DEPARTMENT OF MECHANICAL ENGINEERING

COURSEWORK SUBMISSION

CODE AND TITLE OF COURSEWORK

Course code:
MECH3004

Title:

Vibrations of a cantilever beam: experimental and

theoretical analysis

STUDENT NAME: LOC NGUYEN

DEGREE AND YEAR: BENG MECHANICAL ENGINEERING YEAR 3 (2016)
LAB GROUP: AM3
DATE OF LAB.SESSION: 22/1/2016
DATE COURSEWORK DUE FOR SUBMISSION: 5/2/2016
ACTUAL DATE OF SUBMISSION: 4/2/2016
LECTURER: DR YURIY SEMENOV
PERSONAL TUTORS NAME: DR SHIPLEY

RECEIVED DATE AND INITIALS:

I confirm that this is all my own work (if submitted electronically, submission will
be taken as confirmation that this is your own work, and will also act as student
signature)

Signed: LOC NGUYEN

University College London, Torrington Place LONDON WC1E 7JE

I) Introduction
1) Objectives:
The laboratory experiment aims to show the behaviour of a vertically clamped beam excited
by a range of frequency. By registering the displacement of the excited beam by a laser
sensors and knowing at which frequency the beam is excited, it is possible to plot the
amplitude ratio frequency graphs to determine the natural frequencies experimentally. The
experiment also shows how the mass, inertia and the material of the beam would affect its
behaviour under excitation. From the data, experimental values can be compared to
theoretical analysis results.

2) Theory:
Derivation of equations used in this report can be found at Laboratory Handout1

II) Experimental Procedure:

The experiment consists 3 main parts:
Part 1: Determining the beams natural frequencies without added mass
a) The frequency response graph of the bar was done by exciting the bar within the
range of 1 to 100Hz and collecting corresponding amplitudes. From such graph, the
approximate values of natural frequencies can be found.
b) To determine the natural frequencies (with accuracy of 1Hz), specific chosen
frequencies (close to the approximate values) were used to excite the beam and
their corresponding amplitudes were collected again.
Part 2: Node positions on the bar
a) The bar was then stimulated at the natural frequencies to observe the mode shape.
b) Different number of nodes for each mode shape were observed. Their positions
were found visually by marks on the bar or moving a pen along the bar to find no
vibration points.
c) The distance from the top clamped end to each node were given.
Part 3: Altering the beams natural frequencies by adding mass
a) Extra mass was attached to the node of the 2nd mode (close to the lower end of the
bar) to obtain a new frequency response graph.
b) From such graph, the effect of extra mass can be determined by comparing it to the
without added mass graph in Part 1.

Dr Yuriy Semenov University College London, E466: Vibrating Beam Lab hand-out A

III) Results
1) Determining the beams natural frequencies without added mass:
a) The frequency response graph of the bar without added mass:

Figure 1 Amplitude Ratio of the bar without added mass over a range of 1-100hz Frequency

From the graph, the natural frequency of the bar at the first 3 modes within the range of 1100hz can be determined as 3Hz, 17Hz and 48Hz approximately.

b) Determining the natural frequencies:

After the first run of simulation, the bar was excited again at different frequencies (that are
close to the observed natural frequencies from the Figure 1 to determine more accurate
natural frequencies:

Frequency
(Hz)

Stable
Amplitude at
the bottom
end of the
bar
1st Natural Frequency
2.8
13
2.9
65
3
16
-

Frequency
(Hz)

Stable
Amplitude at
the bottom
end of the
bar
2nd Natural Frequency
17
3.5
18
10
19
2.4
-

Frequency
(Hz)

Stable
Amplitude at
the bottom
end of the
bar
3rd Natural Frequency
46
1.45
47
2.05
48
2.25
49
1.39

Table 1 Frequency Response at closer range of 3 Natural Frequencies

From Table 1, the natural frequencies can be determined as 2.9Hz, 18Hz and 47Hz.

