Structural Analysis and Hydroelasticity of Ships

CHAPTER 1
THE ANALYSIS OF THE GLOBAL SHIP STRENGTHS IN
VERTICAL PLANE WITH 3D-FEM HULL MODELS
There are considered two types of analysis models: the classical 1D-equivalent ship
girder and the 3D-FEM girder full extended over the ship length. It is pointed out that the full
ship length 3D-FEM models makes possible to obtain better results for the global - local
stress distribution at the ship strengths analysis and also it can reveal (locate) the hot spot
domains.
For the numerical analyses there are considered the following two test ships: a
uniform hull structure ship and a bunkering tank, each with several load cases.
1 The global ship strengths analysis based on 1D-girder classical method
In order to ensure the consistency of the theoretical model, in this chapter there is
presented the classical method for ship 1D-girder global strengths analysis, which is used for
the comparison with the method based on 3D-FEM models developed over the full ship
length.
1.1 The ship equivalent 1D-girder still water loads
The ship weight distribution is obtained based on the ship mass distribution with the
following relation:
(1.1) g x (x ) = g (x ) x [ L 2 , L 2] g xi = g i i = 1, n
where: L ship length, g gravity acceleration, (x ) mass distribution, n ship girder elements
over the 1D-beam model.
Obs. In order to simplify the integrals calculation with trapeze method, there are considered
the significant ship hull transversal sections disposed at the middle of the n elements.
L2
n
L x
, x i +1 = x i + x i = 1, n 1 ; x = L n ; f (x )dx = x f (x i )
(1.2) x 1 = +
2 2
i =1
L 2

In order to obtain the ship still water equilibrium position it is necessary to use an
iterative algorithm for given V = c B LBd and xG (from (x ) ), as following:
d i(0 ) = d A (Ti0 ) i = 1, n x (F0 ) ; R (0 ) from ship offset lines

iter = 0
V

(0 )

L2

L 2

(1)

d pp

A (x )dx = x A
(0 )

i =1

(0 )
L
(0 ) x G x B
= d + xF
(0 )
R
2

(0 )
Ti

L2

(0 )

My =

x A (x )dx = x x A
T

L 2

(1)

d pv

(0 )

i =1

(0 )
L
(0 ) x G x B
= d + xF
(0 )
2
R

(0 )
Ti

(0 )

xB =

M (y0 )
V (0 )

Structural Analysis and Hydroelasticity of Ships

iter = k
V (k )

d (pvk ) d (ppk )
L
(k )
(k )
(k )
from ship offset lines
x i + A Ti i = 1, n x F ; R
L
2

L2
L2
n
n
M (yk )
(k )
(k )
(k )
(k )
(k )
(k )
= A T (x )dx = x A T i
M y = x A T (x )dx = x x i A T i
x B = (k )
V
i =1
i =1
L 2
L 2

d (ppk+1) = d (ppk ) +

d i(k ) = d (ppk ) +

(k )
(k )
( k 1)
V(k ) V(k 1) L (k ) x G x (Bk )
L (k ) x G x B
( k +1)
(k ) V V
;
d
d
x

+
x
=
+
+

F
pv
pv
F
k)
k)
(k )
(k )
A(WL
A(WL
2
R
R
2

V V (k ) < 0.004V

The convergence criteria are:

and also the longitudinal trim angle is: (d pv d pp

(k )

(k )

) L.

x G x (Bk ) < 0.001L

Obs. There are noted above (11.3) the following:

L2
n
M yWL
Iy
; x G = x g x (x )dx
; xF =
I y = x 2 b(x )dx = x x i2 b i ; R =
V
A WL
i =1
L 2
L 2
L2

L2

L2

L 2

i =1

L 2

i =1

(1.4) A WL =

b(x )dx = x b i ; M yWL =

(11.3)

L2

g (x )dx
x

L 2

x b(x )dx = x x i bi

where: B maximal breadth, d medium draught amidships, cB block coefficient, b(x) water
plane breadth, xG the longitudinal position of the ship weight centre.
The still water hydrostatic load distribution results from the following relation:
(k )
i = 1, n
(1.5) a cx (x ) = gA T (x ) x [ L 2 , L 2] a cxi = gA T i
The ship still water loads results from the following relation:
(1.6) p cx ( x ) = g x (x ) a cx ( x )
x [ L 2 , L 2] p cxi = g xi a cxi
i = 1, n
The still water shear forces and bending moments results from the following relations:
(1.7) Tc (x ) =

p (x ) dx
cx

L 2

; M c (x ) =

T (x ) dx
c

L 2

1.2 The supplementary ship 1D-girder loads from head cvasi-static waves

There are considered the loads from cvasi-static head waves ( = L ). The amplitude
of the equivalent wave aw=hw/2, with Smith correction, based on Germanischer Lloyd,
I-Part 1, Section 4, A.2.2 Rules, it results from the following expression:
L

h w = + 4.1 c RW [m ] ; L < 90m

25

32

300 L
(1.8) h w = 10.75
c RW [m ] ; 90 L 300m
100

where c RW {1.00 0.90 0.75 0.66 0.60} is the zone navigation coefficient.
In order to take into account of the real ship offset lines, analogue to the case of still
water, there it is used a non-linear iterative procedure with two steps.

