SESG1005 – fluid mechanics

Christophe Dormenval 23825529

Laboratory Report-Boundary Layer Flow

The aim of the experiment was to measure the velocity distributions of laminar and turbulent
boundary layers –using a Pitot tube – and compare our experimental values with analytical
approximations. This would then enable us to examine the accuracy and the restrictions of these
estimates.

Background:

The two fundamental characteristics of a viscous flow are that it will have a continuous velocity and
that a region is observed close to a solid surface, where the velocity of the fluid relative to the
surface is zero. This is also known as the no-slip condition. The flow velocity increases rapidly away
from the surface before reaching the main stream velocity, also called the external flow velocity Ue.
The surface of the object and the main stream velocity limit a region known as the boundary layer (δ)
and its dimensions depends on the surface and form of the object as well as the fluid’s Reynolds
number (Re). The limits of δ can be expressed mathematically by 0 ≤dy/dδ≤ 1 and we also have 0≤U¿
1. Therefore the velocity U will approach the mainstream velocity in an asymptotic way. The
boundary layer is usually very thin leading to a high velocity gradient (dU/dUe). Beyond the boundary
layer, the increase in velocity is very small, leading to a small velocity gradient. Two distinctive flows
can be described: When both Re and background disturbances are low, the flow is smooth and is
called laminar. When Re is high and in flows over rough surfaces, the flows will be unsteady and
know as turbulent.

In this experiment, the boundary layer velocity of a laminar flow is first measured using a Pitot tube.
Using Bernoulli’s equation (P+ρgz+1/2ρU 2=constant), we can determine the velocity across the
boundary layer [appendix 1].The gradient U/Ue is then compared to an analytical approximation for
U y y 3 y 4
laminar flows:
Ue
=2 −
δ δ () ()
+
δ
(Polhausen’s relation), where δ is the boundary layer

thickness and y the position of the Pitot tube against the plate’s surface.

The Experiment is then carried out a second time but by fixing a piece of sand paper to the leading
edge of the plate, the original flow is tripped, leading to a turbulent boundary layer. The estimate
1
U y
velocity gradient for turbulent flow is =( ) where 5≤n≤ 9.
Ue δ
n

The Kármán-Polhausen approximation and The Blasius solution:

Theodore von Kármán introduced an equation to find the boundary layer thickness by using a
mathematical approximation. When applied to a laminar boundary layer on a flat plate with zero
pressure gradient, his equation can be simplified significantly. From these results, P.R.H Blasius was
able to develop solutions of the laminar boundary layer equation for this particular case [appendix 2].
The aim of the experiment is to determine whether our experimental values match the analytical
values obtained from Blasius’ method. The second part of the experiment is the analysis of a

1
 SESG1005 – fluid mechanics
Christophe Dormenval 23825529
Ship Science part 1

turbulent flow over the same smooth surface. This time the experimental values are to be compared
to a rough approximation.

Experimental procedure:

As explained previously and on appendix 1, Pitot’s tube relies on Bernoulli’s equation to find the flow
velocity at different points on the boundary layer. The difference between the stagnation pressure
and static pressure will give us the velocity at a distance y from the plate.

Results:

It is clear from appendix [2] that the extrapolated velocity curves of both flows are tending to zero at
y=[Link] the experimental values didn’t. This can be explained by the play in the thread to adjust
the tube distance to the surface. Another explanation could be that when aiming to set the tube as
close to the surface as possible, it could have been bent slightly and the first values would be
therefore representing the straightening of the tube is bending back into its original shape.

Because the velocity in the boundary layer increases asymptotically to the free stream velocity, we
use the value of 99% Ue to find the boundary layer limit. Therefore, from the graph, we can find the
boundary layer thickness δ (laminar flow) =0.03m (30mm) and δ (turbulent flow) = 0.11m (110mm).

We then use the velocity and δ values to normalize U/Ue and Y/δ. As explained previously, U/Ue can
also be found mathematically and the two are compared graphically in appendix […] and […], for
laminar and turbulent flow respectively. We can observe a significant difference between the values,
in particularly at the end points. The reasons for these differences will be discussed later on.

Finally, to assess the accuracy of the mathematical approximations, we first work out the Reynolds
numbers based on x (the distance from the leading edge of the plate to the measuring station) and
on δ. The Rex values can then be used to find δ, as δ/x=5.83*(Re x) ^ (-0.5) is a theoretical result for
laminar flow and δ/x=0.375*( Rex) ^(-0.2) is an approximation for turbulent flow. However for both
cases our experimental values of δ/x differ significantly from the theoretical ones (by approximately
0.92 and 0.93).

Discussion:

The main idea of this experiment was to compare two ways of finding U/Ue. But it is important to
mention that our main stream velocity value Ue was calculated using data measured at the beginning
of the experiment. Therefore we could assume that the value of Ue might be slightly wrong, as the
room pressure and temperature would have changed throughout the experiment. Indeed from
appendix [3], we can see that our value of U reached 21.98m/s whereas Ue=21.83 m/s. Perhaps it
would have been a good idea to take another set of measurements at the end of the experiment. The
power input of the engines used to generate the velocity could also be a reason for the differences
between our mathematical approximations and experimental values is the. It is likely that this
affected the accuracy of our readings as it was noticed that the difference in pressure would change
unpredictably throughout the experiment. Insufficient output for the engines meant that the velocity
created wasn’t constant.

2
 SESG1005 – fluid mechanics
Christophe Dormenval 23825529
Ship Science part 1

Conclusion: