = − + (5.17)
r ∂r ∂z L 2 r
At r = 0 the shear stress has to be finite. This gives c1 = 0 and thus
∆P r
τrz = Trz = − (5.18)
L 2
It is VERY important that the shear stress in a flow by the pressure difference is linearly
distributed from the wall to the center. The viscous force acting on the fluid at the
wall by the cylindrical pipe becomes
∆P R
τzr |r=R = − <0 (5.19)
L 2
It indicates that the shear force acting on the fluid at the wall by the solid cyclinder is in
the −z direction because the positive direction of τzr |r=R is +z.
3. Calculate vz , vmax , ⟨v⟩, Q(= V̇ ), when the fluid is a Newtonian fluid and forms an incompressible
flow.
The shear stress τrz is expressed as
∂vr ∂vz
τrz =µ + (5.20)
∂z
|{z} ∂r
0
4 Without knowing the differential form of linear momentum balance, this equation can be derived. Because the mass
∂vz ∆P r ∆P r2
µ =− =⇒ vz = − + c2 (5.21)
∂r L 2 L 4µ
∆P R2
r 2
vz = 1− (5.22)
L 4µ R
It is VERY important that the velocity in a flow by the pressure difference is expressed as a
parabolic function. It should be noticed that the velocity can be directly derived from the Naiver-
Stoke’s equation, the linear momentum balance for an incompressible flow of a Newtonian fluid.
Using the velocity, we can also obtain the shear stress Trz .
The maximum velocity occurs at the center of the pipe:
R2 ∆P
vmax = vz |r=0 = (5.23)
4µ L
From this we can get the pressure drop per unit length as
∆P 8µ⟨v⟩ 32µ⟨v⟩
= 2
= (5.25)
L R D2
It is known as the Hagen-Poiseuille equation, which is applied to the steady-state fully-
developed incompressible flow of a Newtonian fluid.
The volumetric flow rate is
πR4 ∆P
Z
V̇ = ⟨v⟩A = vz dA = (5.26)
A 8µ L
τrz = µ dv st
dr = 1 -order
z
n µ dvz
dr (r
1/n
) vz (r1+1/n )
Newtonian 1 0 (constant) 1 2
Shear thinning <1 -0.a 1+0.a >2
Shear thickening >1 +0.b 1-0.b <2
Shear thickenning
Trz vz
Figure 5.7: Velocity profiles for Newtonian and Non-Newtonian fluids. Newtonian 유체의 경우 점도가
일정하므로 속도기울기가 r의 1차 함수이다. 따라서 속도는 r에 대하여 2차 함수로 나타내어 진다. 반면
에 shear thinning 유체의 경우 속도의 기울기가 1차보다 높은 차수로 증가하므로 속도는 2차 보다 높은
차수로 증가한다. 따라서 중심부분이 2차 곡선인 Newtonian 경우 보다 뭉뚝해 진다. Shear thickening의
경우에는 2차보다 뽀쪽해 진다.
5. Estimate the average velocity on cross section if the Newtonian fluid flows in a turbulence range,
instead of a Laminar flow.
As show in Fig 5.8, turbulence causes fluid elements at different radial positions to intermingle.
In other words, in turbulent flow the local velocity fluctuates from moment to moment and thus
the profile becomes more blunted. For a turbulent flow in a pipe at the relatively small Re,
experimental results are well represented by a power-law relationship
1/7
v̄z 60 r
= 1− (5.30)
⟨v⟩ 49 R
Given the relatively uniform velocity in the central part of the tube, there must be a larger
gradient near the wall, if the mean velocity is the same as for laminar flow. This leads to larger
shear stress at the wall in the turbulent flow (τrz = µdvz /dr in (5.20)).
(a) (b)
Turbulent
vz
uz
vz
Laminar Turbulent
t0 time
(c)
vavg vavg
vz vz
Laminar Turbulent
Figure 5.8: (a) Fully developed state state flow on average. The two curves in turbulent flow correspond
to fluid elements starting at the same position at different time. (b) Velocity fluctuations at a fixed
position in a turbulent flow. The instant velocity is vx (solid curve), its-smoothed value is v¯x (dashed
curve), and the fluctuation is uz . (c) Velocity profiles for laminar or turbulent flow in a tube. The
mean velocity in each case is indicated by the thick red dot lines.
h
vx
Tyx =line
ar
>0
· y-component
∂P ∂τxy ∂τyy ∂τzy
0=− + + + + ρby =⇒ P = −ρg cos θy + g(z)
∂y ∂x ∂y ∂z |{z}
| {z } −ρg cos θ
0, τyy = τzy = 0, τxy ̸= f (x)
· z-component
∂P ∂τxz ∂τyz ∂τzz
0=− + + + + ρbz =⇒ P = f (y) (5.33)
∂z ∂x ∂y ∂z |{z}
| {z } 0
0, τxz = τyz = τzz = 0
where P0 is a constant. Since the pressure at the free surface (y = h) is the atmospheric
pressure, we have
P = Patm + ρg(h − y) cos θ (5.35)
· x-component
∂P ∂τxx ∂τyx ∂τzx ∂τyx
0= − + + + + ρbx =⇒ = −ρg sin θ
∂x}
| {z ∂x
| {z } ∂y | ∂z
{z } |{z} ∂y
ρg sin θ
0, P ̸=f (x) 0, τxx = 0 0, τzx = 0
y2
∂vx y ρgh sin θ
µ = ρgh sin θ 1 − =⇒ vx = y− + c2 (5.38)
∂y h µ 2h
By the no-slip boundary condition, the velocity at the plane surface is zero so that
ρgh2 sin θ y
1 y 2
vx = − (5.39)
µ h 2 h
It is VERY important that the velocity in a flow by the gravity is expressed as a parabolic
function.
The maximum velocity occurs at the top surface:
ρgh2 sin θ
vmax = vx |y=h = (5.40)
2µ
5.5. FLUID FLOW DUE TO MOVING PLATES (COUETTE FLOW) 89
v0
y vx
h Tyx=const.
x >0
Figure 5.10: Laminar flow between two plates with the one stationary and the other moving.
· y-component
∂P ∂τxy ∂τyy ∂τzy
0=− + + + + ρby =⇒ P = −ρgy + g(z)
∂y ∂x ∂y ∂z |{z}
| {z } −ρg
0, τyy = τzy = 0, τxy ̸= f (x)
· z-component
∂P ∂τxz ∂τyz ∂τzz
0=− + + + + ρbz =⇒ P = f (y) (5.43)
∂z ∂x ∂y ∂z |{z}
| {z } 0
0, τxz = τyz = τzz = 0
P = −ρgy + P0
· x-component
∂P ∂τxx ∂τyx ∂τzx ∂τyx
0= − + + + + ρbx =⇒ =0
∂x} | ∂x ∂y | ∂z |{z} ∂y
0
| {z {z } {z }
0, P ̸=f (x) 0, τxx = 0 0, τzx = 0
Finally we have
τyx = Tyx = c1 (5.44)
90 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
It is VERY important that the shear stress in a Couette flow is constant, and that the flow
becomes inviscid
∇·τ =0 (5.45)
though the fluid is not inviscid (µ ̸= 0).
3. Calculate vx if the fluid is a Newtonian fluid and forms an incompressible flow.
Using the relation the shear stress and the velocity and (5.44), we have
∂vx ∂vy ∂vx c1
τyx = µ + =⇒ µ = c1 =⇒ vx = y + c2 (5.46)
∂y ∂x
|{z} ∂y µ
0
at y = 0, vx = 0, and at y = h, vx = v0 (5.47)
More cases about the flow of an incompressible Newtonian fluid on plates are shown in Fig. 5.12.
Ex. 24. As shown in Fig. 5.11, an incompressible flow of a Newtonian fluid is formed between two
vertical wide plates. One of the plates are stationary and the other is moving upward with the velocity
of v0 . Calculate the velocity of the moving plate v0 for the total flow rate to be zero.
v0 v0
y
h
τ g
+ =
y Air
h + = vx = U
vx = U
x U U
vx = U(1-y/h)
vr = vz = 0, vθ ̸= vθ (θ), P ̸= P (θ)
It is clear that this velocity satisfies the differential mass balance. The differential form of linear
momentum balance is
∂v
ρ +∇v · v = −∇P + ∇ · τ + ρb (5.54)
∂t
|{z}
0, st-st
· θ-component
1 ∂ c1
r2 τrθ =⇒ r2 τrθ = c1 =⇒ τrθ = 2
0= 2
r ∂r r
Since the fluid is a Newtonian one, we have
vθ 1 ∂vr ∂vθ ∂ vθ
τrθ = µ − + + =µ r (5.55)
r |r {z
∂θ} ∂r ∂r r
0,vr =0
Therefore
∂ vθ c1 ∂ vθ c1 c1 1
µ r = 2 =⇒ = 3 =⇒ vθ = − + c2 r
∂r r r ∂r r µr 2µ r
we know that at infinity there is no fluid moving (no momentum transfer by viscous effect)
at r = ∞, vθ = 0 =⇒ c2 = 0
Thus
ωR2
vθ = (5.56)
r
· r-component
vθ 2 ρω 2 R4
∂P ∂P
ρ − =− =⇒ =
r ∂r ∂r r3
5.7. FLUID FLOW BY A ROTATING SHAFT 93
· z-component
∂P ∂P
0=− + ρbz =⇒ = −ρg (5.57)
∂z |{z} ∂z
−ρg
ρω 2 R4
∇P = er − ρgez (5.58)
r3
Ex. 25. A fluid of constant density and viscosity is in a cylindrical container of radius R, as shown in
Fig. 5.14. The container is caused to rotate about its own axis at an angular velocity Ω. The cylinder
axis is vertical. Find the shape of the free surface when steady state has been established.
vr = vz = 0, vθ = vθ (r) (5.59)
We also know that pressure will depend upon r because of the centrifugal force and upon z because
of gravitational force.
Because the fluid is a Newtonian fluid and forms an incompressible flow, the velocity becomes
c1 1
vθ = − + c2 r
2µ r
we know that at r = 0, vθ should be finite and thus c1 = 0.
at r = R, vθ = RΩ =⇒ c2 = Ω
The r-component
vθ 2
∂P ∂P
ρ − =− =⇒ = ρΩ2 r
r ∂r ∂r
The z-component
∂P ∂P
0=− + ρbz =⇒ = −ρg (5.60)
∂z |{z} ∂z
−ρg
Thus
1
P = −ρgz + ρΩ2 r2 + c (5.61)
2
94 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
t<0 ∞ ∞ ∞
y
Fluid
at rest
x v0 v0 F/A
at t=0 at t=t1 at t=∞
Figure 5.15: Viscous flow of a fluid near a wall suddenly set in motion
vy = vz = 0, vx = vx (y, t) (5.64)
It is clear that this velocity satisfies the differential mass balance. The components of the differential
linear momentum balance are
· x-component
∂vx ∂τyx ∂vx ∂ 2 vx
ρ = =⇒ =ν (5.65)
∂t ∂y ∂t ∂y 2
It says that the time rate of change in vx is dependent upon the kinematic viscosity ν = µ/ρ
· y-component
∂P
0=− + ρby
∂y |{z}
−ρg
· z-component
0=0 (5.66)
The initial and boundary conditions are summarized as
at t ≤ 0, vx = 0 for all y
at y = 0, vx = v0 for all t > 0 (5.67)
at y → ∞, vx = 0 for all t > 0
By solving these, we can obtain the unsteady-state velocity as well as the shear stress.
