1

Translated Report,
November 2000.
Internet:
http://dutw189.wbmt.tudelft.nl/~johan
http://www.shipmotions.nl
Report 0267, J une 1970,
Delft University of Technology,
Ship Hydromechanics Laboratory,
Mekelweg 2, 2628 CD Delft,
The Netherlands.
E-mail: J .M.J .J ournee@wbmt.tudelft.nl
A Simple Method for Determining the Manoeuvring
Indices K and T from Zigzag Trial Data
J.M.J. Journe
Delft University of Technology
Abstract
Nomotos first order model is the simplest mathematical model to describe ship manoeuvres.
Calculated manoeuvring data have been analysed here to determine the relation between the
manoeuvring indices K and T of Nomoto (1960) and zigzag manoeuvring characteristics.
The results have been reflected in graphs, which can be used then to determine these indices
from actual zigzag manoeuvres. This report is a translation in English of a report in Dutch of
the author, J ourne (1970), on this topic.
1 Introduction
The horizontal motions of a vessel due to a
rudder deflection can be described by
using mathematical models. The most
simple system, the first order model of
Nomoto (1960), is a compromise between
the demand for a simple mathematical
model and a fair approximation of the
actual manoeuvres of the ship.
Nomoto has published some methods to
estimate the manoeuvring indices K and
T from (full-scale) zigzag trial data. But,
using these index estimators, it appears
often that the calculated manoeuvres
deviate considerably from those measured
on the ship. Because of this, a method has
been developed to determine these indices
in such a way that the differences between
calculated and measured data, as far as
yaw period and overshoot are concerned,
are as low as possible.
Using Nomotos first order model, a large
number of zigzag manoeuvres have been
calculated here at a practical range of K
and T values. These data have been
analysed and the relation between the
zigzag manoeuvring characteristics and the
Nomoto indices K and T have been
reflected in graphs. Reversed, these graphs
2
can be used then to determine these indices
from actual zigzag manoeuvres.
2 Equation of Motion
The first order model of Nomoto (1960)
reads as follows:
( )
r
K T + ! ! !
Equation (1)
In here:
! ! Yaw angular acceleration
! Yaw angular velocity or
rate of turn
(deg/s
2
or rad/s
2
)
Yaw angle
(deg/s or rad/s)
Actual rudder angle
(deg or rad)
r
Effective rudder angle
(deg or rad)
K Proportionality constant
(1/sec)
T Time constant
(s)
3 Actual and Ideal Kempf Zigzag
Manoeuvres
An ideal Kempf zigzag manoeuvre has
to fulfil the following requirements here:
equal absolute values of rudder angles,
equal absolute values of rudder angle
velocities and
rudder-order when .
It is obvious that an actual zigzag
manoeuvre cant fulfil these specific
demands exactly. It is possible however, to
transform an actual manoeuvre to an ideal
manoeuvre with sufficient accuracy.
Therefore, computed ideal Kempf zigzag
manoeuvres have been analysed here.
4 Analysis of First Order Model
Using Nomotos first order model, a large
number of zigzag manoeuvres have been
calculated at a practical range of K and T
values, with the following parameters:

a
the level of the rudder angle,

r
the rudder angle at which the ship
sails a straight course and

!
the rate of turn of the rudder.
The following parameters have been
obtained from the computed course
histories:

p
t the period,

g
the mean course and

a
the maximum course deviation
relative to
g
.
To avoid transient effects, the third period
has been used to determine these
magnitudes, see figure 1.
Figure 1 Ideal Kempf Zigzag Test
The following relations have been found
by analysing these calculated data:
1
2
2
2
1
1
1
0
C t
C
t t
r
r
a
p p
+

'

,
_

Equation (2)
4
2
3
2
2
1
0
C C T K
t K
a
r
a
r
a
a
a
a

,
_

+ +

Equation (3)
3
r
r
g
t K C
2
1
09 . 1
5

Equation (4)
In here:

a
,
g
,
a
and
r
in degrees

p
t ,
0
p
t and
r
t in seconds.

!
/ 2
a r
t is the time of rudder
action

0
p
t is the period for manoeuvres with

!
and 0
r

0
a
is the yaw amplitude for
manoeuvres with
!
and 0
r

T t
p
/
0
,
a a
/
0
,
1
C ,
2
C ,
4
C and
5
C
are functions of the product T K

3
C is a function of
!
T
These relations are given in the figures 5, 6
and 7 as will be described further on.
5 Transformation of Actual Zigzag
Manoeuvres to Ideal Manoeuvres
The largest deviation of an actual zigzag
test from an ideal one is generally caused
by not fulfilling the requirement
0

ex
, see figure 2.
Figure 2 Manoeuvre Transformation
Because of transient effects, the second
and the third period of the time history has
to be used to determine mean values of
a
,
!
,
a
,
g
and
p
t , being the values
for an ideal manoeuvre.
However, when supposing that 0 ! ! at
1
t t , two additional corrections have to
be used to obtain the values
a
and
p
t for
an ideal manoeuvre:
( ) ( )
( )
1 2 ideal -
0 max ideal - max
t t t t
p p
ex


