UNIVERSITY OF LONDON

MEng EXAMINATIONS 2006

Part III, Mechanical Engineering

for Internal Students of the Imperial College of Science, Technology and Medicine
This paper is also taken for the relevant examination for the Associateship

MACHINE SYSTEM DYNAMICS

Wednesday 26 April: 1400 to1700

This paper contains SIX questions.

Attempt question 1 and FOUR others. 28% of the marks are allocated to question 1 and
18% to each of the other questions. The numbers shown by each part of a question are
for your guidance; they indicate approximately how the examiners intend to distribute
the marks for this paper.
A Data and Formulæ book is provided.

This is a CLOSED BOOK Examination

The rest of this page is intentionally blank

Page 1 of 7 Turn Over

Compulsory question

1. Figure Q1a shows the schematic layout for a wind tunnel which operates up to
airspeeds of 25 m/s. The air temperature is to be monitored by a thermocouple at
the point A, 100 mm from the wall of the tunnel. The thermocouple will be mounted
on the free end of a steel tube of 5 mm outside diameter, 2 mm internal diameter
built into the wall of the tunnel.

(a) Calculate the first natural frequency of the tube. (You may assume that a
cantilever may be modeled as a spring-mass system, the stiffness for a force
at the free end being given by the usual expression and the mass being 24%
of the actual mass) [6%]

(b) Estimate the maximum amplitude of vibration which the tube is likely to
experience in service. Assume that the frequency of vortex shedding from a
body is given by f = 0.2V/D where V is the flow velocity and D is the external
diameter of the body, and that the excitation is equivalent to a sinusoidal
force applied at the free end of the tube, amplitude 5 N and at the vortex
shedding frequency. The equivalent viscous damping ratio for the tube
(including the effects of the airflow) is 0.05. [6%]

(c) Estimate the maximum vibration induced stress experienced by the tube.
[6%]

(d) Estimate the fatigue life of the tube if it was subjected to this maximum stress
level in each vibration cycle. The S-N curve in reverse plane bending for the
steel is shown in Figure Q1b. [4%]

(e) Comment on whether the actual fatigue life is likely to be longer or shorter
than that calculated in (d). [2%]

(f) Suggest possible design changes that would increase the fatigue life, while
minimising the disruption to the airflow in the tunnel. [4%]

S-N Curve
600

550

500
Stress (MPa)

450
Air in Point A
400

350

100 300
mm
250
4 5 6 7 8 9
10 10 10 10 10 10
N

Figure Q1a Figure Q1b

Page 2 of 7 Turn Over

2. A light-weight security CCTV camera is to be mounted on a uniform steel beam
(10 mm deep by 20 mm wide) welded between two large steel girders which are 4
m apart and support the floor above. This floor supports several printing presses
which all run at known constant speeds. Figure Q2 shows the proposed
arrangement in side view. Concerns have been expressed that floor vibration
could excite the beam into resonance and result in poor picture quality or camera
failure, hence approximate natural frequencies are required to ensure proper
function.

Steel beam

Floor girder
CCTV camera

Figure Q2

(a) Assuming that both ends of the steel beam are rigidly welded to the girders,
sketch the first and second modes of the beam in bending in the plane of the
drawing. Which mode is likely to cause more problems to the picture quality
and why? [3%]

(b) Assume the characteristic beam deflection shape equation:

v = C1 cosh λx + C2 . sinh λx + C3 cos λx + C4 sin λx

Where λ 4 = ω2A ρ/EI

Choose and justify suitable boundary conditions, then derive the frequency
equation for the beam in bending to predict the natural frequencies. [6%]

(c) It is proposed to test the natural frequencies by using a calibrated impact

hammer, accelerometer and FFT analyser to measure the natural
frequencies. Suggest a suitable mounting location for the accelerometer and
impact point for the hammer. Draw neat sketches to show the expected Bode
plot and Nyquist plot from these measurements, indicating the natural
frequency in each case. [5%]

(d) (i) Suggest at least 2 reasons why the measured natural frequency may
be different from that predicted by beam theory, assuming that the
exact material properties are known.
(ii) Which one of the reasons is likely to cause the greatest error in the
prediction of natural frequencies and will its effect raise or lower the
natural frequency? [4%]

Page 3 of 7 Turn Over

3. A rigid rotor on a shaft runs in rolling element bearings which are supported by a
flexible casing. The rotor can be considered as a uniform solid cylinder made from
steel with a radius of 400 mm and a length of 500 mm. The bearing centres are
1.6 m apart. Figure Q3 shows the arrangement.

