Department of Mechanical Engineering MP2.

5 Fluid Friction in Pipes

nometer. Smooth pipes have approximate in-

Objectives ternal diameters and length as shown in table 1:
To investigate the effect of surface roughness Table 1: Pipe dimensions
and pipe diameter on the pressure losses in
pipes. Pipe Internal Length, Δx
Diameter, D (mm)
(mm)
Introduction Smooth small 13.6 912
When a fluid is made to flow past a solid Smooth me- 17 912
boundary, the wall will absorb momentum from dium
the fluid. The molecules adjacent to the wall will
Smooth large 26.2 912
come to rest relative to it and the fluid velocity
will vary as a function of the distance from the Rough medi- 17 200
solid boundary. Because of the existence of ve- um Mean effec-
locity gradients in the flow shear stresses will tive diameter
develop between the layers and as a conse- Dme = 14mm
quence energy will be transferred across the flu-
id layers and finally dissipated as heat. In order
to maintain the flow we need to provide energy
to overcome friction. In the case of flow in pipes Specific Objectives
this is readily demonstrated. There will be loss 1. Understand the effect of pipe diameter on
of pressure as the fluid flows which will have to friction factor. Plot friction factors vs. ReD for
be provided externally to maintain the flow. The
the small, medium and large diameter
aim of this experiment is to establish the relation
smooth pipes.
between the pressure drop due to friction and
the mean flow velocity. Using the expression: 2. Understand the effect of surface roughness
on friction factor. Plot friction factors vs. ReD
U 2 for the smooth and rough pipe of the same
p  f Darcy x
2D diameter.
to define the friction factor, the main aim of the Experimental Procedure
experiment is to investigate the behavior of fDarcy
as a function of Reynolds Number : 1. Close the Globe valve and the Ball valve
(light blue and grey circuits). Open the Gate
UD UD valve (dark blue circuit) half a turn.
Re D  
  2. Turn on the cold water supply and wait for
For a discussion of the significance of Reynolds any trapped air to leave the circuit, then close
number see Appendix 1; discussion of the fric- the Gate valve
tion factors follows in Appendix 2. 3. Connect one set of piezometer tubes to tap-
ings 13 (upstream) and 14 (downstream).
Apparatus
4. Use the hand pump if necessary to adjust the
The apparatus consists essentially of a number pressure in the piezometer tubes until the
of smooth tubes of different diameter and a levels are halfway up the scale. The level in
rough tube arranged between two headers, one each of the piezometer tubes should be the
connected to the flow in and the second con- same, if not check for air bubbles or leaks.
nected to the Flow out (see figure 1). Water at
high pressure is supplied from a hydraulic 5. Fully open the gate valve and wait for the
bench while the spent water is returned back to flow to settle. Record the readings on the pi-
the bench. The hydraulic bench is equipped with ezometer into table 2.
means for measuring the flow rate passing 6. Use the gate valve to reduce the flow rate in
through test pipe. The flow to each pipe is con- five suitable steps to give a good spread of
trolled individually by an isolating valve, while results.
the pressure difference can be measured using
7. Repeat for the other smooth pipes and the
two pressure tapings along the individual pipes.
rough pipe.
The pressure drop will be measured using ma-

MP2.5 - 1
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

Results and Discussion 5.4 (pp.128-140 in the 5th edition), also Ch.6.1,6.2
and 7.1-7.3
1. Calculate the values of the Darcy friction fac-
tor fDarcy for all measurements taken in points 2. J.K.Vennard, R.L.Street: Elementary Fluid Me-
5, 6 ad 7 in the Experimental Procedures chanics, 6th ed., SI Version, J.Wiley & Sons, New
above. York 1982, Chapter 9