Structural Analysis and Hydroelasticity of Ships

In this case dm, dpp,dpv, trim become the parameters that can define the position of the
median plane of the equivalent cvasi-static head wave, taking as reference the base plane of
the ship hull.
For the considered loading case there are known: , V, xG , L , the offset lines, the
ship hydrostatics, Bonjean diagram.
Obs. The coordinates system origin is considered at the aft ship x [0, L] .
x
(1.9) x 1 =
, x i +1 = x i + x i = 1, n 1 ; x = L n
2
Step I the floating condition
(1.10)
h
2x i
(0 )
iter = 0 d (m0 ) = 0 d i(0 ) = d (m0 ) w cos
A Ti i = 1, n from Bonjean
2
L
L
L
n
n
M (y0 )
(0 )
(0 )
(0 )
(0 )
(0 )
(0 )
(0 )
V = A T (x )dx = x A T i
M y = x A T (x )dx = x x i A T i
x B = (0 )
V
i =1
i =1
0
0
iter = k

hw
2x i
(k )
cos
A Ti i = 1, n from Bonjean
2
L
L
n
M (yk )
(k )
M (yk ) = x A (Tk ) (x )dx = x x i A T i
x (Bk ) = (k )
V
i =1
0

d (mk ) = d (mk 1) + 0.001 d i(k ) = d (mk )

L

i =1

V (k ) = A (Tk ) (x )dx = x A T i

(k )

and the iteration is made until V ( k ) V .

The solution is refined, using the half domain method, so that at the last iteration m
it is achieved the convergence criteria V V (m ) < 0.001V .

At the end of the first step, it results the following parameters:

d Im = d (mm ) x IF , A IwL
(1.11) x IB = x (Bm )
Step II the trim condition
I
x G > x B trim = 0.00001 or x G < x IB trim = 0.00001
iter = 0

d (m0 ) = d Im

x (F0 ) = x IF

0)
A (wL
= A IwL

d (pp0 ) = d (m0 ) x (F0 ) trim (0 )

d i(0 ) = d (pp0 ) + d (pv0 ) d (pp0 )

V

(0 )

(0 )

= A T (x )dx = x A T i
i =1

trim (0 ) = trim

d (pv0 ) = d (m0 ) + L x (F0 ) trim (0 )

) xL h2

(0 )

(1.12)

2x i
(0 )
cos
A Ti i = 1, n from Bonjean
L
L

M y = x A T (x )dx = x x i A T i
(0 )

(0 )

(0 )

i =1

(0 )

xB =

M (y0 )
V (0 )

x G > x (B0 ) trim = 0.00001 or x G < x (B0 ) trim = 0.00001

iter = k

d (mk ) = d (mk 1) +

V V (k 1)
k)
x (Fk ) , A (wL
( k 1)
A wL

d (ppk ) = d (mk ) x (Fk ) trim (k )

d i(k ) = d (ppk ) + d (pvk ) d (ppk )

V

(k )

) xL h2

i =1

d (pvk ) = d (mk ) + L x (Fk ) trim (k )

= A T (x )dx = x A T i
(k )

trim (k ) = trim (k 1) + trim

(k )

2x i
(k )
cos
A Ti i = 1, n from Bonjean
L
L

M y = x A T (x )dx = x x i A T i
(k )

(k )

(k )

i =1

x G > x (Bk ) trim = 0.00001 or x G < x (Bk ) trim = 0.00001

and it is iterated until trim is changing the sign.
3

(k )

xB =

M (yk )
V (k )

Structural Analysis and Hydroelasticity of Ships

The solution is refined with the half domain method, so that at the last iteration m
x G x (Bm ) < 0.001L .
there are satisfied the convergence criteria: V V (m ) < 0.001V
At the end of the second step there result the following data:
(1.13) d m = d (mm ) , d pp = d (ppm ) , d pv = d (pvm ) , trim = trim (m ) , A Ti = A (Tim ) i = 1, n
The total vertical load from equivalent cvasi-static head wave has the expression:
(1.14) p xi = g xi gA Ti i = 1, n p x (x ) x [0, L]
The total shear forces and bending moments from equivalent cvasi-static wave have
the following expressions:
x

(1.15) T(x ) = p cx (x )dx ; M (x ) = T(x )dx

x [0, L]

Obs. In the above relations the sign make possible to select the hogging (+) and
sagging (-) wave loads cases.
2 The global - local ship strengths analysis based on 3D-FEM models

The new method of ship global - local strengths analysis is based on 3D-FEM models
developed over the full length of the ship.
In compare to the classical method the new approach based on 3DFEM models has the following main advantages:
the real ship 3D structure is taken into account, with the corresponding geometries and
material proprieties;
reduced number of boundary conditions;
the 3D stress and deformations distributions in the ship structure are obtained, pointing
out also the local hot spots domains;
with no restrictions to the ship hull offset lines form, the floating and trim equilibrium
position is obtained at still water and equivalent cvasi-static statistical head waves.
2.1 The 3D-CAD of the ship hull offset lines
In the first step there is developed the ship offset lines CAD, using specialised
programs as Lines-Tribon (Kockums), Multisurf (AeroHydro), etc. This CAD models are
exported as neutral DXF files format.
2.2 The 3D-CAD of the ship hull structure
The second step includes the 3D-CAD ship hull geometry modelling, extended over the
full ship length. This approach is based on the ship offset lines CAD files, which can be carried on
using Tribon (Kockums), general CAD programs as AutoCAD (Autodesk), MicroStation
(Bentley), with export of DXF files format, or directly using the FEM program CAD preprocessing procedures, as those existing at Cosmos/M (SRAC), Marc-Mentat (MSC), etc.

