V65 N1 P2 C804 Korkmaz 2
V65 N1 P2 C804 Korkmaz 2
Muhammet Ertuğrul Su
Fuat Alarçin
ISSN 0007-215X
eISSN 1845-5859
UDC 629.5.035.8
Preliminary communication
Summary
External disturbances and unstable working performance of main engine cause vibration
to propulsion plant of a ship. The most commonly encountered vibration is torsional vibration
as a result of unsteady power loads and inertia forces on the ship main engine-shaft system.
Crankshaft balancing weight as a passive system and torsional damping element as an active
system are used in order to prevent these types of vibration. The objective of the present paper
is to use the modified Proportional-Integrative-Derivative (PID) control system to enhance the
performance of the ship shaft control system and to adjust the torsional vibration for a given
reference input. The conventional PID control system did not give desired results. Therefore
the performance of the traditional PID controller was improved by moving derivative and
proportional blocks on feedback path. All simulation results show that the proposed modified
control algorithm gives a better performance than conventional PID control.
1. Introduction
Vibratory forces generated in ship propulsion systems by some internal sources such as
main engine, shaft, propeller and gearbox as well as by some external sources such as wave,
current and imbalanced ship loads are often unavoidable. These forces have influence on
axial, radial and torsional vibrations. However, their effect on propulsion system can be
minimized or reduced by different isolation methods one of which is the use of springs and
damping elements for isolation between main engine and its foundations.
Hara et al. [1] established a main engine propeller shaft system by building block
approach, and analysed the torsional, axial and lateral vibrations. The three dimensional solid
model of the crankshaft was analysed by using the Finite Element Method (FEM), besides the
equivalent beam model was constructed. Propeller was modelled as a rotor. The properties
like weight, inertia and polar moment of inertia were entered to the equations and the
additional mass of water was also taken into account. Bearings and rigidity of the body were
considered as a spring in the model established. Vibration dampers placed in different
positions were analysed. Then, results of the model were compared with measured results.
Fatih Cüneyd Korkmaz, Control of Ship Shaft Torsional Vibration
Muhammet Ertuğrul Su, Fuat Alarçin Via Modified PID Controller
Because main engine shaft system has many elements having damping effect, it can be
concluded that predictions of these approaches for the model need to be accurate.
Rao [2] modelled the power transmission system with spring and damping element as
four-degrees of freedom system. Damping elements of the system were positioned between
main engine and gearbox to reduce torsional vibration, and between thrust bearing and
propeller to reduce lateral vibration. The springs were placed in between main engine,
gearbox, bearings, propeller and foundations. Radial forces and propeller thrust forces were
thought as moving forces on main engine and propeller, respectively. The equations of model
were solved theoretically and then calculated values of torsional vibration were compared to
practically measured values. Some of the calculated and measured values of frequencies are
very close, while there are differences in some cases, which are recognized as suitable for ISO
vibration standards.
Shu et al. [3] expressed torsional and axial vibrations of the engine crankshaft with
Rayleigh differential method. The engine crankshaft was modelled as a mass spring system to
obtain natural frequency of torsional and axial vibrations. Calculated results were compared
with measured results and it is recognized that torsional vibration has an enormous effect on
axial vibration that causes noise and vibration in engine.
MacPherson et al. [4] investigated methods to calculate the additional water mass of a
propeller for vibration analysis. In this paper, the effect of added mass on the torsional and
axial vibrations of the propeller was examined. In order to calculate the added water mass of
Weldsma BS-VII series of propellers, a new formula and a new prediction was applied. The
authors indicated that this prediction method would be beneficial for vibration analyses of
propulsion systems because the reliability of the method has increased by the contribution of
damping effect.
Grzadziela [5] carried out a study by using Matlab-Simulink, in which a propeller shaft
system had four-degrees of freedom. In this study, the rotation torque of the main engine,
fixed blade propeller torque, axial force and the shaft line bearings hydrodynamics effects
have been taken into consideration. Model was established by using the FEM, the propeller
shaft speed ranging in different frequency values were determined according to different
forms of support. Simulation results showed a maximum 10% error with the measured values
converges.
Zhang et al. [6] investigated the influence of angular and axial forces acting on propeller
by modelling propeller and crankshaft. Frequency changes in torsional vibrations have been
observed between the propeller and 1st order crankshaft journal, besides between 5th and 6th
order piston crank journals. For axial vibrations, on the other hand, changes in frequency
values are expressed almost in the same way. Considering coupled vibrations, a nonlinear
behaviour effect was observed. They determined that large amount of errors would be
experienced, when the propeller crankshaft analysis was assumed to be linear.