Let’s introduce a dimensionless velocity ψ = vx /v0 . Then (5.65) and (5.67) become
∂ψ ∂2ψ
=ν 2 (5.68)
∂t ∂y
5.8. TIME-DEPENDENT FLOW 95
and
Now we use the method of combination of variables (change of variables). This is useful only for
semi-infinite regions, such that the initial condition and the boundary condition at infinity may be
combined into a single new boundary condition. To solve this, we let
y
η≡√ =⇒ ψ(t, y) → ψ(η) (5.70)
4νt
then we have
∂ψ ∂ψ ∂η 1 η dψ
= =−
∂t ∂η ∂t 2 t dη
(5.71)
∂ψ ∂ψ ∂η 1 dψ ∂2ψ 1 d2 ψ
= =√ and 2 = 4νt
∂y ∂η ∂y 4νt dη ∂y dη 2
d2 ψ dψ
2 + 2η dη = 0, =⇒ ψ(η) = c1 + c2 erf(η) (5.72)
dη
and
at η = 0, ψ = 1
(5.73)
at η → ∞, ψ = 0
Its solution is
ψ(η) = 1 − erf(η) (5.75)
that is
y
vx = v0 1 − erf √ (5.76)
4νt
1.0
0.6
y t increases
erf(x)
0.4
0.2
v0
1 2 3 4 5
x
Figure 5.16: Error function and the velocity profiles as time increases
96 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
0 = −∇P + ∇ · τ + ρb
(5.77)
= −∇P + µ∇2 v + ρb if an incompressible flow of a Newtonian fluid
which is called the Stokes flow equation. It is sometimes called the steady creeping flow equation,
because the term ρ(v · ∇v), which is quadratic in the velocity, can be discarded when the flow is
extremely slow. The Stokes flow equation is very important in lubrication theory.
As shown in Fig. 5.17, a solid sphere of radius R is rotating slowly at a constant angular velocity
Ω in a large body of quiescent Newtonian fluid. Let’s determine P , vϕ , and Trϕ . It is assumed that
the sphere rotates sufficiently slowly so that it is appropriate to use the creeping flow equation. From
It is clear that this velocity satisfies the differential mass balance. The Stokes equation becomes
· r-component
∂P
0=− + ρbr (5.79)
∂r |{z}
−ρg cos θ
· θ-component
1 ∂P
0=− + ρbθ (5.80)
r ∂θ |{z}
ρg sin θ
where b = −gez = −g [cos θer + (− sin θ)eθ ] from Fig. 1.6(c). Thus we can see
P = P0 − ρgr cos θ
· ϕ-component
1 ∂ f (θ)
r3 τrϕ =⇒ r3 τrϕ = f (θ) =⇒ τrϕ = 3
0= 3
(5.81)
r ∂r r
The bounday condition for τrϕ to determine f (θ) is
at r → ∞, τrϕ = 0 (5.82)
Thus, we use
1 ∂vr ∂vϕ vϕ
τrϕ = µ + − (5.83)
r sin θ ∂ϕ ∂r r
Applying the boundary conditions (5.85) shows that g(θ) = 0 and f (θ) = −3µΩR3 sin θ. Finally
we have
2
R
vϕ = ΩR sin θ < 0 (5.86)
r
The shear stress acting on the fluid at the wall by the solid Trϕ = τrϕ can be obtained using (5.83)
3
R
τrϕ = −3µΩ sin θ
r (5.87)
=⇒ τrϕ |r=R = −3µΩ sin θ
Thus the viscous force at the wall is in the −ϕ direction due to +r normal direction.
y hL
h0 h(x)
x
Figure 5.18: Flow in a tapered channel with planar walls at small-to-moderate Re.
∂P ∂ 2 vx
0≈− +µ 2
∂x ∂y
(5.90)
∂P
0≈− + ρby
∂y
Later, we will confirm the validity of the approximation.
From the second equation, we have
∂P
P (x, y) = −ρgy + f (x) ⇒ = f ′ (x) (5.91)
∂x
and thus we have from the first equation
1 ∂P
vx = y 2 + a(x)y + b(x) (5.92)
2µ ∂x
Under the no-slip condition, vx (x, ±h) = 0, the velocity finally is expressed as
" 2 #
h2 (x)
∂P y
vx (x, y) = − 1− (5.93)
2µ ∂x h(x)
and thus
∂P 3 µQ
=− 3 (5.95)
∂x 2 h (x)
and " 2 #
3Q y
vx (x, y) = 1− (5.96)
4h(x) h(x)
Equations (5.88), (5.91) and (5.95) lead to
Z
3 µQ 3 µQL 1
f (x) = − dx = +c (5.97)
2 h3 (x) 2 hL − h0 2h2 (x)
5.10. TWO DIMENSIONAL FLOWS 99
Ex. 26. Verification of the Lubrication Approximation. Although the above solutions (5.96),
(5.98), and (5.101) are plausible, we should examine more carefully when they will be accurate. We
have to check the magnitudes of the terms which are omitted in (5.102).
∂vx ∂vx ∂P ∂ 2 vx ∂ 2 vx
ρvx + ρvy =− + µ 2 +µ 2
| {z∂x} | {z∂y} ∂x | ∂x
{z } | ∂y
{z }
T4 T3 T2 T1
(5.102)
∂vy ∂vy ∂P ∂ 2 vy ∂ 2 vy
ρvx + ρvy =− + µ 2 + µ 2 +ρby
| {z∂x} | {z∂y} ∂y | ∂x
{z } | ∂y
{z }
T8 T7 T6 T5
Answer: It has to be proven that T1 is much larger than the other T’s. To facilitate comparisons, we
will average each term over the channel height and consider only absolute values.
∂ 2 vx 3µQ 3µQ
T1 = µ 2
=− 3 ⇒ ⟨T1 ⟩ = (5.103)
∂y 2h 2h3
2 2
∂ 2 vx 3µQ dh y 2 3µQ dh
T2 = µ 2 = 1−6 ⇒ ⟨T2 ⟩ = (5.104)
∂x 2h3 dx h 2h3 dx
2
∂vx 3ρQ dh
T3 = ρvy ⇒ ⟨T3 ⟩ = ⟨T4 ⟩ = 3
(5.105)
∂y 20h dx
Thus 2
⟨T2 ⟩ dh ⟨T3 + T4 ⟩ Re dh
= and = (5.106)
⟨T1 ⟩ dx ⟨T1 ⟩ 5 dx
On the other hand,
⟨T5 ⟩ 3 dh
= (5.107)
⟨T1 ⟩ 2 dx
There is no necessary to check the others, because they are smaller than T5 .
In summary, the lubrication approximation will be accurate if
dh
≪ 1 and Re dh ≪ 1
dx dx (5.108)
100 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
It should be noticed here again that the viscous effect is negligible due to the irrotational motion of
the fluid. Thus, the differential momentum balance cannot give an information on the velocity profiles
as we did up to now. Instead, the differential mass balance can be used for determining the velocity.
As shown in Fig. 5.19, a Newtonian fluid flows incompressiblely and irrotationally around a cylinder
of infinite length. Therefore the viscous transfer does not occur from the surface of the cylinder and
the velocity profile is only caused by geometry of the cylinder. In this case, there are two nonvanishing
components of the fluid velocity: vr and vθ . Let’s determine them with P .
at r = a, vr = 0
at θ = 0 and π, vθ = 0 ∵ symmetry
(5.111)
at r → ∞, |v| = v∞
at r → ∞, P = P∞
Notice that for a flow in which the viscous effect is negligible, no-slip condition cannot be applied.
Thus the tangential component of velocity vθ at wall of the cylinder may not be zero, but the normal
component of velocity at the wall is still zero.
Since the fluid is incompressible, the differential mass balance (equation of continuity) must satisfies
1 ∂ 1 ∂vθ ∂vz
∇·v = (rvr ) + + =0 (5.112)
r ∂r r ∂θ ∂z
|{z}
0, vz =0
Let
∂Ψ
rvr ≡ (5.113)
∂θ
Then (5.112) becomes
∂ ∂Ψ
+ vθ =0 (5.114)
∂θ ∂r
5.10. TWO DIMENSIONAL FLOWS 101
f′ f f ′′ + f ′ /r g ′′
f ′′ g + g + 2 g ′′ = 0 =⇒ 2
=− ≡ λ2 (5.121)
r r f /r g
Therefore we have
g ′′ + λ2 g = 0
f′ f (5.122)
f ′′ + − λ2 2 = 0
r r
We can obtain solutions of these equations as
m(m − 1) + m − λ2 = 0 ⇒ m = ±λ (5.124)
and thus
Ψ = c3 rλ + c4 r−λ c1 cos(λθ) + c2 sin(λθ)
(5.125)
102 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
and thus
ρv 2
P =− − ρgr sin θ + P0 (5.135)
2
where P0 is constant and
" 2 2 #
a2 a2
2 2 2 2
v =v·v = v∞ cos θ 1 − 2 + sin θ 1 + 2 (5.136)
r r
It is clear from (5.133) that the curl of the viscous portion of the stress is zero.
5.10. TWO DIMENSIONAL FLOWS 103
z
r
θ
Moving
fluid v∞
at r = R, vr = 0
at r = R, vθ = 0
(5.138)
at r → ∞, |v| = v∞
at r → ∞, P = P∞
Let
1 ∂Ψ
vr = (5.140)
r2 sin θ ∂θ
Then (5.139) becomes
∂ 1 ∂Ψ
+ vθ sin θ =0 (5.141)
∂θ r ∂r
For this to be true in general
1 ∂Ψ
vθ = − (5.142)
r sin θ ∂r
Instead of having 2 unknowns vr , vθ , we have only one unknown Ψ(r, θ), called the stream function
(see Appendix D).
Now we use the Stoke’s flow equation to solve Ψ. Having both velocity and pressure as dependent
variables in the equation of motion gives difficulty in multidimensional flow. It is therefore convenient
to eliminate the pressure by taking the curl:
where b = −gez = −g [cos θer + (− sin θ)eθ ] from Fig. 1.6(c). Thus we can see
2
3 µv∞ R
P = P∞ − ρgr cos θ − cos θ (5.152)
2 R r
The pressure and shear stress acting on fluid at the surface by the sphere are
3 µv∞
P (R, θ) = P∞ − ρgR cos θ − cos θ
2 R (5.153)
3 µv∞
τrθ (R, θ) = − sin θ
2 R
5.10. TWO DIMENSIONAL FLOWS 105
Note that the normal stress is zero everywhere at the sphere surface from (5.149)
∂vr
τrr (R, θ) = 2µ =0 (5.154)
∂r
Summary
106 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
운동량 전달 및 Reynolds 수
v
흐름(수평)방향 운동량 전달
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
유체의 유동
!"# !" ! 관성력
Reynolds 수 = = =
Fluid $ $" ⁄# 점성력
Inertia force v→
( ρv2 L2 ) 움직이는 유체는 같은 상태로 계속 움직이려
Viscous force
는 관성력과 주변유체의 영향을 받아 주변의
(μvL)
상태에 적응하려는 점성력을 갖는다.
