Dynamics and structural damage of tanker ships running aground

Journal of Hazardous Mate&h,

25 (1990)

61

61-73

Elsevier Science Publishers B.V., Amsterdam

DYNAMICS AGROUND

AND STRUCTURAL

DAMAGE

G.V. ASADI and H. VAUGHAN Department of Mechanical Engineering,

OF TANKER

University

SHIPS RUNNING

of British Columbia

( Canada)

SUMMARY

A finite element method is described which models a ship impacting against another object. The ship is modeled as a beam which has the same flexibility, mass distribution and floatation properties as the ship. The ship structure may collapse locally during the ongoing collision process as also may the struck object. A number of collision scenarios involving the collision of a large tanker-ship with a submerged rigid rock are investigated. The significance of ship speed on the extent and nature of the damage to the hull is established. INTRODUCTION Grounding

incidents,

by their very nature,

docking facility and sometimes crude-oil

spill with its disastrous

demonstrated

recently

being damaged

effects

on local ecology

on the Alaska coast.

and releasing

The parallel

The consequences

of a large

and environment

have been

situation of a LNG tanker ship

upwards of one hundred thousand tonnes of liquid natural

gas in the form of an unconfined not occurred.

occur in shallow waters, often close to a

near populated land masses.

heavy and toxic flammable

gas-cloud

Such ships are usually protected at the sides by collision

has fortunately

barriers

and from

below by a “double bottom” which raises the LNG tanks a meter or so above the bottom plate of the hull. Engineers ticularly

have responded

in the automotive

industry. The literature recent conference

to the demands for improved safety against collision

industry and to a lesser

on “crashworthiness”

proceedings

degree in the general

par-

transportation

is fairly recent and limited although several

and textbooks

have been devoted to it, see for example

111, PI, and ]3]. Some revealing

statistics

on marine damage incidents

gives data on one thousand serious tanker casualties IMCO (International

Maritime

Consultative

are given by Kinkead

by category of casualty,

Organization)

portation

or grounding

(165 collision,

were caused by

It is the major hazard in the trans-

of oils and gases at sea.

The effectiveness

of double bottoms

tankers has been demonstrated although

239 grounding).

by

during the period 1968 to 1980.

The data shows that slightly more than four hundred of those incidents collision

[4]. He

collected

effective

against

above the struck object,

0304-3894/90/$03.50

in the prevention

shallow penetration

of the hull, merely by keeping the cargo

it gives a very limited protection

0 -

of oil outflow from grounded

by Card [S]. A double bottom is a passive system and against

1990 Elsevier Science Publishers B.V.

deep penetration,

as

62

could occur when a ship strikes a sharp reef or ice-flow. stronger hull based on knowledge of estimated beyond the elastic

processes

considerable

and grounding

APPROACHES

In this section we categorize

TO SHIP COLLISION

which have been found to be appropriate

im-

reviewed.

ship damage and ship collisions

for the study of ship collision

the interactive

with a view towards

safety. This work is now briefly

REVIEW OF TECHNICAL DAMAGE

ence frames

work has been done to understand

involved in ship collisions

proved design and increased

cal models

to design a

limit.

In the last decade structural

Then it is necessary

impact forces and using design codes well

AND GROUNDING

into some analytic refer-

for purposes of developing theoreti-

damage.

Firstly

we clarify

minor and major

collisions. A minor collision

is one in which the ship structure

plastic strains to the shell plating or stiffeners. neither

is there any fracture

or plate tearing caused by external

This means that all strength members slightly reduced

level. Significant

which minor collision by Jones

continue to be effective

contributions

of the hull.

the criteria

useful in the preliminary

or restricted

See for example

to the barrier.

A major collision of plate leading Such collisions of structural

cannot

material

destroyed

design of protec-

using classical

[8] is commonly

theories

to the energy absorbed

[ll],

or tearing parts.

and plasticity. the latter being to be used in

method occurs when a ship

but consists

plied to general grounding problems.

of tom plating, per-

in length.

A method

for

[9] and applied by him to some

and [12]. Some dimensional

of small scale tests by Vaughan

erably extended

[7].

for many types of ships.

was first proposed by Vaughan [lo],

distortion

of the damaged

in the collision,

in width but up to one hundred meters

LNG tanker groundings

protection

of elasticity

which cannot be analyzed by Minorsky’s

such incidents

cargo.

used which merely relates the volume

The damage is then often not volumetric

haps only one meter

from a series

paper on collision

energy of the ship. A range of coefficients

method are available

A major collision

Jones’

loss in strength and effectiveness

be analyzed

to the loss in kinetic

runs aground.

under

occurs then the damage is minor

is one in which there is some gross structural

to significant

Instead a method due to Minorsky

examining

penetration

which are built into the sides of specialty ships which carry hazardous is to ensure that if collision

Minorsky’s

or small

although possibly at some

towards understanding

Clearly the design criterion

related

only elastic

damage occurs and the nature of the damage has been developed

[6]. This work has been particularly

tion barriers

sustains

No gross distortion of the structure occurs

effects

were determined

[13] which allowed his method to be ap-

His experiments

were recently repeated

and consid-

by Jones and Jouri [14].