Sestan et al. [7] described the problem of torsional vibration resonance in the propulsion
system of a catamaran ferry. They introduced that in some cases shaft line torsional vibration
resonance could be appear due to non-typical reasons.
In any reference change in manipulated signal has some malfunction in many occasions.
So as to avoid set points and recover system response, Ogata K. [8] modified PID controller
that derivative and proportional blocks moved on feedback path. The negative effects of
conventional PID case of internal zeros were recovered by the modified PID controller.
Kou Yamada et al. [9] presented a design method for modified PID controller for stable
plants. The method suggests that stable plants and admissible sets of PID gains act
independent from each other.
Elksasy et al. [10] modified PID controller as the integral block by adding time delay
and derivative block by subtracting time delay. The purpose of the modified PID in this paper
is to solve instability problem.
In this study, main engine, shaft and propeller system were modelled by Lagrange
Method and the model was solved by using Matlab-Simulink. Analysis of torsional vibration
was emphasized since it causes hazardous vibration. The aim of this paper is to control shaft
torsional vibration by modified PID controller. As a result, modified PID controller was
compared with conventional PID to make better system response.
This paper is organised as follow: In Section 2, the ship model is presented. The PID
controllers are considered in Section 3. The torsional vibration analysis and simulation is
addressed in Section 4. Finally, conclusions are presented in Section 5.
It is the basic formulation of equations to determine the natural frequencies and mode
shapes for two degree of freedom system. M, C and K are inertia, dampings and stiffness
materials respectively. {x} is referred as generalized coordinates, and {Fx} is given as a
generalized forces. Damping and springs elements can be reduced by means of dynamics
loads.
are obtained. kb1 is the stiffness factor and c is damping coefficient. With a view to deriving
the mathematical equations of this system, ϴr > ϴp is assumed.
ϴr and ϴp are the angles of propeller and shaft motion. Mrtor, Jr and Jp are the constants
that indicate reduction gear torque, reduction gear inertia and propeller gear inertia,
respectively. Differential equations of motion for two degree of freedom system are indicated
in matrix form;
From the expression (9), eigenvalues and natural frequencies have been found by
considering data from Table 1.
and .
There are four poles and a zero of shaft torsional motion in uncontrolled system. It is
obviously seen in Fig. 3 that the system has the conjugate complex roots. Therefore, it has
given an unstable result at shaft torsional motion in uncontrolled system.
0
80
System: uchtp
Frequency (rad/sec): 1.11
60 -50
Magnitude (dB): -5.78
Magnitude (dB)
System: uchtp
Frequency (rad/sec): 10.1
40 -100
Magnitude (dB): -53.6
20 -150
Imaginary Axis
0 -200
0
-20
-90
Phase (deg)
-40 -180
-60 -270
-80 -360
-1 0 1 2 3
-80 -60 -40 -20 0 20 40 60 10 10 10 10 10
Real Axis Frequency (rad/sec) Fi
Fig. 3 Eigenvalues of shaft torsional motion g.4 Frequency diagram of shaft torsional motion
The natural frequencies were determined as ωn1=1.112 rad/s for shaft, and ωn2=10.189
rad/s for propeller, and were demonstrated in Fig. 4.
3. PID Controller
3.1 Conventional PID Controller
PID controllers have been generally used successfully in feedback control applications
(Khare Y. B. et al.v [12], Wei T. et al.[13], Xuan K. et al. [14]). The proportional, integrand
and derivative component of a conventional PID controller produces an output signal that is
directly proportional to the error, the difference between the measured variable and its
reference value. The integrand component of a conventional PID controller ameliorates the
final error of system response. The proportional action is generally the main drive in a control
loop and reduces entire error. The derivative block has no effect on final error. However, this
helps reduce overshoot and oscillation.
Typically, the closed loop diagram of feedback system is shown in Fig. 5.
Fig.5 The model of closed loop system with conventional PID controller
In the PID controller block Kp, Kd and Ki are the constant that detonate proportional,
derivative and integral part.
When the desired value called reference signal R(in) is equal to output signal C(out), the
error (e) is zero. It means controller has adjusted system perfectly.
In order to make a fair comparison of closed loop system with conventional PID
controller and modified PID controller the following system is assumed;
Where a and b are the roots that make system stable, K indicates system gain. The PID
controller is obtained as follows;
Gp(s) and Gc(t) are assuming system transfer function and controller block,
respectively. The Laplace transform of PID controller given in (12) becomes;
The general formulation of closed loop transfer function can be given as (15);
As shown in expression (16), it is obviously seen that the model of closed loop system
with conventional PID includes two zeros.