Structural Analysis and Hydroelasticity of Ships

2.3 The 3D-FEM mesh of the ship hull structure
The third step of the ship strengths analysis includes the generation of the 3D-FEM
models, based on the auto-mesh options that are usual included in the FEM programs. In
order to ensure the convergence of the auto-mesh procedure, it is recommended to use
triangular elements rather then quadratic elements.
Obs. At the numerical examples, presented in this chapter, we have developed the 3DFEM model of the ship hull structure using Cosmos/M auto-mesh procedures, with shell3T
triangular thick shell elements.
2.4 The boundary conditions on the 3D-FEM model of the ship hull structure
At the fourth analysis step there are modelled the boundary conditions for the 3DFEM ship hull model full extended over the length, that are of two types:
the symmetry conditions at the nodes disposed in the diametric plane of the ship, the
model being developed only on one side (for head waves loads case);
the vertical support conditions at two nodes disposed at the ship hull structure extremities
(in the diametric plane), noted NDpp at the stern (aft) and NDpv at the bow (fore). At the
vertical equilibrium conditions, at still water or equivalent cvasi-static head waves, the
reactions forces in the two vertical supports become zero.
Obs. Based on the vertical force reactions in the two nodes NDpp, NDpv, disposed at
ship extremities, there are defined the objective functions for the convergence control of the
ship free floating and trim numerical procedures.
2.5 The loading conditions. Numerical analysis based on 3D-FEM models
At the fifth analysis step there are considered the modelling of the loads conditions and
the effective numerical analysis of the 3D-FEM model developed over the full ship length, in
order to obtain the deformations and stress distributions at the ship global-local strengths
analysis.
The loads acting over the ship hull are of three types (considering Cosmos/M
implementation):
the gravity loads from the eigen structures weight and other mass components of the
displacement, except the cargo masses;
the cargo loads, considered as local hydrostatic pressures over the hull structure;
the equivalent cvasi-static head wave pressure loads for the following cases: h w = 0 (still
water) and h w 0 (according the statistical values from Naval Register Rules), using an
iterative procedure for the free floating and trim condition equilibrium, implemented with
eigen GEO macro-commands files in the Cosmos/M FEM program.

Structural Analysis and Hydroelasticity of Ships

2.6. The numerical results evaluation
At the sixth step of the global-local ship strengths analysis, based on 3D-FEM models,
there are obtained the following numerical results:
the free floating & trim equilibrium parameters (draught & trim angle of reference plane);
the global and local deformations of the ship hull structure;
the global and local (hot spots domains) equivalent von Misses stress distributions over
the full ship hull girder length.
Obs. At the numerical analysis of the test ships it will be compared the 3D-FEM
model results with the classical 1D-equivalent girder model results.
3 The numerical global-local strengths analysis for a ship with uniform hull

As first test ship, in this chapter we consider a uniform ship hull structure, cylindrical
over length L, with the main dimensions presented in table 3.1.
The ship structure is divided in 3 main regions: 0-5m aft-pick, 5-77m central part,
77-82m fore-pick. The cargo is introduced only in the central part Lch.
3.1 Test ship main dimensions
L
82 m
B
11 m
D
5m
Lch
72 m
xF
0m
cB
0.991
a0
0.500 m
3.2 Test ship displacement cases
545.4 t
0=steel
3400.2 t
1=full_1
3410.2 t
2=full_2
2854.8 t
cargo
Mpp_supplementary
10.0 t
dm0=dlight
0.601 m
dm1=dfull_1
3.689 m
dm2=dfull_2
3.710 m
1.3 t/m3
cargo
7.7 t/m3
steel
NDpp
186
NDpv
15925

0 the mass only from the hull steel structure;

1 the full loading conditions;
2 the full loading conditions with 10.0t supplementary mass at the aft-pick.

Structural Analysis and Hydroelasticity of Ships

3.1 The classical 1D-equivalent ship girder model at global strengths analysis
Based on the 1D-equivalent ship girder model, for the amidships
section presented in figure 11.3.1, we obtain the characteristic data presented in table 3.3.

Table 3.3 Test ship girder amidships characteristics

A0
0.620 m2
Af0
0.326 m2
I0
2.207 m4
W0
0.676 m3
ReH
235 N/mm2
175 N/mm2
adm
Using eigen programs P_AC & P_ASV, for table 3.2 displacement cases,
considering a uniform cylindrical hull offset lines, there are obtained the loads over the 1Dequivalent ship girder, shear forces and bending moments. In the case of 0 displacement
there is considered only the still water (hw=0) condition. For the 1, 2 displacement cases
there are taken into account the following conditions: still water (hw=0) and Smith equivalent
cvasi-static head waves with height hw=1m and hw=2m.
In table 3.4 there are presented the main numerical results obtained for 1Dequivalent ship girder model as: test ship equilibrium position in still water and waves
conditions (draught medium dm, stern dpp and bow dpv , for reference plane), the maximal
equivalent von Mises stress in the amidships section.