In the late 1970’s

and early 1980’s a considerable

ships in the form of ice breakers,

ice-strengthened

effort went into the design of arctic freighters,

and particularly

ice-break-

63

ing supertankers. financially

At that time the high price of oil made the exploitation

attractive.

A design for a supertanker

posed by Johansson

successful, onstrate

particularly

created

has progressed,

towards safer

tonne displacement. through improved

ships in harsher

designs for ships in which collision

The strongest breaker,

capable

been proposed, greatest

damage

of operating

ELEMENT

and bending moments

We now describe

details

state the equations

Coast Guard is a Class Eight icedesigns

by Glen and Daley [16]. Estimates derived by Vaughan

OF SHIP COLLISION

have of the

of ships of

[17].

AND GROUNDING

a finite element analysis recently developed by the authors in which

some new damage mechanisms the mathematical

instead of “major”.

likely to occur during operation

using simple formulas

ANALYSIS

is “minor”

of collision

They have effectively

in eight feet of level ice at 3 knots. Several

one of which has been assessed

impact forces

were also extremely

All of these ships dem-

understanding

environments.

ship yet proposed for the Canadian

this type can be obtained FINITE

Other designs for ice-breakers

in the range of 7,000

how the designer effects,

general purpose

was built a decade ago to carry ore from the Hudson Bay to North-

which is still in operation.

and impact

to withstand ice impact forces was pro-

[15]. The ship was not built, but an ice-strengthened

ship, the M.V. Arctic, ern Europe,

of the Arctic

for ships have been investigated

since they are derived in complete

[18]. We do not include

detail in [18].

of motion that have been solved, then concentrate

We merely

on the results that

have been obtained. The ship is modelled ment matrix

as a floating beam divided into n stations.

is denoted by [U] and the displacement

The nodal displace-

[u] at any point in the body is then

given by (1)

1~1= N WI where [N] is the interpolation motion

therefore

reduced from twelve to ten, as shown in Fig. 1. The (10 x 10) mass matrix of

the element

is considered

function matrix. The ship is assumed to be axially rigid and

its axial

separately.

The degrees

of freedom

of each element

is

is [m] = [ml] + [ma], where [ml] is due to lateral inertia and [m2] is due to

rotary inertia.

Fig. 1. The beam element

with ten degrees of freedom.

64

In the analysis

it is mathematically

convenient

to represent

rigid-body

motions as

special cases of the dynamic flexural motions. Thus the natural frequency matrix includes as its leading elements

the rigid-body

periodic motions in heave, pitch and roll, respec-

tively. This was made possible by treating each beam element as a free-free consequently

admits rigid-body

The simuhaneous

displacements

treatment

beam which

as well as elastic vibrational motions.

of rigid-body

motions

and flexural

motions

enables

added mass to be included in [m] in the form of a frequency dependent matrix. The three leading elements

of the added mass matrix

correspond

again to heave, pitch and roll

whereas the later elements have to be identified with the appropriate either a vertical,

lateral,

or torsional

mode,

starting with the fundamental

working up to the desired number of higher harmonics. later in this paper.

Added mass coefficients

mode of vibration,

A particular

for transient

water are difficult to assess so in the numerical

modes

example

impulse motions

calculations

discussed

and

is given

in shallow

later, deep water

values for sinusoidal motions have been used The stiffness matrix of the element is denoted by [k]. It is composed (i) the elastic structural

stiffness of the hull, (ii) the hydrodynamic

ancy, (iii) the hydrodynamic

of flexural

vibration and torsion have been included

as well as the rigid body motions in heave, pitch and roll.

The ship, and hence each element, buoyancy,

stiffness due to buoy-

stiffness due to roll. When evaluating the stiffness associated

with (ii) and with (iii) the effects respectively,

of three parts:

restoring

briefly discussed.

buoyancy

torque,

is influenced

by the following forces:

impact or contact

forces.

These

forces

weight, are now

The weight distribution of the ship is obtained from the standard curve

in the form of tonne per meter.

The weight of each element

is thus estimated

and the

nodal values obtained. The buoyance

force on each element is obtained from the hydrostatic

ship. It is assumed that the immersed cross-sectional (different

constants

for each element).

curves of the

area is constant along each element

The buoyancy force is then adjusted to include

vertical motions of the element relative to the water surface - this merely involves adding the relative displacement heeling is calculated

of the ship onto the static draft. The restoring torque caused by

for each element using the small angle formula with the metacentric

height. The angle of heel for each element is easily deduced from the nodal rotations U3 and Ug shown in Fig. 1. The impact concentrated.

forces are treated as surface tractions and may be either distributed or

They are transferred

to the nodal points using the interpolation

The equation of motion of each element may be expressed

[ml

[cl + Eel lG1 + [kl

WI

in the form

= [sl

where [c] is the damping matrix and [q] is the generalized

force matrix.

matrix.