6 Determination of K and T from a
Zigzag Trial
As the ideal values of
a
,
!
,
a
,
g

and
p
t are known, the unknowns in the
three equations (2), (3) and (4) are K , T
and
r
. These equations can be solved in
an iterative way.
Doing this manually, is very time
consuming. Without a large influence on
the accuracy however, these equations can
be simplified by assuming that:
0
25 . 0 4
3
1

<<
C
T K C
a r

Then, equations (2) and (3) reduce to:
r r p p
t T K t t t + 125 . 0 2
0
Equation (5)
and
r
a
a
a
a
t K
2
1
0

Equation (6)
From derivations in the Appendix follows:
4

,
_

,
_

T K
T K t t t
T K
t
T
r r p
p
1
4
125 . 0 2
1
4
0
Equation (7)
In equation (7) is a function of T K
only, see figure 4. Also follows from the
derivations in the Appendix that
a a
/
0
is
a function of T K only, see figure 4.
Thus, with known ideal values of
a
,
a
,
p
t and
r
t , the coefficients K and T
can be found from equations (5), (6) and
(7) by a simple and fast iteration.
The following procedure provides the
coefficients K and T in a very quick and
simple way:
1. As a first guess for
a a
/
0
: ignore
r
t K 5 . 0 in equation (6).
2. At
a a
/
0
, figure 4 provides now
T K and .
3. Equation (7) provides T .
4. From T K and T follows K .
5. Using equation (6), this procedure will
be repeated from step 2 with a new
guess for
a a
/
0
.
The procedure will be terminated as K
and T do not change anymore.
7 Acknowledgement
The author is very grateful to Mr. G. van
Leeuwen for his support when solving
mathematical problems during the analysis
of this first order system.
8 References
Journe (1970)
J .M.J . J ourne, Een eenvoudige methode
ter bepaling van de manoeuvreer-indices
K en T uit zig-zag proeven, Report 267,
1970, Ship Hydromechanics Laboratory,
Delft University of Technology, The
Netherlands.
Nomoto (1960)
K. Nomoto, Analysis of Kempfs Standard
Manoeuvre Test and Proposed Steering
Quality Indices, First Symposium on Ship
Manoeuvrability, DTRC Report 1461,
October 1960.
9 Appendix
From the equation of motion, given in
equation (1), it can be found for
!
and 0
r
(see figure 3):
( )
T t T t
e K e
/
0
/
0
1

+ ! !
Equation (8)
and
( )
( ) T t e T K
e T
T t
T t
/ 1
1
/
0
/
0 0

+

!
Equation (9)
From equations (8) and (9) follows:
( )
( ) T t e T K
e T
T t
T t
/ 1
1
/
0
/
0

+

!
Equation (10)
5
Figure 3 Infinite Rudder Rate of Turn
Values at t
2
:
From equation (8) follows:
( )
T t
a
T t
e K e
/ /
0 2
2 2
1

+ ! !
Ignoring the transient effects yields:
0 2
! !
Thus:
T t
T t
a
e
e
K
/
/
0
2
2
1
1

!
Equation (11)
From equation (9) follows:
( )
( ) T t e T K
e T
T t
a
T t
/ 1
1
2
/
/
0 0 2
2
2

+

!
Because
a
+
0
and
a

2
is:
( )
( )
T t
T t
a
e T
T t e T K
/
2
/
0
2
2
1
/ 1 2

!
Equation (12)
From equations (11) and (12) follows:
T t
T t
e
e
T K T
t
/
/
2
2
2
1
1
2
2

When renaming:
+

T t
T t
e
e
/
/
2
2
1
1
, then is
a function of
T
t
p
0
, because:
2
2
0
t t
p
.
Then the following relation appears for
0
p
t :
+

4
4
0
T K T
t
p
Equation (13)
From the equations above follows that ,
T
t
2
and
T
t
p
0
are functions of the product
T K .
In figure 5,
0
p
t has been plotted against T
with K as parameter. Equation (13) and
this figure show that
K
t
p
4
0

for 0 T
and that
0
p
t for 0 K .
Values at t
1
:
From equation (8) follows:
( )
T t
a
T t
e K e
/ /
0 1
1 1
1

+ ! !
Because 0
1
! is:
( ) 1
/ 0
1

T t
a
e K

!
Equation (14)
Then, from equations (11) and (12)
follows that:
1
1
1
/
/
/
1
2
2

+

T t
T t
T t
e
e
e

Because is a function of the product

T K , also T t /
1
will be a function of
T K .
From equation (10) follows that:
6
( ) T t e T K
T t
a
/ 1
1
/
0 1
1

Thus, because
a
+
0
and
0
1 a
is:
( ) T t e T K
T t
a
a
/ 1 1
1
/
1
0

Equation (15)
Because T t /
1
is a function of T K , also
a
a

0
will be a function of T K .
Figure 6 shows
a
a

0
as a function of
T K . Equation (15) and this figure show
that 1
0

a
a

for 0 T and that 0

0

a
a

for 0 K .
However, from the previous follows too
that both T K and are functions of
a
a

0
.
In figure 4,
T K
1
and have been
plotted against
a
a

0
.
7
Figure 4 Parameters and
T K
1
as a Function of
a
a

0
8
Figure 5 Period
0
p
t as a Function of K and T
9
Figure 6 Transfer Function
a
a

0
as a Function of T K
10
Figure 7 Correction Coefficients