점성유동 비점성유동
(creeping flow) (inviscid flow) 층류 유동 난류 유동
관성력 무시 점성력 무시
(µ=0) 유체입자들이 서로 엉
- v, L: small 유체입자들이 서로 엉
키지 않고 층을 이루
키면서 흐르는 유동
며 흐르는 유동
Re ® ¥
Re <<1
Potential flow
Re 수 = 2100
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
5.10. TWO DIMENSIONAL FLOWS 107
ü Rex < 2✕105 : laminar flow in the boundary layer, " ∝ # !.#
ü 2✕105 < Rex < 3✕106 : transition, " ∝ # $.#
ü 3 ✕ 106 < Rex : turbulent flow in the boundary layer except the laminar sublayer, " ∝ # !.%
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
뉴튼 유체의 층류 유동 (정상상태)
Ø 압력차 유동 Ø 중력 유동 Ø Couette 유동
g t F
.+
P1 P2 t
t
t
1 ) .+
.&'(
. = .&'( 1 − . = .= >
2 2 ℎ
3$ − 3) 4 = t ×4*
;4:< = t ×4*
7.
= −6 1 = 2 4* 7.
71 = −6 1 = 2 4*
71
∆32 ) ∆" 32' (
.&'( = = ;<2 )
46: # )! .&'( =
46
Hagen-Poiseuille Eq.
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
108 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
비뉴튼 유체의 층류 유동
Shear force는 중심에서 r방향으로 선형적으로 증가함 ! d&
= $
" d'
Shear force 속도
' $%$/'
& = &!"# 1 −
*
Newtonian: n=1
Shear thinning: 0<n<1
Shear thickening: n>1
Newtonian Shear Shear
thinning thickening
ü Newtonian: 점도가 일정하고 shear force가 일차이므로 속도는 반지름에 대하여 2차식임
ü Shear thinning: 중심에서 벽으로 갈수록 속도기울기가 커져서 점도가 줄어듦으로 Newtonian
유체보다 상대적으로 관성력이 커져서 벽쪽의 속도가 커진다. 따라서 속도의 차원은 2차보다
커짐.
ü Shear thickening: 중심에서 벽으로 갈수록 속도기울기가 커져서 점도가 커짐으로 Newtonian
유체보다 상대적으로 점성력이 커져서 벽쪽의 속도가 줄어든다. 따라서 속도의 차원은 2차보
다 적어짐.
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
뉴튼유체의 난류 유동
Ø 난류 유동
(a) (b)
Turbulent
vz
uz
vz
Laminar Turbulent
t0 time
(c)
vavg vavg
vz vz
Laminar Turbulent
$/(
&̅ 60 '
= 1−
& 49 *
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
5.10. TWO DIMENSIONAL FLOWS 109
속도의 중첩원리
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
1.0
0.6
y t increases
erf(x)
0.4
0.2
v0
1 2 3 4 5
x
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
110 CHAPTER 5. MOMENTUM TRANSFER AND VARIOUS FLOWS
윤활근사법에 의한 근사 1차원 유동 해석
- 윤활근사 (lubrication approximation)은 1886년 Reynolds 가 lubrication 유동을
해석하면서 시작되었음.
y hL ℎ! ≪ 1 & Re×ℎ! ≪ 1
h0 h(x)
x u(x) +)" +)"
⟹ )" ≫ )# & ≫
+, +-
L
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
z
r FPstat FPmov Fτ
P∞ P∞
θ
!! 가 매우 적어서
R
Re ≪ 1인 흐름
P∞ + ρgR
3μv∞ 3μv∞
Pstat = P∞ - ρgR cosθ Pmov = - cosθ τrθ = - sinθ
2R 2R
∇×( = 0인 흐름
Moving FPstat FPmov Fτ
fluid v∞
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
Chapter 6
Until Ch. 5, we have used the conservation laws to analyze fluid flows macroscopically (conservation
of mass, momentums, and mechanical energy) and microscopically (equation of continuity, motion,
and mechanical energy-Bernoulli). In this chapter, we will discuss how to expand them to analyze
forces acting on a solid, energy losses during fluid flow, pumps, drag, and flow in packed and fluidized
beds.
v2
P2
v1
P1 h2
h1
P+1/2ρv2+ρgh = consant at points on a streamlne
Figure 6.1: Streamline Bernoulli Theorem: the total energy of a fluid flowing from Cross section 1 to
Cross section 2 remains constant, though one energy form can be converted into another.
Most of fluid flow problems encountered in engineering involve streams that are influenced by solid
boundaries and therefore contain boundary layers. This is especially true in the flow of fluids through
111
112 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
pipes and other equipment, where the entire stream may be in boundary layer flow. To extend the
Bernoulli equation to cover these practical situations, two modifications are needed. The first, usually
of minor importance, is a correction of the kinetic energy term for the variation of local velocity with
position in the boundary layer; the second, of major importance, is the correction of the equation for
the existence of fluid friction, which appears whenever a boundary layer forms.
where ⟨v⟩ is the mean velocity on the cross section and α is called the kinetic energy correction factor :
⟨v 3 ⟩
α= 3 (6.5)
⟨v⟩
An error of about 6% is ordinarily negligible, given the other approximations being made. For laminar
flow, using (5.22) and (5.24) gives
Z R r 2 3
α=8 1− dA A = 2 (6.8)
0 R
where ff is the Fanning friction factor of the fluid, which will discussed later.
equivalent to the loss in mechanical energy is generated. Thus the energy loss (or heat generation)
should be added in the Bernoulli equation (6.6)
2 2
ρ⟨v1 ⟩ ρ⟨v2 ⟩
P1 + α1 + ρgh1 = P2 + α2 + ρgh2 + hf (6.10)
2 2
Here hf represents the frictional heat(energy loss) generated per unit volume of fluid when the fluid
flows the pipe between cross-section 1 and cross-section 2 and is always positive. There exist the
conversion or loss of mechanical energy to heat by viscous action, sudden change in diameter, fitting,
and so on. Therefore
hf = hf s + hf c + hf e + hf f (6.11)
hf s , hf c , hf e , and hf f are energy loss by skin friction due to viscous action, by sudden contraction,
sudden expansion, and fittings and valves, respectively. This will be discussed as follows.
Answer: PB =PC and use the Bernoulli equation between a point on A and at a point on B, and
between a point onC and a point on D.
PA + ρT gd1 = PB and PC = PD + ρM gd2 =⇒ PA = Patm + ρM gd2 − ρT gd1 (6.12)
Ex. 28. A siphon(흡입관) may be used to draw a liquid above its level in an open tank and discharge
it below that level. Once initiated, the flow is sustained by the height difference. Assume that viscous
losses are negligible, and that the kinetic energy correction factor is one.
• After an empty tube has been inserted into the tank, how much suction must be applied at
position C to start the flow?
• Find the velocity of the fluid exiting the siphon. Here the velocity is assumed to be the same
on the cross section.
• Show that during operation the pressure at position B is less than atmospheric pressure.
• Lowering the outlet will increase the flow rate. However, cavitation might eventually make the
siphon unreliable. That can be avoid if the pressure everywhere exceeds Pv , the vapor pressure
of the liquid. Find the maximum velocity vmax without the cavitation.
Answer: The pressure at a point on B is PB = P0 − ρgHB . Because the tube between B and C is
occupied by air before the flows begins, this is also value of PC needed to start the flow.
PC = PB + ρair g(HB + HC ) ≈ PB (6.13)
This suction pressure at C becomes P0 after removing the suction. Applying the Bernoulli equation
at a point on the surface of tank and at a point on C gives
1 1 2 p p
PS + ρvS2 + ρghS = PC + ρvC + ρghC =⇒ vC = 2g(hS − hC ) = 2gHC (6.14)
2 2
114 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
1 2 1 2
PB + ρvB + ρghB = PC + ρvC + ρghC =⇒ PB = P0 − ρg(HB + HC ) (6.15)
2 2
because vB = vC . Though the pressure at B is lower than the pressure at C, the gravity makes the flow
downward. From (6.16), it is clear that lowering the outlet HC increases the velocity, but decreases
PB . Applying the Bernoulli equation at a point on the surface of tank and at a point on B gives
1 1 2 p
PS + ρvS2 + ρghS = PB + ρvmax + ρghB =⇒ vmax = 2[(P0 − Pv )/ρ − gHB ] (6.16)
2 2
Ff = f (⟨v⟩, ρ, µ, D) (6.17)
where we neglect pipe roughness. By the dimensional analysis (see Sect. E.3), we are able to signifi-
cantly reduce number of experiments to obtain the relation (6.17) between the friction force and other
variables. The dimensional analysis has enabled us to relate the original five variables in terms of only
two dimensionless parameters in the form
Ff /D2
2 = f¯(Re) (6.18)
ρ⟨v⟩
and thus
1 2
Ff = Cf (Re) × ρ⟨v⟩ × Aw (6.19)
2
It is clear that the skin friction coefficient is a function of the Reynolds number only. Of course, if we
consider pipe roughness, the friction coefficient should be also a function of the roughness.
6.2. ENERGY LOSS BY SKIN FRICTION FOR FLOW IN PIPES 115
More commonly, the Fanning friction factor 1 ff (cf. fD = 4ff the Darcy friction factor ) has been
used by chemical engineers. Very similar to (6.19), the friction force can be expressed as
1 2
Ff = ff (Re) × ρ⟨v⟩ × Aw
| {z } 2 |{z} (6.20)
area on which friction acts
| {z }
Fanning friction factor
magnitude of the kinetic energy
Thus, the Fanning friction factor is also a function of the Reynolds number only and can be determined
by experiments as shown in Fig 6.4.
Ff /Aw τ
ff (Re) ≡ 1 2 = 1 2 (6.21)
2 ρ⟨v⟩ 2 ρ⟨v⟩
The symbols in Fig. 6.4 show representative data for ff obtained over a range of Re spanning four
orders of magnitude. In this log–log plot, there is a linear decline in the friction factor at relatively
low Reynolds numbers. At a critical value of the Reynolds number 2,100 a sharp rise occurs, followed
by a decline that is more gradual than the earlier one. What happens at the critical Reynolds number
is a transition from laminar to turbulent flow. In laminar flow all elements of the fluid move in
straight lines that are parallel to the tube wall, whereas in turbulence the pattern is chaotic, with
eddies of varying size continually forming and decaying. Such eddies greatly increase the shear stress.
The critical Reynolds number can be increased significantly if special precautions are taken to avoid
vibrations or other flow disturbances. Thus, where the laminar–turbulent transition will occur is not
entirely predictable, and in piping design it is prudent to avoid values of Re between 2,000 and 4,000.
Figure 6.4: Friction factor for smooth cylindrical tubes. The curves labeled Laminar, Prandtl–Kármán,
and Blasius are based on Eqs. (6.22), (6.23), and (6.24), respectively. The dashed lines are extrap-
olations. The symbols show data for water: circles, D = 1 cm from Koury (1995); filled and open
triangles, D = 5 cm and 10 cm, respectively, from Nikuradse (1932).
civil and mechanical engineers prefer to use the Darcy fiction factor.
