Ecole Centrale de Nantes

Fluid Mechanics II
Labwork III Report

INTERNAL FLOWS IN PIPES

Youssef El Iskandarani
[email protected]

El Mehdi Es-sabar
[email protected]

08/05/2022
El Iskandarani, Es-sabar Flow measurement and head loss

1. Introduction
The aim of this work is the measurement of velocities and head loss for laminar and turbulent
regimes along a pipe. Those two studies will both be conducted for laminar and turbulent regimes.
The tools used in the experiment are: a thermometer that is used in order to measure the temper-
ature due to its relation with the effect of viscosity and density on the oil that is used as a working
fluid in the experiment. A pitot tube is also used to measure the stagnation pressure through the
experiment in order to quantify the local velocities. To measure the distances of the pipe surface,
a micrometer was used in the experiment.

The control of the flow’s regime will be done through changing the flow rate in our system.
This is parctically done using a valve that we either open or close in order to allow more or less
oil in our system. The experimental facility used in this work includes a convergent device that
facilitates the laminarization of the flow. Because of that, obtaining a fully turbulent flow will be
done using an increase in the system’s flow rate, but also by introducing a screw in our pipe which
induces turbulence.

After quantifying the pressure drop along the pipe, we will be able to calculate the head losses
using relevant relationships. In order to characterize the flow regime, the Reynolds number will be
calculated using a measured flow rate. This will allow us to come up with a value of the friction
factor. A comparison will be made with theoretical results in order to test the accuracy of meas-
urements and the theory used.

Next, we will tackle the establishment of velocity profiles through the pipe for the turbulent and
laminar regimes. This will also be done by measuring the stagnation pressure in different radial
locations of our pipe. A comparison will then be done with theoretical data, but also between
laminar and turbulent results in order to draw conclusions regarding the two.

1
El Iskandarani, Es-sabar Flow measurement and head loss

2. Experimental Facility
The experiment is divided into several sections. First, it involves a pump that allows the flow
circulation and is also linked to the pressure chamber. A long pipe of 20 pressure outlets along it
is used to connect the pressure chamber with the storage tank. So, circulation of the fluid works
as follows, the pump operates by pumping the fluid towards the pressure chamber that contains
a convergent nozzle and a manometer. Then, the fluids move from the pressure chamber to the
storage tank though a long pipe and repeats as follows. The working fluid used in this experiment
is oil and the information that governs it are read from the lab Manuel. A valve is also present
in order to allow the control of flow and fluid pressure. To convert from laminar to turbulent
flow through the experiments a screw is placed before the chamber. The long pipe of internal
diameter of 0.019m has 18 pressure outlet connected to the pressure gauge to be able to read the
pressure along them. A Pitot or Prandtl tube is located near the pipes exist to allow the pressure
measurements of different points through the pipe. As opening the valve wider, the mass flow rate
changes through the experiment and should be measures. To proceed with the measurement a
weighing bench is placed on the storage tank. A thermometer provided through the experiment to
obtain the temperature of the working fluid.

Figure 1: Experimental facility

The pressure drop depends on internal diameter. With time in service, the interior of the
pipe becomes covered with dirt, scale and it is often practical to take into account the expected
diameter changes. Also, during the usage of the pipe, roughness may be expected to increase due
to corrosion depending on the pipe material and nature of the fluid.

2
El Iskandarani, Es-sabar Flow measurement and head loss

3. Results
This section of our work exposes the different results onvolved with our study. the first set of
results is related to the head loss calculation through the pipe in both the turresubresubulent and
laminar regimes along with the quantification of the friction factor. The second set of data is the
velocity profiles that were also established for both the turbulent and laminar regimes.

3.1 Study of friction head loss

This first subsection of our results tackles the head loss calculation in our pipe for both the
laminar and turbulent regimes. The calculation of the friction factors will be done, along with the
comparison to theoretical calues.

