Case Studies in Mechanical Systems and Signal Processing 1 (2015) 3237

Contents lists available at ScienceDirect

Case Studies in Mechanical Systems and Signal

Processing
journal homepage: www.elsevier.com/locate/csmssp

Practical demonstration of a large-scale active vibration

isolation system
Tiejun Yang a,*, Yao Sun a, Liubin Zhou b, Michael J. Brennan c, Zhigang Liu a
a

College of Power and Energy Engineering, Harbin Engineering University, Harbin, Peoples Republic of China
Wuhan Second Ship Design and Research Institute, Wuhan, Peoples Republic of China
c
Departamento de Engenharia Mecanica UNESP, Ilha Solteira, Brazil
b

A R T I C L E I N F O

A B S T R A C T

Available online 14 April 2015

Small experimental test-rigs are often used to investigate active vibration control concepts
in the laboratory because of ease of construction and implementation. However, in
marine applications, there is a large gap between this type of experiment and full-scale
implementation onboard a ship. In this article a large-scale laboratory based active
vibration control system is demonstrated. It involves a oating raft system attached to a
hull-like structure by way of four hydraulic actuators, which are placed in parallel with
eighteen passive resilient isolators. The exible hull-like structure is supported on twenty
six pneumatic springs to simulate a oating ship. A decentralized feedforward control
strategy was implemented resulting in the reduction of vibration levels on the exible
hull-like receiving structure of up to 36 dB at some tonal excitation frequencies. The
passive isolation results in broadband control and is most effective at higher frequencies.
2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords:
Floating raft
Active vibration isolation
Flexible hull-like structure

1. Introduction
Active vibration control has been studied for many years, and is now a fairly mature research topic with several text books
being written on the topic [13]. There have been some successful demonstrations of active control, for example the active
control systems described in Refs. [48], but there are only a few current commercial systems, for example the active mount
from Paulstra-vibrachoc [9]. Of particular interest to the authors of this paper is the active control of machinery vibrations
in marine applications [1016].
Active vibration control concepts are often investigated on small experimental test-rigs in the laboratory, but in marine
applications, there is a large gap between this type of experiment and full-scale implementation onboard a ship. Since vibroacoustic behavior of a ship structure is often of major interest, a large exible structure is needed to emulate the hull, rather than
rigid foundations found in many laboratory-based experimental test-rigs. There is thus a need to carry out large-scale
laboratory experiments in controlled conditions to investigate some of the practical issues. This was the motivation for the work
described in this article. Established control techniques were applied to a bespoke large-scale test-rig of a typical marine set-up
involving a oating raft structure attached to a hull-like structure by resilient mounts. Four hydraulic actuators were used in a
decentralized feedforward control system and some experiments were performed. The exible hull-like structure is supported

* Corresponding author at: College of Power and Energy Engineering, Harbin Engineering University, No. 145, Nantong Street, Nangang District, Harbin,
Heilongjiang 150001, Peoples Republic of China. Tel.: +86 451 82589199x306/18645066270; fax: +86 45182589199x311.
E-mail address: yangtiejun@hrbeu.edu.cn (T. Yang).
http://dx.doi.org/10.1016/j.csmssp.2015.03.001
2351-9886/ 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/
4.0/).

T. Yang et al. / Case Studies in Mechanical Systems and Signal Processing 1 (2015) 3237

33

on pneumatic springs and the whole rig behaves dynamically as a section of a oating ship. Details of the experiments
conducted and the results of these experiments are given in this article.
2. Description of the test-rig and the control system
A photograph of the experimental test-rig is shown in Fig. 1(a). It consists of a 6.2 m  5.2 m  0.4 m oating raft,
constructed by 16 I-shaped steel beams, which are bolted together. It is supported on a exible hull structure by 18 BE-400
type resilient isolators mounted on 18 steel pillars. The 7 m  6 m  0.7 m exible hull structure is made from steel plates
which are welded together and supported by 26 pneumatic isolators. The details of the test-rig are given in Table 1.
From modal testing, it was found that there were 25 modes below 150 Hz [17]. A Bruel and Kjaer (B&K) type 3628 shaker,
positioned at the center of the oating raft as shown in Fig. 1(b), was used as a primary vibration source to excite the raft and
simulate
a primary uncontrolled vibration source. Note that this is an intermediate stage of experimentation. In the next
[(Fig._1)TD$IG]

Fig. 1. Details of the active vibration isolation oating raft system connected to the hull structure. (a) Photograph of the experimental test-rig. (b) Schematic
of the test-rig showing the primary B&K shaker and one of the hydraulic secondary actuators.