Fig.3.1 Test ship transversal section, uniform hull

M [KNm]
0

3000

T [KN]

32.8

49.2

65.6

82

49.2

65.6

82

x [m]

-80000

-2000

x [m]

-40000

32.8

-60000

1000
0

16.4

-1000

-3000

16.4

-20000

2000

7
Fig.3.3.a Shear forces in the 1D-girder
2 at
=3410.2t
hw=0m

hw=1m

hw=2m

hw=0m

hw=1m

hw=2m

Fig.3.3.b Bending moments in the

1D-girder at
2

=3410.2t

Structural Analysis and Hydroelasticity of Ships

3.2 The 3D-FEM model used for global - local strengths analysis

The new approach of ship global-local strengths analyse is based on 3D-FEM models,
developed over the full length of the ship .
In the first step there is developed the 3D-CAD model of the ship hull using FEM
Cosmos/M program. In figure 3.4 there are presented the surfaces primitives of the 3DCAD model. Using the auto-mesh options in the Cosmos/M program, there is generated the
3D-FEM model with 41140 shell3T triangular thick shell elements and 16404 nodes.
The material used is the isotropic steel A naval class, with ReH=235N/mm2.
The boundary conditions consist in:
the symmetry condition at the diametric plane of the ship, the model been developed only
one side,
the vertical support conditions at nodes NDpp=186 and NDpv=15925, where at the
vertical equilibrium condition of the ship the reaction forces become zero.
Over the eigen weight of the ship there is considered the equivalent Smith cvasi-static
head wave pressure load for the following cases: hw=0, hw=1m, hw=2m, (table 11.3.2), using an
iterative procedure for the free floating and trim condition equilibrium, implemented with eigen
GEO macro-commands files in Cosmos/M FEM program.
In figure 3.6 there are presented the bottom and side shells of the test ship with the
water pressure distribution at case wave hw=2m, obtained from the free floating & trim
equilibrium condition.
In figure 3.7.a,b there is presented th e von Mises stress distribution over the test
ship full-length model and in amidships section.

Fig.3.4 Primitives of the 3D-CAD Cosmos/M ship model

Structural Analysis and Hydroelasticity of Ships

Fig.3.6 Bottom & side shell with wave pressure (h =2m)

w

Fig.3.7.a 3D-FEM model, von Mises stress distribution (KN/m )

9

Structural Analysis and Hydroelasticity of Ships

Fig.3.7.b Von Mises stress distribution amidships section (KN/m )

3.3 Numerical results. Conclusions

From the numerical results there are obtained the following conclusions:
The differences between draughts values at the 3D-FEM and 1D-girder ship models are
under 1%.
It results maximum 5.4% stress difference between the classical 1D-equivalent girder and
the 3D-FEM model full extended over the ship length, where the ship geometry,
structures and loads are more precisely idealised.

10

Structural Analysis and Hydroelasticity of Ships

4 The numerical global-local strengths analysis for a bunkering tanker ship

As second test ship, in this report we consider the analysis of a small size bunkering
tank.
4.1 The 3D-CAD offset lines tanker model

In order to design the tanker ship lines, we have used program MultiSurf, which
includes points and master curves definition (6 CCurve), parametric surfaces definition
(5 CloftSurf) and also the intersection contours for the ship frames definition (71 IntSnake
Curves).
In order to obtain an accurate 3D-CAD Model, the ship hull surface has been divided
in 3 main regions: from 0-10 m aft-pick, 10-40 m central cylindrical part, 40-50 m fore-pick.
Table 4.1 Tanker main dimensions

1544 t

M_steel

294 t

M_eq+pers

170 t

M_cargo

1080 t

Dw

1250 tdw

50 m

10 m

5,5 m

dm

3.687 m

xF

-0.23 m

cB

0.812

Fig.4.1.a Tanker hull model

12

Structural Analysis and Hydroelasticity of Ships

4.2 The classical ship equivalent 1D-girder at global strengths analysis

Using the classical approach of the ship included in the standard Naval Register Rules,
for the tanker amidships section presented in figure 4.3 we obtain the characteristic data presented in table 4.2.

Fig.4.3 Tanker hull section amidships

M [KNm]
0.00E+00
0
-5.00E+03

T [KN]
2.00E+03
1.50E+03

x [m]
10

20

30

40

50

1.00E+03

-1.00E+04

5.00E+02
0.00E+00
-5.00E+02

10

20

30

40

-1.50E+04

50

-2.00E+04

x [m]