3.2 Modified PID Controller
The conventional PID controller is the simplest form of controllers that utilize the
derivative and integration operations in the compensation of control systems. The flexibility
of this controller makes it easier to be used in many applications and many control problems
(Hagiwara T. et al. [15], Visioli A., [16],). However, it has also some drawbacks such as
instability which does not response desired values in nonlinear systems. So as to overcome
this problem, PID controller was modified. The model of closed loop system with modified
PID is shown in Fig. 6. Position of integral action which affects the difference of reference
signal and feedback signal has remained unchanged on forward path. However, derivative and
proportional actions have moved on feedback path to affect only the output signal (Ogata K.
[8]). The aim of the control system is to activate the shaft vibration stabilizer to minimize
torsional motion. In a practical, data collection hardware device (sensor) and software
program are used. In such a system the torsional angle sensing device and the drive signal
generated the control system. The signal is compared with the potentiometer signal, and the
difference in signal forms the basis of the feedback control system.
BRODOGRADNJA/SHIPBUILDING 65 (1) 2014 22
Control of Ship Shaft Torsional Vibration Fatih Cüneyd Korkmaz,
Via Modified PID Controller Muhammet Ertuğrul Su, Fuat Alarçin
Fig.6 The model of closed loop system with modified PID controller (Ogata K. [8])
Transfer function of closed loop system formulation is calculated according to (15) and
given below;
There are two zeros at the system with conventional PID controller. It is hard to adjust
system response due to these zeros. Their effect is occurred as earlier peak or higher
overshoot. The proposed modified PID controller recovers from these effects and ameliorates
system response by adding proportional and derivative blocks of PID on feedback path
instead of on forward path. Therefore better solution of system response is achieved in
modified PID in comparison with conventional PID.
Ship main engine, shaft and propeller system was analysed with uncontrolled,
conventional PID and modified PID controlled structures.
Gains of the PID controller (Kp, Ki and Kd) are obtained using Ziegler-Nicholes method
(Ogata K. [8]). Modified and conventional PID controller’s gains are indicated in Table 3.
Table 3 The PID controller parameters
Fig. 8 shows the torsional angular displacement of the shaft system in case of using
conventional PID controller, modified PID controller and uncontrolled model. Shaft torsional
displacement has been decreasing up to 0.2 rad point at the uncontrolled condition, so the
displacement is insufficient in the uncontrolled situation. System has responded quicker at the
implementation of the conventional PID controller; however shaft vibrations oscillate
randomly as in the uncontrolled situations at the first 8 seconds. Then displacement curve
proceeds to 0.05 rad with little amplitudes. After performing modified PID controller the
reference input has been rapidly achieved. It can be stated that modified PID are best solution
for shaft angular displacement.
0.3 1.5
Uncontrolled Uncontrolled
Conventional PID Conventional PID
0.25 Modified PID Modified PID
1
0.2
Displacement (rad)
velocity (rad/s)
0.5
0.15
0.1
0
0.05
-0.5
0
-0.05 -1
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15
Time (s) Time (s)
Fig. 8 Shaft torsional angular displacement at shaft Fig. 9 Shaft torsional angular velocity change at shaft
Shaft torsional velocity is shown in Fig. 9. Values of angular velocity have been
changed from 1.5 rad/s to -1 rad/s in the first second at the uncontrolled situation. Then
velocity come close to 0 rad/s 50 seconds later and these oscillations follow almost the same
line as conventional PID. Conventional PID controller has not affected the system adequately.
The response of system has significantly fast and amplitudes directions of the line
turned to zero after 2 seconds in modified PID. It is obvious that modified PID controller
affects the rising time of system effectively.