116 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
Ex. 29. At Re = 1×104 , ff in turbulent flow is approximately five times what it would be if laminar
flow could be maintained. Explain why.
Answer: As shown in Fig. 5.8, the variation in velocity with radial position is parabolic in a laminar
flow, whereas in a turbulent flow the profile is more blunted. Thus, the higher shear rate at the wall
is what increases τ = Ff /Aw in turbulent flow (τrz in (4.76)) and ultimately leads to such an increase
in ff according to (6.20).
Turbulence can be costly in terms of energy usage. In a horizontal pipe, the required pumping power
equals Q∆P and ∆P is proportional to τ (or ff ). For certain other purposes, however, turbulence is
advantageous. For example, the increased mixing due to the turbulent eddies facilitates heat transfer
between the fluid and the pipe wall. That is helpful whenever a process stream must be heated or
cooled.
Wall Roughness
No matter how carefully fluid conduits are manufactured, there are always at least microscopic ir-
regularities on their surfaces. Irregularities can increase over time, as a result of corrosion, mineral
deposits, or the growth of microorganisms. Wall roughness tends to increase the friction factor at a
given Reynolds number, and ff is found to be governed by roughness alone when Re is sufficiently
large. The effects of roughness are noticeable only in turbulent flow. In that regime, where most of
the velocity variation occurs very near the wall, even small protrusions or indentations can affect the
key part of the velocity profile. In laminar flow the profile is much less sensitive to what happens in
the immediate vicinity of the wall.
What is needed to quantify the effects of roughness, at the very least, is the height of the irregular-
ities. This new length scale, denoted as k, gives rise to an additional dimensionless group k/D. Thus,
dimensional analysis indicates that ff = ff (Re, k/D). The effect of wall roughness on the friction
factor is shown in Fig. 6.5.
Figure 6.5: Friction factor for circular pipes. k is the roughness parameter.
6.2. ENERGY LOSS BY SKIN FRICTION FOR FLOW IN PIPES 117
P2
P1 P2 Ff g
Ff
P1
Ex. 30. [Pressure drop for water in process pipes] It is desired to find the values of ∆P for
water flow at 20 ◦ C through either of two pipes. One is entirely horizontal and the other has horizontal
and vertical segments. In each, the flow rate is Q = 8.0 × 10−3 m3 /s, the diameter is D = 0.1 m, and
the total length is L = 30 m. The second pipe has a bend, such that flow in the first 20 m is horizontal
and that in the last 10 m is upward. Here ρ = 1.00 × 103 kg/m3 and ν = 1.00 × 10−6 m2 /s. Compare
the pressure drop for each pipe.
된다.
118 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
Ex. 31. For the steady-state laminar flow of an incompressible Newtonian fluid in a circular pipe,
show that the friction factor can be obtained theoretically as
16
ff = (6.32)
Re
∆P D 32µ⟨v⟩ D 16µ
ff = 2 = 2 2 = (6.33)
L 2ρ⟨v⟩ D 2ρ⟨v⟩ ρ⟨v⟩D
That is true whether the flow is laminar or turbulent. What is special about turbulent flow is that the
velocity profile very close to the wall has a nearly universal form. Because the profile near the wall is
insensitive to the channel shape, so is the relationship between ff and ReH .
Laminar Flow
Similar to the friction factor for a laminar flow in a circular pipe, the friction factor for a laminar flow
in noncylindrical conduits can be predicted from first principles as
c
ff = (6.37)
ReH
the c usually has a value other than 16. For a parallel-plate channel c = 24, and for an equilateral
triangular channel c = 40/3. For a rectangular conduit with α = a/b ≤ 1
∞ −1
24 X tanh(λn /α) 1
c= 1 − 6α , λn = n+ π (6.38)
(1 + α)2 n=0
λ5n 2
6.3. ENERGY LOSS BY OTHER FRICTIONS 119
The frictional energy loss hf e from a sudden expansion of cross section is proportional to the
kinetic energy of the fluid in the small conduit and can be written as
2
ρ⟨va ⟩
hf e = Ke (6.40)
2
where Ke is called the expansion loss coefficient. In this case the calculation of Ke can be made
theoretically and a satisfactory results obtained. Consider the control volume defined by sections AA
and BB in Fig. 6.7. Gravitational forces do not appear because the pipe is horizontal. The wall friction
is negligible because the wall is relatively short and there is almost no velocity gradient at the wall
between the sections. The conservation of linear momentum (6.41) gives
X
Ṗout − Ṗin = F =⇒ ṁ(βb ⟨vb ⟩ − βa ⟨va ⟩) = Pa Sa − Pb Sb (6.41)
where β is call the momentum correction factor 3 , which is analogous to the kinetic energy correction
3 This can be determined in a very similar way to (6.4) and defined as
⟨v 2 ⟩
β= (6.42)
⟨v⟩2
120 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
When the cross section of the conduit is suddenly reduced, the fluid stream cannot follow around
the sharp corner and the stream breaks contact with the wall of the conduit. A jet is formed, which
flows into the stagnant fluid in the smaller section. The jet first contracts and then expands to fill the
smaller cross section, and downstream from the point of contraction the normal velocity distribution
eventually is reestablished. The cross section of minimum area at which the jet changes from a
contraction to an expansion is called the vena contracta. The flow pattern of a sudden contraction is
shown in Fig. 6.8. Section CC is drawn at the vena contracta. Vortices appear as shown in the figure.
The frictional energy loss hf c from a sudden contraction of cross section is proportional to the
kinetic energy of the fluid in the smaller conduit and can be written as
2
ρ⟨vb ⟩
hf c = Kc (6.45)
2
where Kc is called the contraction loss coefficient. Experimentally, for laminar flow, Kc < 0.1 so that
hf c is negligible. For turbulent flow, Kc is given by the empirical equation:
Sb
Kc = 0.4 1 − (6.46)
Sa
Ex. 32. A pump draws a solution of specific gravity 1.84 from a storage tank shown in Fig. 6.10. Its
efficiency is 60%. The velocity in the suction line is 0.914 m/s. Friction losses in the entire system are
29.9 J/kg. (a) What is the power delivered to the fluid by the pump? and (b) What pressure must
the pump develop? Here the flows is assumed to be turbulent so that the correction factors are nearly
unity.
5.0 cm
15.2 m
7.5 cm
we have
1 ρvb 2
Wp = + ρg(hb − ha ) + hf (6.52)
η 2
The power delivered to the pump is the produce of Wp and the volume flow rate:
Ab vb ρvb 2
P = V̇ Wp = + ρg(hb − ha ) + hf = 2.24 kW (6.53)
η 2
(b) In order to get the pressure developed by the pump, we choose two points just before and after
the pump and apply the Bernoulli’s equation.
ρ
Paf − Pbe = (vbe 2 − vaf 2 ) + ηWp = 330 kN/m2 (6.55)
2
because
hbe = haf , and hf = 0 (6.56)
the total developed head (TDH) is a head(or pressure) increased by a pump only. The Bernoulli
equation can be written between stations a and b in Fig. 6.11 as
Pa va 2 ηWp Pb vb 2 hf (= 0)
+ Za + + = + Zb + + (6.58)
ρg 2g ρg ρg 2g ρg
where hf is introduced due to fluid friction (any conversion of mechanical energy to heat in a flowing
system). Since the only friction is that occurring in the pump itself and is accounted for by η, hf = 0.
Then (6.58) becomes
vb 2 va 2
Pb Pa ηWp
TDH = ∆H = Hb − Ha = + Zb + − + Za + = (6.59)
ρg 2g ρg 2g ρg
Power requirement
The power supplied to the pump drive from an external source is denoted by PB :
Qρg∆H
PB = QWp = (6.60)
η
Pa′ va′ 2 Pa va 2 hf
+ Z a′ + = + Za + + (6.61)
ρg 2g ρg 2g ρg
Since va′ = 0 and Za′ = 0, this becomes
Pa Pa ′ va 2 hf
= − Za − − (6.62)
ρg ρg 2g ρg
and letting Pa /ρg = Pv /ρg + NPSH gives
Pa′ − Pv va 2 hf
Za ≈ − − − NPSH (6.63)
ρg 2g ρg
where Pv is the vapor pressure. In general, the velocity term can be neglected because it contributes
very little to the NPSH compared to the others.
tτt t
tp+tτn
v
Fluid
Figure 6.12: The surface force acting on a falling sphere by a surrounding fluid.
where tP and tτ are portions of the stress vector caused by pressure and viscous forces, respectively.
It should be noticed that these are counteractive to the force acting on the fluid by the sphere. In
addition, we used the pressure and viscous stress vectors expressed as
because of its opposite direction to pressure and using (4.52) and (4.60).
As seen from (5.152), the movement of surrounding fluid makes a change in the pressure acting
on an object placed in the moving fluid4 (Fig. 6.13):
" 2 #
3 µv∞ R
P (r, θ) = P∞ − ρgr cos θ + − cos θ (6.66)
| {z } 2 R r
Pstat | {z }
Pmov
Thus, contributing to the pressure force are the sum of static pressure variations, which ultimately
z
P∞ - ρgR
r
FPstat FPmov Fτ
R P∞ P∞
Form drag Wall drag
Moving
v∞ P∞ + ρgR
fluid
3μv∞
3μv 3μ
3μvv∞
Pstat = P∞ - ρgR cosθ Pmov = - cosθ τ= sinθ
2R 2R
Figure 6.13: Pressure and viscous forces acting on a sphere by a creeping flow past the sphere.
create the buoyancy force, and pressure variations caused by the fluid motion. Finally the total surface
force acting on the sphere by the fluid can be written as
Z Z Z
FS = −nPstat dS + −nPmov dS + τ · n dS
S S S
= FP stat + FP mov + Fτ (6.67)
| {z } | {z }
Fbuoy = surface force due to FD = surface force due to
the difference in static pressure the movement of a fluid
4 When air blows, the pressure on a sphere can be written as
3 0.018 × 10−3 Pa · s × v∞
P (R, θ) = P∞ − 1.293 kg/m3 × 9.8 m/s × R cos θ − cos θ
2 R
v∞
≈ 101.3 × 103 − 12.7R cos θ − 2.7 × 10−5 cos θ [Pa]
R
6.5. BUOYANCY AND DRAG FORCES 125
a FP1 V
water (ρ)
A P1 Fbuoyancy = FP = ρgV
ρp
Pside
h FP3
Fbody
P2
FP2
https://www.britannica.com/science/Archimedes-principle
Figure 6.14: Buoyancy forces acting on a cylinder fully immersed in water and on a boat partially
immersed in water.
n
n
Moving tio
tio rf ic
ric
fluid
in
f
in
Sk
sk
Wall drag
Stationary Moving
fluid Stationary Form drag
sphere Form drag
sphere
Stagnation
Drag = Form drag + Wall drag point
Skin friction appears in the unseparated boundary layer. When the fluid reaches the backside of the
sphere, its momentum prevents it from making the sharp turn around the sphere, and it separates
from the sphere and proceeds outward into the bulk of the fluid if the velocity of the fluid is enough
fast. Behind the sphere is a backwater zone of strongly decelerated fluid, in which large eddies, called
vortices, are formed. This zone is known as the wake. The eddies in the wake are kept in motion by
the shear stresses between the wake and the separated current. They lead to a large pressure lose in
the fluid and thus form large form friction drag.