Table 1
Details of the experimental test-rig.
Floating raft

Dimensions
Number of I-shaped beams

6.2 m  5.2 m  0.4 m

16

BE-400 resilient isolators

Number
Vertical stiffness

18
1.61e6 N/m

Hull structure

Dimensions
Weight

7 m  6 m  0.7
4.8 t

Pneumatic isolators

Number
Vertical stiffness
Horizontal stiffness

26
2.3e4 N/m
4.7e4 N/m

T. Yang et al. / Case Studies in Mechanical Systems and Signal Processing 1 (2015) 3237

34

Table 2
Details of the active oating raft vibration isolation system.
Transducers

Type
Number

B&K 2258a-100
10

High-pass lters

Type
Number

YE3762A
4

Low-pass lters

Type
Number

YE3770A
4

Signal source and analyzer

Type

Bruel and Kjaer type 3560D (multi-function analyzer PULSE 3560D)

Controller

Type
Number of A to D
Number of D to A

TMS320VC33 DSP system

4
4

Charge ampliers

Type
Number

Bruel and Kjaer type 2692-014

4

Primary actuator

Make
Type
Output force

Bruel and Kjaer type 3628

Electrodynamic
1000 N

Secondary actuators

Number
Type
Output force
Locations

4
Hydraulic
8.9 kN
(2 m, 2 m), (2 m, 5 m), (4 m, 2 m), (4 m, 5 m) from the origin shown in Fig. 1

stage multiple sources and then a diesel engine will be tted to the test rig. Four hydraulic actuators, which were used as
secondary sources and could each generate a force of 8.9 kN, are located symmetrically between the raft and the hull,
as shown in Fig. 1(b) and Table 2. The whole system weighs about 8 tons.
The active vibration isolation is depicted in Fig. 2. Fig. 2(a) is a photograph of the system in which the primary shaker,
which was suspended on a crane, and the control instrumentation, can be seen. A block diagram of the control system is
shown in Fig. 2(b). The vibration of the system was measured using 8 accelerometers. Four accelerometers (Nos. 14) called
error sensors were used to measure the vibration of the hull structure at the positions where the four hydraulic secondary
actuators were located, and four accelerometers (Nos. 58) were placed on the raft at the positions of the hydraulic actuators
to determine the effect of the control system on the raft.
A B&K 3560D system with Pulse software was used as a signal source for the primary shaker and for subsequent
analysis of the resulting vibration. Four hydraulic actuators with a four-channel valve driver acted as the secondary
sources. A TMS320VC33 based DSP board with 4-channel A to D and 4-channel D to A converters was used as the
controller. The system also included four-channel high-pass and low-pass lters. The transducers and instrumentation
are listed in Table 2.
The signal used to drive the primary shaker through a power amplier was also used as a reference signal for the adaptive
controller. Four hydraulic actuators were operated simultaneously by feeding the acceleration from their locations on the
hull structure (measured by 4 error sensors) back to the adaptive controller. The frequency responses between the electrical
input to one of the hydraulic actuators and the four error sensors outputs, which are called error paths in the active control
system, are shown in Fig. 3. It can be seen that, except at some frequencies close to the natural frequencies of the whole
system, in which there is strong coupling between the raft and the hull as described in [17], the response of the structure
at the point where it is excited (point response) is almost 10 dB greater than at the other positions (transfer responses).
This means that a decentralized control system could be implemented, reducing the complexity drastically. Four
independent adaptive x-LMS based lters [18], which are described by the following equations, were used as controllers,
with each actuator operating independently.
8
L1
X
>
>
>
wi;l xn  l
a
> yi n wTi nxn
>
>
>
l0
>
>
>
>
b
< ei n di n  si n  yi n
wi n 1 wi n mi x0i nei n
c
(1)