3.1.1 Laminar flow

Figure 2 illustrates the pressure evolution with the length o our pipe. The different measurements
are the ones done in the corresponding probes. The uncertainty of each measurement is also
illustrated.

Figure 2: Pressure readings along the pipe - Laminar flow

As the position along the pipe changes from point 1 to 17, the pressure slope drops due to
friction forces. So, the pressure at the outlet is always smaller than the pressure at the inlet. The
convergent in the pressure chamber cause the nonlinearity in the graph, that are seen between
point 1 and 14, producing parallel stream line in small displacements without having the need to
place a very long pipe. Dispersion of streamlines is occurred due to the small pipe diameter of
the experiment; were they takes a period of time to become parallel with the needed formation.
Through point 14 to 18, the changes show linearity due to pressure change and constant density
and gravity in head losses.

Pressure drop & Head loss

At the entrance of the pipe, the flow is not fully established and the reading values are therefore
not accurate. From the related figure, we can even see how the uncertainty is higher for those
particular readings. The probes chosen are then among the very mast ones, i.e. 14 and 17. Those
latter probes will allow us studying the highest portion of the pipe where the flow has higher
chances to be fully established. From the pressure figure, the slope of the pressure drop is nearly

3
El Iskandarani, Es-sabar Flow measurement and head loss

constant after probe 14, unlike before it where it fluctuates. The pressure drop is then found as:

∆P = P17 − P14 = 0.0305 − 0.1175

−→ ∆P = -0.087 b

The head loss is then found using the following equation:

∆H 1 P17 − P14
=−
L ρg L
−→ ∆H = 1.07m

Friction factor
We first find Ud using the measured mass flow rate of the flow found to be 0.46512kg/s. The mean
velocity is calculated as follows:
ṁ
ṁ = ρ.U.A −→ U = 2
ρ.π D4
−→ U = 1.9691m/s

Then, we can calculate the friction factor of our pipe using the following equation:
P14 − P17 D 2g
λ=
ρg L Ud2

The friction factor is then found to be:

−→ λexperimental = 0.045125

Reynolds number
We calculate the Reynolds number using the following equation. The value of the kinematic
viscosity is taken from the labwork file graph at a recording temperature of 19.5°C. It is found to
be ν = 1.525 × 10−5 .
U.D
Re =
ν
−→ Re = 2459.7

Even though the Reynolds number is greater than 2300, we still have a laminar flow due to the
convergent device used in the facility. The values are then plotted in the graph, i.e. log(Re) = 3.39.
The result is shown in the figure below. In order to plot our point, we used interpolation as follows:
66.5mm
(3.39 − 3.1) × = 42.29mm (distance f rom origin of plot)
3.556 − 3.1

4
El Iskandarani, Es-sabar Flow measurement and head loss

Figure 3: Pressure readings along the pipe - Turbulent flow

Then , again by using interpolation, we found that log(100λ) = 0.42, meaning:

−→ λtheoretical = 0.0263

By using the theoretical formula λ = 64/Re, we also get λ = 0.0260 which represents a 1.15%
difference.

When comparing this value with our measured value, we find a difference of

Difference 100% = 41.7%

which is very high. The accuracy of our measurements is unfortunately unsatisfactory. The
significant error may mostly be due to the inaccuracy in measuring the mass flow rate, in the
uncertainty of pressure readings, or to human error. We however think it would mostly be due
to an inaccurate mass flow rate measurement as our pressure reading produced very satisfactory
results for the next sections, for the turbulent flow and for the establishment of velocity profiles.
Also, since we adopted the same measuring methodology throughout the entirety of the experiment,
we do not believe human error is the main cause of this large difference.

5
El Iskandarani, Es-sabar Flow measurement and head loss

3.1.2 Turbulent flow

Figure 4 illustrates the different pressure readings done in our pipe for the turbulent regime along
with the uncertainty involved with each.