When a fluid flows past a parallel thin flat plate, there appears only skin friction drag on the two
sides of the plate. For a time after the fluid leaves the plate, the layers and velocity gradient presist.
Soon, however, the gradients fade out, the boundary layers intermingle and disappear, and the fluid
once more moves with a uniform velocity. There occurs no form drag by eddies. When the fluid flows
past the vertical plate, there occurs large vortex, leading to large form drag.
Form drag
Figure 6.16: Flow past a flat plate: (a) parallel with plate (skin friction) (b) perpendicular to plate
(form).
The form drag is dependent upon the shape of an object. More generally, the drag is dependent
6.5. BUOYANCY AND DRAG FORCES 127
of the velocity, density, and viscosity of the fluid, and the characteristic length (e.g. diameter):
FD = f (v, ρ, µ, Dp ) (6.70)
By the dimensional analysis (see Sect. E.3), we are able to significantly reduce number of experiments
to obtain the relation (6.70) between the drag and other variables. The dimensional analysis has
enabled us to relate the original five variables in terms of only two dimensionless parameters in the
form
FD /D2
= f¯(Rep ) (6.71)
ρv 2
where Rep = ρvDp /µ indicates the Reynolds number of the solid particle. It is clear that the drag
coefficient is a function of the Reynolds number only.
1 2
FD = CD (Re) × ρv × AP
|2 {z } (6.72)
| {z } |{z}
drag coefficient projection area
magnitude of the kinetic energy
where AP is the area of the solid projected to the direction of flow. Ap (cylinder) = LDp , Ap (sphere) =
πDp 2 /4.
In addition, the drag coefficient is a function of the Reynolds number only:
FD /AP
CD ≡ 1 2
= CD (Rep ) (6.73)
2 ρv
For regular shapes such as spheres or cylinders at low fluid velocities, the flow patterns and drag
forces can be estimated by the linear momentum balances. For high velocities and irregular shapes
they are most easily determined by experiments. Figure 6.17 shows the drag coefficient measured
experimentally for regular shapes as a function of the Reynolds number.
Figure 6.17: Drag coefficient measured experimentally for spheres, disk, and cylinder as a function of
Reynolds number.
128 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
Figure 6.18: Types of behavior for the flow around a sphere. Regions of turbulent flow are shaded in
gray
Sphere
Figure 6.18 shows types of behavior for the steday-state incompressible flow of a Newtonian fluid past
a sphere.
Regime I: Rep ≪ 1 (Creeping flow), the viscous forces predominate compared to the inertia forces and
affect at considerable distance from the cylinder. No eddies occur.
where the drag has been calculated from the surface force by viscous and pressure forces obtained
in Sect. 5.9 (see more details in p. 125 of Bird’s book or Ch. 8 of Deen’s book.). More exactly,
it can be obtained from the surface integration of the viscous and pressure forces (5.153) and
(6.69).
Regime II: 1 <Rep < 103 , At Rep =10, a pair of vortices appears behind the cylinder. The type of flow
persists up to about Rep =40, when there appear two separation points (about 85◦ from the
front stagnation point) at which the streamlines separate from the solid surface. With further
increase in Rep , the vortices separate regularly from alternate sides of the cylinder (vortex street
in Figure 6.19).
Regime III: 103 <Rep < 2 × 105 (Constant CD region), This constancy of CD implies that viscous stresses
(skin friction drag) are negligible and that the drag (form drag) is proportional to v 2 . As it was
hypothesized by Newton that the drag on any object should be proportional to the fluid density,
the square of the velocity, and a characteristic area, this is called the Newton’s-law regime.
CD
FD /AP = cρv 2 , c= = constant (6.75)
2
6.5. BUOYANCY AND DRAG FORCES 129
When Rep > 2500, the wake is no longer characterized by large eddies. The flow on the surface
of the body from the front stagnation point to the separation point is laminar, and the shear
stress in this interval is appreciable only in a thin boundary layer. The drag coefficient remains
constant, approximately 0.445 for a sphere.
Regime IV: As Rep increases, the front boundary layer becomes turbulent and the separation point moves
toward the rear of the sphere (about 140◦ )and the wake shrinks. The remarkable drop in the drag
coefficient is the result of this decrease in the size of the wake and the corresponding decreases
in form drag. This is called Eiffel phenomena.5
This is why golf balls have lots of dimples on their surface to make turbulent flow (300∼500,
0.175 mm depth) .
Figure 6.20: Dimples of a golf ball (300 ∼ 500, 0.175 mm depth) decrease the form friction by decreasing
the areas of wakes.
Disk
Some of experimental equations are available for the drag coefficients of disks.
20.4
CD = 1 + Re0.792 , Re ≤ 133
Re (6.76)
CD = 1.17, Re > 133
The Eiffel phenomenon does not appear.
5 The phenomenon of the sudden change of sphere drag was first observed in a rather amusing way. Prandtl in
Göttingen and Eiffel in Paris measured the drag of the sphere: Prandtl obtained a value for the drag coefficient which
was more than twice that obtained by Eiffel. They exchanged information, and one of the young engineer in Prandtl’s
laboratory said, “Oh, M. Eiffel forgot a factor of two. He calculated the coefficient referred to ρv 2 , not 12 ρv 2 .’’ This
remark somehow reached Paris and the elderly M. Eiffel became very angry. He then measured the drag for a wider range
of Reynolds numbers and discovered that a sudden decrease in the drag coefficient occurred beyond a certain Reynolds
number. But he did not find the physical reason for the sudden change. It was Prandtl who gave the explanation
mentioned above. He put a fine wire ring around a sphere a short distance in front of the separation points of the
laminar layer. He found out that the total drag was reduced by the presence of the wire because laminar separation
was prevented
- Theodore Von Karman, “Aerodynamics: Selected Topics in the Light of Their Historical Development,” Dover
Publications, New York (2004).
130 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
Cylinder
As shown in Fig. 6.17, a nearly linear decline in a log–log plot is followed by a leveling off. The Eiffel
phenomenon occurs, CD declining suddenly from 1.2 to 0.3 at Re = 2 × 105 . Thereafter, CD increases
gradually.
For Re below the Eiffel threshold,
4 × 10−4 Re
6.8 1.96
CD = 1.18 + + − , Re < 2 × 105 (6.77)
Re0.89 Re1/2 1 + 3.64 × 10−7 Re2
Flat plates
Consider a thin plate parallel to the flow (The same category is a long cylinder aligned end-on, for
which A|| = DL ≫ A⊥ = πD2 /4) as shown in Fig. 5.4. Since A⊥ is negligible, pressure cannot act
in the flow direction and form drag will be absent. Thus, a thin plate parallel to the flow experiences
only skin friction (or wall) drag.
As will discussed in Sect. 6.2, the skin friction drag can be expressed as
1
FD = Cf (Re) × ρv 2 × LW (6.78)
2
where L and W are the length and width of the flat plate. It should be clearly noticed that Cf is an
average value since the Reynolds number varies with the distance from the entrance, x (Rex = ρvx/µ).
As derived in Ch. 9 of Deen’s book, if the boundary layer is entirely laminar and ReL > 100, then
The value of B depends on where the transition occurs, with B = 1050 for Ret = 3×105 and B = 8700
for Ret = 3 × 106 .
Ex. 33. [Drag on a cylinder in water] calculate the drag per unit length on a long cylinder due
to water flow perpendicular to its axis. The diameter is D = 0.10 m, the relative velocity is U = 1.0
m/s, ρ = 1.0 × 103 kg/m3 , and µ = 1.0 mPa·s.
The projected area is AP = DL and, from Eq. (6.77), CD = 1.18. The drag per unit length is then
Ex. 34. [Comparative drag on a cylinder and a flat plate] Compare the drag on a long cylinder
to that on a flat plate of equal surface area, for the conditions in the above example. Imagine that
the cylinder is flattened into a plate, giving a length in the flow direction that equals half the cylinder
circumference (Lp = πD/22) and a width that equals the cylinder length (W = Lc ). It is desired to
compare FD /W for the plate to FD /Lc for the cylinder.
Answer: For D = 0.1 m, the length of the equivalent plate is Lp = π(0.1)/2 = 0.157 m and the plate
Reynolds number is
ρLp U (1)(0.157)(1.0 × 103 )
Re = = = 1.57 × 105 (6.83)
µ 1.0 × 10−3
6.5. BUOYANCY AND DRAG FORCES 131
Because this is less than 3 × 105 , the boundary layer is entirely laminar. Accordingly,
Cf = 1.328 Re−1/2 = 1.328 (1.57 × 105 )−1/2 = 3.35 × 10−3 (6.84)
Allowing for boundary layers on both sides of the plate, the drag per unit width is
FD
= ρU 2 Lp Cf = (1 × 103 )(1)2 (0.157)(3.35 × 10−3 ) = 0.53 N/m (6.85)
W
This is only about 1% of 59 N/m, the value of FD /Lc from the above example. Assuming that the
flat-plate drag roughly equals the friction drag on the equivalent cylinder, this suggests that almost
the entire drag on a cylinder at large Re is form drag. Pressure measurements and more precise
calculations have confirmed that.
ρp v FD(drag) = CD 1/2ρv2 Ap
Fg(gravity) = mg = ρp gV
Figure 6.21: Force balance of a falling particle (움직이는 고체에 작용하는 힘=중력+부력+항력= ma).
즉 움직이는 고체의 속도의 변화는 중력, 부력, 항력의 합에 의하여 결정된다. 힘이 합이 0이면 고체의
속도는 일정하게 된다.
The Newton’s second law of the particle (or the force balance) says that
ρv 2
dv ρ
m = Fg − Fbuoy − FD = mg 1 − − CD Ap
dt ρp 2
(6.86)
CD ρv 2 Ap
dv ρ
=⇒ =g 1− −
dt ρp 2m
For simplicity, the positive velocity of the particle corresponds to the direction of the gravitational force
(downward). For a centrifugal field the gravitational acceleration is changed to centrifugal acceleration
rω 2 . What if we drop a ball at a very high place? From (6.86) It can be seen that the acceleration
decreases with time and approaches zero because g is constant while the drag always increases with
velocity. Thus the particle quickly reaches a constant velocity (dv/dt = 0), which is called the terminal
velocity vt . In other words, the terminal velocity is the velocity of a particle whose gravitational force
is equal to sum of the buoyancy force and the drag force.
Let the velocity of a particle relative to the fluid be v. A reference frame is chosen in which the
sphere is stationary and the fluid velocity is v. At the situation of the terminal velocity (no change in
the velocity) by taking dv/dt = 0 in (6.86), we have
s
2g(ρp − ρ)m
vt = (6.87)
CD ρρp Ap
and thus s
4g(ρp − ρ)Dp
vt = (6.89)
3CD ρ
Generally, the drag coefficient is determined by experiments. However for the following limiting
cases it can be obtained directly.