>
T
0
0
0
0
>
>
n

x
n;
x
n

1;
.
.
.
;
x
n

L

1
d
x
> i
i
i
i
>
>
H
1
>
X
> 0
>
T
>
ci;h xn  h
e
>
: xi n ci nxn
h0

where n is iteration number, si(n) (i = 14) is impulse response function of the ith error path, x(n) is the reference input
signal and yi(n) is the ith controller output signal, * denotes linear convolution, L is the tap number of the adaptive control
lters, H is the tap number of the estimated error path lters, mi is the convergence coefcient of the adaptive lter, x(n),

[(Fig._2)TD$IG]

T. Yang et al. / Case Studies in Mechanical Systems and Signal Processing 1 (2015) 3237

35

Fig. 2. Details of the active control system. (a) Photograph of the active vibration control experimental test-rig showing the instrumentation. (b) Block
diagram of the active control system.

[(Fig._3)TD$IG]

Fig. 3. Frequency response function (FRF) between the input to one of the secondary actuators and the four error sensors (point response: blue solid line; transfer
response 1: red dashed line; transfer response 2: green dashed line; transfer response 3: green solid line; line 1: mode 11 at 28 Hz; line 2, mode 12 at 32 Hz; line
3, mode 13 at 37 Hz; line 4, mode 20 at 58 Hz; line 5, mode 22 at 67 Hz; line 6 mode 23 at 107 Hz; line 7, mode 25 at 144 Hz; line 8, mode 26 at 162 Hz).

[(Fig._4)TD$IG]

T. Yang et al. / Case Studies in Mechanical Systems and Signal Processing 1 (2015) 3237

36

Fig. 4. Reduction of the outputs from the error sensors at discrete frequencies due to control.

wi(n), ci(n) are reference input signal vector, control lter vector and error path lter vector respectively, which can be
expressed as
xn xn; xn  1; . . . ; xn  L 1T
wi n wi;0 ; wi;1 ; . . . ; wi;l ; . . . ; wi;L1 
ci n ci;0 ; ci;1 ; . . . ; ci;h ; . . . ; ci;H1 

(2)

(3)

(4)

x
[(Fig._5)TD$IG] 0 (n) is the reference signal input after being ltered by error path lter.
110

110

Acceleration (dB re 1e-6 m/s )

100

24Hz

95
90
85
80
75
70

100
95
90
85
80

50

100

150
Frequency (Hz)

200

250

70

300

(c)

50

100

(d)

105
Acceleration (dB re 1e-6 m/s )

71Hz

100

100Hz

95
90
85
80
75
70

150
Frequency (Hz)

200

250

300

200

250

300

110

105
2

71Hz
54Hz

75

110

Acceleration (dB re 1e-6 m/s )

(b)

105

54Hz
Acceleration (dB re 1e-6 m/s )

(a)

105

130Hz
155Hz

100
95
90
85
80
75

50

100

150
Frequence(Hz)

200

250

300

70

50

100

150
Frequency (Hz)

Fig. 5. Square root of the mean squared accelerations measured by the four error sensors. (a) Excitation frequencies 24 Hz & 54 Hz. (b) Excitation frequencies
54 Hz & 71 Hz. (c) Excitation frequencies 71 Hz & 100 Hz. (d) Excitation frequencies 130 Hz & 155 Hz (raft without control, green dashed line; hull without
control, red thin solid line; hull with control, blue thick solid line).

