Inviscid (Frictionless) Flow

ME 306 Fluid Mechanics II • Continuity and Navier-Stokes equations govern the flow of fluids.
• For incompressible flows of Newtonian fluids they are

Part 1 𝛻⋅𝑉 =0
𝐷𝑉 𝜕𝑉
𝜌 =𝜌 + 𝑉 ⋅ 𝛻 𝑉 = 𝜌𝑔 − 𝛻𝑝 + 𝜇𝛻 2 𝑉
𝐷𝑡 𝜕𝑡
Potential Flow • These equations can be solved analytically only for a few problems.
• They can be simplified in various ways.
These presentations are prepared by
𝜇𝑤𝑎𝑡𝑒𝑟 = 10−3 Pa s
Dr. Cüneyt Sert • Common fluids such as water and air have small viscosities. 𝜇𝑎𝑖𝑟 = 2 × 10−5 Pa s
Department of Mechanical Engineering
Middle East Technical University • Neglecting the viscous term (zero shear force) gives the Euler’s equation.
Ankara, Turkey
csert@metu.edu.tr 𝐷𝑉
𝜌 = 𝜌𝑔 − 𝛻𝑝
𝐷𝑡
You can get the most recent version of this document from Dr. Sert’s web site.
• Still difficult get a general analytical solution for the unknowns 𝑝 and 𝑉.
Please ask for permission before using them to teach. You are not allowed to modify them.
1-1 1-2

Inviscid Flow (cont’d) Inviscid and Irrotational Flow

• Bernoulli Equation (BE) is Euler’s equation written along a streamline. • To simplify further we can assume the flow to be irrotational.
2
𝑝 𝑉 𝜉 = 2𝜔 = 𝛻 × 𝑉 = 0 (irrotational flow)
+ + 𝑧 = constant along a streamline
𝜌𝑔 2𝑔
Vorticity (ksi) Angular velocity
Exercise: Starting from the Euler’s equation derive the BE.

• Question: What’s the logic behind irrotationality assumption?

Inviscid flow away from the
object (negligible shear forces) • Irrotationality is about velocity gradients.

𝜕𝑤 𝜕𝑣 𝜕𝑢 𝜕𝑤 𝜕𝑣 𝜕𝑢
𝜉= − 𝑖+ − 𝑗+ − 𝑘=0
𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦
Airfoil or
1 𝜕𝑉𝑧 𝜕𝑉𝜃 𝜕𝑉𝑟 𝜕𝑉𝑧 1 𝜕 𝑟𝑉𝜃 1 𝜕𝑉𝑟
𝜉= − 𝑖 + − 𝑖 + − 𝑖
𝑟 𝜕𝜃 𝜕𝑧 𝑟 𝜕𝑧 𝜕𝑟 𝜃 𝑟 𝜕𝑟 𝑟 𝜕𝜃 𝑧
Adapted from
http://web.cecs.pdx.edu/~gerry/class/ME322 Viscous flow close to the object and in the
wake of it (non negligible shear forces)
1-3 1-4
 Inviscid and Irrotational Flow (cont’d) Inviscid and Irrotational Flow (cont’d)
• One special irrotational flow is when all velocity gradients are zero. • In an inviscid flow net shear force acting on a fluid element is zero.
• An example is uniform flow such as 𝑢 = 𝑈, 𝑣 = 0, 𝑤 = 0. 𝑦 𝑢=𝑈 • Only pressure and body forces act on the fluid element. But they cannot cause
𝑣=0 rotation because
• In many flow fields there will be uniform-like flow regions. 𝑥 𝑤=0
𝑧
• pressure forces act perpendicular to the element’s surface.
Away from the body, flow has small velocity • body forces act through element’s center of gravity.
gradients (uniform-like flow), small shear
forces and can remain irrotational.
In an inviscid flow, a fluid element
Uniform that originates from an irrotational
approach flow region will remain irrotational.
Airfoil
velocity
(irrotational) Uniform
Close to the body velocity gradients upstream
are high, shear forces are high and flow Airfoil
flow becomes rotational. (irrotational)

Exercise: Sketch the developing flow inside a pipe with uniform entrance and Exercise: Show how a fluid element will rotate inside the developing flow region
show the uniform and non-uniform flow regions. of a pipe with uniform entrance.
1-5 1-6

Inviscid and Irrotational Flow (cont’d) BE for Irrotational Flow

• In general, flow fields are composed of both Note: There are other Exercise: Repeat the exercise of Slide 1-3 (derive BE) for irrotational flow.
factors that can cause
• irrotational regions with negligible shear forces rotation, but they are
not as common as • In an irrotational flow BE is valid between any two points of the flow field, not
• and rotational regions with considerable shear forces
viscous effects. necessarily two points on the same streamline.