2. Plot all the points of fDarcy vs ReD for smooth 3. J.E.A. John, W.L. Haberman: Introduction To
and rough pipes (point 7 in the Experimental Fluid Mechanics, 3rd edition, Prentice-Hall Inter-
Procedures above). State relative roughness national, London 1988, Chapter 6
as e/Dme, where e is equal to 0.5 mm and Dme 4. J.R.Welty, Ch.E.Wicks, R.E.Wilson: Fundamen-
is the mean effective diameter. Comment tals of Momentum,Heat and Mass Transfer, 3rd edi-
briefly on your results and possible sources tion, J.Wiley &Sons, New York 1984, Chapter 14
of error.
5. P.M.Gerhart, R.J.Gross: Fundamentals of F;luid
3. Calculate the friction factor using the Blasius Mechanics, Addison-Wesley, Reading (Ma.) 1985,
equation fBlasius for the smooth pipes (see fur- Ch.7
ther discussion of the friction factors in Ap-
6. R.W.Fox, A.T.McDonald: Introduction to Fluid
pendix 2).
Mechanics, J.Wiley&Sons, New York 1985, Ch.9
0.316 part B
.

4. Plot fDarcy and fBlasius vs. ReD for the smooth

pipes and comment on the results and possi-
ble sources of error.

Useful Data
Pressure:
1kN/m2 = 1kPa = 0.145 psi
1 mm Hg = 0.0394 in.Hg = 133.3 Pa
1 in.Hg = 3386 Pa
1 mm H2O = 9.807 Pa

101.32 kPa = 760 mm Hg = 14.70 psi=

1 atm (1kgf/m2) = 10.33 m H2O

1 bar = 100 kPa = 14.504 psi = 10,197 m H2O

Kinematic viscosity of water:

at 0oC  (m2/s): 1.785 x 10-6

5oC 1.518 x 10-6

10oC 1.306 x 10-6

15oC 1.139 x 10-6

20oC 1.003 x 10-6

25oC 0.893 x 10-6

Literature (general)
1. B.S. Massey : Mechanics of Fluids, 5th ed., Van
Nostrand Reinhold, Wokingham 1986, Ch.5.1 to

MP2.5 - 2
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

Figure 1 Schematic diagram of the test facility

MP2.5 - 3
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

Table 2: Smooth small diameter pipe

Internal Diameter (D)= Flow Cross Sectional Area (A)= Length (L)
Measurement Flow Rate (Q) Piezometer Readings Flow Velocity ReD fDarcy fBlasius
3 -1 -1
Number (m .s ) Upstream Downstream Difference (m.s )
Taping Taping (h)
(mm) (mm) (m)
1
2
3
4
5
6
Table 3: Smooth medium diameter pipe
Internal Diameter (D)= Flow Cross Sectional Area (A)= Length (L)
Measurement Flow Rate (Q) Piezometer Readings Flow Velocity ReD fDarcy fBlasius
3 -1 -1
Number (m .s ) Upstream Downstream Difference (m.s )
Taping Taping (h)
(mm) (mm) (m)
1
2
3
4
5
6

MP2.5 - 4
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

Table 4: Smooth large diameter pipe

Internal Diameter (D)= Flow Cross Sectional Area (A)= Length (L)
Measurement Flow Rate (Q) Piezometer Readings Flow Velocity ReD fDarcy fBlasius
3 -1 -1
Number (m .s ) Upstream Downstream Difference (m.s )
Taping Taping (h)
(mm) (mm) (m)
1
2
3
4
5
6
Table 5: Rough medium diameter pipe
Internal Diameter (D)= Flow Cross Sectional Area (A)= Length (L)
Measurement Flow Rate (Q) Piezometer Readings Flow Velocity ReD fDarcy fBlasius
3 -1 -1
Number (m .s ) Upstream Downstream Difference (m.s )
Taping Taping (h)
(mm) (mm) (m)
1
2
3
4
5
6