2) Node positions on the bar:

Three mode shapes were observed at the range of 1-100Hz frequency. The positions of
nodes measured from the top end are given as following:

Mode Node Positions (mm)

1st Node 2nd Node
1
2
3

148
92

162

Table 2 Node positions

3) Determining beams natural frequencies with added mass

After the first experiment, a mass was attached at the node of the 2nd mode shape of the
bar. The new plot of amplitude ratio over frequency is presented below:

Figure 2 Amplitude Ratio of the bar with added mass over a range of 1-100hz Frequency

Details of how adding mass would effect the frequency response graph will be discussed in
later section.

IV) Theoretical results and Discussion, Comparing of the

experimental results
1) Theoretical natural frequencies of the beam without added mass
The nth natural frequency of the beam can be determined using the following equation:
!" = (%&)( "

)*

+&,

Where:

)* is the flexural rigidity of the beam () is Youngs modulus and * is 2nd moment of
inertia).
+ is mass per unit length of the beam, which can also be found as:
+ = bh

& is length of the beam.

%& is the corresponding result to satisfy the shape function (2) from hand-out A2
! is the angular frequency (rad/s).

The moment of inertia of the beam can be calculated by using following equation:
* =

35
12

The material of the beam was given as Aluminium. Its density, Youngs Modulus can be
found in The Engineering ToolBox3
The set of variables used for this equation is shown below:

L (m)
1.9
b (m)
0.0762
h (m)
0.012
4
1.1 x 10-8
* (m )
69
) (GPa)
3
2712
8 (Kg/m )
2.48
+ (Kg/m)

Mode 1 Mode 2 Mode 3

%& 1.875
4.694
7.855
%
0.987
2.471
4.314

Table 3 Set of used variables and aL / a values

2
3

Dr Yuriy Semenov University College London, E466: Vibrating Beam Lab hand-out A
http://www.engineeringtoolbox.com

From given equations and set of values above, first three theoretical natural frequencies can
be calculated and compared with the experimental results:

Mode Shape
1
2
3

Theoretical 9 (Hz) Experimental 9(Hz)

2.71
2.9
17
18
47.6
48

Percentage Difference (%)

6.57
5.66
0.93

Table 4 Theoretical Frequency compared with Experimental Frequency

v Discussion
From Table 4, the percentage difference between experimental and theoretical
results of all 3 modes are below 10%. The biggest differences lie in 1st and 2nd mode
shape with 5-7% difference while 3rd mode has less than 1%. This difference can be
considered as perceivable but also insignificant.
The reason of these difference can be explained as following:

The deflection characteristics of the beam is calculated by using Euler-Bermoulli

Beam Theory which assumes the plane section remains plane after deformation
meaning there is no shear force.
Another beam theory would give more accurate results is Timoshenkos beam
theory which includes a rotation between cross section and the bending line.
Another source of errors can be impurities in material of the beam which can slightly
change all of its mechanical properties and alter the behaviour of the beam.
The beam is assumed to be vibrated without any damping. However, this can not be
the case in reality. Damping sources can be listed as:
o Structural Damping energy dissipation in a vibrating structure
o Vicious drag in mechanical system of actuator/vibration source
o Coulomb Damping from the friction between parts inside vibration source
Finally, the errors in measurement and calculation also need to be considered. The
list of errors can be listed as:
o The dimension of the beam is given without any of measurement at the point
of experiment.
o The delay from the laser sensor to the receiver.
o Other errors in truncation of decimal point during calculation.

2) Theoretical Node Positions

From hand-out A4, the mode shapes can be plotted using the following equation and data
from Appendix A
: 2 = ;< cos %2 cosh %2

cos %& + cosh %&

(sin %2 sinh %2)
sin %& + sinh %&

Dr Yuriy Semenov University College London, E466: Vibrating Beam Lab hand-out A

Where:
;< is an arbitrary constant and can be set equal to 1
% and %& are known from previous part.

Shape Function against Beam Length with A1 = 1

2.5
2
Shape Function

1.5
1
0.5

First Mode

0
-0.5 0

0.2

0.4

0.6

0.8

1.2

1.4

1.6

1.8

Second Mode
Third Mode

-1
-1.5
-2
-2.5

Beam Length, x (m)

Figure 3 Shape function over the beam length for first three mode shapes.

