Manesh Bhudia Group 1

Laminar and Turbulent Flow in a Pipe

Manesh Bhudia, Student, UCL Civil Engineering

Abstract – Pipes are integral when transporting fluids and play Poiseuille’s equation (1) describes the pressure drop in laminar
an important part in our daily lives. They can used to transport flow through a long cylindrical pipe with a constant cross
millions of barrels of oil from nation to nation, or deliver clean section.
water from treatment plants to the taps in our homes. This 𝜋𝑅 4
𝑄= (𝑝 − 𝑝𝐵 ) (1)
report will explore how energy is lost within pipes as a fluid 8𝜇𝐿 𝐴
flows through it.
Where Q is the flow rate through the pipe, μ is the viscosity of
Keywords – Empirical formula, Kinematic viscosity, Laminar water, L is the length of pipe and p is the pressure at a given
flow, Poiseuille equation, Reynolds number, Turbulent flow. point. The manometers are used to measure the pressure at each
point. The pressure head is converted to elevation head. If no
1. Introduction energy losses were incurred the head in manometers A and B
would be the same as the height of the head tank. However head
Due to its ability to with stand high pressures, its excellent
is lost as water flows through the pipe. The head at A is
thermal conductivity and reliability; copper is one of the most
significantly lower than that of the head tank because there is a
used material for plumbing in Europe. Copper pipes are used to
significant change in the cross sectional area at the inlet. As well
transport gas, drinking water and used widely in heat exchange
as this there are two elbows to change the direction of flow
systems. [1] This report will investigate the relationship between
before point A. Further head is lost between points A and B due
the head loss due to friction and the average flow velocity in a
to frictional losses due to viscosity.
small copper pipe during laminar and turbulent flow. As well as
determine the lower critical velocity, the co-efficient of viscosity Poiseulle’s equation can be manipulated to help calculate a
of water and the roughness of the copper pipe using value for the kinematic viscosity of a fluid (2), by using the
experimental data. 𝑄 (𝑝𝐴 −𝑝𝐵 ) 𝜇
following equations: = , ℎ𝑓 = and 𝑣 = .
𝜋𝑅2 𝜌𝑔 𝜌
2. Experimental Setup and Methodology
𝑔𝑅 2 ℎ𝑓
Figure 1 shows the experimental set up. Water will discharge 𝑣= ( )( ) (2)
8𝐿 𝑉
from the constant head tank, from which it will flow through a
copper pipe. The head will be measured at two points along the Where v is the kinematic viscosity of water, g is the acceleration
4.85mm diameter pipe, 1500mm apart from each other using due to gravity, ℎ𝑓 is the head loss due to friction and V is the
manometers. The flow rate through the pipe will be measured flow velocity through the pipe. A graph of hf against V can be
using a rotameter. A rotameter is a mechanical device that plotted as shown in figure 2. The gradient of the laminar flow
establishes an equilibrium position as a result of the downward can be used to calculate the kinematic viscosity, as the
force of gravity on a metal float and the upward force from the remaining values are constants.
flow of the water. Using a calibration relationship the flow rate
can be determined from this reading. [2]
A graph of the head loss due to friction agianst Velocity
By gradually closing valve C the flow rate was reduced by 3mm 800
intervals on the rotameter scale. At each interval readings of the 700
rotameter and both manometers were taken, in order to calculate
600
the head loss through the pipe.
500
hf (mm)

400
300
y = 0.1987x - 1.3087
200
100
0
0 200 400 600 800 1000 1200
Velocity (mms-1) Turbulent Laminar

Fig.2 A graph of the head loss due to friction against the flow velocity through the
pipe, for laminar and turbulent flow.

The gradient of the graph when the flow was laminar was
0.1987s. Hence the kinematic viscosity of water can be found
Fig.1 A diagram that illustrates the setup of the pipes in the laboratory. using Poiseuille’s equation:

9.81×2.4252
3. Results and Discussion 𝑣= ( ) (0.1987) = 0.955 mm2s-1
8×1500

I. Cross verify methods of calculating the kinematic viscosity of The empirical formula (3) can also be used to calculate the
water kinematic viscosity of water at a given temperature:
The kinematic viscosity of water can be calculated using 𝑣 = 1.78 (1 + 0.03368(𝑡℃) + 0.000221(𝑡℃)2 )−1 𝑚𝑚2 𝑠 −1
Poiseuille’s equation and using the empirical formula.
Manesh Bhudia Group 1

The temperature of the water remained 18℃ throughout the 𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 𝑖𝑛 𝑅𝑒𝑎𝑑𝑖𝑛𝑔 0.5𝑚𝑚
𝐸𝑅𝑜𝑡𝑎 = = = 3.85%
duration of the experiment, hence the kinematic viscosity of 𝑀𝑒𝑎𝑑𝑖𝑎𝑛 𝑅𝑒𝑎𝑑𝑖𝑛𝑔 13𝑚𝑚
water can be calculated to be:
The error in Q would be significantly different to the error in the
𝑣 = 1.78 (1 + 0.03368(18) + 0.000221(18)2 )−1 𝑚𝑚2 𝑠 −1 rotameter reading due to the large additive constant (AC). The
= 1.061 𝑚𝑚2 𝑠 −1 error in flow rate can be given as:

