Ocean Engineering 27 (2000) 87–96

www.elsevier.com/locate/oceaneng

The calculation of zig-zag maneuver in regular

waves with use of the impulse response
functions
Seung-Keon Lee
Pusan National University, Pusan, South Korea

Received 30 April 1997; received in revised form 24 June 1997; accepted 20 August 1997

Abstract

The traditional governing equations for sway–yaw maneuvering motion are a set of ordinary
differential equations with constant coefficients. But, as is well known, integro–differential
equations with impulse response functions are more strict governing equations that can handle
the frequency dependence of hydrodynamic forces.
In this paper, the two types of equation are compared and used to calculate the 10°–10°
zig-zag maneuver in waves. Differences between the solutions are discussed.  1999 Published
by Elsevier Science Ltd. All rights reserved.

Keywords: Zig-zag maneuver; Regular waves; Impulse response

1. Introduction

The maneuvering motion of a ship in waves contains some difficulties from the
viewpoint of hydrodynamics. The changes of yaw angle and forward speed result
in a change in the encounter frequency of waves. As a result, the hydrodynamic
forces exerted on the ship vary continuously. Meanwhile, the response of the ship
to the rudder deflection is very slow and it becomes difficult to determine the suitable
frequency of hydrodynamic forces when we treat the problem with traditional, ordi-
nary differential equations.
Many researchers have pointed out that it is not correct to treat such problems
with differential equations of constant coefficients. Cummins (1962), Bishop et al.

0029-8018/00/$ - see front matter  1999 Published by Elsevier Science Ltd. All rights reserved.
PII: S 0 0 2 9 - 8 0 1 8 ( 9 7 ) 1 0 0 1 0 - 5
88 S.-K. Lee / Ocean Engineering 27 (2000) 87–96

(1974), Tick (1959), Wehausen et al. (1976) and Hamamoto and Hasegawa (1987)
have preferred integro–differential equations using the impulse response functions
of hydrodynamic forces.
Fujino (1975) has developed a sway–yaw integro–differential equation with
impulse responses obtained through Fourier transformation of PMM data of added
masses and damping coefficients over wide range of frequency. He also studied the
differences between the ordinary differential and the integro–differential calculation,
when a stepwise rudder is applied in both cases. He showed that the response of the
ship is almost the same in both approaches.
Leonardo Perez y Perez (1974) also has determined the sway–yaw–roll equation
with linear impulse response functions of wave-exciting forces. He used the strip
method to calculate the wave forces in the frequency domain, and later used these
to calculate the impulse response functions.
In this paper, linear integro–differential sway–yaw equations with rudder deflec-
tion in regular waves are derived, and the maneuvering motion of Todd’s series 60
model is calculated. The differences between the integro–differential equation and
the ordinary differential equation are discussed by comparing the response of a typi-
cal maneuver like the zig-zag test. Here, the indirect method using Fourier transform-
ation is adopted to obtain the impulse responses of radiation forces and wave-excit-
ing forces.

2. Equation of sway–yaw linear motion

By using the coordinate system of Fig. 1, the sway–yaw equations with constant
coefficients are written as follows:
(m ⫺ Yv̇)v̇ ⫺ Yvv ⫹ (mU ⫺ Yr)r ⫺ Yṙṙ ⫽ Y␦␦ ⫹ Y␻c cos ␻et ⫹ Y␻s sin ␻et (1)
(Izz ⫺ Nṙ)ṙ ⫺ Nrr ⫺ Nvv ⫺ Nv̇v̇ ⫽ N␦␦ ⫹ N␻c cos ␻et ⫹ N␻s sin ␻et

Fig. 1. Coordinate system.

 S.-K. Lee / Ocean Engineering 27 (2000) 87–96 89

Meanwhile, the strict sway–yaw equation can be built as the Volterra Functional
forms. Here, the kernels are the impulse responses of the hydrodynamic forces.

冕 冕
t t

m(v̇ ⫹ Ur) ⫹ Kv(␶)v(t ⫺ ␶) d␶ ⫹ Kr(␶)r(t ⫺ ␶) d␶

⫺⬁ ⫺⬁

冕 冕
t t

⫽ K␦(␶)␨(t ⫺ ␶) d␶ ⫹ K␨(␶)␨(t ⫺ ␶) d␶
⫺⬁ ⫺⬁

冕 冕
t t

Izzṙ ⫹ Mv(␶)v(t ⫺ ␶) d␶ ⫹ Mr(␶)r(t ⫺ ␶) d␶

⫺⬁ ⫺⬁

冕 冕
t t

⫽ M␦(␶)␨(t ⫺ ␶) d␶ ⫹ M␨(␶)␨(t ⫺ ␶) d␶ (2)

⫺⬁ ⫺⬁

The impulse response functions on the left-hand side of Eq. (2), having singular-
ities at ␶ ⫽ 0, can transformed into Eq. (3) which has only regular functions:
(m ⫺ Yv̇(⬁))v̇ ⫺ Yv(⬁)v ⫺ Yṙ(⬁) ⫹ (mU ⫺ Yr(⬁))r