• Sometimes rotational regions will be very thin such as high speed external flow
over an airfoil. 1 2

• But still neglecting them and assuming the flow to be totally irrotational would 3 Inviscid, irrotational
yield unrealistic results. flow over an object

Assuming external flow over a body to

𝐹drag = ? be inviscid and irrotational everywhere
will result in zero air drag, which is not
𝑝 𝑉2 𝑝 𝑉2 𝑝 𝑉2
correct. This is known as d’Alambert’s + +𝑧 = + +𝑧 = + +𝑧
paradox. 𝜌𝑔 2𝑔 1
𝜌𝑔 2𝑔 2
𝜌𝑔 2𝑔 3

Exercise: What will happen if we assume pipe flow with uniform entrance to be
inviscid and irrotational?
1-7 1-8
 Velocity Potential (𝜙) Potential Flow
𝜕𝜙 𝜕𝜙
• For an irrotational flow : 𝛻 × 𝑉 = 0 • For a 2D flow in the 𝑥𝑦 plane : 𝑉 = 𝛻𝜙 → 𝑢= , 𝑣=
𝜕𝑥 𝜕𝑦
• As studied in ME 210, curl of the gradient of any scalar function is zero. 𝜕𝜙 1 𝜕𝜙
• For a 2D flow in the 𝑟𝜃 plane : 𝑉 = 𝛻𝜙 → 𝑉𝑟 = 𝜕𝑟
, 𝑉𝜃 = 𝑟 𝜕𝜃
𝛻 × (𝛻𝜙) = 0

• Using these two equations we can define a velocity potential function (𝜙) as • If the irrotational flow is also incompressible (In ME 306 we’ll NOT study
compressible irrotational flows)
𝑉 = +𝛻𝜙
Continuity Equation : 𝛻⋅𝑉 =0
Some books use a minus sign so that 𝜙
decreases in the flow direction, similar to Phi: A scalar function 𝛻 ⋅ 𝛻𝜙 = 0
temperature decreasing in the heat flow called velocity potential
direction. But we use plus in this course. 𝛻2𝜙 = 0 Laplace’s equation

𝛻 2 = 𝛻 ⋅ 𝛻 : Laplace operator
• In an irrotational flow field, velocity vector can be expressed as the gradient of a
scalar function called the velocity potential.
• For an incompressible and irrotational flow, velocity potential satisfies the
Laplace’s equation. These flows are called potential flows.

1-9 1-10

Velocity Potential (cont’d) Potential Flow Exercises

𝜕2 𝜙 𝜕2 𝜙
Exercise : Using Cauchy Riemann equations show that streamfunction also satisfies
In the 𝑥𝑦 plane : + 𝜕𝑦 2 = 0 the Laplace’s equation for incompressible, potential flows.
𝜕𝑥 2
• 𝛻2𝜙 = 0
1 𝜕 𝜕𝜙 1 𝜕2 𝜙 Exercise : Show that constant streamfunction lines (streamlines) are always
In the 𝑟𝜃 plane : 𝑟 𝜕𝑟
𝑟 𝜕𝑟 + 𝑟2 𝜕𝜃2 = 0
perpendicular to constant velocity potential lines for incompressible, potential flows.
Streamfunction

• Note that the relation between 𝑉 and 𝜙 is similar to that of 𝑉 and 𝜓. Exercise : Draw constant velocity potential lines of the following flow fields for which
streamlines are shown. Constant velocity potential lines and streamlines drawn
together form a flow net. What’s the ‘‘heat transfer’’ analogue of a flow net?
Cauchy Riemann Equations
Flow near a corner Flow over a cylinder
In the 𝑥𝑦 plane In the 𝑟𝜃 plane

𝜕𝜙 𝜕𝜙 𝜕𝜙 1 𝜕𝜙
𝑢= 𝑣= 𝑉𝑟 = 𝑉𝜃 =
𝜕𝑥 𝜕𝑦 𝜕𝑟 𝑟 𝜕𝜃
𝜕𝜓 𝜕𝜓 1 𝜕𝜓 𝜕𝜓
𝑢= 𝑣=− 𝑉𝑟 = 𝑉𝜃 = −
𝜕𝑦 𝜕𝑥 𝑟 𝜕𝜃 𝜕𝑟
1-11 1-12
 Potential Flow Exercises (cont’d) Potential Flow Exercises (cont’d) Outer region
Exercise : The two-dimensional flow of a nonviscous, incompressible fluid in the Exercise : A horizontal slice through a tornado is 𝑦
vicinity of a corner is described by the stream function modeled by two distinct regions. The inner or core Inner 𝑟
region (0 < 𝑟 < 𝑅) is modeled by solid body rotation. region
𝜓 = 2𝑟 2 sin(2𝜃) 𝜃
The outer region (𝑟 > 𝑅) is modeled as an irrotational
where 𝜓 has units of m2/s when 𝑟 is in meters. Assume the fluid density is region of flow. The flow is 2D in the 𝑟𝜃-plane, and the 𝑥
1000 kg/m3 and the 𝑥𝑦 plane is horizontal. components of the velocity field are given by
a) Determine, if possible, the corresponding velocity potential. 𝜔𝑟 0<𝑟<𝑅
b) If the pressure at point 1 on the wall is 30 kPa, what is the pressure at point 2? 𝑉𝑟 = 0 , 𝑉𝜃 = 𝜔𝑅2 Inner
𝑟>𝑅 region Outer region
Reference: Munson’s book. 𝑟 0
𝑦 where 𝜔 is the magnitude of the angular -0.2
velocity in the inner region. The ambient
pressure (far away from the tornado) is 𝑝 − 𝑝∞ -0.4
1 equal to 𝑝∞ . Obtain the shown nondimen- 𝜌𝜔 2 𝑅2 -0.6
𝑟 sional pressure distribution.
0.5 m -0.8
𝜃
𝑥 Reference: Çengel’s book. -1
2 0 1 2 3 4 5
1m 1-13 𝑟/𝑅 1-14