100mm 1.6m

x
θ

k k
rotor

Figure Q3

The casing stiffness in the vertical direction for each bearing k = 1.4 kN/mm.

(a) Set up the equations of motion for the rotor in its bearings using the
coordinate system indicated in Figure Q3. Consider only motion in the
vertical plane, as the system is identical in the horizontal plane. You may
ignore axial motion and assume the shaft is massless and rigid. [5%]

(b) Hence estimate the natural frequencies for the rigid body modes of the rotor
in its bearings. Neglect the mass of the shaft and bearings. [8%]

(c) Estimate the natural frequencies using separate single degree of freedom
models and explain the differences (if any) from (b). Hence deduce the
approximate mode shapes for each natural frequency calculated in (b). [5%]

Page 4 of 7 Turn Over

4. (a) A digital spectrum analyser stores 2048 points of a time domain waveform. If
the sampling (digitisation) rate is 102.4 kHz, what is the maximum signal
frequency which can be measured correctly, and what is the frequency
resolution obtained? [2%]

(b) What precautions are usually taken to ensure that signals above the
maximum frequency identified in (a) do not corrupt the results obtained from
the analyser? Describe how the results are corrupted if these precautions are
not taken. [3%]

(c) In a vibration test on a small component, modes are excited at 7130 Hz,
15200 Hz and 15230 Hz. The amplitudes of the three modes measured
using an accelerometer are similar at the measurement point and all the
modes are lightly damped. Sketch the spectrum (amplitude vs frequency) of
the acceleration response at the measurement point which would be
obtained from the analyser described in (a). [2%]

(d) How could the resolution of the spectral analysis be improved? What is the
disadvantage of using better resolution? [4%]

(e) Sketch the spectrum from 15000 to 15500 Hz which would be obtained in the
test described in (c) using Fourier analysis at a frequency resolution of 1 Hz.
How many points would be captured in this test if the sample rate of 102.4
kHz is retained? [3%]

(f) The component of (c) is rather small and there is a concern that the
accelerometer has significantly affected its response. Describe two
alternative transducers that could be used to obtain the vibration response
and state clearly how they work. [4%]

Page 5 of 7 Turn Over

5. (a) Explain why it is important to consider the disturbance response of a control
system and state what simple input function may be used to evaluate the
response of a system to a disturbance. [4%]

(b) Figure Q5(a) shows the response of a system with a proportional controller
to a ramp input. It is decided that the steady state error is unacceptable. A
trial modification to the controller produces the response of Figure Q5(b).
Suggest how this may have been achieved, justifying your answer with
reference to typical effects of different controllers. [3%]

(c) An alternative modification to the system of Figure Q5(a) produces the

response of Figure Q5(c). Suggest how this may have been achieved, again
justifying your answer with reference to typical effects of different controllers.
[3%]

(d) Discuss the pros and cons of adopting the solutions of Figures Q5(b) and
Q5(c). [4%]

(e) Discuss whether adding (or increasing) derivative control would affect the
responses of Figures Q5(a), (b) and (c). [4%]
2.5 2.5

(a))
(a (b)
2 2
Amplitude
Amplitude

1.5 1.5

1 1

0.5 0.5

0 0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
Time (sec) Time (sec)

2.5

(c)
2
Amplitude

1.5

0.5

0
0 0.5 1 1.5 2 2.5

Time (sec)

Figure Q5. Dotted line is input (desired) response; solid line is actual response.

Page 6 of 7 Turn Over

6. A servomechanism is designed to maintain the absolute vertical position of a
platform in a vehicle constant as it is driven over a rough road. The mass of the
platform is m and it is mounted on a hydraulic ram which supplies a force equal
to K times the error between the desired position of the platform and its actual
position. The viscous damping coefficient provided by the ram is c. After a period
of driving along a smooth road with the platform in its correct position, a sudden
bump in the road gives the platform a velocity v.

(a) Write down the equation of motion for the system and state the initial
conditions. [4%]

(b) If m = 5 kg, c = 30 Ns/m and K = 200 N/m, calculate the maximum

displacement of the platform from its desired position if the initial velocity, v =
1 m/s. [10%]

(c) The root locus plot for a system under proportional control is shown in Figure
Q6. Describe the nature of the roots as K increases and discuss the stability
of the system. Also state the order of the system. [4%]

Figure Q6. NB There is another root corresponding to K→ ∞ on the real axis at - ∞.

Page 7 of 7 Turn Over