65

Physical characteristics are: cross-sectional element.

of each station of the ship are taken as input data. These data

area; breadth,

Mass and stiffness

(excluding obtained

contact forces) with respect

frames

matrices

inertias,

mass and length of each

for each element are generated

and external

forces

are included. The global equation of motion for the ship is then

to a fixed external

is accomplished

requirement

depth, two principal

by a series

that the displacement

frame.

Transference

of transfer

matrices.

between

local and gIoba1

Assemblage

at a node shared by two elements

is based on the

must be the same for

both of those elements. The global matrices

of the ship are of the form

[Ml = .p, [Tl’s 14

[‘Us (3)

WI = %;, PI’s [Us Pls [cl = where

.T,iTI’s [cl, [Tls [Tls

stiffness

is the transfer

matrix

and damping matrices

for element

for element

s, [ml,,

and

Ms

[cl,

are the mass,

‘s’. The global equation of motion for the ship

is then [Ml

(4)

[=I + [Cl 161 + WI PI = IQ3

where

[Ql= ,p,ITI's [q]s

When assembling

the global equation of motion (4) there is a matrix overlap of size 5

x 5 between adjacent at the common number

elements

of elements.

Equations

Clearly

the matrices

usual way by determining trix is a linear

the eigenvalues,

coordinates.

combination

damping.

requirement

constructing

This necessitates

They are uncoupled in the

the modal matrix,

the assumption matrices,

usually

supply.

records

to as

(9 dissipation

a and p can then be found from the logarithmic response

referred

we write

The main cause of damping is due to hydrodynamic tional

and then trans-

that the damping ma-

[Cl = 4Ml+ /%I ficients

where n is the

[U] and [Q] are columnar.

coupled equations.

of the mass and stiffness

Accordingly

on the displacements

in (4) are of size S(n+l),

[M], [C] and [K] are square,

(4) are a set of simultaneous

forming to principal proportional

due to the assemblage

node. Consequently

in heave and pitch.

Such records

of energy. Damping coef-

decrement

of the ship vibra-

are available

but in limited

66

uncoupled

The

difference

method.

in the following

equations

incrementally

The grounding and collision

using a second

forces are interactive

The nature of the contact

either explicitly or implicitly.

Examples

coefficients

involving one or

force relative to ship position must be considered

are : the ship and object are

rigid; the ship is rigid but the object crushes with a specified object is rigid but the ship becomes

order finite

and are dealt with

way. The position of initial contact must be specified,

more nodal positions. specified,

are integrated

constitutive

behaviour;

the

crushed or torn. In the latter case, certain strength

of the Minorsky type involving crushing strength both laterally and longitudi-

nally must be used together with tearing strengths obtained from [13]. During the integration process,

the position of the ship must be checked and the contact force updated to its

new value. In the case of separation

or rebound the force must be cancelled.

The extent

of the contact force is allowed to move as the motion proceeds so that new nodal positions may become

involved as time increases.

Some typical grounding scenarios

which have been examined

using this method are

now presented. SIMULATION

RESULTS

We consider thousand

tonne.

OF TANKER

SHIP GROUNDINGS

a very large tanker ship of length 330 meters The mass and buoyancy

and displacement

260

curves are shown in Fig. 2. Other required

geometry is given in Table 1. This ship is referred to as the ‘Standard Tanker Ship’ about which various parameters Added mass coefficients

are changed. The yield strength of the hull is taken as 280 MFa. in heave, sway and surge are taken as 0.75,

tively. The natural frequencies

(first six) were calculated

0.6 and 0.1 respec-

and are identified

in Table 2.

The ship is assumed to be running ahead when it strikes a smooth rock, 1.5 meters below water level, 17 meters to the port side. It is assumed that the smooth rock does not tear the hull plating, but volumetric

damage of the Minorsky type does occur.

I

1000

t

AP

Fig. 2. (a) Massdistribution

FP

and (b) 3uoyancy

AP

distribution

FP

for standard ship

67

TABLE

1 - Specification

STATION No.