Superposition of Elementary Potential Flows 1. Uniform Flow

• Laplace’s equation is a linear PDE. • Consider uniform flow in the 𝑥𝑦 plane in +𝑥 direction.
• Superposition can be applied to both velocity potential and streamfunction. 𝑢=𝑈 , 𝑣=0
• Let’s find the equation for velocity potential.
Potential flow 1 Potential flow 2 Potential flow 3 𝜕𝜙 𝜕𝜙
𝑢= → 𝑈= → 𝜙 = 𝑈𝑥 + 𝑓(𝑦)
+ = 𝜕𝑥 𝜕𝑥

𝜕𝜙 𝜕𝜙 𝑑𝑓
𝑣= → 0= → =0 → 𝑓 = constant
𝜕𝑦 𝜕𝑦 𝑑𝑦

• 𝜙1 + 𝜙2 = 𝜙3 , 𝜓1 + 𝜓2 = 𝜓3 , 𝑉1 + 𝑉2 = 𝑉3
• Taking 𝑓 = 0 for simplicity
• To obtain complicated flow fields we can combine elementary ones such as 𝜙 = 𝑈𝑥
• Uniform flow • Constant 𝜙 lines correspond to constant 𝑥 lines, i.e. lines parallel to the 𝑦 axis.
• Line source/sink
• Vortex Exercise : Show that streamfunction equation is 𝜓 = 𝑈𝑦
1-15 1-16
 1. Uniform Flow (cont’d) 2. Line Source at the Origin
• Constant 𝜙 and constant 𝜓 lines are shown below. • Consider the 2D flow emerging at the origin of the 𝑥𝑦 plane and going radially
outward in all directions with a total flow rate per depth of 𝑞.

𝑦 View from the top

𝑦
𝑥 𝑥 𝑦
Streamlines
𝜓 = 𝜓1
𝜓 = 𝜓0 𝑥

𝜙 = 𝜙1
𝜙 = 𝜙0
𝑞 : Flow rate per depth Constant
(Strength of source [m2 /s]) 𝜙 lines
𝑈
𝑞
Exercise : Determine the equations of 𝜙 and 𝜓 𝑦 • Conservation of mass: 𝑞 = 2𝜋𝑟 𝑉𝑟 → 𝑉𝑟 = , 𝑉𝜃 = 0
2𝜋𝑟
for uniform flow in a direction making an angle 𝑥
of 𝛽 with the 𝑥 axis. 𝛽 • 𝑉𝑟 decreases with 𝑟, i.e. effect of source diminishes with 𝑟.
• Origin is a singular point with 𝑉𝑟 → ∞, which is not physical, so don’t get too close.
1-17 1-18

2. Line Source (cont’d) 2. Line Source (cont’d)

• Let’s find the equation for velocity potential. • Consider a line source that is located NOT at the origin.
𝜕𝜙 𝑞 𝜕𝜙 𝑞 • Equations for 𝜙 and 𝜓 change as follows
𝑉𝑟 = → = → 𝜙= ln(𝑟) + 𝑓(𝜃) 𝑦
𝜕𝑟 2𝜋𝑟 𝜕𝑟 2𝜋 𝑞 (𝑥, 𝑦)
𝜙= ln 𝑟1
2𝜋
1 𝜕𝜙 1 𝑑𝑓 𝑑𝑓 𝑟
𝑟1
𝑉𝜃 = → 0= → =0 → 𝑓 = constant 𝑞
𝑟 𝜕𝜃 𝑟 𝑑𝜃 𝑑𝜃 𝜓= 𝜃 𝜃1
2𝜋 1 𝑏
𝜃
• Taking 𝑓 = 0 for simplicity or equivalently using 𝑥 and 𝑦 coordinates 𝑎
𝑥
𝑞 𝑞
𝜙= ln(𝑟) 𝜙= ln 𝑥−𝑎 2+ 𝑦−𝑏 2
2𝜋 2𝜋
• Constant 𝜙 lines correspond to constant 𝑟 lines as shown in the previous slide. 𝑞 𝑦−𝑏 Some useful relations
𝜓= arctan
2𝜋 𝑥−𝑎 𝑥 = 𝑟 cos(𝜃)
𝑞 𝑦 = 𝑟 sin(𝜃)
Exercise : Show that the streamfunction equation is 𝜓 = 2𝜋 𝜃
𝑟= 𝑥2 + 𝑦2
𝑦
• To study a line sink for which the flow is radially inward towards a point, simply 𝜃 = arctan
𝑥
use a negative 𝑞 value.
1-19 1-20
 3. Irrotational Vortex 3. Irrotational Vortex (cont’d)
• Studied in ME 305 as free vortex. Its velocity components are 𝑉𝜃 = 𝐾/𝑟 • Strength of a vortex is not given by 𝐾. Instead we use its circulation Γ.
𝐾 • Circulation is the line integral of the tangential component of the velocity vector
𝑉𝜃 = , 𝑉𝑟 = 0
𝑟 around a closed curve. It is related to the rotationality of the flow.