冕 冕 冕
⬁ ⬁ ⬁

⫹ Kv*(␶)v(t ⫺ ␶) d␶ ⫹ Kr*(␶)r(t ⫺ ␶) d␶ ⫽ Y␦␦ ⫹ F␨(␶)␨(t ⫺ ␶) d␶

0 0 0

(Izz ⫺ Nr(⬁))ṙ ⫺ Nr(⬁)r ⫺ Nv̇(⬁)v̇ ⫹ Nv(⬁)v

冕 冕 冕
⬁ ⬁ ⬁

⫹ Mv*(␶)v(t ⫺ ␶) d␶ ⫹ Mr*(␶)r(t ⫺ ␶) d␶ ⫽ N␦␦ ⫹ M␨(␶)␨(t ⫺ ␶) d␶ (3)

0 0 0

In Eq. (3), the impulse response functions due to rudder deflection are treated as
constant, because no data exist that can be used to calculate the rudder impulse
responses.

冕 冕
t t

K␦(␶)␨(t ⫺ ␶) d␶ ⫽ ␦ K␦(␶) d␶ ⫽ ␦·Y␦ (4)

⫺⬁ ⫺⬁

冕 冕
t t

M␦(␶)␨(t ⫺ ␶) d␶ ⫽ ␦ M␦(␶) d␶ ⫽ ␦·N␦

⫺⬁ ⫺⬁
90 S.-K. Lee / Ocean Engineering 27 (2000) 87–96

3. Calculation of impulse response functions

3.1. Impulse response functions of radiation forces

The direct method to obtain the impulse responses of the moving ship is to solve
the boundary-value problem in the time domain. However, the procedure is very
complicated and difficult. Another way to get the impulse response functions is to
use the Fourier transform of the frequency responses of radiation forces. The latter
method is indirect but simple.
In this paper, the PMM data of Todd’s series 60 model by van Leeuwen (1964)
are used to calculate the impulse responses resulting from sway velocity and yaw
angular velocity. That is,

冕 冕
⬁ ⬁
1 1
Kv*(␶) ⫽ ⫺ 兵Yv(␻) ⫺ Yv(⬁)其 cos ␻z d␻ ⫽ ⫺ 兵(m ⫺ Yv(␻)) (5)
␲ ␲
⫺⬁ ⫺⬁

⫺ (m ⫺ Yv(⬁))其 sin ␻z d␻
Kr*(␶), Mv*(␶) and Mr*(␶) can be calculated in the same way.
Fig. 2 shows the impulse response functions of the given model at Froude number
0.2. Here, the added masses and damping coefficients are fitted with rational func-
tions to obtain the exact Fourier transformation. The impulse response functions
obtained from Fourier cosine transformation coincide exactly with the results of
Fourier sine transformation.

Fig. 2. Impulse response function of radiation forces.

 S.-K. Lee / Ocean Engineering 27 (2000) 87–96 91

Fig. 3. Offset of series 60 (Cb ⫽ 0.7).

3.2. Impulse response functions due to waves

The impulse response functions resulting from an impulsive wave with unit ampli-
tude can be calculated through Fourier transformation of the responses of the model
to the various sinusoidal waves in the frequency domain. In this paper, the OSM
(ordinary strip method) is used to calculate the frequency responses of wave forces
of a given model.
For the 21 two-dimensional sections in Fig. 3, the added mass and wave-making
damping coefficient are calculated by the close-fit method (Frank and Salvesen,
1975). Fig. 4 shows the response characteristics of sway force and yaw moment
caused by regular waves coming with an encounter angle ␮ ⫽ 150° and amplitude
␨a ⫽ 1.0 m.
The impulse response functions caused by the waves are

冕
⬁
2
K␨(t) ⫽ Re兵 Hf(␻)· ei␻t d␻其 (6)
␲
0

冕
⬁
2
M␨(t) ⫽ Re兵 HM(␻)· ei␻t d␻其
␲
0

where Hf(␻) and HM(␻) are the frequency response of sway force and yaw moment,
respectively. The impulse response functions obtained in such a way are presented
in Fig. 5.
In usual maneuvering motion, the heading angle changes successively and the
encounter angle of incident waves also varies continuously. The hydrodynamic forces
on the ship caused by the waves are a function of encounter frequency, ␻e. However,
92 S.-K. Lee / Ocean Engineering 27 (2000) 87–96

Fig. 4. Frequency responses of wave-exciting forces.

Fig. 5. Impulse response functions of wave-exciting forces.

in this paper, we have neglected this fact and fixed the encounter angle at ␮ ⫽ 150°
for the following calculation, to simplify the problem.
 S.-K. Lee / Ocean Engineering 27 (2000) 87–96 93

Fig. 6. 10°–10° zig-zag with rudder deflection only.