The values of v calculated using Poiseuille’s equation and the 𝑄 − 𝐴𝐶 3825 − 2765
𝐸𝑄 = × 𝐸𝑅𝑜𝑡𝑎 = × 3.85% = 1.07%
empirical formula are in agreement with values used in industry 𝑄 3825
of 1.0546mm2s-1. [3] There is considerably more error in the
value obtained through Poiseuille’s equation. Using the error in Q and the error in R, the error in V can be
calculated because V= Q/𝜋𝑅 2 :
II. Identify the lower critical point and compare with theoretical
values. 𝐸𝑉 = 𝐸𝑄 + 2𝐸𝑅 = 1.07% + 2(1%) = 3.07%

The critical velocity is the flow velocity of the water at which There was also an error in the measurement of Ha and Hb as the
the flow changes from laminar to turbulent or vice versa. At the standpipes could only be measured to the nearest 0.5mm. The
lower critical point the flow changes from turbulent to laminar. percentage error in hf is given as:
Plotting log(hf) against log(V) produces a graph (illustrated in
figure 3) with two straight lines, one for laminar flow and the 𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 0.5 0.5
𝐸ℎ𝑓 = + = +
other for turbulent. 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝐻𝑎 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝐻𝑏 1219 1179
= 0.08%

The error arisen in the kinematic viscosity can be calculated

using the percentage errors in V, hf and R. The error is the
summation of the error in the measured values:

𝐸𝑣 = 2𝐸𝑅 + 𝐸𝑉 + 𝐸ℎ𝑓 = 2(1%) + 3.07% + 0.08%

= 5.15%

This would result in an absolute error of ±0.049mm2s-1

(0.955mm2s-1× 5.15%) in the value of kinematic viscosity
calculated using Poiseuille’s equation.

IV. Calculating the friction factor.

Fig.3 A graph of log(hf) against log (V), for laminar and turbulent flow. 2𝑔 𝑑
Using the equation 𝑓 = ℎ𝑓 ( 2 ) ( ) to obtain a value of friction
𝑉 4𝐿
Using the graph in figure 3, the lower critical point can be factor for each flowrate, a graph of friction factor against
identified as the point at which the two lines intersect as Reynolds number can be plotted.
indicated in red. The lower critical point occurs when log (V) is
2.57 and when log (hf) is 1.86. The value of Reynolds number A Graph of Friction Factor Against Reynold's Number
that corresponds to this point is 1720. This is lower than the 0.025
expected range of 2000-4000. [4] The critical value at which
turbulence begins is sensitive to disturbances. A sudden 0.02

adjustment in flow when nearing the critical point may have Laminar Rough zone
Friction Factor

resulted in the value of Reynolds number being lower than 0.015

expected.
0.01

Theoretically the gradient of the laminar region of the graph will

0.005
be 1 and the gradient of the turbulent region will be between 1.7 Transition zone
and 2. If the pipe is relatively smooth the gradient will be nearer
0
to 1.7. [4] The experimental results are in agreement with the 100 1000 10000
Reynolds Number
theoretical predictions. The gradient of the turbulent region is
1.97, from which we can draw the conclusion that the copper
pipe was relatively rough. Fig.4 A graph of friction factor against Reynolds number for various flow rates
through the copper pipe.
III. Experimental error in the kinematic viscosity of water.
The graph begins with a slope with a negative gradient which
Due to the number of measurements taken to obtain a value of represents the laminar region. The graph the line then transitions
kinematic viscosity errors can accumulate. It is important to and plateaus where the flow is turbulent. The plateau occurs
calculate the magnitude of the error arisen to be able to justify when the friction factor is 0.009. By comparing this to the graph
the value obtained. The error occurring in the median value of of friction factor and relative roughness against the Moody
rotameter reading when flow was laminar (13mm) was Diagram [4] a value of relative roughness can be identified. The
calculated. This value was chosen as it best represents the error value of relative roughness corresponding to a friction factor of
in calculating viscosity. 0.009 is 0.008.
𝑘
There was an error of 1% in the measurement of the radius of Using the know values of relative roughness ( ) and the internal
𝑑
the pipe. The smallest increment on the rotameter was 1mm diameter (d) of the pipe the absolute roughness (k) can be
therefore it could be read to the nearest 0.5mm which lead to an calculated using the formula:
error in the rotameter reading of: Absolute roughness (k)
𝑘
= Relative roughness ( ) × Internal Diameter of pipe (d)
𝑑
Manesh Bhudia Group 1

= 0.008 X 4.85mm carried out. The value of velocity through the pipe can be taken
= 0.0388mm as 0.607ms-1.