• Using Cauchy Riemann relations we get 𝑦 Closed

𝜙 = 𝐾𝜃 , 𝜓 = −𝐾 ln(𝑟) curve 𝐶 Γ= 𝑉 ⋅ 𝑑𝑠 [m2 /s]
𝐶
𝑑𝑠 𝑉
• Compared to a line source, streamlines and constant potential lines are interchanged. Differential vector
along the path of
Closed path of
integration
integration
𝑦
Streamlines
Similar to a line source/sink, 𝑥
origin is a singular point,
𝑥
where the velocity shoots to
Constant 𝑉 = 𝑢𝑖 + 𝑣𝑗
𝜙 lines infinity. • For the 2D flow in the 𝑥𝑦 plane
Γ= (𝑢𝑑𝑥 + 𝑣𝑑𝑦)
shown above
𝑑 𝑠 = 𝑑𝑥 𝑖 + 𝑑𝑦𝑗 𝐶
1-21 1-22

3. Irrotational Vortex (cont’d) 3. Irrotational Vortex (cont’d)

Exercise: Calculate the circulation of an irrotational vortex for the following curve 𝐶 Exercise: A liquid drains from a large tank through a small opening as illustrated. A
vortex forms, whose velocity distribution away from the opening can be
𝑦 approximated as that of a free vortex. Determine an expression relating the surface
Curve 𝐶 shape to the strength of the vortex Γ.
𝑉𝜃 = 𝐾/𝑟
Reference: Munson’s book
𝑥
𝑅

𝑧
𝑝𝑎𝑡𝑚 𝑟
• Irrotational vortex is irrotational everywhere except the origin. All the circulation is 𝑦
squeezed into the origin, which is a singular point. 𝑥

• Circulation Γ = 2𝜋𝐾 can be understood as the strength of the vortex. It’s in m2 /s. 𝑧 𝑟 =?

Γ Γ
𝜙= 𝜃 , 𝜓=− ln(𝑟)
2𝜋 2𝜋

• Direction of the vortex is determined as Γ > 0 : CCW (+𝑧) rotating vortex

Γ < 0 : CW (−𝑧) rotating vortex
1-23 1-24
 Exercises for Elementary Potential Flows Source in a Uniform Flow
Exercise : Elementary components of a potential flow of water is shown below. (Flow Past a Half Body)
Find the velocity and pressure at point A if the pressure at infinity is 100 kPa. Exercise: Study the flow obtained by the combination of uniform flow in 𝑥
direction and a source at the origin. Obtain the location of the stagnation point(s)
and draw the stagnation streamline.
𝑦
𝑈 𝑈 = 3 m/s 𝑦
𝛼 𝑈
𝛼 = 30°
A 𝑞 = 10 m2 /s 𝑥
𝑞
𝑎 𝑏 𝑎 = 0.8 m
𝑥 𝑏 = 0.6 m
Source (𝑞) Sink (−𝑞)
Uniform flow: 𝑢 = 𝑈 , 𝑣 = 0 , 𝜙 = 𝑈𝑥 , 𝜓 = 𝑈𝑦
𝑞 𝑞 𝑞
Exercise : For the previous problem determine the equations of velocity potential Source: 𝑉𝑟 = , 𝑉𝜃 = 0 , 𝜙 = ln 𝑟 , 𝜓= 𝜃
2𝜋𝑟 2𝜋 2𝜋
and streamfunction by superimposing elementary flows. Find the velocity at point
A by differentiating both 𝜙 and 𝜓 equation.
1-25 1-26

Flow Past a Half Body (cont’d) Flow Past a Half Body (cont’d)
𝑦 Exercise : Consider the top part of a half body. Draw speed vs. 𝜃 and pressure vs. 𝜃
𝑞/2𝜋𝑟 using the following values: 𝜌 = 1000 kg/m3 , 𝑈 = 5 m/s , 𝑞 = 10 m2 /s and
𝑈 𝑈 𝑝∞ = 100 kPa.