-1.00E+03

-2.50E+04

-1.50E+03
-2.00E+03
hw=0

hw=1

hw=2

-3.00E+04

hw=3

Fig.4.4.b Shear forces in equivalent ship 1D-girder

hw=0

hw=1

hw=2

hw=3

Fig.4.4.c Bending moments in equivalent ship 1D-girder

12

Structural Analysis and Hydroelasticity of Ships

Table 4.2 Ship girder section amidships characteristics

0.658 m2
0.293 m2
3.089 m4
0.951 m3

A0
Af0
I0
W0

From MultiSurf program we export the 3D-CAD ship hull model as an ASCII file.
Using eigen program codes, P_MSF3DA & P_DYN we obtain the offset ship lines file.
For the tanker offset lines, with the mass distribution presented in figure 4.4.a,
using eigen programs P_AC & P_ASV, there are obtained the loads over the ship equivalent
1D-girder, shear forces figure 4.4.b and be nding moments figure 4.4.c, in the following
cases: still water (hw=0) and Smith equivalent cvasi-static head waves, with height hw=1m ,
hw=2m and hw=3m.
In table 4.3 there are presented the main numerical results obtained at this section:
draught amidships dm, after dpp and fore dpv pick, trim angle (for reference plane) and the
maximal equivalent von Mises stress in the amidships section.
4.3 The 3D-FEM model used for global - local tanker strengths analysis

In the following there is presented the tanker ship global - local strengths analysis
based on 3D-FEM models developed over the full length .
From MultiSurf program we export the 3D-CAD ship hull model (shell surface), using
eigen programs P_MSF3DA & P_MSFCOS. There is obtained a GEO type file, which
includes macro commands for the generation of quad surfaces, representing the tanker hull
shell ( also with water pressure distribution at case h =3m).
Using the auto-mesh options in the Cosmos/M program, there is generated the 3DFEM model with 18973 shell3T triangular thick shell elements and 7661 nodes
(figure 11.4.6.a,b).
The material used is the isotropic steel A naval class, with ReH=235N/mm2.
The boundary conditions consist in the symmetry conditions at the diametric plane of
the ship, the model been developed only one side, and the vertical support conditions at nodes
NDpp=6238 and NDpv=7088, where at the vertical equilibrium condition of the ship the
reaction forces become zero.
Same as in chapter 4.2, over the eigen weight of the ship there is considered the
equivalent Smith cvasi-static head wave pressure load for the following cases: hw=0, hw=1m,
hw=2m, hw=3m, using an iterative procedure for the free floating and trim condition
equilibrium, implemented with eigen GEO macro-commands files in Cosmos/M program.
In figure 4.7.a,b there is presented the von Mises stress distribution over the tanker
ship full-length model and in amidships sections (with a bulkhead).
In table 4.3 there are presented the main numerical results at FEM approach:
draught amidships dm, after dpp and fore dpv pick, trim angle and the maximal equivalent von
Mises stress in the amidships section with & without frames and the maximal equivalent von
Mises stress over the tanker ship length, pointing out the stress hot spots.

13

Structural Analysis and Hydroelasticity of Ships

Fig.4.7.a 3D-FEM tanker ship hull model, von Mises stress distribution (KN/m 2

Fig.4.7.b Von Mises stress distribution amidships (with bulkhead) (KN/m2 )

14

Structural Analysis and Hydroelasticity of Ships

4.4 Numerical results

Table 4.3 Numerical results from 3D-FEM and 1D-Equivalent girder analyses
3D-FEM Tanker Model

ech_CM_no

VMfull_CM

CASE

dm [m] dpp [m] dpv [m] trim [rad] dpp+hw/2

hw=0 m

3.687

3.849

3.521

-0.00656

3.849

13.04

68.15

hw=1 m

3.710

3.859

3.557

-0.00605

4.359

20.71

77.20

hw=2 m

3.733

3.870

3.593

-0.00554

4.870

28.44

86.14

hw=3 m

3.756

3.880

3.629

-0.00502

5.380

36.25

95.00

frames

Equivalent 1D - Girder Tanker Model

CASE

dm [m] dpp [m] dpv [m] trim [rad] dpp+hw/2 ech_CM

hw=0 m

3.684

3.805

3.563

-0.00483

3.805

8.94

_1D/_3D
(no frames)
0.686

hw=1 m

3.720

3.862

3.576

-0.00573

4.220

14.72

0.711

hw=2 m

3.731

3.879

3.579

-0.00599

4.879

21.03

0.739

hw=3 m

3.754

3.916

3.588

-0.00657

5.416

27.34

0.754

From numerical results there are derived out the following conclusions:
The differences between draughts values at the 3D-FEM and 1D-girder ship models are
under 1%.
It results that the classical ship equivalent 1D-girder method cannot put in evidence the
stress hot-spot domains (2,6 times bigger stress).
It results maximum 31.4% stress differences for the amidships domain without frames
and bulkheads, between the classical method and the 3D-FEM method, where there are
idealised more precisely the tanker ship geometry, structures and loads.

5 The numerical results evaluation

From the numerical analysis of global-local strengths of the two test ships: uniform
hull and small size bunkering ta nker, there are obtained the following conclusions:

At both test ships the differences between draughts values at the 3D-FEM and 1D-girder
ship models (for reference plane) are not significant (under 1%).
For the uniform ship hull structure it results maximum 5.4% stress difference between the
classical 1D-equivalent girder and the 3D-FEM model extended over the full length.
For the bunkering tanker ship, with non-uniform structure over the ship hull length,
without taking into account the transversal frames & bulkheads, it results maximum
31.4% stress difference between the classical 1D-equivalent girder and the 3D-FEM
model.
It results that the classical ship equivalent 1D-girder method cannot put in evidence the
stress hot-spot domains. For the bunkering tanker ship structure the hot-spot stress is 2,6
times bigger in the amidships bulkhead as at adjacent structures domains.
Using the 3D-FEM models, it makes possible simultaneously to obtain the local and the
global stress distribution over the structure, and also to put in evidence the stress hot-spot
domains.
The method of 3D-FEM ship hull models full extended over the length can now be easily
developed using general FEM programs and standard PC, Intel Pentium 4 computers,
including a large number of elements and nodes per model (over 200.000).