This value of absolute roughness is significantly larger than the The volume flow rate can then be calculated:
expected value of 0.001mm-0.002mm for a new copper pipe. [5]
There is a significant discrepancy between the two values, the 𝜋0.0032
𝑄 = 𝑣𝐴 = 0.607 × = 4.29 × 10−6 𝑚3 𝑠 −1 (5)
value obtained experimentally is approximately twenty times the 4
magnitude of the theoretical value. This may largely be due to
The volume flow rate could also be calculated using a single
the copper pipe tested not being new. The roughness of pipes
equation that combines equations 4 and 5 with Haaland’s
increase with age because the interior surface of the pipe may
equation:
corrode or become dirty. This would lead to more irregularities
on the surface of the pipe allowing eddies to develop at lower
𝜋 2𝑔ℎ𝑓 0.5 𝑘 2.51𝑣
velocities. 𝑄 = − 𝑑 2.5 ( ) 𝑙𝑜𝑔 { + }
2 𝑙 3.71 × 𝑑 𝑑1.5 (2𝑔ℎ𝑓 /𝑙)0.5
−6 3 −1
V. Application of experimental results = 4.22 × 10 𝑚 𝑠

“A fluid with relative density σ=0.7 and kinematic viscosity

ν=0.6mm2 /s is used as a cooling agent in an electrical
equipment cooling system. The fluid circulates in a 2m long 4. Concluding remarks
copper pipe of 3mm in diameter. A pump is used to create a
In conclusion I believe the data obtained from the laboratory
head difference of 0.5m between entry and exit of the system.
experiment has verified that the numerous formulae used to
Use results of your small pipe experiment to calculate coolant
calculate the kinematic viscosity of water. The actual values
flow rate through the system.”
obtained experimentally agree with values obtained from books
Applying the steady flow of energy equation to the situation and journals, with discrepancies of up to 5% accounting for
gives the following: error. I believe that if this laboratory experiment was repeated a
more accurate value of viscosity could be obtained and the mean
𝑝1 𝑉12 𝑝2 𝑉22 value would be in closer agreement with the theoretical values
𝑧1 + + = 𝑧2 + + + ℎ𝑝 used in industry.
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔
𝑝 𝑣.2 4. References
Where z is the elevation head, is the pressure head, is the
𝜌𝑔 2𝑔
velocity head and hp is the head added by the pump. Assuming
the pipe lies horizontally z1=z2. Applying the continuity of flow
equation results in v1=v2 because the area of the pipe remains
constant. Hence: [1] The Copper Development Association, "Pipe Systems,"
2014. [Online]. Available:
𝑝1 𝑝2 𝑙 𝑉2 http://www.copperalliance.org.uk/applications/pipe-
ℎ𝑝 = − = ℎ𝑓 = 4𝑓 (3) systems.
𝜌𝑔 𝜌𝑔 𝑑 2𝑔

Where hf is the head loss due to friction, f is the friction factor, l [2] Sensors Online, "Flow The Basics of Rotameters," 2012.
is the length of the pipe, d is the diameter of the pipe, v is the [Online]. Available:
flow velocity and g is the acceleration due to gravity. http://www.sensorsmag.com/sensors/flow/the-basics-
rotameters-1068.
The diameter of the pipe (d) is known and having obtained a
value for the absolute roughness (k) of copper pipe
experimentally the relative roughness (k/d) for the pipe can be [3] International Towing Tank Conference, "Testing and
found. Extrapolation Methods, General Density and Viscosity of
Water," 1999. [Online]. Available:
𝑘
Relative roughness ( ) http://ittc.sname.org/2002_recomm_proc/7.5-02-01-03.pdf.
𝑑
= Absolute roughness (k) ÷ Internal Diameter of pipe (d)
= 0.0388mm ÷ 3mm [4] B. Massey, "Mechanics of Fluids," in 8th edition, 2006, pp.
= 0.0129 247-50.

Using this value the corresponding friction factor can be [5] The Engineering ToolBox, "Major loss in Ducts, Tubes
obtained using Moody’s diagram and an iteration process can be and Pipes," 2009. [Online]. Available:
used to calculate a more accurate value for velocity. The friction http://www.engineeringtoolbox.com/major-loss-ducts-
factor corresponding to 0.0129 is 0.01. This value of friction tubes-d_459.html.
factor can be used in the first iteration. The flow velocity
through the pipe could be found by rearranging equation (3):

ℎ𝑝 2𝑔𝑑 0.5 × 2 × 9.81 × 0.003

𝑉= √ = √ = 0.607𝑚𝑠 −1 (4)
4𝑓𝑙 4 × 0.01 × 2

The Reynolds number can be calculated as the kinematic

viscosity is known.

𝑉𝑑 0.607𝑚𝑠 −1 × 0.003𝑚
𝑅𝑒 = = = 3035
𝑣 0.6 × 10−6 𝑚2 𝑠 −1
The second iterative value of friction factor can be calculated
using Blasisus’s formula. This gives a friction factor of 0.0106.
This change can be considered insignificant. If the change was
larger, then a further iteration through equation 4 could be