Stagnation Exercise: A 64 km/h wind blows toward a hill that can be approximated with the
𝑠
𝑞 𝑥 streamline top part of a half body. The maximum height of the hill approaches 60 m.
𝜓 = 𝑞/2
a) What is the air speed at a point directly above the origin (at point 2)?
b) What is the elevation of point 2?
Stagnation point c) What is the pressure difference between point 2 and point 1 far from the hill?
𝑟𝑠 = 𝑞/2𝜋𝑈 Movie Reference: Munson’s book
Flow Over Half Body
𝑦
• Flow outside the stagnation streamline resembles a flow over 64 km/h 𝑔
a body with a blunt nose.
2 60 m
• Equation of the half body is given by the equation of the
stagnation streamline. 𝑥
1
1-27 1-28
 A Source and a Sink in Uniform Flow Flow Past a Rankine oval (cont’d)
(Flow Past a Rankine oval) 𝑦
• Superposition of 𝑦 𝑈
𝑈
• a source of strength 𝑞 at 𝑥 = −𝑐,
• a sink of strength −𝑞 at 𝑥 = 𝑐 and 𝑐 𝑐
𝑥 𝑥
• uniform flow of magnitude 𝑈. 𝑞 −𝑞

𝑐 𝑐

𝑞 𝑞
• 𝜙 = 𝜙𝑢𝑛𝑖 + 𝜙𝑠𝑜𝑢 + 𝜙𝑠𝑖𝑛𝑘 = 𝑈𝑥 + ln 𝑟1 − ln(𝑟2 ) 𝑦
2𝜋 2𝜋 Exercise : Determine the location of the stagnation points of the shown Rankine
A oval. Determine its length and thickness of the oval. Plot the variation of speed
𝑞 𝑞
• 𝜓 = 𝜓𝑢𝑛𝑖 + 𝜓𝑠𝑜𝑢 + 𝜓𝑠𝑖𝑛𝑘 = 𝑈𝑦 + 𝜃1 − 𝜃 𝑟1 and pressure (with respect to 𝑝∞ ) on it as a function of 𝜃.
2𝜋 2𝜋 2 𝑟2
𝜃1
𝜃2
𝑐 𝑐 𝑥

1-29 1-30

Doublet A Doublet in Uniform Flow (Flow Past a Cylinder)

• Superposition of • Superposition of
• a source of strength 𝑞 at the orgin (moved from – 𝑥 axis to the origin), • a doublet of strength 𝑑 at the origin.
• a sink of strength −𝑞 at the origin (moved from +𝑥 axis to the origin) , • uniform flow of magnitude 𝑈 in +𝑥 direction.
• Consider the limiting case of the source and sink of Slide 1-29 approaching to the
origin. Skipping the details we can get 𝑦
𝑈
𝑑 𝑑
𝜙𝑑𝑜𝑢𝑏𝑙𝑒𝑡 = cos 𝜃 , 𝜓𝑑𝑜𝑢𝑏𝑙𝑒𝑡 =− sin 𝜃
2𝜋𝑟 2𝜋𝑟
𝑥
𝑑
where 𝑑 is the strength of the doublet.
𝑦
• Velocity field is given by Streamlines
𝑑
𝜕𝜙 𝑑 Constant 𝜙 lines
• 𝜙 = 𝜙𝑢𝑛𝑖 + 𝜙𝑑𝑜𝑢𝑏𝑙𝑒𝑡 = 𝑈𝑥 + 2𝜋𝑟 cos(𝜃)
𝑉𝑟 = =− cos 𝜃
𝜕𝑟 2𝜋𝑟 2
𝑥 𝑑
1 𝜕𝜙 𝑑 • 𝜓 = 𝜓𝑢𝑛𝑖 + 𝜓𝑑𝑜𝑢𝑏𝑙𝑒𝑡 = 𝑈𝑦 − 2𝜋𝑟 sin(𝜃)
𝑉𝜃 = =− sin 𝜃
𝑟 𝜕𝜃 2𝜋𝑟 2
1-31 1-32
 Flow Past a Cylinder (cont’d) Flow Past a Cylinder (cont’d)
𝑈
Important results are as follows 𝑦 • Pressure distribution on the cylinder is 𝛽
(using BE with 𝑉∞ = 𝑈 and 𝑝∞ )
• Velocity components are
𝑅 1
𝜌𝑈 2
𝑅2 𝑅2 𝑥 𝑝𝑐𝑦𝑙 = 𝑝∞ + 1 − 4 sin2 (𝜃)
𝑉𝑟 = 𝑈 1 − 2 cos 𝜃 , 𝑉𝜃 = −𝑈 1 + 2 sin 𝜃 𝑑 2
𝑟 𝑟 0 Experimental
• Pressure on the cylinder is symmetric
with 𝑅 = 𝑑/𝑈 with respect to both 𝑥 and 𝑦 axis. 𝑝𝑐𝑦𝑙 − 𝑝∞
-1
• Pressure does not create any drag 𝜌𝑈 2
• Stagnation points are located at (−𝑅, 0) and (𝑅, 0). 2
force (in 𝑥 direction) or any lift force -2
• Stagnation streamline is a circle of radius 𝑅. (in 𝑦 direction). Potential
2 -3
• Velocity distribution on the cylinder is 0 𝜋/2 𝜋
𝑦 𝑝𝑐𝑦𝑙 𝛽
𝑉𝜃 𝑐𝑦𝑙 2𝜋
𝑉𝜃 𝑐𝑦𝑙 = −2𝑈 sin(𝜃) 1 𝐹𝑑𝑟𝑎𝑔 = − 𝑝cyl cos 𝜃 𝑅𝑑𝜃 = 0
𝑈 0 𝑅𝑑𝜃
2𝜋 𝐹𝑑𝑟𝑎𝑔 𝜃
𝑥
𝐹𝑙𝑖𝑓𝑡 = − 𝑝cyl sin 𝜃 𝑅𝑑𝜃 = 0
0 0
0 𝜋/2 𝜋 𝐹𝑙𝑖𝑓𝑡
𝜃 1-33 1-34

