VIBRATION AND RADIATED NOISE OF A SMALL SHIP [622549]

VIBRATION AND RADIATED NOISE OF A SMALL SHIP

IGNACY GLOZA

Polish Naval Academy
Śmidowicza, 81 – 919 Gdynia, Poland
[anonimizat]

Extensive measurements “in situ” both of the sound intensity and the vibration were
made of M25, a small ship (length 8 m , beam 4 m, displacement 2 tons) powered by 20 kW
a direct-drive low-speed dies el engine. A small ship create s a series of harmonics which
amplitudes and frequencies are connected with ship speed. Th e underwater sound
measurement was performed for anchored conditi on. In this paper, two different methods of
measurement were used, which provide comple mentary information. A static method to
measure noise from an anchored ship was used when only the main engine was running. In
addition to the radiated noise measuremen ts, vibration measurements were conducted
aboard this ship. The coherence function was performed to a ssociate each component of
underwater noise with the vibra ting part of the engine which ge nerates it. The calculation of
the sound intensity was made to locate the ma in source of noise on board. Underwater noise
from small ships elevates the natural ambient by 10 – 20 dB in many area ; the effects of this
noise on the biological environmen t have been rarely reported.

INTRODUCTION
Noise radiated by ships into the water environment is an important contribution to the ocean ambient noise. Therefore controlling acoustic signature on vessels is now a major consideration for researchers, naval architects a nd operators. Noise is not a major problem for
all vessels. For small ships the problem is cau sed by closely packed hi gh powered equipment,
confined in a small metal or plastic vessel. Shipboard noise problems are generally created by
poor or improper ship acoustical design. The noisiest piece of equipment on any ship is
usually a diesel engine. As a reciprocating machin e, the diesel is very loud and also creates a
great deal of vibration.
Donald Ross studied the underwater noi se of big commercial sh ips and trends in ship
sizes and powering. He also wrot e a book that described the f undamental general features of
surface ship noise [1]. Many European resear chers in France, Germany, Norway, Sweden
Finland and Poland have also cont ributed to ship noise and propeller investigation. In recent

years sophisticated numerical and experimental research on propeller cavitation has been
conducted [2].
A lot of the older pub lished ship underwater noise meas urements were made with third-
octave bandwidth analysis, which is too wide for separation of the individual spectral
components of ship radiation. The data were near ly always acquired in shallow water, so the
levels may not be well representative of free- field values, especially at low frequencies.
In the 1970 S the US navy initiated a new pract ical and theoretical program for accurate
narrow-band measurements of underwater ship nois e [5]. Because of to the high cost of ship
time as well as measurement facilities, it is perhaps not surprising that few detailed
measurements of merchant ship s and warships are available.

1. METHODS OF MEASUREMENTS

Static measurement of a small surface ship was conducted while she was moored to
buoys in the centre of the range , allowing investigation of th e contribution from individual
machines and machinery systems to the noise signature. Within the range terminal,
hydrophone and vibration signals, either recorded or direct, were processed by the noise
analysis of data from all aspects covering frequencies 1 Hz to 10 kHz.
In our case the first step was to determine the characteristic frequencies of the main
engine, by analyzing the spectrum of engine vi brations. The second st ep was to identify, in
the underwater environment, the underwater no ise coming from the ship 2 m below the sea
surface. The last step consisted in measuri ng the sound intensity level around the ship. It
allowed to determine the location of machiner y which radiated the highest level of noise.
These measurements were carried out in the Polish Navy Test and Evaluation Acoustic
Ranges in Gdynia, that is located in the southe rn part of the Baltic Sea. The basin was well
protected from wind and waves, but the weathe r was not specially good, so the ambient noise
level was average. It was a windy and raining day. During the ship measurements, the mean
wave hight was less than 1 m, with wind speeds less than 6 m/s. The bottom depth was 6m. Measurements were made with an 8 meter l ong ship – with beam 3m – call signed M25.
Its main engine was the only working mach inery on board. The ship was in the middle
of the basin, bound to the ground by three long hawsers, the propeller was stopped and only the main engine was running. At the time of the hydr oacoustical measurements in 2007, M25
was rather old, but her engine, hull and shaft we re observed to be in excellent condition. No
fouling or damage was evident on the hull or engine. This ship was powered by a four-
stroke four-cylinder diesel engine, that vibrated with firing rate equal to 6.2 Hz.
In order to measure vibrations and propaga tion of the waves through the ship, three
piezoelectric accelerometers were used. The fi rst one was fixed by the magnetic connection
directly to the engine. The two others were stuck to the hull: one in the middle of the boat and
the other one was located on the bow. They were all connected to the conditioning amplifiers
which were able to amplify low signals coming from the piezoelectric accelerometers.
The pressure signals using two hydrophones were measured, they were connected to a
wide range measuring amplifie rs. The hydrophones were joined together at a distance of
14cm which enables measurements of the sound in tensity. Fixed at the extremity of a 3 meter
long boom, they were moved around the ship, 2 meters in depth, to carry out sixteen measurement points: seven on port side, one in the bow, seven on the st arboard side and one
astern. Two railings were attached on the boards to have an accurate position of the hydrophones in relation to the boat. Figure 1 show s the locations of the different measuring