• Very low Rep : The drag coefficient at very low Reynolds numbers can be expressed by the
Stokes’ law as 24/Rep as seen in (6.74) and the Stokes terminal velocity becomes
s
4g(ρp − ρ)Dp ρDp vt gDp2 (ρp − ρ)
vt = =⇒ vt = (6.90)
72ρµ 18µ
In liquids, this expression applies for Dp as small as a few nm. For particles that small
in gases, the continuum approximation breaks down and a correction factor involving the
Knudsen number is needed (see p. 64 in. Deen’s book).
• Very high Rep : For 1,000<Rep <200,000, the drag coefficient is approximately 0.44 as
shown in Fig. 6.17. The terminal velocity becomes
s
g(ρp − ρ)Dp
vt = 1.75 (6.91)
ρ
Ex. 35. Estimate the terminal velocity for 80- to 100-mesh particles of limestone (ρp = 2, 800 kg/m3
falling in water at 30 ◦ C. (ρ = 995.7 kg/m3 , µ =0.801 cP).
Answer: Since DP for 100-mesh = 0.147 and DP for 80-mesh = 0.175, the average of the particles is
0.161 mm.
First assume that Rep < 1, then the terminal velocity becomes from (6.90)
9.8 × (0.161 × 10−3 )2 (2800 − 995.7)
vt = = 0.032 m/s (6.92)
18 × 0.801 × 0.001
From this value, we can calculate Rep
995.7 × 0.161 × 0.001 × 0.032
Rep = = 6.4 (6.93)
0.801 × 0.001
Thus, we did a wrong guess. Let’s guess again that Rep = 2.5, then CD ≈ 20 from Fig. 6.17. The
terminal velocity from (6.89) becomes
r
4 × 9.8(2800 − 995.7) × 0.161 × 0.001
vt = = 0.138 m/s (6.94)
3 × 20 × 995.7
From this we can get Rep = 2.76. Therefore our guess is correct.
6.5. BUOYANCY AND DRAG FORCES 133
Ex. 36. A metal empty sphere (D=4.5 mm and m=0.05 g) is placed in a liquid with ρ=0.95 g/cm3 .
Determine a viscosity of the liquid and the moving direction of the sphere when the terminal velocity
is 4.0 mm/s.
950
Fg = 0.00005 × 9.8 = 0.00049 N, Fb = 0.00005 × 9.8 = 0.00044 N (6.95)
1048
where the density of the sphere is ρp = m/V = 0.05/(π/6 × 0.453 ) = 1.048 g/cm3 . Since Fb < Fg , the
sphere is moving down which corresponds to the gravitational direction.
Now assume that the Reynold number of the particle is very low, the terminal velocity becomes
Practically, the distance traveled during the acceleration can be roughly estimated by multiplying vt
and t0 , which is the time scale for the acceleration (see Eq. (6.101)). As an example, for a 1 mm sand
grain in air, t0 = 0.75 s, and vt = 7.4 m/s so that vt t0 = 5.6 m. The transient could be ignored in
calculating a settling time if the sand were falling from a height much greater than that.
A complication in applying Newton’s second law (6.86) is that, as the neighboring fluid is displaced
by an accelerating sphere, it too accelerates. The acceleration of the fluid tends to slow the acceleration
of the sphere, just as if the sphere mass were larger than its actual value. This can be accounted for
by replacing the sphere mass with an effective or virtual mass m∗ . The increment, or added mass, is
calculated most easily for potential flow, and the objective is to evaluate it for a solid sphere of radius
R. Assume that the sphere moves at a velocity U (t) in the z direction and that the inviscid fluid of
density ρ is otherwise at rest. The fluid-dynamic force on the sphere is (see Problem 9.5 of Deen’s
book )
2 dU
Fz (t) = − πR3 ρ (6.99)
3 dt
The negative sign confirms that displacing the surrounding fluid slows the sphere, and the coefficient
of dU/dt shows that the added mass is half the mass of the displaced fluid. For a sphere of volume V
the added mass is ρV /2, or half the mass of the displaced fluid. Accordingly, m∗ = (ρp + ρ/2)V .
ρv 2
dv ρ
m∗ = Fg − Fbuoy − FD = mg 1 − − CD Ap
dt ρp 2 (6.100)
=⇒ v(τ ) = vt tanh τ
where
t (ρp /ρ + 1/2)vt
τ= , and t0 = (6.101)
t0 |ρp /ρ − 1| g
Noting that tanh−1 (0.95) = 1.83, this indicates that the acceleration is 95% complete at τ = 1.83 or
t = 1.83 t0 .
134 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
Flow
v
S0 Vp Sp
L Deq
Φs (Vp const.)
(total S
const.)
Flow Dp
v
S0
L
Figure 6.22: Procedure to estimate pressure drop of the flow through a bed of solids
where the void volume in the bed is the same as the total volume of the n channels.
πDeq 2
S0 Lε = n L (6.105)
4
Combining eqs. (6.103) through (6.105) gives
2Φs Dp ε
Deq = (6.106)
3(1 − ε)
6.6. FLOW IN A PACKED BED 135
Ex. 37. McCabe 7.1: A partial oxidation is carried out by passing air with 1.2 mole percent hydro-
carbon (Mw = 30) through 40-mm tubes packed with 2m of 3-mm catalyst pellets. The air enters at
350◦ C and 2.0 atm with a superficial velocity of 1 m/s. What is the pressure drop through the packed
tubes? Here the ΦS = 1 and µ = 3 × 10−5 Pa·s. How much would the pressure drop be reduced ny
using 4-mm pellets? Assume ε = 0.4.
Answer: The only variable we don’t have in (6.113) is the density of the air with 1.2 mole percent
hydrocarbon. Since
PM 2 × 30
P = cRT =⇒ ρ = = = 1.174 g/l (6.114)
RT 0.082 × 623
Thus, the pressure is
150 × 3 × 10−5 × 1 (1 − 0.4)2 1.75 × 1.174 × 12 (1 − 0.4)
2
∆P = 2 + = 18, 466 N/m (6.115)
12 × 0.0032 0.43 1 × 0.003 × 0.43
When Dp = 0.004, ∆P = 12, 798 N/m2 , 31% reduction.
136 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
Ex. 38. The pressure drop through a particle bed can be used to determine the external surface
area and the average particle. Data for a bed of crushed ore particles show ∆P/L = 84 (lbf /in2 )/ft
for airflow at a superficial velocity of 0.015 ft/s. The measured void fraction is 0.47 and the estimated
sphericity Φ is 0.7. Calculate the average particle size and the surface area per unit mass if the densities
of air and solid are 0.00123 and 4.1 g/cm3 . The viscosity of air is 0.018 cP.
Figure 6.23: Pressure drop and bed height versus superficial velocity for a bed of solids
1. Up to Point A: As the superficial velocity v0 increases, the pressure drop increases by the Ergun
equation (6.113), but the particles do not move. At the velocity of point A, the pressure drop
counterbalances the gravity on the particles.
2. Point A to Point B: The bed expands slightly with the grain still in contact due to closed
packing and keep ∆P constant.
6.7. FLOW IN A FLUIDIZED BED 137
3. Point B to Point C: True fluidization begins. The pressure drop keep constant but the height
of the bed increases as increasing the superficial velocity.
If the flow rate is gradually reduced, the pressure drop remains constant, and the bed height decreases,
following the line BC. However,, the final bed height may be greater than the initial value due to less
tight packing.
P2 Fb
S
P1 S + Fb = P2 S + Fe
Fe Fb(buoyance) = mg ρ/ρp
P1
Fe(gravitation) = mg
vOM
Fg Fb
∆P = − = ρp L(1 − ε)g − ρL(1 − ε)g = g(ρp − ρ)L(1 − ε) ∵ m = ρp SL(1 − ε) (6.120)
S S
150µv0M (1 − εM ) 1.75ρv0M 2
+ = g(ρp − ρ) (6.121)
Φs 2 Dp 2 εM 3 Φs D p ε M 3
Ex. 39. For a fluid bed shown in Fig. 6.23, find the ratio of the terminal velocity to the minimum
fluidization velocity.
(6.26) should be used for deriving the Ergun equation. The ρg term in (6.26) indicates the buoyant force in the packing
bed.
138 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
Case 1: Very small particles (< 300 µm): Rep < 1, Comparing (6.90) with (6.122), we have
vt 8.33(1 − εM )
= (6.124)
v0M Φs 2 εM 3
For spheres with εM ≈ 0.45,
vt ≈ 50 v0M (6.125)
We can operate the fluidized bed with v0 = 50 v0M without particles carried out with the exit
gas.
Case 2: Very large particles (Rep > 1000): Using (6.123) and (6.91), we have
vt 2.32
= (6.126)
v0M εM 3/2
For spheres with εM ≈ 0.45,
vt ≈ 7.7 v0M (6.127)
which is much lower ratio than for fine particles. This is a slight disadvantage in the use of
coarse particles, but the optimum particle size depends on other factors such as chemical reactor
efficiency, heat- and mass-transfer rates, grinding costs, and the gas velocity.
Summary
6.7. FLOW IN A FLUIDIZED BED 139
%%
v1 #= 층류 ≈ 2, 난류 ≈1
P1 h2 %%
h1 - Frictional energy loss: 마찰에 의한 열 발
P+1/2ρv2+ρgh = consant at points on a streamlne
생으로 역학적에너지가 감소함.
$ %! " $ %# "
!! + #! + $'ℎ! = !# + ## + $'ℎ# + ℎ$
2 2
- Skin Friction
- Friction from sudden
expansion/contraction
- Friction from fittings and valves
- $% "
ℎ$ = ℎ$& + ℎ$' + ℎ$( + ℎ$$ = 4,$ + /( + /' + /$
. 2
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
Skin Friction
- Skin friction: 유체는 고체표면에서 미끄러지지 않으므로 유체가 고체표면을 흐를
때 마찰이 발생함.
$% # .%
ℎ!& = #& , where #& = 0.4 1 −
2 .$
베르누이 방정식(마찰+펌프)
• 펌프의 효율 및 파워 : 펌프가 한 일중에 유체로 전달된 일
# #
$ %$ $ %%
6$ + 7$ + $8ℎ$ + 9:( = 6% + 7% + $8ℎ% + ℎ!
2 2
- 9: 펌프 효율
- :( : 유체의 단위부피당 펌프가 한일 [J/m3]
- ̇ ( [W]
펌프파워 = =:
6",+"- 6. 6$/ %$ ! # 6. %$ #
NPSH = − ⟹ 펌프위치 = ℎ$ − ℎ$! = + − + NPSH +
$8 $8 $8 28 $8 28
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
6.7. FLOW IN A FLUIDIZED BED 141
부력 및 항력
표면힘 = 정압력힘 + 동압력힘 + 점성힘
유체의 흐름방향
대체된 유체의 무게
(또는 고체 움직임의 반대 방향)
z 부력 항력
rfluidgVdisplaced r 유체 P∞ - ρgR
FPstat FPmov Fτ
R P∞ P∞
Form drag Wall drag
Moving
v∞ P∞ + ρgR
fluid
3μv∞
3μv 3μ
3μvv∞
Pstat = P∞ - ρgR cosθ Pmov = - cosθ τ= sinθ
2R 2R
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
부력 (Buoyancy Force)
• 물체가 유체에 잠겨 있을 때 높이에 따른 정압력 차이로 발생함.
