Professional Documents
Culture Documents
Abstract
In this study,, an identification procedure is proposed to estimate the hydrodynamic parameters of a range of ship
maneuvering models, by coupling the dynamic ship motion model with mathematical programming techniques. In
order to assess efficiently the hydrodynamic parameters, a sensitivity analysis is first performed to identify the
most sensitive coefficients. Different Mathematical Programming Techniques have been used and compared in the
determination of optimal hydrodynamic parameters.
parameter
The proposed procedure has been validated through turning circle and zigzag manoeuvres based on experimental
data of sea trials of the 190,000-dwt
dwt oil tanker. Comparisons between experimental and computed data show that
the RMSD (Root-Mean-Square Deviation) of ship trajectory decreases from 68.0m to 5.8m in Turning Circle test,
and RMSD of ships heading angle decreases from 17.3deg to 6.6deg in Zigzag.
Keywords: Ships; waterway; maneuverability; hydrodynamic forces; parameters
ters identification.
identification
Rsum
Une procdure didentification de paramtres est propose pour estimer les paramtres hydrodynamiques des
bateaux dans la poursuite de trajectoires optimales dfinies au pralable pour des navires inconnues ou connus
(Esso Bernicia,
cia, Mariner Class Vessel,..). Elle est base sur le couplage des techniques doptimisation avec un
modle mathmatique de mouvements
vements de navires en manuvre. Ainsi, une analyse de sensibilit, est effectue
dans un premier temps pour identifier les coefficients hydrodynamiques les plus sensibles. Ensuite,
Ensui les valeurs
optimales de ces coefficients
efficients hydrodynamiques sont identifies en utilisant les techniques doptimisation.
doptimisation
La procdure propose a t valide en utilisant des donnes exprimentales des
des essais en mer de l'Esso Bernicia
190000dwt pour les tests classiques de Giration et Zigzag.
Zigzag Les comparaisons entre ces donnes et les rsultats
numriques montrent que l'cart
'cart quadratique moyen diminue de 68.0m 5.8m pour la trajectoire du navire dans le
cas du test de Giration et de 17.3deg 6.6deg pour langle de cap dans le cas du test de Zigzag.
Zigzag
Mots-cl: Navires; voies navigables ; manoeuvrabilit; forces hydrodynamiques ; identification de paramtres.
Tran K., Ouahsine A., Hissel F. and Sergent P./ Transport Research Arena 2014, Paris
Nomenclature
x, y
ships position
u
v
r=
c
n
X
Y
G
Fobj
function objective
1. Introduction
Itt is still tedious to perform real ship manoeuvres in an open sea or to carry out fine simulations using complex 3D
CFD calculations. Even though their fast calculations, system-based
system based simulations need numerous tests to adjust the
manoeuvring hydrodynamic coefficients (hull, rudder, propeller,...) in order to achieve quantitative agreement
with the experimental
al measurements. In this context, manoeuvrability turns out to be an essential ability to
perform a safe navigation of any ship against the danger of collisions and stranding.
A literature review (Rawson and Tupper,
Tupper 2001),( Arak et al., 2012 showed that there
ere exist three methods for ship
manoeuvrability prediction: the experimental based method, the Computational Fluid Dynamics (CFD) based
method and the system based method. (i) Manoeuvring experiments are based on the verification of
manoeuvrability tests to fulfill the International Maritime Organization (IMO) criteria, the determination of
hydrodynamic manoeuvring coefficients. Unfortunately experimental determination of the hydrodynamic
coefficients can be tedious and expensive. (ii) The CFD based methodd allows determining not only the motion of
the ships but also the flow field around ships by solving a set of RANS (Reynolds-Averaged
(Reynolds Averaged Navier-Stokes)
Navier
equations (Ji et al., 2012).. Lately, CFD based method has been extended for free-running
free running simulations, including
incl
measurable data from Experimental Fluid Dynamics for ship manoeuvres in calm sea and in the presence of
waves. However, CFD method needs few hours or even days depending on the turbulence and propulsion
modeling and the mesh size.. (iii) Elsewhere, system-based
based method is a major simulation task to predict ship
manoeuvrability. Computation time is much smaller than that of CFD based method,
method, and requires approximately
few minutes of computation on a Personal Computer for one free-running
free
trial while.
System-based
based methods have been extensively investigated by researchers (Sutulo et al., 2002, Neves and
Rodrigues, 2009). The course keeping stability of the ship can be investigated on the basis of the stability of the
solutions of the linear equations of motion,
motion, if only first order terms of this expansion are considered.
considered However, the
manoeuvres at high rudder deflection angles require the consideration of nonlinear hydrodynamic and inertial
components. This leads to the utilization of nonlinear hydrodynamic models, which include higher order terms of
Taylors series expansion of the hydrodynamic external forces and moments (Dan,
Dan, 20010).
The herein proposed paper presents an efficient procedure to determine optimal hydrodynamic coefficients by
using a mono-objective
bjective optimization based on a mathematical programming (MP) technique suitable for highly
nonlinear problems such us ship manoeuvring simulation from sea trials. Accurate modeling of a ship trajectory is
achieved effectively using three steps. The first step deals with the modeling of the nonlinear dynamic ship
motion, to this end a 3DOF mathematical model based on the Lagrangian dynamic motion of a 3D rigid body has
been developed taking into account the nonlinear hydrodynamic forces acting on the ship hull, propeller and
rudder. Then a sensitivity analysis is carried out to identify the most significant hydrodynamic coefficients which
affect the ship trajectory. The interest of applying sensitivity analysis is to reduce the number of coefficients to be
optimized.. In the present investigation, it has been found that only 14 coefficients were sensitive for the prediction
of the trajectory of the ESSO 199,000-dwt
199,000
oil tanker benchmark (Manoeuvring
anoeuvring committee,
committee 2005).
The structure of the paper is as follows. The
T nonlinear hydrodynamic model in given in section 2. Section 3
presents the mathematical programming based system identification for manoeuvring of large tankers. Section 4 is
devoted to the numerical results and discussions.
discussions Conclusions are drawn in section 5.
Tran K., Ouahsine A., Hissel F. and Sergent P / Transport Research Arena 2014, Paris
According to the Newton second law, the nonlinear transient equations of motion in the ship moving coordinate
system Oxyz can be written in the form (Tran,
(
2012).
u v r xG r 2
=g
v + u r + xG r
=g
2
2
L k z r + xG ( v + u r ) = gL
(1)
X ,
Y ,
N are the non-dimensional
non
forces and moment respectively, where m is the mass of the
=
=
mg
mg
mgL
ship; X and Y the external forces; N is the external moment; I z is the moment of inertia about the z axis; r =
is heading/yaw, (u,v) are the components of the velocity in the x-axis
x
and y-axis
axis directions, respectively.
I
dimensional radius of gyration.
gyration
k z = z L is the non-dimensional
m
=
= u L u + u u u u + vr vr + v v v v + c c 2 c c + c c c c
+ gL(1 td ) + u u + u u u u + vr vr + v 2 2 vv
gL
= v L v + uv uv + v v v v + c c c c + ur ur + c c c c
+ g s L T + ur ur + uv uv + v v + v v L v v + c c c c
gL
= r r + uv uv + v rL v r + c c c c + urL ur + c c c c
+ g s L T + ur L ur + r L2 r + uv uv + vr L vr + c c c c
gL3
(2)
Tran K., Ouahsine A., Hissel F. and Sergent P./ Transport Research Arena 2014, Paris
where is the ship drift angle, u , u u ,.., v , uv ,..., r , uv ,..., c c are the non-dimensional
dimensional derivatives of
ship hydrodynamic coefficients, which have to be identified.
2
2
c = nu cun
+ n 2 cnn
(3)
(4)
where u , v, r are the ship velocity components , x0 , y0 , the ship position components in the reference frame
O0 x0 y0 z 0 , is the actual rudder angle and n is the shaft velocity.
The complete manoeuvrability of ship is carried out by assembly and solving the set of ODE given by (1),
( the yaw
rate = r and the rates
, n
u
= g + v r + xG r 2
v + xG r
= g u r
L2 k z2 r + xG v = gL xG u r
x0
= u cos v sin
y 0
= u sin + v cos
=r
= c
60 ( nc n)
n
=
Tm
(5)
velo
The
Tm is the coefficient of propeller, c is the commanded rudder angle, nc is the commanded shaft velocity.
equation (6)) can be rewritten in a more convinient form, as:
x T = f ( t , x )
(6)
x t +t
xt +t xt the discret implicit Euler solution of (7), with respect to t and x , reads:
t
xt + t = xt + t f (t + t , xt + t )
(7)
The latter system of of nonlinear equations (10) was solved using a Newton-Raphson
Raphson (NR) iterative procedure
and the 4th order implicit Runge-Kutta
Kutta scheme .
Tran K., Ouahsine A., Hissel F. and Sergent P / Transport Research Arena 2014, Paris
The optimization (or the mathematical programming problem) can be stated as follows (souli et al. 1999):
Find
= {1 , 2 ,..., N }
which minimizes Fobj ( )
(8)
(9)
j = 1,2,..., N g
lk ( ) = 0,
k = 1,2,..., N l
where is an N -dimensional
dimensional vector called the design vector which contains the Design
esign Variables (DV), in
our case they are represented by the ship hydrodynamic coefficients to be determined, Fobj ( ) is called the
Objective Function (OF), and g j ( ) and lk ( ) are known as inequality and equality constraints,
constraints respectively.
The number of variables N and the number of constraints N g and/or Nl need not be related in any way.
In this work, a single OF is used in the optimization problem in order to identify ship hydrodynamic
coefficients. Therefore, for the turning circle manoeuvre for instance, the OF is chosen as
a
1/2
Fobj = Si2
i =1
(10)
where p is the number of sampling points, Si2 depending on ship hydrodynamic coefficients , reads:
) (
2
(11)
Superscripts num and exp indicate the computed and experimental data respectively, ( xi , yi ) are the
coordinates of the point i . For the case of a zigzag test, the expression of Fobj reads:
p
Fobj = inum iexp
i =1
1/2
)
2
(12)
where i is the ships heading angle, which depends also on ship hydrodynamic coefficients . The
optimization problem stated in (88) and (9) is solved by using the so-called mathematical programming
techniques. In the present investigations we use the Sequential Quadratic Programming (SQP) algorithm
(Zhang. J et al. 2003) and the BFGS (Broyden-Fletcher-Goldfarb-Shanno)
(Broyden
algorithm (Dai, 2002).
3.2. Normalization and sensitivity analysis
Normalization of DV consists in a linear transformation of the original variables into new transformed
variables , given by:
= A+B
(13)
Tran K., Ouahsine A., Hissel F. and Sergent P./ Transport Research Arena 2014, Paris
The normalization of the objective function (OF) is accomplished by dividing the objective function at each
iteration by Fobj0 F obj = Fobj , where Fobj0 is the value of the objective function at the first iteration. Once the
0
Fobj
normalization is applied, the gradients of the OF have to be adapted to the new set of design variables.
Fobj
F obj 1 Fobj
=
A
F 0
obj
(14)
The
he numerical procedure for the identification of hydrodynamic coefficients based on MP techniques, can be
summarized into three main steps:
1- Compute the ship trajectory with original coefficients from dynamic ship equations given by (6).
(
2- Filter the most sensitive coefficients S among all others , based on the sensitivity analysis.
3- Calculate optimal values opt for only sensitive coefficients S , by carrying out the optimization
procedure using SQP or BFGS algorithms.
4. Numerical Applications, Results and Discussion
The proposed procedure has been validated through turning circle and zigzag manoeuvres accordingly to the
IMO and based on experimental data of sea trials of the ESSO 190,000 -dwt
dwt oil tanker (ITTC 2005). The
associated physical and geometrical
geometrica parameters are given in Table 1.
Table 1.. Parameters of the ESSO 190,000-dwt
190,000
oil tanker (ITTC 2005)
Physical and geometrical parameters
Value
304.8 m
B : beam
47.17 m
18.46 m
220000 m3
Lpp /B
6.46
B/T
C B : block coefficient
2.56
U 0 : design speed
16 knots
n : nominal propeller
80 rpm
0.83
The associated input data related to the turning circle the zigzag manoeuvre tests are presented in Table 2.
x0
Table 2.. Input data for turning circle and zigzag (ESSO 190,000-dwt
190,000 dwt oil tanker model).
Input data
Turning circle test
zigzag test
(0,0)
m
(0,0) m
, y : initial ships position
0
0 deg
0 deg
5.3 m/s
7.5 m/s
0 deg
0 deg
2.7 deg/s
2.7 deg/s
57 rpm
80 rpm
57 rpm
80 rpm
-35 deg
[-20,+20]
20,+20] deg
Tran K., Ouahsine A., Hissel F. and Sergent P / Transport Research Arena 2014, Paris
In the present application, there is 35 hydrodynamic coefficients to control the manoeuvrability of the ESSO
190,000 -dwt oil tanker model (ITTC 2005). These coefficients (see Table 5) were used as as initial guess for
the numerical processing. Hence, the identification procedure starts from original reference values of all
hydrodynamic coefficients in the dynamic ship motion Eqs.(7),
Eqs.( which will be firstly analyzed through a
sensitivity analysis procedure to select the most important parameters. The sensitivity analysis consists of
filtering the largest gradient corresponding to each i , which is accomplished using the following
followin criteria:
F obj
>
i
(15)
the gradient Fobj . By using the developed manoeuvring model, the sensitivities are computed and filtered
using the above criterion. Variations of gradients are shown in Fig.3 for the turning circle and zigzag
manoeuvres. Tthis figure shows that only a few coefficients are of a great sensitivity (greater than 10% ). It
shows also, that only 6 coefficients are common for the two turning circle and zigzag manoeuvres (see table
5);; this indicates that it is important to include the physical knowledge of the hydrodynamic problem, in the
identification process in order to insure a good result.
The proposed procedure has been validated using Eq.(12) for the turning circle test. Figure 4 (left) shows
the computed ship trajectory before optimization, i.e. when starting with the initial reference values of
hydrodynamic coefficients, the total cumulative error on the ship trajectory was S = 68m .
We defined p =40 as the total number of experimental sampling points in the Eq.(1
(12). The identification
procedure has been carried out by using SQP and BFGS algorithms with a convergence criterion based on
the error of the OF as 1E 04 and a maximum number of iterations of 50 . Convergence has been achieved
ach
in
Tran K., Ouahsine A., Hissel F. and Sergent P./ Transport Research Arena 2014, Paris
21 iterations for the SQP algorithm with a final optimal relative value of 0.084 and only in 10 iterations for
the BFGS algorithm with a final optimal relative value of 0.12 (see Table 3).
Table 3.. Summary of final optimal solutions obtained by both SQP and BFGS in turning circle test.
BFGS
SQP
Total nb of iterations
21
10
1E-04
1E-04
Final error on F obj
Final optimal value of
Fobj
(m)
0.084
0.12
5.8
8.0
From Table 3,, we can see that the SQP algorithm is more precise than the BFGS algorithm because it led to
a minimal cumulative error on the trajectory of 5.8 m only. The results are shown in Fig.4
Fig. (right-a, b).
4.2. Identification of hydrodynamic coefficients using a zigzag test
The second application used to show the efficiency of the proposed procedure is a zigzag test. In this case,
we define
ine an OF based on the Eq.(12).
Eq.(
The OF (10)) used in the turning circle test, cannot be used in the
present application because in the zigzag test there will be large variations in the heading angle during the
dynamic ship motion. As previously, at first the
the initial reference values of hydrodynamic coefficients, given
in Table 5, are used in the developed ship manoeuvring model to evaluate the predicted heading angle.
Figure 10 shows the calculated ship heading angle using our manoeuvring model. It is interesting
intere
to notice
that before optimization, i.e. when starting with the initial reference values of hydrodynamic coefficients,
coefficients the
total cumulative error on the ship trajectory was = 17.3 deg .
Fig.5 Left: Comparison of the heading angle using initial reference hydrodynamic coefficients.
Right: Comparison of the heading angle after optimization by:: a) SQP algorithm, b) BFGS algorithm.
In this application p=53 is chosen as the total number of experimental sampling points in the Eq.(12). The
identification procedure has been carried out by using SQP and BFGS algorithms with a convergence
criterion based on the error on the OF as 1E 04 and a maximum number of iterations of 50. Convergence
has been achieved in 15 iterations for the SQP algorithm with a final optimal OF relative value of 0.365 and
only in 7 iterations for the BFGS algorithm with a final optimal relative value of the OF of 0.389. A
summary of the final optimal solutions obtained by both SQP and BFGS
BFGS algorithms is given in Table 4.
Table 4.. Summary of final optimal solutions obtained by both SQP and BFGS in zigzag test.
SQP
BFGS
Total nb of iterations
15
7
Final error on F obj
Final optimal value of
Fobj
(deg)
1E 04
1E 04
0.365
0.389
6.6
7.1
Figure 5 (right-a,b) shows the optimal solution obtained at the end of the optimization process,
process where Table
4 shows that minimal cumulative error on the heading angle is 6.6 deg for the SQP algorithm and is 7.1 deg
for the BFGS algorithm. The optimal hydrodynamic coefficients obtained at the end of the optimization
process aree summarized in Table 5.
5. We can observe that both SQP and BFGS algorithms predicted correctly
Tran K., Ouahsine A., Hissel F. and Sergent P / Transport Research Arena 2014, Paris
the experimental ship heading angle. The comparison of the obtained heading angle shows that the result
obtained using the SQP algorithm is more precise than the BFGS one, since the total cumulative error
between the experimental trajectory and the predicted one is very small relatively compared to the length of
the ESSO 190,000 -dwt
dwt oil tanker.
From the result of coefficient identification for turning circle and zigzag manoeuvres, we can choose finally
a set of identified hydrodynamic coefficients for ESSO 190,000 -dwt
dwt oil tanker model, which may be used
for both two tests as well ass for other manoeuvring simulations. This set has 35 hydrodynamic coefficients,
which includes the mean optimal values of 6 shared most sensitive coefficients, the optimal values of 8
independent most sensitive coefficients of each test, and 21 other original
original coefficients that influence weakly
the gradient of the objective function..
function.
190,00 dwt oil tanker model.
Table 5. Reference and Identified hydrodynamic coefficients for the ESSO 190,00-dwt
Design
Variables
1
Hydrodynamic
coefficients
Original values
(Reference)
-0.0500
Optimal Values
(turning test)
-
Optimal Values
(zigzag test)
-
Identified
values
-0.0500
vr
1.0200
1.0200
-0.0200
-0.0200
c c
-2.16
-2.16
0.0400
0.0300
0.0300
-0.0200
-0.0240
-0.0160
-0.0200
-0.0728
-0.0878
-0.0878
vv
-2.4000
-2.4000
vr
-0.3000
-0.3000
10
vv
0.3000
0.3000
11
uv
-1.2050
-1.2050
12
uv
-0.4510
-0.4510
13
-0.0500
-0.0500
14
-0.3780
-0.3780
15
ur
0.1820
0.1598
0.1420
0.1509
16
ur
-0.0470
-0.0416
-0.0380
-0.0398
17
vr
0.3780
0.3780
18
v v
-1.5000
-1.5000
19
vr
-0.1200
-0.1200
20
c c
0.2080
0.1761
0.1761
21
uv
0.0000
0.0000
22
uv
-0.2410
-0.2823
-0.2910
-0.2867
23
c c
0.1520
0.1684
0.1684
24
c c
-0.0980
-0.0805
-0.0800
-0.0803
25
vv
0.0125
0.0125
26
c c
-2.1600
-2.1600
27
c c
0.6880
0.6880
Tran K., Ouahsine A., Hissel F. and Sergent P./ Transport Research Arena 2014, Paris
10
28
c c
-0.1910
-0.1910
29
cc
0.3440
0.3440
30
ur
0.2480
0.2480
31
ur
-0.2070
-0.2105
-0.2105
32
u u
-0.0377
-0.0457
-0.0457
33
-0.0045
-0.0054
-0.0054
34
u u
-0.0061
-0.0073
-0.0073
-0.0073
35
c c
-0.0930
-0.1000
-0.1000
5. Conclusion
A model based on the coupling between ship manoeuvring simulation model and mathematical
programming techniques by the use of SQP and BFGS algorithms,
algorithms is proposed. The model was validated
using experimental data of sea trials of the ESSO 190,000-dwt
190,000
oil tanker model
del for the turning circle and
zigzag manoeuvres, using 35 hydrodynamic coefficients.
coefficients. The results show that only 10 distinct
hydrodynamic parameters were identified to be sensitive in the turning circle and zig-zag
zig
manoeuvres
respectively, and that 14 hydrodynamic
odynamic parameters may be used for both the turning circle, zig-zag
zig
tests and
other manoeuvring simulations.. They also show that the SQP algorithm predicted accurately the
experimental trajectories for the turning test, with a total cumulative error on trajectory
rajectory of 5.8 m, starting
from an initial error of 69 m (before minimization). In the zigzag test of ship heading, the SQP algorithm
gave a final cumulative error 6.6 deg, compared to the starting initial error of 17.3 deg.
6. References
Araki. M., Sadat-Hosseini.
Hosseini. H., Sanada. Y., Tanimoto. K., Umeda N. and Stern F.. Estimating maneuvering
coefficients using system identification methods with experimental, system-based,
system based, and CFD free-running
free
trial data
[J], Ocean Engineering, 2012, Vol.51,
51, 63-84.
6
Dai. Y.H.. Convergence Properties of the BFGS Algorithm [J], Society for Industrial and Applied Mathematics,
Mathematics
2002, Vol.13, N.3, 693-701.
Dan O., Radoslav N., Liviu C. Sandit P. P., Identification of hydrodynamic coefficients for manoeuvring
simulation model of a fishing vessel. [J], Ocean Engineering, 2010, Vol. 37, N.8-9, 678687.
687.
ITTC. The manoeuvring committee. The final report and recommandations to the 24th ITTC
Proceeding of the 24 4th International Towing Tank Conference, 2005, Edinburgh, Vol 1, 137-198
137
.
Ji. S.C., Ouahsine. A., Smaoui. H. and Sergent. P.. 3-D
D Numerical Simulation of Convoy-generated
Convoy
Waves in a
Restricted Waterway [J], Journal of Hydrodynamics, 2012, Vol.24, N.3, 420-429.
Neves. M.A.S. and Rodriguez. C.A.. A Coupled Non-Linear
Non Linear Mathematical Model of Parametric Resonance of
Ships in Head Seas [J], Journal of Applied Mathematical Modelling,
Modelling 2009, Vol.33, 2630-2645.
2645.
Rawson. K.J. and Tupper. E.C.. Basic Ship Theory [M], Butterworth-Heinemann,, 2001, Vol.2, Fifth Edition, 368
Pages.
Souli. M., Zolesio. J.P. and Ouahsine. A.. Optimization for Non-smooth
Non ooth Geometry in Two-dimensions
Two
[J], Journal
of Computer Meth in Appl. Mech. And Eng,
Eng 2000, Vol.180,109-119.
Sutulo. S., Moreira. L. and Soares. C.G.. Mathematical models for ship path prediction in manoeuvring simulation
systems [J], Ocean Engineering, 2002,
02, 29, 1-19.
1
Tran.K.T.. Simulation and Identification of Hydrodynamic Parameters for a Freely Manoeuvring Ships , Ph.D
thesis, France, University of Technology of Compigne,
Compigne 2012.
Zhang. J. and Zhang. X.. A Robust SQP Method for Optimization with Inequality Constraints [J], Computational
Mathematics,, 2003, Vol.21, N.2, 247-256.
247