From Figure 3 and above equation, the theoretical node positions can be found by using
Bisection Iteration Method in Matlab (Appendix A).
1st Node Position (cm)
Mode
1
2
3

2nd Node Position (cm)

% Error

Theoretical
149
96

Experimental
148
92

0.68%
4.35%

% Error
Theoretical
165

Experimental
162

1.85%

Table 5 Theoretical Node positions comparing to Experimental Node Positions

v Discussion:
Again, the experimental node positions were given without any of remeasurement which
can explain in the slight difference between theoretical and experimental results. However,
as all the percentage differences are below 5% where 1st node of 2nd mode and 2nd node of
3rd mode have even below 2% difference. The experimental result can be used as
confirmation of theoretical result.
The locating node position process would require to run the experiment like in previous
part, so they both can share the same sources of error. The most likely error is
measurement error as the node positions were not located by using high accuracy device
(laser sensor)

3) Theoretical natural frequencies of the beam with added mass

Figure 4 Comparison of frequency response graphs between beam without added mass and with added mass respectively.
st
rd
(red lines show original natural frequency, yellow lines show the shift in 1 and 3 mode)

From Figure 4, it can be seen that the 1st and 3rd modes natural frequency have been
shifted to the left (decreased). Back to the natural frequency equation:
!" = (%&)( "

)*

+&,

From equation, + is in the denominator. The shift in frequency can be explained that as +
increases, the natural frequency should decrease. Assuming the added mass shares the
same material with the beam bar, its Youngs Modulus would stay the same. While 2nd

moment of inertia can be changed due to the added mass make a crossbar shape with the
beam bar, its effect is considered as negligible (the added mass is small compared to the
whole beam bar).
There is no change in the natural frequency of 2nd mode. This can be explained by the
location of attached extra mass - node of 2nd mode which is stationary in 2nd mode.
Because of it, the added mass has no effect on 2nd modes natural frequency and only
would effects the other modes.

v Discussion
By adding mass to one specific node changing its inertia, it provides a control on the
natural frequencies and a bigger range of working frequency. As from Figure 4, without
added mass, the beam has to work in 2.9 18Hz to avoid resonance. While with added
mass, the working range is now extended to 1 18Hz. Such way can help a specific
building avoid reaching its resonance at a specific frequency of vibration induced by wind
blowing to the structure or any other source of vibrations. However, the added mass may
have no effect on the resonance amplitude or can be negligible; another way of safety
would be required at 2nd natural frequency and higher.
This method of adding mass to specific node is very important if one can know the
frequency of wind or other sources of vibration and can help the structure prevent from
collapsing.

4) Changing the beam material from Aluminium to Steel

By changing the material of the beam bar and from equation below, natural frequencies of
the beam should be affected by the difference in mass, density or Youngs Modulus
between two materials:
!" = (%&)( "

)*

+&,

As there is no relationship between those above variables so the effect of changing material
is not predictable. To study this effect, the beam material is assumed to be Stainless Steel
instead of Aluminium with following properties5 and natural frequency would be
recalculated.

L (m)
1.9
b (m)
0.0762
h (m)
0.012
4
1.1 x 10-8
* (m )
200
) (GPa)
3
7700
8 (Kg/m )
2.48
+ (Kg/m)

Mode 1 Mode 2 Mode 3

%& 1.875
4.694
7.855
%
0.987
2.471
4.314

Table 6 Set of used variables and aL/a values

http://www.engineeringtoolbox.com

Mode
1

Natural Frequency (Hz)

% Difference
Stainless Steel Aluminium
2.74
2.71

17.2

17

48

47.8

1.03

Table 7 Comparing natural frequencies of a same beam bar with two different materials

v Discussion
From Table 7, It can be seen the effect of changing material is almost negligible with the
difference between the theoretical natural frequencies of two materials being only 1.03%.
The positions of the node should also not be changed as the equation describing the mode
shapes below doesnt get affected by changing of material properties such as mass or
Youngs Modulus.
: 2 = ;< cos %2 cosh %2

cos %& + cosh %&

(sin %2 sinh %2)
sin %& + sinh %&

V) Conclusion
The experiment successfully demonstrated the behaviour of a top-clamped beam being
vibrated and showed the importance of knowing natural frequency, mode shapes and node
positions. By knowing how they would affect the behaviour of the beam, a method of
control of natural frequency and vibration was discussed and successfully showed a
bigger working range of frequency. Theoretical analysis also agreed well with experimental
data and show how Euler-Bernoulli beam theory can model the behaviour of simple
structures at first few mode shapes. However, for more complex and bigger structures,
Finite Element Analysis is a better choice to achieve a higher accuracy.