Flow Past a Cylinder (cont’d) Flow Past a Cylinder (cont’d)

• As seen from the above exercise potential flow theory predicts ZERO DRAG FORCE
𝑅𝑒𝐷 = 1.5 𝑅𝑒𝐷 = 26
on the cylinder.
• Actually this is the case for any closed body, irrespective of its shape.
• This result is not physical and it is known as d’Alembert paradox (1752).
• In a real viscous flow
• shear stresses inside the boundary layer will cause a frictional drag force.
• viscous action will cause separation & the pressure at the front and back of the
cylinder would not be symmetric.

Movie Movie
Flow with Separation Potential vs. Viscous

𝑅𝑒𝐷 = 2000 ‘‘An Album of Fluid Motion’’, by M. Van Dyke 𝑅𝑒𝐷 = 10000
1-35 1-36
 Flow Past a Cylinder (cont’d) Flow Past a Rotating Cylinder
Exercise : When a small circular cylinder is placed in a uniform stream, a • Superposition of
𝑦
stagnation point is created on the cylinder. If a small hole is located at this point, 𝑈
• a doublet of strength 𝑑 located at the origin,
the stagnation pressure, can be measured and used to determine the approach
velocity, 𝑈 (similar to a Pitot tube). • CCW rotating irrotational vortex of strength Γ 𝑑 Γ
𝑥
located at the origin,
a) Show how 𝑝𝑠𝑡𝑎𝑔 and 𝑈 are related. Pressure far away is 𝑝∞ .
• uniform flow of magnitude 𝑈.
b) If the cylinder is misaligned by an angle 𝛼, but the measured pressure is still
interpreted as the stagnation pressure, use potential flow theory to determine an • This will result in
expression for the ratio of the true velocity, 𝑈, to the predicted velocity, 𝑈′. Plot
this ratio as a function of 𝛼 for the range 0° < 𝑎 < 20°. 𝑅2 Γ
𝜙 = 𝑈𝑟 1 + cos 𝜃 + 𝜃
Reference: Munson’s book. 𝑟2 2𝜋
with 𝑅 = 𝑑/𝑈
𝑈 𝑈 𝑅2 Γ
𝜓 = 𝑈𝑟 1 − 2 sin 𝜃 − ln 𝑟
𝑟 2𝜋
𝛼

𝑅
• This is the potential flow that resembles the flow over a rotating cylinder of
radius 𝑅.
Stagnation point
1-37 1-38

Flow Past a Rotating Cylinder (cont’d) Flow Past a Rotating Cylinder (cont’d)
𝑦 • Streamlines and stagnation points for different circulation values.
𝑈

Γ
𝑑 𝑥
𝑅 𝑅
Γ
Γ=0 <1
4𝜋𝑈𝑅

Exercise : For the flow shown above, obtain the following results
Γ
𝑉𝜃 𝑐𝑦𝑙 = −2𝑈 sin 𝜃 + White’s book
2𝜋𝑅
2
𝜌𝑈 2 2Γ Γ
𝑝𝑐𝑦𝑙 = 𝑝∞ + 1 − 4 sin2 𝜃 + sin 𝜃 −
2 𝜋𝑈𝑅 2𝜋𝑈𝑅
Γ Γ
=1 >1
4𝜋𝑈𝑅 4𝜋𝑈𝑅
Integrate the above pressure distribution to get the following results

𝐹𝑑𝑟𝑎𝑔 = 0 , 𝐹𝑙𝑖𝑓𝑡 = −𝜌Γ𝑈 (per unit depth)

1-39 1-42
 Magnus Effect Magnus Effect (cont’d)
• Magnus Effect: A rotating body in a uniform flow will have a net lift force on it (1853). Exercise : Magnus effect acts not only on cylinders but also on other rotating bodies
such as spheres. It can be used to explain how a spinning ball moves in a curved
• Direction of the lift force depends on the direction of 𝑈 and Γ.
trajectory. A football player wants to make a penalty kick as sketched below. Will a
CW or a CCW spin do the trick?
Top Bottom
Γ Γ
𝑉𝜃 𝑡𝑜𝑝 = −2𝑈 + 𝑉𝜃 𝑏𝑜𝑡𝑡𝑜𝑚 = 2𝑈 +
2𝜋𝑅 𝑈 2𝜋𝑅

Opposite signs Γ 𝜌Γ𝑈 Same signs

Velocity is low. Velocity is high.
Pressure is high. Pressure is low.
?