Figure 4: Pressure readings along the pipe - Turbulent flow

Same exact procedure is repeated for the turbulent flow and as can be seen through the data
obtained in the graph that the turbulent flow shows higher pressure through the point in the pipe
than laminar flow. The turbulent flow increases the loss of energy in the friction form which then
generates heat. As the pipe length increases, friction pushes against the flow, resulting in pressure
loss.

Pressure drop & Head loss

Similarly to the laminar flow case, we use the pressure readings at probes 14 and 17. The pressure
drop is then found as:

∆P = P17 − P14 = 0.1655 − 0.535

−→ ∆P=-0.3695 b

The head loss is then found using the following equation:

∆H 1 P17 − P14
=−
L ρg L
−→ ∆H = 4.6m

Friction factor
We first find Ud using the measured mass flow rate of the flow found to be 1.052632kg/s. The
mean velocity is calculated as follows:
ṁ
ṁ = ρ.U.A −→ U = 2
ρ.π D4
−→ U = 4.4630m/s

Then, we can calculate the friction factor of our pipe using the following equation:
P14 − P17 D 2g
λ=
ρg L Ud2

6
El Iskandarani, Es-sabar Flow measurement and head loss

The friction factor is then found to be:

−→ λexperimental =0.037363
Reynolds number
We calculate the Reynolds number using the following equation. The value of the kinematic
viscosity is taken from the labwork file graph at a recording temperature of 21.5°C. It is found to
be ν = V = 1.475 × 10−5 .
U.D
Re =
ν
−→ Re = 5575.1
The values are then plotted in the graph, i.e. log(Re) = 3.75. The result is shown in the figure
below. In order to plot our point, we used extrapolation as follows:
66.5mm
(3.75 − 3.1) × = 94.8mm (distance f rom origin of plot)
3.556 − 3.1

Figure 5: Pressure readings along the pipe - Turbulent flow

Then , again by using extrapolation, we found that log(100λ) = 0.56, meaning:

−→ λtheoretical = 0.0363
By using the theoretical formula λ = 0.3164/Re0.25 , we also get λ = 0.0366 which represents a
0.82% difference.

In contrary to the laminar flow case, the experimental and theoretical are in convergence with
each other, i.e. respectively 0.03736 and 0.03645 which represents a percent difference of:
Difference 100% = 2.44%
Our measured results are satisfactory for this case. As discussed above in the laminar case section,
we believe the factor that induced the large error in the laminar case may mostly be an incorrect
mass flow rate measurement.

7
El Iskandarani, Es-sabar Flow measurement and head loss

3.2 Study of local velocities

This section of our work discusses rhe establishment of velocity profiles in our pipe for both the
turbulent znd laminar regimes. Those results will be compared to theoretical ones ThosiThosin
order to assess the accuracy of measurements. Conclusions about the different regimes will also be
drawn by comparing the two standard and nondimensional velocity distributions.

3.2.1 Laminar flow

This first part discusses the velocity profile establishment for the laminar regime. The figure below
illustrates the stagnation pressure readings for this first case along the radial distance pressuripres-
surin our pipe.

Figure 6: Pressure readings for velocity calculations

- Laminar flow -

As the fluid flows along the pipe there will be a pressure drop that occurs as a result of res-
istance to flow. The resistance of flow in the pipe causes frictional forces that act on a fluid as it
moves through the pipe. The main elements of fluid flow resistance are fluid velocity along the pipe
and fluid viscosity. Another factor is the increase in fluids kinetic energy that results in pressure
drop. As the flow velocity increase means more kinetic energy, which causes a pressure drop to
match that energy.

Based on this set of data of pressure readings, we calculated the corresponding velocities using
Bernoulli’s equation as follows:
ρ 2
Preading − P∞ = .V
2 ∞
with Pinf being the reading of probe 18. Our velocities are then calculated as follows:
r
2
V∞ = .(Preading − P18 )
ρ
Figure 7 illustrates the result of establishing the velocity profile for the laminar regime.