15

Structural Analysis and Hydroelasticity of Ships

CHAPTER 2
THE SIMPLIFIED LINEAR ANALYSIS OF THE WHIPPING
DYNAMIC RESPONSE AND THE BOTTOM SLAMMING
At the linear analysis there is taken into account the vertical sides hypothesis above
the still water line. In this conditions it can be modelled only the bottom slamming (impact
plus momentum) and the transitory induced dynamic response, whipping. This simplified
linear analysis is useful for the qualitative study of the bottom slamming and the associated
whipping phenomenon.
2.1 The Whipping and Slamming phenomena
Due to the ship oscillations with large amplitudes in a rough sea, it occurs the bottom
slamming, after a complete emersion of the fore-pick from water, and the side slamming at
ships with flare forms in the fore-pick domain.
Besides the slamming local effects, as damages to the structural panels in the ship
fore-pick, the slamming phenomenon is inducing in the elastic structure of the ship a
transitory dynamic response, which initially will induce vibrations on all the ship eigen
modes. Due to the structural damping, in the transitory ship dynamic response will dominate
the component on the fundamental ship vibration eigen mode (the 2-nodes ship vibration).
Def. The transitory vibration induced in the ship girder from slamming it represents
the whipping phenomenon.
In the following, we are briefly presenting the slamming phenomenon, in order to be
able to analyse the transitory-whipping ship dynamic response induced in the ship girder.

Fig.2.1 The mechanism of phenomenon occurrence

16

Structural Analysis and Hydroelasticity of Ships

Szebehely has experimentally obtained the following conditions necessary for the
bottom-slamming occurrence:
a complete emersion of the fore-pick from water;
the existence of a critical relative speed ship-water surface;
a certain phase difference between the fore-pick motion and the external wave.
Base on real ships measurements, laboratory analysis and statistics, Ochi has
demonstrated that only the first two conditions presented by Szebehely are necessary to occur
the bottom slamming.
From Ochi and Kawakami, for the mechanism of slamming occurrence at x abscise
section in the ship fore-pick, we can consider the following process phases (fig.2.1):
(I)
0 1 <o>: the time between the moment of complete emersion and the moment when
the contact between the ship hull-water surface is produced (the beginning of the
bottom slamming);
(II) 1 2 <1>: the impact time, on which it is considered that the section has a maximal
immersion of 0.1d(x) and the pressure is increasing from zero to the maximal value.
The load on this period represents the pressure from the impact slamming
and 1 is about 0.1s;
(III) 2 3 <2>: the time between the moment of maximum impact pressure and the
moment of the equilibrium floating position. At the final of this period it is considered
the end of the bottom slamming. On this period it is acting the component due to the
hydrodynamic masses variation (base on the hydrodynamic momentum theorem) and
the hydrostatic component, due to the variation of the Archimedes force. In this period
the load is named momentum slamming
(IV) 3 4 <3>: in this time continue the immersion. If there are flare forms in the forepick it occurs side slamming. On this period the load is calculated using also the
momentum slamming, being a continuation of the III-rd period, until the deck is in
water or the relative speed ship hull wave becomes vr=0.
Obs. In fig. 2.1 we have noted: zr,vr the relative displacement and speed between shipwave; D(x),d(x) are the height and the draught of the transversal ship section at abscise x.

Fig.2.2 The time variation of the slamming pressure

17

Structural Analysis and Hydroelasticity of Ships

Structural Analysis and Hydroelasticity of Ships

2.2 The bottom slamming phenomenon occurrence conditions
It is considered the following hypothesis: in the case of linear analysis the
deformations produced by the transitory ship dynamic response are very small in compare to
the ship oscillations displacements (the stabilised response on rigid hull ship oscillations
modes). The encounter ship-wave circular frequency has the values e=0.50.7 rad/s,
representing the domain with maximum steady state ships oscillations.
It has to be checked if the extreme bow section is emerging complete from water, and
in this case which is the bow domain length where the slamming occurs l s .
Obs. Different to Ochi we are recording all the slams that occur, indifferent of the
slam intensity. This is equivalent to the case with vertical critical speed vcr=0.
At an "x" section, according to Ochi (vcr=0) the impact slamming is produced if:
(2.1)

z r (x, t) d(x) ;

z r (x, t) w st (x, t) v* (x, t) ;

st
w (x, t) = w r (x) pr (t)
st

r =0

st

where w (x,t) is the displacement steady ship dynamic response.