• 중력에 항상 반대 방향으로 작용.
• 부력의 세기 = 잠긴 부분에 해당하는 유체 무게 = !!"#$% "#%$&'"()*%
a FP1 V
water (ρ)
A P1 Fbuoyancy = FP = ρgV
ρp
Pside
h FP3
Fbody
P2
FP2
https://www.britannica.com/science/Archimedes-principle
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
142 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
항력 (Drag force)
• 물체가 유체에 잠겨 있을 때 유체의 흐름(또는 물체의 움
Skin friction drag
Moving
cti
tio
fluid
n
f
ki
in
S
sk
Wall drag
Stationary Moving
fluid Stationary Form drag
sphere Form drag
sphere
Stagnation
Drag = Form drag + Wall drag point
!! ⁄)%
• 항력 = 항력계수 ✕ 운동에너지 ✕ 투영면적 #! (Re) =
+, $ ⁄2
"# !
!! = #! (Re) × × )%
$ !! cylinder = +,! !! (sphere) = 2 ,! " ⁄4
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
종말속도
"#
부력 Ø 뉴튼의 운동법칙 ' = 중력 − 부력 − 항력
"$
!
= )! *+ − )*+ − ," #$ -% .!
rp 항력
Ø 종말속도 (terminal velocity)
입자의 자유낙하 시 입자의 일정 낙하속도( "#⁄"$ = 0 )
중력
중력 = 부력+항력 2 )! − ) *'
#& =
))! ," .!
충전층 압력강하
Ø 충전층 부분에서 압력 강하 ü 상당 채널 지름 (equivalent diameter)
e 충전층 전체 표면적 (Sp´np)
= n개 채널의 전체 표면적 (pDeqL´n)
v D
구형화 2Φ( 0! ?
0+, =
3(1 − ?)
S
L
채널화 ü 구형도 (sphericity)
(S, V 일정)
Vp=Vsphere
Dp
#
D C-./010 6⁄0!
P1 P2 Φ( = = ≤1
C! C! ⁄+!
S
L
ü 평균 속도 (채널)
< )#̅ % #
∆8 = 94;* #̅ =
0+, 2 ?
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
144 CHAPTER 6. APPLICATION OF MOMENTUM TRANSFER
2Φ' (& 5 +
v D (#$ = +̅ =
3(1 − 5) 5
S
L
q 층류 흐름 (low Re) : Kozeny-Carman empirical equation (l1=2.1)
)& − ) >5* (
+)* ≈ Φ' (& ++ ≈ 7.7+)*
1.75)
If $! = 0.45
화학공학부 울산대학교
SCHOOL OF CHEMICAL ENGINEERING UNIVERSITY OF ULSAN
Appendix A
∂P ∂P ∂P
∇P = ex + ey + ez (A.1)
∂x ∂y ∂z
Divergence
∂τxx ∂τxy ∂τxz ∂τyx ∂τyy ∂τyz ∂τzx ∂τzy ∂τzz
∇·τ = + + ex + + + ey + + + ez
∂x ∂y ∂z ∂x ∂y ∂z ∂x ∂y ∂z
(A.4)
Curl
∂vz ∂vy ∂vx ∂vz ∂vy ∂vx
∇× v= − ex + − ey + − ez (A.5)
∂y ∂z ∂z ∂x ∂x ∂y
Laplacian
∂ 2 vx ∂ 2 vx ∂ 2 vx ∂ 2 vy ∂ 2 vy ∂ 2 vy ∂ 2 vz ∂ 2 vz ∂ 2 vz
2
∇ v= + + ex + + + ey + + + ez (A.7)
∂x2 ∂y 2 ∂z 2 ∂x2 ∂y 2 ∂z 2 ∂x2 ∂y 2 ∂z 2
145
146 APPENDIX A. SUMMARY OF DIFFERENTIAL VECTOR OPERATIONS
Divergence
1 ∂ 1 ∂vθ ∂vz
∇·v = (rvr ) + + (A.10)
r ∂r r ∂θ ∂z
1 ∂ 1 ∂τrθ ∂τrz τθθ 1 ∂ 2
1 ∂τθθ ∂τθz
∇·τ = (rτrr ) + + − er + 2 r τθr + + eθ
r ∂r r ∂θ ∂z r r ∂r r ∂θ ∂z
(A.11)
1 ∂ 1 ∂τzθ ∂τzz
+ (rτzr ) + + ez
r ∂r r ∂θ ∂z
Curl
1 ∂vz ∂vθ ∂vr ∂vz 1 ∂ ∂vr
∇× v= − er + − eθ + (rvθ ) − ez (A.12)
r ∂θ ∂z ∂z ∂r r ∂r ∂θ
Laplacian
1 ∂2T ∂2T
1 ∂ ∂T
∇2 T = r + 2 2 + (A.13)
r ∂r ∂r r ∂θ ∂z 2
1 ∂ 2 vr ∂ 2 vr
∂ 1 ∂ 2 ∂vθ
∇2 v = (rvr ) + 2 − + er
∂r r ∂r r ∂θ2 r2 ∂θ ∂z 2
1 ∂ 2 vθ ∂ 2 vθ
∂ 1 ∂ 2 ∂vr
+ (rvθ ) + 2 + 2 + eθ (A.14)
∂r r ∂r r ∂θ2 r ∂θ ∂z 2
1 ∂ 2 vz ∂ 2 vz
1 ∂ ∂vz
+ r + 2 2 + ez
r ∂r ∂r r ∂θ ∂z 2
Divergence
1 ∂ 1 ∂ 1 ∂vϕ
r 2 vr +
∇·v = 2
(vθ sin θ) + (A.17)
r ∂r r sin θ ∂θ r sin θ ∂ϕ
1 ∂ 2
1 ∂ 1 ∂τrϕ τθθ + τϕϕ
∇·τ = r τrr + (τ rθ sin θ) + − er
r2 ∂r r sin θ ∂θ r sin θ ∂ϕ r
1 ∂ 1 ∂ 1 ∂τθϕ τϕϕ cot θ
r3 τθr +
+ 3 (τθθ sin θ) + − eθ (A.18)
r ∂r r sin θ ∂θ r sin θ ∂ϕ r
1 ∂ 1 ∂ 1 ∂τϕϕ τθϕ cot θ
r3 τϕr +
+ 3 (τϕθ sin θ) + + eϕ
r ∂r r sin θ ∂θ r sin θ ∂ϕ r
Curl
1 ∂ 1 ∂vθ 1 ∂vr 1 ∂ 1 ∂ 1 ∂vr
∇× v= (vϕ sin θ) − er + − (rvϕ ) eθ + (rvθ ) − er
r sin θ ∂θ r sin θ ∂ϕ r sin θ ∂ϕ r ∂r r ∂r r ∂θ
(A.19)
Laplacian
∂2T
1 ∂ ∂T 1 ∂ ∂T 1
∇2 T = r2 + sin θ + (A.20)
r2 ∂r ∂r r2 sin θ ∂θ ∂θ 2
r2 sin θ ∂ϕ2
1 ∂2 ∂ 2 vr
2 2
1 ∂ ∂vr 1
∇ v = 2 2 r vr + 2 sin θ + 2 2 er
r ∂r r sin θ ∂θ ∂θ r sin θ ∂ϕ2
∂ 2 vθ
1 ∂ 2 ∂vθ 1 ∂ 1 ∂ 1 2 ∂vr 2 cot θ ∂vϕ
+ 2 r + 2 (vθ sin θ) + 2 2 + 2 − 2 eθ
r ∂r ∂r r ∂θ sin θ ∂θ r sin θ ∂ϕ2 r ∂θ r sin θ ∂ϕ
∂ 2 vϕ
1 ∂ ∂vϕ 1 ∂ 1 ∂ 1 2 ∂vr 2 cot θ ∂vθ
+ 2 r2 + 2 (vϕ sin θ) + 2 2 + + eϕ
r ∂r ∂r r ∂θ sin θ ∂θ r sin θ ∂ϕ2 r2 sin θ ∂ϕ r2 sin θ ∂ϕ
(A.21)
If a fluid is incompressible (ρ = constant), the volume of the fluid keeps constant during a deformation
(i. e., isochoric deformation). However the converse is not true. A compressible body can also undergo
an isochoric motion (or incompressible flow). Now we calculate the ratio of volumetric flux through
an element to the original volume of the element shown in Fig. B.1
(x, y, z)
Volumetric Flux (net volume per unit area and unit time)
Original volume
[(vx |x+∆x − vx |x )∆y∆z + (vy |y+∆y − vy |y )∆x∆z + (vz |z+∆z − vz |z )∆x∆y]
= (B.1)
∆x∆y∆z
In the limit ∆x, ∆y, ∆z → 0, it becomes
[(vx |x+∆x − vx |x )∆y∆z + (vy |y+∆y − vy |y )∆x∆z + (vz |z+∆z − vz |z )∆x∆y]
lim
∆x,∆y,∆z→0 ∆x∆y∆z
∂vx ∂vy ∂vz (B.2)
= + +
∂x ∂y ∂z
=∇·v
For an incompressible flow, the divergence of velocity should be zero:
Volumetric increase by flux
∇·v =
Original volume (B.3)
=0
149
150 APPENDIX B. INCOMPRESSIBLE AND IRROTATIONAL FLOW
As shown in Fig. B.2, let’s consider an element deformed by translational and rotational motions
in the xy plane during time ∆t The average angular velocity (the average rotation at a point) about
In the xz and yz planes the rotations of a point about the y and x axises are given by
1 ∂vx ∂vz 1 ∂vz ∂vy
ωy = − and ωx = − (B.8)
2 ∂z ∂x 2 ∂y ∂z
where ω is called the vorticity vector. As the derivation of the word ‘vorticity’ from ‘vortex’ (a swirling
body of fluid) suggests, ω is a measure of rotational motion. If a flow is irrotational, then
∇ × v=0 (B.10)
Appendix C
∂ρ
+ ∇ · (ρv) = 0 =⇒ ∇ · v = 0
∂t
∂v
ρ + (∇v) · v = −∇P + ∇ · τ + ρb ∵ T = −P I + τ
∂t
If a fluid is Newtonian
∂v
ρ + (∇v) · v = −∇P + µ∇2 v + ρb ∵ τ = µ(∇v + ∇vT )
∂t
• Mass balance
∂vx ∂vy ∂vz
+ + =0
∂x ∂y ∂z
• Momentum balance
– x-component
∂vx ∂vx ∂vx ∂vx ∂P ∂τxx ∂τyx ∂τzx
ρ + vx + vy + vz =− + + + + ρbx
∂t ∂x ∂y ∂z ∂x ∂x ∂y ∂z
– y-component
∂vy ∂vy ∂vy ∂vy ∂P ∂τxy ∂τyy ∂τzy
ρ + vx + vy + vz =− + + + + ρby
∂t ∂x ∂y ∂z ∂y ∂x ∂y ∂z
– z-component
∂vz ∂vz ∂vz ∂vz ∂P ∂τxz ∂τyz ∂τzz
ρ + vx + vy + vz =− + + + + ρbz
∂t ∂x ∂y ∂z ∂z ∂x ∂y ∂z
• Components of stress
Txx Txy Txz −P + τxx τxy τxz
Tyx Tyy Tyz = τyx −P + τyy τyz
Tzx Tzy Tzz τzx τzy −P + τzz
151
152APPENDIX C. EQUATION OF CONTINUITY AND MOTION FOR AN INCOMPRESSIBLE FLOW
If a fluid is Newtonian
∂vx ∂vx ∂vy ∂vx ∂vz
2µ µ + µ +
∂x ∂y ∂x ∂z ∂x
τxx τxy τxz
τyx
∂vx ∂vy ∂vy ∂vy ∂vz
τyy τyz = µ
+ 2µ µ +
τzx τzy τzz ∂y ∂x ∂y ∂z ∂y
∂vx ∂vz ∂vy ∂vz ∂vz
µ + µ + 2µ
∂z ∂x ∂z ∂y ∂z
• Momentum balance
– x-component
2
∂ 2 vx ∂ 2 vx
∂vx ∂vx ∂vx ∂vx ∂P ∂ vx
ρ + vx + vy + vz =− +µ + + + ρbx