8
El Iskandarani, Es-sabar Flow measurement and head loss

Figure 7: Local velocity profile

- Laminar flow -

At laminar regime of flow, the velocity (Vmax) is highest on the pipe axis, while as moving
toward the wall the velocity converges to zero. It is called a no slip boundary condition. Also,
because of the effects of friction on the fluid near the wall, the flow is very slow if not zero due
to intermolecular forces. This actually makes calculations easy by giving us a smooth gradient of
velocity that traces back to zero at the walls of the pipe. As can be seen the velocity profile is
parabolic in shape with the maximum velocity in the center being about triple the velocity near
the walls.

Figure 8: Nondimensional velocity profiles (Measured vs. Theoretical)

- Laminar flow -

The blue line shows the measured data of local velocity with respect to maximum velocity,
and the red line represents the theoretical data, that is calculated by expression of the velocity

9
El Iskandarani, Es-sabar Flow measurement and head loss

field. As shown in the graph, the measured and the theoretical data are close enough to each
other which represents how good the accuracy of our results are. Do not forgot the convergent
in the pressure chamber, that causes the slight deviation in the results allowing the measured
data to be higher than the theoretical one’s due slight velocity increase formation. The laminar
regime of flow points out on an important aspect, which the parabolic development of velocity field.

The theoretical expression of the velocity field is as follows:

V (r) r2
=1− 2
Vaxis R
with Vaxis found to be 4.5821m/s. The calculated percent difference between the measurements
and theoretical values is found to be:
Difference 100% = 3.80%
which is very satisfactory. Our measured values are therefore in accordance with the theory. This
somehow eliminates the fact that the large difference in the laminar flow calculations was due to
pressure readings or to human error.

3.2.2 Turbulent flow

This second part of our work discusses the results we got for the turbulent regime. The figure
below illustrates the pressure measurements for this regime in different radial positions along the
pipe.

Figure 9: Pressure readings for velocity calculations

- Turbulent flow -

As opposed to the laminar pressure distribution, the turbulent one is characterized by a flatter
curve resulting in pressures that are closer to the maximum one when we move away from the
pipe’s center. The pressure then suddenly decreases as we reach the wall. The uncertainty related
to each measurement is also shown in the figure above.

As discussed beforehand for the laminar case, the velocities are calculated based on the recorded
pressure data using the following equation derived from the Bernoulli’s one:
r
2
V∞ = .(Preading − P18 )
ρ

10
El Iskandarani, Es-sabar Flow measurement and head loss

Figure 10: Local velocity profile

- Turbulent flow -

In the turbulent regime of flow, the velocity profile is no longer parabolic. The velocity of pipe
in the turbulent regime is higher. The graph of the turbulent regime shows a more flatter velocity
profile along the radius. In turbulence, the velocity profile is flatter in the central part of the pipe
and then it drops rapidly extremely close to wall surface due to diffusivity of the flow.

Figure 11 shows the nondimensional velocity profile. The measured and theoretical curves are
plotted in order to evaluate the accuracy of our experimental procedure.

Figure 11: Nondimensional velocity profiles (Measured vs. Theoretical)

- Turbulent flow -

The empirical equation for the velocity profile is used. This latter is characterized by its com-
plexity and non-uniformity when compared to laminar flows. Figure 11 shows that our data is in

11
El Iskandarani, Es-sabar Flow measurement and head loss

satisfactory accordance with the theoretical ones. The measured data will always be bigger than
the theoretical one like in laminar flow due to the placement of convergent in the pressure chamber.
The results of the velocity are flatter than the laminar ones also due to flow diffusivity.

The theoretical expression of the velocity field is as follows:

V (r) r 1
= (1 − ) n
Vaxis R
with n = 6 based on our Reynolds number (Re = 5575.1). Vaxis is also found to be 5.8169m/s.
The calculated percent difference between the measurements and theoretical values is found to be:

Difference 100% = 2.07%

which is again very satisfactory. Our measured values are therefore in accordance with the theory.