MP2.5 - 5
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

element should be equal to zero in any flow sit-

Appendix 1 - Flow Similarity and uation:
Reynolds Number
FG + Fp + F + FE + F + FI = 0
For the two different flow situations to be simi-
lar, three basic similarity laws must be obeyed: In many practical flow situations not all the
forces are present or some are negligibly small
1. Geometric similarity - the two solid bounda-
when compared with others.
ries of flow must be identical in shape and can
differ only in size (sometimes problems occur For flows of incompressible viscous fluids in
here - surface roughness can be difficult to scale long, completely filled pipes, Osborne Reyn-
down or up exactly). Flow patterns should be olds reasoned that to compare two flows, it is
geometrically similar. The ratio of some charac- sufficient to take into account the viscous and
teristic dimension L of the two solid boundaries inertia forces. (Gravity effects will be identical
is called the scale ratio Lr . For long circular and surface tension nonexistent provided there
pipes, fully filled with fluid, the pipe diameter D is no free surface, elasticity effects can be ne-
is usually taken as the characteristic dimension . glected for incompressible flows, pressure ef-
It follows that if the scale ratio is Lr, then area fects are only important for large changes of hy-
drostatic pressures, etc.)
ratio is Lr2 and volume ratio is Lr3 .
F 1 + FI 1 = 0 and F 2 + FI 2 = 0
2. Kinematic similarity = geometric similarity +
identical ratio of velocities at all corresponding For the two flows "1" and "2" to be similar, the
points of the two flow situations. three laws of similarity demand that the remain-
ing acting forces will act in the same directions
3. Dynamic similarity = geometric similarity + and be correspondingly proportional:
kinematic similarity + identical ratio of all forces
acting on corresponding fluid elements in the if F 1 = k F 2 , then FI 1 = k FI 2
two flow situations.
F 1 FI 1
Forces that may act on a fluid element can be since = ,
expressed in general terms as follows (where L F 2 FI 2
will be some characteristic dimension, A - char-
acteristic area, V - characteristic velocity, T - also
FI 1
=
FI 2
=
inertia forces
= constant
characteristic time): F 1 F 2 viscous forces
Gravity: FG= mg= L3g This ratio of (inertia forces)/(viscous forces) is of
Pressure: Fp= (∆p)A=(∆p)L2 great importance in comparing the two flow
situations in which these forces govern the flow.
Viscosity:
This ratio is now universally known as the
F= µ(dux/dy)A= µ(V/L)L2 = µVL Reynolds number (Re).
µ =  is the dynamic viscosity Referring back to the general notation of forces
  is the kinematic viscosity acting on a fluid element, the ReL , based on
dux /dy is the local velocity gradient in the some characteristic dimension L, can be written
direction normal to ux as:
Elasticity: Re L =
inertia forces
=
V 2 L 2 = rV L = V L
FE= EvA= EvL2 viscous forces µVL µ 
(Ev is the bulk modulus of elasticity) In any consistent system of units, Reynolds
number will be dimensionless. It always advisa-
Surface tension: F= L
ble to check whether it really is for the values
 is the surface tension you put in, thus detecting any inconsistencies.
Inertia: When fluid flows in a long circular pipe (and
FI= - ma = - L3(L/T2 )= - rL4T-2= - (L/T)2 L2= - the entry effects are no longer important), pipe
 V 2 L2 diameter D becomes the characteristic dimen-
sion L. Average flow velocity u becomes the
a is the resultant acceleration
characteristic velocity V. ReD based on the
in general terms a = L/ T2 , V = L/T
pipe diameter is:
According to Newtonian laws of mechanics, the
vector sum of all these forces acting on a fluid Re D = V D = V D
µ 
MP2.5 - 6
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