Shaft torsional acceleration is shown in 20
Uncontrolled
acceleration (rad/s2)
uncontrolled model. Conventional PID 5
0.45 0.4
uncontrolled
Uncontrolled
Conventional PID
0.4 0.3 Conventional PID
Modified PID
Modified PID
0.2
0.35
0.1
0.3
displacement (rad)
Velocity (rad/s)
0.25
-0.1
0.2
-0.2
0.15
-0.3
0.1
-0.4
0.05 -0.5
0 -0.6
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Time (s) Time (s)
Fig. 11 Propeller torsional angular displacement Fig. 12 Propeller torsional angular velocity
0.5 0
Uncontrolled
Conventional PID
0.4 -50
Modified PID
Magnitude (dB)
Uncontrolled
0.3 -100
Conventional
Modified
acceleration (rad/s 2 )
0.2 -150
0.1
-200
90
0
0
Conventional
Phase (deg)
-0.1
-90
-0.2 Uncontrolled
-180 Modified
-0.3
0 5 10 15 20 25 30 35 40 45 50 -270
Time (s) -4 -3 -2 -1 0 1 2 3
10 10 10 10 10 10 10 10
Frequency (rad/sec)
Fig. 13 Propeller torsional angular acceleration Fig.14 Frequency diagram of displacement of shaft
motion
5. Conclusions
In this study, control of a ship shaft torsional vibration by modified PID controller was
proposed. Lagrange equation was used to obtain mathematical model of dynamic system. The
conventional and modified PID controller was applied to the system. It was deduced that
displacement, velocity and acceleration of main engine shaft propeller systems can be
substantially reduced by controller. The simulation results of modified PID controller for
whole ship propulsion system in case of some interior and exterior effect have been
demonstrated and compared with conventional PID along with the uncontrolled system in
Table 4. It was observed that conventional PID controller has enhanced displacement of shaft
torsional at the rate of %26. Modified PID controller improved displacement of shaft torsional
in the ratio of %81. As a result, it is obviously clear that proper performance has been
achieved in Modified PID controller.
Table 4 The shaft variable values at 10 second
Shaft Modified PID Conventional PID Uncontrolled
Displacement (rad) 0.039 0.1541 0.2110
Velocity (rad/s) 0.0183 0.0190 0.0497
Acceleration(rad/s2) 0.0045 0.1247 0.0289
REFERENCES
[1] T. Hara, T. Furukawa, K. Shoda: Vibration Analysis of Main Engine Shaft System by Building Block
Approach. Bulletin of the M.E.S.J, 23 (02). Tokyo 1995.
[2] T. V. Rao: A Diagnostic Approach to the Vibration Measurements and Theoretical Analysis of the a
Dredger Propulser System. Yanam 2005.
[3] G. Shu, X. Liang, Y. Wang: Investigation on Coupled Torsional-axial Vibration of Crankshaft Based on
Rayleigh Differential Method. Tianjin, Sae international 2006.
[4] D. Macpherson, V. R. Puleo, M. B. Packard: Estimation of Entrained Water Added Mass Properties for
Vibration Analysis. SNAME New England Section 2007.
[5] A. Grzadziela: Modelling of Propeller Shaft Dynamics at Pulse Load. Polish Maritime Research 15, 52-
58. 2008.
[6] S. Zhang, Q. Zhang: Coupled Torsional and Axial Nonlinear Vibration Model of the Crankshaft with a
Propeller. Shanghai, Intl. Conf. on Sys. Simulation and Scientific Computing.2008.
[7] A. Šestan, N. Vladimir, N. Vulić, B. Ljubenkov: A Study into Resonant Phenomena in the Catamaran
Ferry Propulsion System. Transactions of FAMENA, Vol. 36 (1), pp. 35–44. 2012.
[8] K. Ogata: Modern Control Engineering. 4th Edition, New Jersey: Prentice-Hall.1990.
[9] K. Yamada, N. Matsushima, T. Hagiwara: A Design Method for Modified PID Controllers for Stable
Plants and Their Application. ECTI Transactions on Electrical Eng., Electronics and Communıcatıons 5
(1) 2007.
[10] M. S. M. Elksasy, H. H. Gad: A New Technique for Controlling Hybrid Stepper Motor through Modified
PID Controller. International Journal of Electrical & Computer Sciences IJECS-IJENS 10 (02).2010.
[11] W. W. Seto: Mechanical Vibrations. Schaum pub., New York 1964.
[12] Y. B. Khare, Y. Singh: PID Control of Heat Exchanger System. Int. Journal of Computer Applications, 8
(6) 2010.
[13] T. Wei, W. Zhongfeng, Z. Jinguo, W. Mengxiao: A Relay Feedback Based Auto tuning PID Controller
and Its Application to Pulp Consistency Control. Proceedings of the 29th Chinese Control Conference
Beijing, China 2010.
[14] K. Xuan, X. Bao, C. Li, W. Li, G. Liu, J. Wang, L. Wang: Application Of Integral-Separated PID
Algorithm In Orbit Feedback. Proceedings of ICALEPCS, Grenoble, France 2011.
[15] T. Hagıwara, K. Yamada, Y. Ando, I. Murakamı, S. Aoyama, S. Matsuura: A Design Method For
Modified PID Control Systems For Multiple-Input, multiple-Output Plants To Attenuate Unknown
Disturbances. World Automation Congress 2010.
[16] A. Visioli: Modified anti-windup scheme for PID controllers. IEE Proc.-Control Theory Appl., 150 (1)
2003.
[17] MAN Diesel: Project Guide for Marine Plants. Ausburg 2006.