See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269873875

An Approximate Method for Cross Curves of Cargo Vessels

Article  in  Marine technology · April 2001

CITATIONS READS
2 373

1 author:

Huseyin Yilmaz
Yildiz Technical University
14 PUBLICATIONS   155 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Environmental Effect on Maritime Activities" Research Works in JICA project View project

Characteristics of Roll Damping of Monohull View project

All content following this page was uploaded by Huseyin Yilmaz on 04 November 2019.

The user has requested enhancement of the downloaded file.

Marine Technology, Vol. 38, No. 2, April 2001, pp. 92–94

An Approximate Method for Cross Curves of Cargo Vessels

Hüseyin Yılmaz1 and Mesut Güner1

In this study, a formula is presented to estimate cross curves of cargo vessels and to predict statical
stability at the preliminary design stage of the vessel. The predictive technique is obtained by regression
analysis of systematically varied cargo vessel series data. In order to achieve this procedure, some cargo
vessel forms are generated using Series-60. The mathematical model in this predictive technique is
constructed as a function of design parameters such as length, beam, depth, draft, and block coefficient.
The prediction method developed in this work can also be used to determine the effect of specific hull form
parameters and the load conditions on stability of cargo vessels. The present method is applied to a cargo
vessel and then the results of the actual ship are compared with those of regression values.

Introduction that at the initial stage of design, stability cross curves (KN-
T), which are a function of ship draft, are proposed to be
THE analytical, statistical, and empirical formulations for estimated by using approximate statistical methods. As can
ships have generally been based on theoretical or experimen- be seen from the geometrical features of an inclined ship
tal information. However, the naval architect needs mostly given in Fig. 1, the righting moment arm (GZ) value is cal-
some practical techniques to obtain appropriate designs and culated as follows in terms of KN value:
some indication of an acceptable level of performance in as-
suring vessel safety at the preliminary design stage. There- GZ = KN − KG sin⌽ (1)
fore, the stability of ships can be considered crucially impor-
tant for the naval architect. Within this context, approximate A number of cargo vessels have been derived from Series-
expressions for the ship stability computations can be ob- 60 for the regression modeling of cross curves. The lengths of
tained by analyzing a large amount of computer calculations. those vessels are kept as a constant and their actual cross
The development of ship stability methods has had a long curves are used in the regression analysis method. A total of
period of evolution. During that period, many studies have 72 different hull forms are generated by incrementing the
been carried out on various aspects of the subject. For ex- L/B ratio by 0.5 where it provides a sufficient number of
ample, the Prohaska method [1] has been used in stability sample stability data for reliable statistical evaluation. For
calculations for a long time. There are some papers concern- each L/B ratio, the B/T value is varied by 0.25 increments,
ing the application of regression analysis for calculating the and then for each B/T ratio the block coefficient (CB) is in-
stability of ships, such as Campanile and Cassella [2], Ku- cremented by 0.01. All calculations are considered up to the
pras [3], and Yılmaz and Kükner [4]. Campanile and Cassella main deck.
have investigated the reduction of form stability for Series-60 The parameter range of generated vessels is set as follows:
ship forms in waves and have established some graphs for 5.0 < L Ⲑ B < 7.5
the practical calculation of the righting moment arm. Kupras 2.25 < B Ⲑ T < 3.0
has modeled KN-⌽ values based on design parameters for 0.60 < CB < 0.80
cargo ships at certain heeling angles. Yılmaz and Kükner 1.3 < D Ⲑ TC < 4.0
used a similar method as shown in this study to develop a
mathematical model for computation of cross curves of fish- where D/TC is the depth-draft ratio.
ing vessels at the preliminary design stage. The actual cross-curve computations of the vessels are per-
The objective of this paper is therefore to derive a math- formed by a well-established stability software. The results
ematical model for estimating cross curves of cargo vessels by obtained from the implementation of the software are used to
use of regression analysis. The mathematical model based on establish approximated stability expression.
hull form parameters is established for cross curves. There- Since the (KN-⌽) values have been modeled in terms of
fore, the systematically varied hull series [5] has been chosen design parameters, it is necessary to use the values of KN in
for the application of regression analysis to obtain an ap- the nondimensional form. Hence, the KN values should be
proximate expression for the stability characteristics of cargo nondimensionalized by the molded beam B, which is one of
vessels. the most effective design parameters in a vessel’s cross-curve
calculations. At the beginning of the regression modeling, the
(KN-⌽) curve has been fitted for the fifth-degree polynomial
Modeling of cross curves (KN-⌽) values by using polynomial regression as follows:
The prediction of ship stability at the preliminary design 5
stage is very difficult since the statical stability of a ship
cannot be easily predicted without knowing the ship’s hull
KN Ⲑ B = 兺a ⌽
i=1
i
i
(2)
geometry and her center of gravity. Therefore it is intended
where ⌽ is the heel angle in radians and ai is the polynomial
coefficient which is determined by the application of the
1
Assistant professors, Department of Naval Architecture, Yıldız least-squares method. In order to determine ai coefficients as
Technical University, 80750 Beşiktaş, Istanbul, Turkey. a function of design parameters, the following multiple linear
Manuscript received at SNAME headquarters May 24, 2000. regression model can be used:

92 APRIL 2001 0025-3316/01/3802-0092$00.31/0 MARINE TECHNOLOGY

 Table 1 List of regression coefficients for KN calculation

Fig. 1 Geometric representation of a heeled vessel

5 5

ai = 兺兺b
i=1 j=0
ij Xj (3)

where Xj is the jth design parameter and bij is the coeffi-

cient of the jth design parameter for the ith coefficient of the
polynomial expression, ai. The selection of design parameters
(Xj, j ⳱ 1,2,..,5) can be achieved by systematic trial in which
a parameter or combination of some parameters should give
the highest possible correlation with the computed righting
arm curve defined by KN-⌽ values. The selected parameters
used in this study are given in the following forms:

X0 = 1
X1 = B Ⲑ T
X2 = CB
X3 = 共B Ⲑ T兲2
X4 = 共CB兲2
X5 = 共B Ⲑ T兲 CB Length (L) : 120 (m)
Breadth (B) : 20 (m)
It is assumed that the draft is changed at different loading Depth molded (D) : 12 (m)
conditions. Therefore, the bij coefficients given in equation (3) Design draft (T) : 8 (m)
have been determined at each loading condition. Computation draft (TC) : 8 (m)
The other design parameter variations with the loading Block coefficient (CB) : 0.70
condition have been considered as constant since it is difficult
to carry out the calculations at the initial design stage. Thus, The computation procedure for the righting arm curve (KN-
the variations of bij coefficients are defined with the free- ⌽) of the proposed cargo vessel should be as follows:
board variation (D/TC) only. For the inclusion of loading
variation, the following third-degree polynomial regression 1. Calculate regression parameters, Xi
with a maximum error of 2% was used: X0 = 1
5 5 3 X1 = B Ⲑ T = 2.5
bij = 兺兺兺c
i=1 j=0 k=1
ijk␦
k
(4) X2 = CB = 0.70
X3 = 共B Ⲑ T兲2 = 6.25
where ␦ is D/TC. X4 = 共CB兲2 = 0.49
Using the regression coefficients (cijk) given in Table 1, the X5 = 共B Ⲑ T兲 CB = 1.75
righting arm curve can be easily computed for a given set of ␦ = 共D Ⲑ Ti兲 = 1.5
design parameters. 2. Calculate bij coefficients
bij = cij0 + cij1 ␦ + cij2 ␦2 + cij3 ␦3
Application of the method 共i = 1,2,..,5 j = 1,2,..,5兲
The presented regression model has been applied to a cargo b10 = −2.328721 + 4.109446 * 共1.5兲
vessel, whose principal dimensions are: − 1.458433 * 共1.5兲2 + 0.135208 * 共1.5兲3

Nomenclature
ai, bij, cijk⳱ regression coefficients of KN-⌽ GM ⳱ metacentric height L ⳱ length of between perpendiculars
B ⳱ molded breadth GZ ⳱ righting moment arm T ⳱ design draft
BM ⳱ transverse metacentric radius KB ⳱ height of center of buoyancy TC ⳱ calculating draft
CB ⳱ block coefficient KG ⳱ vertical center of gravity ␦ ⳱ D/TC ratio
CP ⳱ longitudinal prismatic coefficient KN ⳱ form stability lever in keel ⌽ ⳱ heel angle
D ⳱ depth