For the two flows of viscous incompressible flu- Darcy: fD Fanning: fF

ids in long circular, completely filled pipes to be
similar, their ReD numbers have to be identical. u 2 L u 2 L
p  f D p  4 f F
N.B. For flows in which other forces are im- 2 D 2 D
portant, these are taken into account and various for laminar flow ( ReD = uD/ < 2300 ), from
other dynamic similarity criteria can be devel- theoretical solution by Hagen and Pouseuille :
oped. As an example let us consider the flows in
which compressibility effects become important 64 16
fD  fF 
(such as flows of gases with high velocities). Re D Re D
Universally known Mach number (M) becomes
an important criterion of dynamic similarity : for turbulent flow the friction factors depend on
surface roughness. Numerous formulae have
inertia forces V 2 L 2 been developed from theoretical and experi-
M = =
elasticity forces Ev L
2 mental studies. Of particular interest is the theo-
retical solution by Blasius for very smooth
= V2 = V2 = V pipes:
Ev c2 c
 0.316 0.079
fD  fF 
where c is the velocity of sound. The proof that Re D 0.25 Re D 0.25
Ev/= c is a bit lengthy but simple and can be
The most accurate values for rough pipes can be
found in most thermodynamics or fluid me-
obtained from a chart of f vs ReD. These charts
chanics textbooks, in chapters dealing with
compressible flows. are referred to by various names, as they were
developed over many years by Stanton (1914),
Appendix 2 - Friction Factors and the Nikuradse (1933) , Colebrook (1939) and others.
Moody Chart In 1944 Moody prepared a chart for use with
commercial pipes. Apart from the charts, vari-
Because of their practical importance, pressure
ous algebraic formulae, usually accurate to with-
losses both in laminar and particularly in turbu-
in a few percent, are to be found in textbooks.
lent flow in pipes have been extensively studied
theoretically and experimentally. Full treatment A sample of the Moody chart is enclosed here
of these studies is way beyond the scope of this for plotting the results of fD obtained from ex-
instruction; it is dealt with in the Fluid Mechan- perimental ∆p and the definitions given above.
ics course.
N.B. The easiest way to find out whether the
For practical engineering use it is convenient to chart refers to fF or fD (often simply f is stated)
be able to predict pressure losses ∆p (kN/m2), is to look at the formula given on the chart for
or head losses hL (which are expressed in meters the laminar flow line. f = 64/Re means the
of water head and related to ∆p by: ∆p = ghL ), chart refers to fD , f = 16/Re means the chart
in terms of the Reynolds number. refers to fF

To that end we use the friction factor f, which is Friction in Non-circular conduits: All the above
a function of ReD . Two parallel definitions of calculations and use of charts can be applied to
flows in closed ducts of reasonably any non-
the friction factor are in common use: the Darcy
circular cross-section.). An "equivalent diame-
friction factor fDarcy (often quoted as fD or simp-
ter" is used, calculated according to:
ly as f ), preferred by mechanical engineers, and
the Fanning friction factor fF, often used by Deq  4   Hydraulic Radius Rh 
chemical and process industries engineers. The
relation between the two friction factors is sim-  cross-sectional area of flow 
 4 
ple:  wetted perimeter 
fD = 4 fF, but if proper care is not taken, using the
The accuracy is deteriorating for very odd
values obtained from the chart for one of them
shapes, such as rectangles with one side more
in calculations relating to the other will lead to
than 8 times greater than the other side, etc., and
serious errors. Choose just one definition for
is not always good for laminar flow.
your calculations.
One often encountered example is the flow in an
Definitions of friction factors and calculation of
annular area between two concentric pipes (of-
pressure drop (or head loss):
ten a design for a simple heat exchanger):

MP2.5 - 7
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

 The Moody Chart (Darcy friction factor fD):

Cross-sectional area of flow =
4
D
2
o  Di2 
(from R.W.Fox, A.T.McDonald: Introduction to
Fluid Mechanics, J.Wiley&Sons, New York 1985)
Wetted perimeter =  Do  Di 

4
D 2
o  Di2  Relative roughness e/D is calculated from abso-
Deq =4   Do  Di lute roughness e , given for commercial pipes
  Do  Di  as:
Accuracy should be reasonable, although some Pipe material: e (mm)
textbooks warn against using the calculation if riveted steel 1-10
the outer diameter is less than 3 times the inner concrete 0.3-3
wood stave 0.2-1
diameter.
cast iron 0.25
galvanized steel 0.15
asphalted cast iron 0.12
commercial steel or wrought iron 0.045
drawn tubing 0.0015

MP2.5 - 8
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

MP2.5 - 9
Department of Mechanical Engineering MP2.5 Fluid Friction in Pipes

This page is intentionally blank

MP2.5 - 10