positions. For measurements 1 to 8 acceleromet ers were fixed on the port side, and for
measurements 9 to 16, they were connected on the starboard side.

Fig.1 Scheme of the small ship with two frames and different detectors

The methods of measurements the rotational and translation components of the vibration
or structure borne sound levels on a stationary vessel and moving ship are a mixture of analog
and digital techniques. The resu lting spectra were made dig itally both by a Brüel & Kjær
PULSE analyzer and a computer. A simultane ous on-board vibration monitoring system
provided additional measurements of tonals from inside our boat, because an accelerometer
was mounted on the diesel engine.

2. RESULTS OF RESEARCH

A general quantitative description of the vibr ation and the ship’s radiated noise should be
provided by power spectral analysis.
To understand how this ship generates noise, we should first analyze the spectrum of
vibrations from the main engine. It is shown below on figure 2:

Fig.2 Spectrum of vibrations from the main engine

It was done to identify the engi ne parts responsible for the most important vibrations. In fact,
mechanical unbalance, impact, friction, and pressu re fluctuation genera te vibratory forces.
The dominant noise of a diesel engine is normally the piston slap. It is caused by the impact
of the piston on the cylinder wall [6]. The tabl e below describes the ch aracteristic vibration
frequencies of the engine.

Tab.1 Table of main vibration frequencies of a diesel engine

Vibration frequencies Cause of vibrations
60 . 2s
cfrn kf⋅= Cylinder firing rate
60s
cknf⋅= Crankshaft
60psz
vkz n zfm⋅⋅⋅=⋅ Engine valves
60ps
pskz nf⋅⋅= Piston slap
60ps
prkbznf⋅⋅⋅= Piston rings

where k is the number of the harmonic (it is a whole number),
ns is the rotational speed of the engine in rpm,
zp is the number of pi stons in the engine,
zz is the number of valves for one piston,
b is the number of piston rings for one piston,
m indicates if the motor is a two or four-stroke engine.

During the measurements, the engine should run at about 750 rpm or 12.5 Hz, because
the engine spectrum contains a peak at 12.3 Hz, it means that the accurate run of the engine
was n s =738 rpm (12.3 · 60). We could then obtain the theore tical vibrating frequencies due to
the engine parts:

Tab.2 Table of fundamental vibrati on frequencies for the main engine

Fundamental frequency Cause of vibrations
fcfr = 6.2Hz Cylinder firing rate
fc = 12.3 Hz Crankshaft
fv = 24.6 Hz Engine valves
fps = 49.3 Hz Piston slap
fpr = 98.6 Hz Piston rings

Lots of harmonics are associated with t hose frequencies. We cannot know exactly the
vibration level of each harmonic, but the sche me below summarizes the contribution of each
source in the spectrum.