Exercise : Watch the following movies

𝑈
https://www.youtube.com/watch?v=2OSrvzNW9FE
Exercise: Determine the direction of
? Γ http://www.youtube.com/watch?v=23f1jvGUWJs
the lift force. http://www.youtube.com/watch?v=2pQga7jxAyc
http://www.youtube.com/watch?v=wb5tc_nnMUw
1-41
(Suprising applications of Magnus effect)
(Magnus force on Veritasium channel)
(Enercon's rotor ship. Audio in German)
(Roberto Carlos knows the Magnus force)
1-42

Magnus Effect (cont’d) Kutta Condition (Lift on an Airfoil)

• Warning: Although potential flow theory can predict the direction of lift force • Magnus effect applies not only to cylinders but any closed shape.
due to Magnus effect correctly, it may give quite inaccurate values for its
• Consider the flow over a slender body with a sharp trailing edge, such as an airfoil.
magnitude. We’ll come back to this in the next chapter.
• An airfoil is designed to generate small drag and high lift force.
Exercise: In 1920s Anton Flettner built a series of rotor ships that are propelled
by rotating cylinders driven by electric motors. Read about Flettner’s ship at
rexresearch.com/flettner/flettner.htm and understand how it works.

𝑠1
𝑠2

• There are two stagnation points, 𝑠1 and 𝑠2 .

• Experiments show that the streamlines leave the trailing edge smoothly as shown
above, known as the Kutta condition.
One of Flettner’s original rotating cylinder ships https://en.wikipedia.org/wiki/E-Ship_1

1-43 1-44
 Lift on an Airfoil (cont’d)
• Potential flow theory will predict an
unphysical location for point 𝑠2 .
• It is impossible for streamlines to make
such a sharp turn at the trailing edge.
Simulating Flows Near Walls using Mirror Images
• Consider a line source located at a distance 𝑏 to a solid wall.
• Fluid cannot pass across the wall and therefore it is a streamline.
𝑠2
𝑠1 • The effect of the wall can be simulated by using another source, which is the
mirror image of the original one with respect to the wall.

• If we add the correct amount of CW vortex to this flow field we can bring point 𝑠2
down to the trailing edge and obtain the correct streamline pattern.

+ =
Γ
Wrong Correct
Çengel’s book
• The magnitude, Γ, of the necessary vortex can be used to calculate the lift force
generated on the airfoil.
𝐹𝑙𝑖𝑓𝑡 = 𝜌 Γ 𝑈 Kutta-Joukowski Law (1902)
1-45 1-46

Simulating Flows Near Walls using Mirror Images (cont’d) Superposition Exercises
Exercise : Consider a source of strength 𝑞 located close to two walls forming a 90o Exercise : We want to study the potential flow over the following bodies. If
corner. possible, which elementary flows need to be superimposed to get the desired
shape?
a) How many and where the mirror images need to be placed to simulate
existence of the walls?
𝑦
b) Locate the stagnation point(s).
𝑈 𝑈 𝑈
c) Draw the streamlines.
Source (𝑞)
𝑎

𝑥
𝑎
𝑈 𝑈

Exercise : Repeat the previous exercise by replacing the source with a clockwise
vortex of strength Γ.

1-47 1-48
 Numerical Solution of Potential Flow Numerical Solution of Potential Flow (cont’d)
• Obtaining complicated flow fields by superposing elementary ones is limited. • Boundary conditions are (study in the given order)
• To study potential flows on arbitrary geometries one can perform a numerical 2 Top wall is a streamline. 𝜓 should be constant
solution. there. In order to have 10 m2/s flow rate per
At the inlet 𝑢 = 10. depth between the top and bottom walls
• Consider a flow inside an expanding duct (coordinates are in meters). 3
𝜕𝜓
Therefore 𝜕𝑦
= 10. 𝜓top = 10
(0,2) (3,2) 𝜓 varies linearly from 0 to 10.
10 m/s
(0,1) 𝜓left = 10𝑦 − 10
(1,1) 5 m/s
𝑦
(3,0)
𝑥 (2,0) At the exit 𝑢 = 5.
4 𝜕𝜓
Therefore 𝜕𝑦
= 5.
𝜓 varies linearly from 0 to 10.
Bottom wall is a streamline. 𝜓
• Potential flow inside the duct can be obtained by solving Laplace’s equation 1
should be constant there. Let’s set 𝜓right = 5𝑦
𝜕2𝜓 𝜕2 𝜓 it to zero.
+ =0 , with proper boundary conditions
𝜕𝑥 2 𝜕𝑦 2 𝜓bottom = 0