∂t ∂x ∂y ∂z ∂x ∂x2 ∂y 2 ∂z 2
– y-component
2
∂ 2 vy ∂ 2 vy
∂vy ∂vy ∂vy ∂vy ∂P ∂ vy
ρ + vx + vy + vz =− +µ + + + ρby
∂t ∂x ∂y ∂z ∂y ∂x2 ∂y 2 ∂z 2
– z-component
2
∂ 2 vz ∂ 2 vz
∂vz ∂vz ∂vz ∂vz ∂P ∂ vz
ρ + vx + vy + vz =− +µ + + + ρbz
∂t ∂x ∂y ∂z ∂z ∂x2 ∂y 2 ∂z 2
• Mass balance
1 ∂ 1 ∂vθ ∂vz
(rvr ) + + =0
r ∂r r ∂θ ∂z
• Momentum balance
– r-component
vθ 2
∂vr ∂vr vθ ∂vr ∂vr ∂P 1 ∂ 1 ∂τθr ∂τzr τθθ
ρ + vr + − + vz =− + (rτrr ) + + − +ρbr
∂t ∂r r ∂θ r ∂z ∂r r ∂r r ∂θ ∂z r
– θ-component
∂vθ ∂vθ vθ ∂vθ vr vθ ∂vθ 1 ∂P 1 ∂ 1 ∂τθθ ∂τzθ
ρ + vr + + + vz =− + 2 r2 τrθ + + +ρbθ
∂t ∂r r ∂θ r ∂z r ∂θ r ∂r r ∂θ ∂z
– z-component
∂vz ∂vz vθ ∂vz ∂vz ∂P 1 ∂ 1 ∂τθz ∂τzz
ρ + vr + + vz =− + (rτrz ) + + + ρbz
∂t ∂r r ∂θ ∂z ∂z r ∂r r ∂θ ∂z
If a fluid is Newtonian
∂vr vθ 1 ∂vr ∂vθ ∂vr ∂vz
2µ µ − + + µ +
∂r r r ∂θ ∂r ∂z ∂r
τrr τrθ τrz
v θ 1 ∂vr ∂vθ v r 1 ∂v θ ∂vθ ∂vz
µ − r + r ∂θ + ∂r
τθr τθθ τθz = 2µ + µ +
τzr τzθ τzz r r ∂θ ∂z ∂θ
∂vr ∂vz ∂vθ ∂vz ∂vz
µ + µ + 2µ
∂z ∂r ∂z ∂θ ∂z
• Momentum balance
153
– r-component
vθ 2
∂vr ∂vr vθ ∂vr ∂vr
ρ + vr + − + vz
∂t ∂r r ∂θ r ∂z
1 ∂ 2 vr ∂ 2 vr
∂P ∂ 1 ∂ 2 ∂vθ
=− +µ (rvr ) + 2 − + + ρbr
∂r ∂r r ∂r r ∂θ2 r2 ∂θ ∂z 2
– θ-component
∂vθ ∂vθ vθ ∂vθ vr vθ ∂vθ
ρ + vr + + + vz
∂t ∂r r ∂θ r ∂z
1 ∂ 2 vθ ∂ 2 vθ
1 ∂P ∂ 1 ∂ 2 ∂vr
=− +µ (rvθ ) + 2 + 2 + + ρbθ
r ∂θ ∂r r ∂r r ∂θ2 r ∂θ ∂z 2
– z-component
∂vz ∂vz vθ ∂vz ∂vz
ρ + vr + + vz
∂t ∂r r ∂θ ∂z
1 ∂ 2 vz ∂ 2 vz
∂P 1 ∂ ∂vz
=− +µ r + 2 + + ρbz
∂z r ∂r ∂r r ∂θ2 ∂z 2
• Momentum balance
– r-component
vθ 2 + vϕ 2
∂vr ∂vr vθ ∂vr vϕ ∂vr
ρ + vr + − +
∂t ∂r r ∂θ r r sin θ ∂ϕ
2
∂ 2 vr
∂P 1 ∂ 1 ∂ ∂vr 1
+ µ 2 2 r 2 vr + 2
=− sin θ + 2 2 + ρbr
∂r r ∂r r sin θ ∂θ ∂θ r sin θ ∂ϕ2
– θ-component
vϕ 2 cot θ
∂vθ ∂vθ vθ ∂vθ vr vθ vϕ ∂vθ 1 ∂P
ρ + vr + + + − =− + ρbθ
∂t ∂r r ∂θ r r sin θ ∂ϕ r r ∂θ
∂ 2 vθ
1 ∂ 2 ∂vθ 1 ∂ 1 ∂ 1 2 ∂vr 2 cot θ ∂vϕ
+µ 2 r + 2 (vθ sin θ) + 2 2 + 2 − 2
r ∂r ∂r r ∂θ sin θ ∂θ r sin θ ∂ϕ2 r ∂θ r sin θ ∂ϕ
– ϕ-component
∂vϕ ∂vϕ vθ ∂vϕ vr vϕ vθ vϕ cot θ vϕ ∂vϕ 1 ∂P
ρ + vr + + + + =− + ρbϕ
∂t ∂r r ∂θ r r r sin θ ∂ϕ r sin θ ∂ϕ
∂ 2 vϕ
1 ∂ ∂vϕ 1 ∂ 1 ∂ 1 2 ∂vr 2 cot θ ∂vθ
+µ 2 r2 + 2 (vϕ sin θ) + 2 2 + +
r ∂r ∂r r ∂θ sin θ ∂θ r sin θ ∂ϕ2 r2 sin θ ∂ϕ r2 sin θ ∂ϕ
Appendix D
The stream function can be used to visualize a velocity field that has been found either analytically
or numerically. It is applicable to incompressible flows that are bidirectional. There is little need for
Ψ in a unidirectional flow, and Ψ is undefined if there are three velocity components or if ρ varies.
Many flows of technological or scientific interest are unidirectional or bidirectional, provided that the
coordinate system is chosen properly. Bidirectional flows fall into two broad categories, according to
the type of symmetry that causes the third velocity component to be absent. In planar flow, vz = 0
and the other velocity components are independent of z. Planar stagnation flow (Example 5.2-1) is
an example. Depending on what is being modeled, a planar problem might be analyzed using either
Cartesian (x, y) or cylindrical (r, θ) coordinates. In axisymmetric flows there is rotational symmetry
about the z axis, such that all quantities are independent of a cylindrical or spherical angle.
For a two-dimensional incompressible flow shown in Fig. D.1, the continuity equation is
∂vx ∂vy
+ =0 (D.1)
∂x ∂y
Let
∂Ψ
vx ≡ (D.2)
∂y
Then (D.1) becomes
∂ ∂Ψ
+ vy =0 (D.3)
∂y ∂x
for this to be true in general
∂Ψ
vy = − (D.4)
∂x
Instead of having 2 unknowns vx , vy , we have only one unknown Ψ(x, y), called the stream function.
155
156APPENDIX D. STREAM FUNCTION AND STREAMLINE FOR INCOMPRESSIBLE FLUIDS
What makes the stream function valuable for flow visualization is that constant values of Ψ correspond
to streamlines. A streamline is a curve that is tangent everywhere to the velocity vector. Consider a
path in the xy plane such that Ψ=constant. Along this path, dΨ=0, and the equation becomes
dy vy
= (D.6)
dx Ψ=constant
vx
3. Streak Line: The streak line through the point x(0) at time t represents the positions at time
t of the material particles that at any time τ ≤ t have occupied the place x(0) ., i.e., the locus of
points of all the fluid particles that have passed continuously through a particular spatial point
in the past. Dye steadily injected into the fluid at a fixed point extends along a streakline.
For one particular type of flow v ̸= f (t), the streamline, the path line, and the streak line coincide.
Appendix E
Dimensional Analysis
To group the variables into dimensionless parameters that are less numerous than the original vari-
ables. By combining the variables into a smaller number of dimensionless parameters, the work of
experimental data reduction is considerably reduced.
z =x+y (E.1)
If we let
z x
z⋆ ≡ x⋆ ≡ (E.2)
y y
then (E.1) becomes
z ⋆ = x⋆ + 1 (E.3)
E.1 Dimensions
Fundamental dimensions: Mass, Length, Time
157
158 APPENDIX E. DIMENSIONAL ANALYSIS
∇⋆ · v ⋆ = 0 (E.6)
and
∂v⋆ 1 1 ⋆
+ (∇⋆ v⋆ ) · v⋆ = −NEu ∇⋆ P ⋆ + ∇⋆ 2 v ⋆ + g (E.7)
∂t⋆ NRe NF r
Here NEu , NRe , and NF r are the Euler number, the Reynolds number, and the Froude number defined
as
P0 ρv0 L0 v0 2
NEu ≡ , NRe ≡ , N F r ≡ (E.8)
ρv0 2 µ g0 L0
Buckingham pi theorem
The number of dimensionless groups used to describe a situation involving in variables is equal to
n − r, where r is the rank of the dimensional matrix of the variables and n is the number of variables.
Ex. 40. Determine the dimensionless groups formed from the variables involved in the flow of fluid
external to a solid body. The force exerted on the body is a function of v, ρ, µ, and D(a significant
dimension of the body).
F = F (v, ρ, µ, D)
π1 = v a ρb D c F
(E.11)
π2 = v d ρe D f µ
Thus
F F/D2
π1 = =
ρv 2 D2 ρv 2 (E.13)
µ
π2 = = 1/Re
ρvD
The dimensional analysis has enabled us to relate the original five variables in terms of only two
dimensionless parameters in the form
F/D2
= f¯(Re) (E.14)
ρv 2
1. Row-switching transformation
2. Row-multiplying transformation
3. Row-addition transformation
160 APPENDIX E. DIMENSIONAL ANALYSIS
1 0 0 1/2 1/2
r1 +r3 /2→r1
−−−−−−−−−−−−−−→ 0 1 0 0 −1
r2 −2r3 →r2 ,−r3 /2→r3
0 0 1 1/2 −1/2
(E.16)