3.2.3 Laminar VS. Turbulent flow

This section exposes a comparison between laminar and turbulent velocity profiles. This com-
parison was done using the measured and theoretical values, but also with the nondimensional
quantities as shown in figures 12 and 13.
Figure 12 illustrates the theoretical and measured velocity profiles for both the laminar and
turbulent regimes versus the rdial distance in the pipe. The pipe walls are plotted in green.

Figure 12: Velocity profiles comparison (Measured & Theoretical)

- Laminar vs. Turbulent -

Figure 12 shows a comparison between the laminar and turbulent velocity profiles. For the
turbulent regime of flow, the velocity is higher on the pipe axis and on the rest of regions in the
duct. This is mostly due to the higher flow rate in the turbulent case, but also due to the fact that
it has a higher energy. A turbulent profile is when the velocity is greater than a critical threshold.

12
El Iskandarani, Es-sabar Flow measurement and head loss

Also, when going towards the pipe’s wall, the slope is sharper for the laminar profile. However, it
is extremely steep for the turbulent flow in regions very close to the wall as this latter is zero due
to the no-slip condition.

Figure 13: Nondimensional velocity profiles comparison (Measured & Theoretical)

- Laminar vs. Turbulent -

Figure 13 shows the nondimensional velocity profiles plotted against the nondimensional radial
position along the pipe. Both curves evolve in the same manner as in figure 12 when moving away
from the pipe’s center and when getting close to the edge of the pipe. The maximum values are
however the same, i.e. 1, since in both regimes the velocity reaches its maximum value at the
center. This type of comparison also allows us to say that when moving towards the pipe’s edge,
velocities are closer to the maximum value for the turbulent regime than for the laminar one. This
means that velocity decreases faster when moving away from the center for the leminar regime.

13
El Iskandarani, Es-sabar Flow measurement and head loss

4. Conclusion
The aim of this work was to study the flows inside a pipe in turbulent and laminar regimes by
evaluating the head losses and velocity profiles induced by each. We started our study by meas-
uring the pressure in different positions throughout our system, and used the ones where the flow
has more chances to be fully established for our readings. Those readings allowed us to derive the
head loss through the system for each regime. With our measurement of the flow rate, we were
able to quantify the friction factor for each regime, and to compare the results with theoretical
data. We thereby found that the percent difference between our result and the theory was 41.7%
and 2.44% for the laminar and turbulent regimes respectively. The very high difference recorded
for the laminar case can be induced by different causes. However, it is thought that is caused by
a faulty mass flow rate reading as our pressure recordings resulted in a satisfactory result for the
turbulent case and for velocity profiles, but also as we adopted the same measurement methodology
throughout the whole experiment. Indeed, proper measurement is a researcher’s skill that should
be polished and perfectionned.

Moreover, for each regime, we established the velocity profile by recording the stagnation pres-
sure in different radial positions in our pipe. The results converged with the theory, as we found
mean percent differences of 3.8 and 2.07% for the laminar and turbulent regimes respectively.
When comparing the two regimes, we found that the turbulent flow had a higher maximum velo-
city than the laminar one. This is due to the higher flow rate. A turbulent flow indeed has more
energy than a laminar one. We also found that the velocity profile for a laminar regime had an
approximately parabolic shape, while the turbulent regime has a curve that is flatter. When mov-
ing towards the pipe’s edge, velocity for the turbulent regime are also higher than for the laminar
one. With our nondimensional comparison, we were also able to say that when moving away from
the pipe’s center, velocities are closer to the maximum value for the turbulent regime than for the
laminar one. This means that velocity decreases faster when moving away from the center for the
laminar regime. The velocity then suddenly decreases to zero at the boundary for the turbulent
case due to the no-slip condition.

As to sum up, our experience with this labwork allowed us to draw conclusions about turbulent
and laminar regimes in a pipe with their influence on head loss and velocity profiles. Our personal
experience also underlined the importance of vigor in the experimental methodology and in data
recording as this is a crucial step in a researcher’s or scientist’s journey.

14