Fig.3 Scheme describing the contribution of sources for some vibration frequencies. The bandwidth is
0.1 Hz

The fundamental frequency of cylinder firi ng rate is associated with a mess of
harmonics. But the corresponding level is not th e highest. We have to focus on the highest
levels, which are normally 24.6 Hz and 49.3 Hz. You can notice a peak at 50 Hz in figure 4:

Fig.4 Zoom of figure 3 (frequency around 50 Hz). The bandwidth is 0.05 Hz

This peak is connected w ith the alterna ting current which powers all electric
equipment. As it was very difficult to connect the grounds of all electric equipments together,
a DC component of 50 Hz appear ed during the measurements. Even if this frequency is close
to the 49.2 Hz vibrating frequency, it does not cause any problems when calculating the
spectrum of vibrations; the re solution frequency was chosen 0.05 Hz, what means that is

narrow enough to separate two spectrum lines with very close central frequencies. This effect
was also clearly visible in the spectrum obtai ned when the calibration of accelerometers was
done .
Three accelerometers were used for the experimentation connected with transmission of
vibration inside the ship. Figure 5 shows the vibration spectr um calculated for these three
signals.
Fig.5 Vibration spectra from three accelerometers

The vibrations generated by the main engine spread out over the ship and should be
attenuated. Machinery noise originates as m echanical vibration of the many unbalanced and
diverse parts of a moving vessel. Results of m easuring vibration are presented in the table
below:
Tab.3 Level of vibration for three accelerometers at different frequencies

Frequencies
(Hz) Vibration level on
the main engine
(dB ref 1 μm · s-2) Vibration level in the
middle of ship
(dB ref 1 μm · s-2) Vibration level on
the bow of ship
(dB ref 1 μm · s-2)
6.2 78 58 50
12.3 89 85 90
24.6 117 113 98
49.3 121 113 90
50 111 105 108
73.9 109 85 86
86.2 92 99 86
123 93 99 85

This vibration is coupled to the sea via the hu ll of the vessel. Various paths, such as the
mounting of the machine, connect the vibrating part to the hull. Dominant machine vibration
originates here in the following ways: as reci procating parts, such as the explosion in
cylinders of reciprocating engine, piston slap s or noise of valves and rings, mechanical
friction in bearings and journa ls. Consequently, the level of vibration of the main engine

should be the highest of the three levels. Inde ed, figure 5 shows that the black spectrum is
overall higher than the others. In addition, the red level is in the middle of the hull higher than
the blue one.
However, you can see in the table above that there are some inversions for some
frequencies. Actually, at frequency 12.3 Hz, the level registered on the bow is higher than the
others. You can also see that the red spectru m is the highest on frequencies 86.2 Hz and 123
Hz.
This phenomenon is due to the resonance of poi nts of the hull. Indeed, some parts of the
hull resonate at certain frequencies, and the leve ls of vibration are high er in these parts than
the level of the vibrating source. However, at a distance in the sea, the sound radiated by
these vibrational forces depends not only on their magnitude, but also on how such forces are
transmitted to the hull and coupled to the water. A notable example is the resonant excitation
of large sections of the hull by machinery vi bration, what is called “hull drone”[6]. The
disposition of the accelerometers could be a second explanation for this phenomenon.
Effectively, in order to compar e the vibration on different places of the boat, it is better to
record acceleration on three axes. Furthermor e, the three accelerometers should be screwed
on the engine and on the hull in order to record the most accurate signal. But they were only
fixed by magnetic connection and stuck to the hull, which leads to less accurate
measurements.

Fig.6 Spectrum of underwater pressure, position 3

Figure 6 shows the spectrum of underwater pressure measured by one of the two
hydrophones at point 3 which was in front of the engine. You can see that some frequencies
noticed on the vibrations spectrum are presente d here: 12.3 Hz, 24.6 Hz, 49.3 Hz and lots of
harmonics of 6.2 Hz. You can also notice th at hydrophone signals do not contain the 50 Hz
frequency connected with the alternating current.
The pressure was also recorded all around th e ship, this was done for illustrating the
nature of the ship spectra. A frequency-time anal yzer are often used for speech analysis; this
kind of analyzer is called a sound spectrograph and was firs t described by Koenig and Dunn
and is widely used for the analysis of speech [6].
It gives a plot of frequency against time and shows the intensity of the sound in the

analysis bandwidth by changing th e color of the record as you can see in the picture below.