1-49 1-50

Numerical Solution of Potential Flow (cont’d) Numerical Solution of Potential Flow (cont’d)
• Finite Difference Method can be used to get the numerical solution. • Discretized form of the Laplace’s equation at node (𝑖, 𝑗) is
• First we discretize the problem domain into a set of nodes. 𝜓𝑖+1,𝑗 − 2𝜓𝑖,𝑗 + 𝜓𝑖−1,𝑗 𝜓𝑖,𝑗+1 − 2𝜓𝑖,𝑗 + 𝜓𝑖,𝑗−1
+ = 0
• Following mesh has 55 nodes with Δ𝑥 = Δ𝑦 = 1/3. Δ𝑥 2 Δ𝑦 2
• 27 of the nodes are at the boundaries and 𝜓 is known at these nodes.
𝜕2𝜓 𝜕2𝜓
• 28 of the nodes are inside the domain and 𝜓 needs to be calculated at them. 𝜕𝑥 2 𝜕𝑦 2
𝑖,𝑗 𝑖,𝑗

5 point computational stencil • For Δ𝑥 = Δ𝑦, discretized equation for node (𝑖, 𝑗) becomes
𝑖, 𝑗 + 1
𝜓𝑖+1,𝑗 + 𝜓𝑖−1,𝑗 + 𝜓𝑖,𝑗+1 + 𝜓𝑖,𝑗−1 − 4𝜓𝑖,𝑗 = 0
Δ𝑦 𝑖, 𝑗
𝑖 − 1, 𝑗 𝑖 + 1, 𝑗 • This equation needs to be written for all non-boundary nodes.
• For nodes with boundary neighbors, some 𝜓 values are known and they need to be
𝑖, 𝑗 − 1 transferred to the right-hand-side of the equation.
Δ𝑥
• At the end we’ll get a system of 28 equations for 28 unknowns and solve it.
1-51 1-52
 Numerical Solution of Potential Flow (cont’d) Numerical Solution of Potential Flow (cont’d)
𝜓 = 10
• Following solution is obtained using a mesh with Δ𝑥 = Δ𝑦 = 0.2. • After obtaining 𝑢 and 𝑣, pressures
can be calculated using the Bernoulli 𝑉𝑖𝑛 8
𝝍 = 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 1 0.00 10.00 10.00 10.00
equation. 𝑝𝑖𝑛
8.00 8.02 8.04 8.07 8.12 8.20 8.30 8.41 8.52 8.62 8.71 8.79 8.85 8.91 8.95 9.00 6
6.00 6.03 6.06 6.12 6.22 6.37 6.58 6.82 7.05 7.26 7.44 7.59 7.71 7.82 7.91 8.00
4.00 4.03 4.07 4.13 4.26 4.48 4.84 5.24 5.61 5.93 6.19 6.41 6.59 6.74 6.88 7.00
White’s book 4
2.00 2.02 2.05 2.09 2.20 2.44 3.08 3.69 4.22 4.65 5.00 5.28 5.50 5.69 5.85 6.00 2
𝝍 = 0.00 0.00 0.00 0.00 0.00 0.00 1.33 2.22 2.92 3.45 3.87 4.19 4.45 4.66 4.84 5.00 • For irregular geometries
0.00 1.00 1.77 2.37 2.83 3.18 3.45 3.66 3.84 4.00 modifications need to be done.
Red ones are 0.00 0.80 1.42 1.90 2.24 2.50 2.70 2.86 3.00 𝑝 − 𝑝𝑖𝑛
0.00 0.63 1.09 1.40 1.61 1.77 1.89 2.00 𝐶𝑝 = 2
boundary values Black ones are 0.00 0.44 0.66 0.79 0.87 0.94 1.00
𝜌𝑉𝑖𝑛 /2
calculated 0.00 0.00 0.00 0.00 0.00 0.00
White’s book 1.0
0.8 0.75
• After obtaining the 𝜓 values at the nodes, velocity components can be obtained 0.6 Upper
0.4 wall
𝜕𝜓 𝜓𝑖,𝑗+1 − 𝜓𝑖,𝑗−1 0.2
𝑢= → 𝑢𝑖,𝑗 = • Red node is NOT at ∆𝑥 distance from 0.0
𝜕𝑦 2∆𝑦 the central node. What can be done? -0.2
𝜕𝜓 𝜓𝑖+1,𝑗 − 𝜓𝑖−1,𝑗 -0.4 Lower wall
𝑣=− → 𝑣𝑖,𝑗 = − -0.6
𝜕𝑥 2∆𝑥 -0.8
• Different formulas need to be used at the boundary nodes.
1-53 1-54