4. Examination of calculation results

In this paper, our interest is restricted to linear sway–yaw motion, and so the
author has confined the simulations to 10°–10° zig-zag maneuvers.
First, the usual calculation by Eq. (1) and the elaborate calculation by Eq. (2) in
calm sea were performed, and the results are compared in Fig. 6. The hydrodynamic
derivatives and the principal dimensions for the model are listed in Tables 1 and 2,
respectively. The Runge–Kutta–Gill method was used to simulate the zig-zag motion.
The time step dt is 0.01 s, and the rudder is deflected in a stepwise manner. As can
be seen from Fig. 6, no visible differences can be found between the two calculations.
Next, for simulation of the zig-zag maneuver in regular waves, the wave-exciting

Table 1
Linear maneuvering derivatives at ␻ ⱌ 0, ␻ ⫽ ⬁ (Froude number ⫽ 0.2)

Derivative ␻⫽0 ␻⫽⬁

m⬘ ⫺ Yv̇⬘ 0.0229 0.01542

Yv⬘ ⫺ 0.0222 ⫺ 0.052552
Yṙ⬘ ⫺ 0.00039 0.000186
m⬘ ⫺ Yr⬘ 0.0076 0.011
Izz⬘ ⫺ Nṙ⬘ 0.0012 0.00092
Nr⬘ ⫺ 0.0034 ⫺ 0.0093
Nv̇⬘ ⫺ 0.00048 ⫺ 0.000073
Nv⬘ ⫺ 0.0057 0.0037
Y␦ ⬘ ⫺ 0.00211
N␦⬘ 0.001
94 S.-K. Lee / Ocean Engineering 27 (2000) 87–96

Table 2
Principal dimensions of the model

Lpp 2.258 m
B 0.323 m
d 0.129 m
W 65.714 kgf
kzz⬘ 0.25

forces were calculated by the convolution integral; the results are presented in Figs.
7 and 8. Here, the convolution integral was carried out for 5 s at time intervals of
0.01 s, with Simpson’s rule.
In the figures, dotted lines represent the values of wave forces fitted to the sinus-
oidal functions. Later, these are used as the regular exciting forces in the ordinary
differential equation. By doing this, the exciting forces in the ordinary differential
equation and in the integro–differential equation become equal.
The 10°–10° zig-zag simulation results obtained with Eqs. (1) and (2) in regular
waves are shown in Figs. 9 and 10. We can see that the difference between the two
calculations is very small.

5. Conclusion

For the 2.258 m model of Todd’s series 60 (Cb ⫽ 0.7), two governing equations
for the maneuvering motion in waves — i.e., the ordinary differential equation and

Fig. 7. Wave-exciting force used in the simulation.

 S.-K. Lee / Ocean Engineering 27 (2000) 87–96 95

Fig. 8. Wave-exciting moment used in the simulation.

Fig. 9. 10°–10° zig-zag in regular waves.

the integro–differential equation — are examined from the viewpoint of a linear

physical system. The results of examination are as follows.
1. Confined to the linear maneuvering motion in waves, the usual constant coef-
ficients equation gives good agreement with the exact integro–differential equ-
ation.
96 S.-K. Lee / Ocean Engineering 27 (2000) 87–96

Fig. 10. 10°–10° zig-zag in regular waves.

2. The calculation in this paper was for a model which has good directional stability
and ordinary hull form. Therefore, the same calculations must be carried out for
a blunt or directionally unstable ship, to make conclusion (1) more general.
3. Extension of the present topic to nonlinear maneuvering motion, such as turning
in waves, needs more strict considerations.

References

Bishop, R.E.D., Burcher, R.K., Parkinson, W.G., 1974. Oscillatory testing for the assessment of ship
maneuverability. In: Proceedings of 10th Naval Hydrodynamics Symposium.
Cummins, W.E., 1962. The impulse response function and ship motions. Schiffstechnik 9.
Frank, W., Salvesen, N., 1975. The Frank close-fit ship-motion computer program. NSRDC Report.
Fujino, M., 1975. The effect of frequency dependence of the stability derivatives on maneuvering motion.
I.S.P. 22 (256).
van Leeuwen, G., 1964. The lateral damping and added mass of an oscillating ship model. Report no.
23, Ship-building Laboratory, University of Technology, Delft.
Leonardo Perez y Perez, 1974. A time-domain solution to the motion of a steered ship in waves. Journal
of Ship Research 18 (1).
Hamamoto, M. and Hasegawa, K., 1987. Estimation of hydrodynamic derivatives derived from transient
maneuvering tests. Journal of the Society of Naval Architects of Japan 161.
Tick, L.T., 1959. Differential equations with frequency dependent coefficients. Journal of Ship Research,
3 (2).
Wehausen, J.V., Frank, T., Loeser, D.T., Scragg, C.A. and Sibul, O.J., 1976. Transient manoeuvre testing
and equation of maneuvering. In: Proceedings of 11th Symposium on Naval Hydrodynamics, London.