T. Yang et al. / Case Studies in Mechanical Systems and Signal Processing 1 (2015) 3237

37

3. Experimental tests
To demonstrate the active control system for at least two rotating machines on the oating raft (but at the same location),
several two-frequency combinations of 24 Hz & 54 Hz, 54 Hz & 71 Hz, 71 Hz & 100 Hz, and 130 Hz & 155 Hz were used as the
input signals to the primary shaker.
The details of the vibration reductions at each frequency component are shown in Fig. 4. It can be seen that reductions in
the vibration of the hull structure (as measured by the error sensors) of up to 36 dB were achieved at each frequency except
at 54 Hz, where the vibration at error sensor 3 was increased slightly when the 24 Hz & 54 Hz combination of excitation
frequencies were used. The square roots of the mean squared acceleration at the four error sensor positions with and without
control are plotted in Fig. 5 for the excitation frequencies given above. Also plotted is the vibration of the machinery raft
without control (as measured by sensors 58). It can be seen that although the oating raft reduces the vibration
transmission from the raft to the hull structure greatly, especially for frequencies greater than about 100 Hz, the responses
at the excitation frequencies are still distinct. It is only with active control that all the responses at the excitation frequencies
were suppressed.
4. Conclusions
A practical demonstration of a large-scale active vibration system has been described in this article. Decentralized
feedforward control was applied using four hydraulic actuators which were attached between a oating raft and a exible
hull-like structure. There were also eighteen rubber isolators in parallel with the actuators. Up to 36 dB reduction in the
vibration of the hull-like structure was achieved at some tonal excitation frequencies, while the passive isolators were
effective at controlling broadband vibration transmission, especially at high frequencies. It should be noted that although
vibration suppression has been demonstrated at the control positions, further work is needed to estimate the positioning
of additional sensors to determine the effect of active control on the global vibration of the hull structure.
Acknowledgement
The authors gratefully acknowledge the nancial support from the National Natural Science Foundation of China
(No. 51375103).
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]

Fuller CR, Elliott SJ, Nelson PA. Active control of vibration. London: Academic Press; 1996.
Hansen C, Snyder S. Active control of noise and vibration. London: E&FN Spon; 1997.
Gawronski WK. Advanced structural dynamics and active control of structures. New York: Springer; 2004.
Moriyuki S, Eiichi Y, Tetsuya K, Hisao W, Hideaki S. Active vibration isolation of a diesel engine generator with linear voice coil motors. In: Proc. of
CIMAC congress; 2004.
Swinbanks MA. The active control of vibration & shock. In: Engineered adaptive structures 4th conference; 2004.
Brennan MJ, Elliott SJ, Huang X. A demonstration of active vibration isolation using decentralized velocity feedback control. Smart Mater Struct
2006;15:1922.
Yang TJ, Suai ZJ, Sun Y, Zhu MG, Xiao YH, Liu XG, et al. Active vibration isolation system for a diesel engine. Noise Control Eng J 2012;60:26782.
Orivuori J, Zazas I, Daley S. Active control of frequency varying disturbances in a diesel engine. Control Eng Pract 2012;20:120619.
New technologies-active isolation, Home page of Paulstra-vibrachoc Inc. http://www.paulstra-vibrachoc.com/Active_isolation_GB.pdf.
Mitsuhashi K, Biwa T, Mizuhara S. Application of active vibration isolating system to diesel engine mounting. In: 18th international congress on
combustion engines conf. diesel engines volume. 1989. p. 281300.
Winberg M, Hansen C, Claesson I, Li X. Active control of engine vibration in a Collins class submarine. Blekinge Institute of Technology Research Report.
Report no: 11; 2003.
Yang TJ, Zhang XY, Xiao YH, Huang JE, Liu ZG. Adaptive vibration isolation system for marine engine. J Mar Sci Appl 2004;3:305.
Daley S, Johnson FA, Pearson JB, Dixon R. Active control for marine applications. Control Eng Pract 2004;12:46574.
Winberg M, Johansson S, Hakansson L, Claesson I, Lago TL. Active vibration isolation in ships: a preanalysis of sound and vibration problem. Int J Acoust
Vib 2005;10:17596.
Daley S, Zazas I, Hatonen J. Harmonic control of a smart spring machinery vibration isolation system. J Eng Mater Technol 2008;222:10919.
Yang TJ, Li XH, Zhu MG, Du JT, Liu XG, Liu ZG. Active vibration control experimental investigation for a diesel engine generator in marine applications.
J Vib Eng 2013;26:1608 [in Chinese].
Zhou LB, Yang TJ, Yuan WP, Shi H, Liu ZG. The vibration characteristics of a large exible vibration isolation structure with nite element analysis
and modal test. Adv Mater Res 2011;199200:85864.
Burgess JC. Active adaptive sound control in a duct: a computer simulation. J Acoust Soc Am 1981;70:71526.