Fig.7 Pressure spectrogram of sixteen meas urements carried out around the small ship

To understand figure 7, you have to associate each point of measurement with a location
around the boat (see figure 1 for accurate inform ation).One more time, the three main
frequencies 12.3 Hz, 24.6 Hz and 49.3 Hz appear clearly all around the ship. It is also
interesting to see the variati on of pressure level with hydr ophone position. Around our points
of measurement (2-3-4) and (14-15-16), the level of the underwater pressure is overall higher.
The engine was indeed in fr ont of points 2-3 and 13-14. Mo reover, the high level of
broadband noise for those points of measurement is certainly connected with the engine noise.

Fig.8 Coherence function between the underwater pressure measured on position 3 and
the vibrations of the main engine
In this measurement, the engine had to be the main source of noise. But in order to
confirm this point, the coherence function was us ed between vibrations recorded on board and

the underwater pressure signals. The two signals was recorded at the same time on position 3
and are shown in figure 8 with the coherence function.
The coherence function nearly takes the 1 value for the thr ee frequencies: 12.3 Hz, 24.7
Hz, 49.3 Hz. It means that those frequencies ar e connected with the noise generated by the
ship, and are not connected with environmental noise.

Fig.9 Calculation of sound intensity on location 3, port side

The sound intensity on each point of measurement was calculated. The figure 9 represents the
sound intensity in dB ref 1pW/m2 calculated on location 3. You can notice that the three
discrete lines on the curve which occur for the characteristic frequencies.
In figure 10 the sound intensity for the different locations and for those three
frequencies are shown.

Fig.10 Calculation of sound intensity on port side

Those results enable to locate the engine betw een locations of recording 2 and 3 exactly
2.6 m from the stern. This is where the sound intensity has its maximum for the three

frequencies. You can then observe a decrease of the sound intensity when our measurements
are made forward to the bow.
The sound intensity is maximum when the noise source is on the axis of the probe, then
decreases when the source moves away
There are still a number of unresolved problems in the sound intensity measurements.
For example, more research is needed on the influence of flow. Flow generates a ‘false’
intensity signal with an unknown sign that depends on the particulars of the turbulence, and
this quantity is simply adde d to the sound intensity[3].

3. SUMMARY

This paper presents vibration and underwater noise radiated by a stationary small ship.
It is possible to detect a mach ine like a diesel engine and th e whole boat in the background of
the shallow sea’s natural noises.
In order to verify different analysis, not only the software Pulse was used but to check it
Matlab programs were performed an d both gave the same results.
A small ship with only one main engine wa s tested here, but the same experimentation
was carried out with multi device warships.
In order to reduce generated noise, different pro cesses have been developed to locate
and analyze the inboard sources of noise. Mo st of them are based on registering the
underwater pressure with hydrophones. Those anal yses also provide important information
for military applications. It can be useful to esta blish if a ship has mechanical failures such as
problems with main engine generators or propellers.

REFERENCES

1. D. Ross, Mechanics of Underwater Noise, Pergamon, and New York 1976.
2. E. Kozaczka, Underwater ship noise . Symposium on Hydroacoustics, Gda ńsk-Jurata 2000.
3. S. J. Malinowski, I. Gloza, Underwater No sie Characteristics of Small Chips, Acta
Acoustica united with Acoustica, Vol. 88 (2002),
4. F. Cervera, H. Estellees, F. Galvez, F. Belm ar, Sound intensity in the near field above a
vibrating flat plate. “Noise Contro l Engineering Journal” 45 (1997), 193-199
5. P.T. Arveson, D. T. Vendittis, Radiated noise characteristics of a modern cargo ship. J.
Acoust. Soc. Am., 107 (1), 118-129, 2000.
6. R. J. Urick, Principles of Underwater Sound, Mc Graw-Hill, New York 1975. Chap.10.
7. Finn Jacobsen, Sound Intensity , XLVII Ot warte Seminarium z Akustyki OSA’2000.