From relations from chapter 17 it results:
w st (x, t) = w 1 (x) cos e t + w 2 (x) cos e t
4
4
(2.2)
w 1 (x) = w r (x) p1str ; w 2 (x) = w r (x) pst2 r
r =0

(2.3)

r =0

(x, t) = a (x) cos e t + a 2 (x) cos e t

*
v

a 1 (x) = a w (x) f s (x) cos(kxcos) ; a 2 (x) = a w (x) f s (x) sin (kxcos)

From relations (19.1),(19.2),(19.3) the bottom slamming occurrence condition
becomes:

(2.4) f l (x) = [w1 (x) - a1 (x)] + [w 2 (x) - a 2 (x)] d(x) 0

First there is tested the expression (19.4) for the ship bow extreme section (x=L) and
in the case that one slam occur, there is determined the last section xs where the slamming
phenomenon still occurs f l (x s ) = 0 ; x s [0, L] .
The length of the domain where the slamming occurs is: l s = L x s .
Obs. In order to improve the precision of the bow excitation forces calculation, the
x[xs,L] domain will be divided into a supplementary number of sections that are usually
used for the steady ship dynamic response analysis.
2

18

Structural Analysis and Hydroelasticity of Ships

2.3 The motion equations system

From the ship hull dynamic equilibrium conditions, chapter 17, the motion equations
system has the following expression:
(2.5)

[a ]{&p&(t)}+ [b]{p& (t)}+ [c]{p(t)} = {Fh (t)}

Fhs (t) = f h (x, t)w s (x)dx ; s = 0,4

0

For the hydrodynamic force there are considered the simultaneously phenomena:
impact slamming and momentum slamming.
(2.6) Fh (x, t) = Fimp (x, t) + Fmom (x, t)

(2.7)

The total dynamic response has two components: steady and transitory.
{p(t)} = pst (t) + ptr (t)

} {

2.3.1 The hydrodynamic force from momentum slamming

Using the generalisation of the A strip theory relations (Gerritsma & Beukelman), it
results:
(2.8)

D zr
D
D
N33 (x, t) zr + g[A(x, t) A0 (x)]
m33 (x, t)

Dt
Dt
Dt
instantaneous
values
have
the
following

Fmom (x, t) =

where
the
(2.9)
m33 (x, t) = m0 (x) + m33 (x, t) |d-z r the additional hydrodynamic mass

expressions:

N33 (x, t) = N0 (x) + N33 (x, t) |d-z r the hydrodynamic damping coefficient
A(x, t) = A 0 (x) b0 (x) zr (x, t) + A(x, t) |d-z r the immerse transversal area
I(x, t) = d(x) zr (x, t) d(x) w st (x, t) + *(x, t) the immersion
The hydrodynamic coefficients m0(x),N0(x)=f(d,b0,cT,e) are computed using the Lewis
conformal transformation, for the still water position and the encounter ship-wave circular
frequency e. Because the excitation from the bottom-slamming phenomenon occurs almost
instantaneous, it is considered having a high frequency (e) and it results:

(2.10) m33
(x, t) |d-z r = m33
d z r , b |d-z r , cT |d-z r m0 (x) ; N33
(x, t) |d-zr 0

(2.11)

The immerse transversal area is decomposed in:

A 0 (x) = b 0 (x)d(x)c T (x) ; A(x, t) |d z r = b |d-z r (d z r )c T |d-z r
A(x, t) |d-z r = A(x, t) |d-z r A 0 (x) + b 0 (x)z r (x, t)

Obs. The terms from relations (2.9), (2.10), (2.11) will be calculated for 11
discreet values zr=0; 0,1d; 0,2d; .... 0,9d; d; following that at the numerical simulation in the
time domain to be used the linear interpolation procedure.
From relations (2.8),(2.9) it results:
2

Fmom (x, t) = m0 (x) D z2r + [N0 (x) - u s m0 (x)] Dzr + g b0 (x) zr (x, t)
Dt
Dt

(2.12)
D

mz |d-z r Dzr gA |d-z r

Dt
Dt

Analogue to relation (19.7) the total vertical displacement is decomposed as following:

(2.13) w(x, t) = w st (x, t) + w tr (x,t)
(2.14) z r (x, t) = w st (x, t) + w tr (x, t) v* (x, t)

19

Structural Analysis and Hydroelasticity of Ships

From relations (2.12),(2.13) it results:

sf
Fmom (x, t) = H st0 (x, t) + Fw 0 (x, t) H 0tr (x, t) + Fmom
(x, t)
st

(2.15)

st

st

Dw
D w + [ (x)
+ g b0 (x) w st
N0
u s m0 (x)]
2
Dt
Dt
tr
tr
2
Dw + g (x) tr
D w + [ (x)
tr
(x)
(x,
t)
=
m
(x)
b0 w
]
N0
us m0
H0
0
2
Dt
Dt
2 *
D v*
D v
[
]
+ g b0 (x) v*
(x,
t)
=
m
(x)
+
(x)
(x)

N0
u s m0
Fw0
0
2
Dt
Dt
D
D zr
sf
+ gA |d-z r
Fmom (x, t) = m33 |d-z r
Dt
Dt

H0 (x, t) = m 0 (x)

2.3.2 The hydrodynamic force from impact slamming

(2.16)

According Ochi the impact slamming force has the expression:

2
(x, t)

Fimp (x, t) = K imp (x, t) Dzr

Dt

t t
t t
1
K imp (x, t) = g k1 (x)G(x) e1- 1 = Kimp (x) e1- 1
2
1

1
In the above expression k1(x) is the maximum pressure coefficient in the keel base
point at impact, 1 the impact time and G(x) the factor of the transversal section form,
between the keel base point and 0,1d(x).
Obs. The time depending term from expression (2.16) has been introduced by
Kawakami, taking into account that after t21 the impact slamming force becomes zero.

2.3.3 The total hydrodynamic force

(2.17)

From relations (19.15),(19.16) it results:

Fh (x, t) = H st0 (x, t) + Fw 0 (x, t) H 0tr (x, t) + Fsf (x, t)
sf
Fsf (x, t) = Fimp (x, t) + Fmom
(x, t) bottom slamming excitation

Analogue to chapter 17 , it results the generalized hydrodynamic forces vector:

{Fh (t)} = {Hst0 (t)} + {Fw0 (t)} {H0tr (t)} + {Fsf (t)}
(2.18)

{H (t)} = [A
{H (t)} = [A

{ }
{ }
{ }
( )]{&p& (t)}+ [B ( )]{p& (t)}+ [C ( )]{p (t)}

(e)] &p& (t) + [Bh (e)] p& (t) + [Ch (e)] p (t)
st

st
0

tr
0

st

tr

st

tr

tr

where c is the calculation circular frequency.

From relations (19.5),(19.18) there are obtained two equations systems:
(19.19) [a + A h (e)] &p&st + [b + B h (e)] p& st +[c + C h (e)] pst = {Fw 0 (t)}
which, same to the system for the steady ship dynamic response, it has a analytical solution,
and the second one:
tr
tr
tr
[
A ] &p& (t) + [B] p& (t) + [C] p (t) = {Fsf (t)}
(2.20)
[A] = [a ] + [Ah (c)] ; [B] = [b] + [Bh (c)] ; [C] = [c] + [Ch (c)]
which is the motion equations system at the ship transitory dynamic response - whipping.

{ }

{ }

{ }

20

Structural Analysis and Hydroelasticity of Ships

From relations (2.18),(2.20) because the ship dynamic response whipping has the
main component on the fundamental ship elastic girder vibration mode, it will be considered
the following calculation circular frequency c=2. In analyses Bishop and Guedes S. have
considered (c ).
The generalized excitation force from bottom slamming has the expression:
L

(2.21) Fsf r (t) = Fsf (x, t) w r (x)dx ; r = 04,

xs

with the integral calculated using a numerical method based on the finite element technique.
From relations (2.15),(2.16),(2.17) it results:
2

Fsf (x, t) = sf (x, t) Kimp (x, t) Dzr sf (x, t) m33

(x, t) |d-z r D z2r
Dt
Dt

(2.22)

us

(x, t) |d-z r D zr m33

(x, t) |d-z r zr D zr
m33

+
gA |d-zr
Dt
x
zr
t Dt

z r (x, t) w st (x, t) v* (x, t) ; sf (x, t) =

1
0

yes s.b.
no s.b.

2.34 The transitory-whipping ship dynamic response

In order to solve the system (2.20) it is used the direct time domain integration
method for t[0,Ts]. The simulation time is Ts=3Te where Te=2/e, and the time step is
t=Ts/8000. After that, the resulted time functions are supplementary analysed using the Fast
Fourier Transformation (FFT) procedure.
As time domain integration procedure it is used the -Newmark (=1/2) algorithm,
having a high numerical stability:
iter t = 0 ptr (0) = 0 ; p& tr (0) = 0 &p&tr (0) = [A-1]{Fsf (0)}

}{

}{

iter t ptr (t) ; p& tr (t) ; &p&tr (t)

iter t + t It is solved the linear algebraic equations system in : &p&tr (t + t)

t
t tr
[A ] + [B] + [C] &p& (t + t) = {Fsf (t + t)} {Fsf (t)} +
2

2
2

(2.23)
t
t tr
+ [A ] [B] [C] &p& (t) {[C]t} p& tr (t)
2

2
1 tr
tr
tr
tr
p& (t + t) = p& (t) + &p& (t) + &p& (t + t) t
2
1 tr
tr
tr
tr
tr
2
p (t + t) = p (t) + p& (t) t + &p& (t) + &p& (t + t ) (t )
4
...iter t = Ts
Analogue to chapter 17, using relations (2.7) from the modal analysis, it is obtained
the time function of the total ship dynamic response = steady + transitory (whipping), for
displacements w and loads T,M at any section "x".
As example for the numerical simulation we consider a tanker test ship with
50000t, at ballast condition (d=6.75m), us=7.5m/s , e=0.53rad/s and Ts=25.3s. At ballast
2

} {

} [{

} {

} {

}]

} {

[{

21

} {

}]

Structural Analysis and Hydroelasticity of Ships

condition it results that in order to occur the bottom slamming the linear wave height has to be
hwmin=7.21m .
In fig.2.a,b there are considered h w=8m ( l s =7.48m) and hw=9m ( l s =14.08m), for

c=25.72rad/s. There is presented the time variation of the total dynamic bending moment
amidships Mv(L/2,t). From the spectral analysis of time functions, it results that the whipping
induced bending moment represents 18% in case (a) and 34% in case (b), from the steady
oscillation bending moment component.

Fig.2.a. Time record of the bending moment amidships

Fig.2.b. Time function of the bending moment amidships

22