VI) Appendix A Excel Data and Matlab Code

: 2 Mode 1

: 2

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9

0
-0.00950341
-0.037073127
-0.081299494
-0.140776631
-0.214106612
-0.29990508
-0.39680818
-0.503480677
-0.618625124
-0.740991956
-0.869390394
-1.002700029
-1.139882983
-1.279996534
-1.422206115
-1.565798584
-1.710195704
-1.854967751
-1.999847201

: 2 Mode 3

: 2

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8

0
-0.147394726
-0.496168477
-0.911124699
-1.271882861
-1.484413798
-1.491120158
-1.276764303
-0.868590017

: 2 Mode
2

: 2

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9

0
-0.055916504
-0.203224663
-0.411509911
-0.651152605
-0.894085547
-1.114669291
-1.29058121
-1.403623363
-1.440366518
-1.39256292
-1.257278048
-1.036710717
-0.737690468
-0.37086014
0.050431445
0.511487733
0.997824726
1.496879546
1.999814359

0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9

-0.330193516
0.250090268
0.77448492
1.152247403
1.314544559
1.225152945
0.885023405
0.329674438
-0.380570535
-1.176511999
-2.000484689

v Matlab Code Bisection Method

function m = bisection(f, low, high, tol)
disp('Bisection Method');
% Evaluate both ends of the interval
y1 = feval(f, low);
y2 = feval(f, high);
i = 0;
% Display error and finish if signs are not different
if y1 * y2 > 0
disp('Have not found a change in sign. Will not continue...');
m = 'Error'
return
end
% Work with the limits modifying them until you find
% a function close enough to zero.
disp('Iter
low
high
x0');
while (abs(high - low) >= tol)
i = i + 1;
% Find a new value to be tested as a root
m = (high + low)/2;
y3 = feval(f, m);
if y3 == 0
fprintf('Root at x = %f \n\n', m);
return
end
fprintf('%2i \t %f \t %f \t %f \n', i-1, low, high, m);
% Update the limits
if y1 * y3 > 0
low = m;
y1 = y3;
else
high = m;
end
end
% Show the last approximation considering the tolerance
w = feval(f, m);
fprintf('\n x = %f produces f(x) = %f \n %i iterations\n', m, y3, i-1);
fprintf(' Approximation with tolerance = %f \n', tol);

v Matlab Code Mode shape 2 bisection method between 1.4

and 1.5
clear all
aL=4.694;
a=aL/1.9;

fun=@(x) (cos(a*x) - cosh(a*x)) + ((cos(aL) +

cosh(aL))/(sin(aL)+sinh(aL)))*(sinh(a*x)-sin(a*x));
low = 1.4;
high = 1.5;

tolerance = 1e-6;
x = bisection(fun, low, high, tolerance);

v Matlab Code Mode shape 2 bisection method between 0.9

and 1
clear all
aL=7.855;
a=aL/1.9;
r=0;
fun=@(x) (cos(a*x) - cosh(a*x)) + ((cos(aL) +
cosh(aL))/(sin(aL)+sinh(aL)))*(sinh(a*x)-sin(a*x));
low = 0.9;
high = 1;
tolerance = 1e-6;
x = bisection(fun, low, high, tolerance);

v Matlab Code Mode shape 2 bisection method between 1.6

and 1.7
clear all
aL=7.855;
a=aL/1.9;
r=0;
fun=@(x) (cos(a*x) - cosh(a*x)) + ((cos(aL) +
cosh(aL))/(sin(aL)+sinh(aL)))*(sinh(a*x)-sin(a*x));
low = 1.6;
high = 1.7;
tolerance = 1e-6;
x = bisection(fun, low, high, tolerance);