APRIL 2001 MARINE TECHNOLOGY 93

 b10 ⳱ 1.01030
b11 ⳱ −0.32306
b12 ⳱ −0.58242 b13 ⳱ 0.05861
b14 ⳱ 0.41972 b15 ⳱ 0.00160
b20 ⳱ −1.17982 b21 ⳱ 0.94881
b22 ⳱ −0.10926 b23 ⳱ −0.15424
b24 ⳱ −0.29547 b25 ⳱ 0.16178
b30 ⳱ 1.80998 b31 ⳱ −1.43292
b32 ⳱ 1.27342 b33 ⳱ 0.15611
b34 ⳱ −0.08859 b35 ⳱ −0.34924
b40 ⳱ −1.10720 b41 ⳱ 0.71352
b42 ⳱ −1.0782 b43 ⳱ −0.02472
b44 ⳱ 0.13501 b45 ⳱ 0.244862
b50 ⳱ 0.22244 b51 ⳱ −0.09857
b52 ⳱ 0.24750 b53 ⳱ −0.01418
b54 ⳱ −0.01978 b55 ⳱ −0.05311
3. Calculate the ai coefficients
ai = bi0 X0 + bi1 X1 + bi2 X2 + bi3 X3 + bi4 X4 + bi5 X5 Fig. 3 Comparison of KN-T values
a1 = b10 X0 + b11 X1 + b12 X2 + b13 X3 + b14 X4 + b15 X5
a1 = 1.010301 − 0.3230642 * 共2.5兲 − 0.5824233 * 共0.70兲 The actual and suggested methods of computing results
+ 0.05861723 * 共6.25兲 + 0.4197221 * 共0.49兲 of (KN-⌽) values for the sample cargo vessel are shown in
+ 0.001608094 * 共1.75兲 Fig. 2.
= 0.3697794 For other loading conditions, the KN values have been
computed by changing the draft values. The KN-T values of
a1 = 0.3697794 the actual and suggested methods of computing results for
The rest of the ai coefficients have been obtained in the same different heeling angles have been plotted in Fig. 3.
way:
Conclusions
a2 ⳱ 0.2900721 a3 ⳱ −0.5597836
a4 ⳱ 0.2619391 a5 ⳱ −0.04202612 The present method and the computational tool given in
this paper can be successfully applied to the calculation of a
4. Calculate KN values cargo vessel’s cross curves at the preliminary design stage.
The accuracy of the method is based on the hull form param-
KN = 共a1⌽1 + a2⌽2 + a3⌽3 + a4⌽4 + a5⌽5兲 B eters as described herein. It can be seen from the computa-
KN = 共0.3697794 ⌽ + 0.2900721 ⌽2 − 0.5597836 ⌽3 tional results that the present method gives very high accu-
+ 0.2619391 ⌽4 − 0.04202612 ⌽5兲 20 racy for Series-60 and for other different cargo hull
geometries as well, and the level of error is acceptable. If the
where ⌽ is in radians. approximate KG value is known for a cargo vessel, the stati-
cal stability calculation of the vessel can be executed at the
preliminary design stage.
One of the other important results of this study is that it
provides the designer with appropriate design parameters for
a cargo vessel from the point of view of vessel stability.

References
1. Prohaska, C. W., “Influence of Ship Form on Transverse Stability,”
TINA, 1951.
2. Campanile, A. and Cassella, P., “Form Stability Reduction Among
Waves for Series-60 Hulls,” Ocean Engineering, Vol. 16, 1989, pp. 431–
462.
3. Kupras, L. K., “Optimization Method and Parametric Study in Pre-
contracted Ship Design,” International Shipbuilding Progress, Vol. 23,
1976, pp. 138–155.
4. Yılma, H. and Kükner, A., “Evaluation of Cross Curves of Cargo
Vessels in Preliminary Design Stage,” Ocean Engineering, Vol. 26, No. 10,
1999, pp. 979–990.
5. Todd, F. H., “Some Further Experiments on Single-Screw Mer-
chant Ship Form—60 Series,” Transactions, SNAME, Vol. 1, 1953, pp.
516–589.
6. Yılmaz, H., “The Determination of Practical Stability Criteria De-
pending Upon Ship Design Parameters,” MS thesis (in Turkish), Yıldız
Fig. 2 Comparison of KN-⌽ values Technical University, Istanbul, 1994.

94 APRIL 2001 MARINE TECHNOLOGY

View publication stats