STATION LENGTH meter

Tanker

32 38 42 :;

33 33 33 33 33

42 42 42 41 40

22 22 22 22 22 22 22 22 22 22

2 - Natural Frequencies

of the Ship Frequency

Mode Number

Ship’ at the Stations

STATION DRAFT meter

STATION 3READTI-I meter

33 33 33 z;

1

2 3 4 5 6 7 8 9 10

TABLE

of the ‘Standard

STATION MOMENT SHAPE FCT. Iv m4 5 7 8 :

Fig. 3 shows the extent of damage

8 8 8 7 5

360 400 600 600 600 600 600 600 400 400

Mode Shape

( Hz)

sustained

ii-l4

500 600 800 800 800 800 800 800 600 600

0.154 0.186 0.426 1.638 1.751 4.307

1

2 3 4 5 6

OF INERTIA I

Heave Pitch Roll 1st Vertical 1st Lateral 1st torsional by the hull for forward velocities

m/set, 3 m/set, 5 m/set and 7 m/set. For the low speeds the hull receives impulse near the bow which changes the angular momentum rock but the combined near midships catastrophic importance

forward motion and angular motions cause a secondary

and separation

from the rock does not occur.

inflicting

considerable

show the force and

structuraf

damage to the ship, a process

only by

which can last for up to

half the length of the ship. structural

for an identical

bled. The reduced damage Next we consider

and are shown in Fig. 4 for the higher speeds. It

kinetic energy of such a large ship can be absorbed

The effect of increased age estimates

The figures

collision are rather

clearly

that the initial

one minute and damage

estimates

speed on the nature and extent of the damage. Collision

contact times have also been determined is evident

of the ship. It then clears the

which bring it to rest. For high speeds the damage

of collision

of 1

a damaging

scenario

strength has been examined.

Fig. 5 shows the dam-

except that the ship structural

steel has been dou-

is .evident in each case.

what happens when the ship plating is ruptured,

ship runs over a sharp rock. The damage inflicted

as is likely if the

to the standard tanker for the range of

68

0

(a)

SO

100

f-p.

150 SHXP

200

LENGTH

250

300 &.P.

(M)

A, -----

Z

(b)

VERTIC. HOR:ZO.

DAMAGE DAMAGE

f.P.’

0

(cl

~~-L

f.P.

I_~~~;:

.CI

3c

[

” I

(d)

c”, .

I

1

,

100

150 swrr

LENGTH

I

1

1

200

25D

300

(Ml

L.C.

Fig. 3. Extent of damage to standard hull without rupture (a) V = lm/sec, (b} v=3 mkec, (c) v = 5 rnkec, (d) v = 7 mlsec.

69

80

v=Sm/s

z

z

40

E

d

i,\, 0

0

50

25

30 set

'IS

set

Fig. 4. Collision velocities

force on the standard hull versus time.

considered

previously, is shown in Fig. 6. It exceeds that shown in Fig. 4 for the

simple reason that a torn plate absorbs less energy than a dented one, primarily because the torn plate has little membrane Finally we illustrate maneuvering.

strength.

what happens to a ship when it grounds on a smooth rock during

We take a forward speed of 2 m/set and fore and aft rotational speeds of 2

m/set and 0.5 m/see. The rock is 15 meters below the surface, the hull, with centre of collision

extends over 20 meters of

50 meters from the bow. The force and damage calcula-

tions are shown in Fig. 7. As a results of the initial impact force, from the rock and the damage depth subsequently been included

to show that the interaction

fundamentally

different

from a normal head-on

The number of variables

decreases

and response

and scenarios

comprehensive

investigation

with time. This example

due to a rotational

has

collision

is

collision.

which may be examined

great. In this paper we have merely demonstrated oped and that it can be used to investigate

the ship moves away

are obviously very

that a working model has been devel-

many real ship-grounding

is included in the thesis

situations.

A more

[18] but ongoing case studies are

also being made. CONCLUDING

REMARKS

The consequences releases

of what happens when a tanker ship runs aground and spills oil or

liquid natural gas is a subject of much current interest. By making mathematical

models of a grounding ship and determining design

criteria

can be established

safety standards can be increased be shown to be structurally

the interactive

and subsequently

forces that the ship receives,

incorporated.

By such processes

and ships that might otherwise have been breached

sound under specified

A finite element model developed at the University of British Columbia ship grounding been illustrated.

incidents

has been described

Such illustrations

and a number

clarify the complex

ship motion and grounding force. This clarification

can

operating conditions.

of realistic

interaction

which models situations

have

which exists between

can assist in the improved structural

design of LNG and tanker ships so that spill incidents will be less likely to occur, even if grounding does occur.

70 1

VERTIC.

---I-

2

HORIZO.

OAMJGE DAMAGE

_

0

..-

(=I

I

I

I

50

100

150

F.P.

I

200

2;o

LE?rGTW (M)

SHIP

A ----.

(b)

VERTIC.

OAwh;E

HORIZO.

OAMAGE

F.P.

fc) A

F.P. 4

SHIP

LE’GfH

A.P.

(M) -----

VEFITICdL OhMatE HORIZONTAL DAMAGE

2

0 0 td)

F.P.

SO

100

150 SHIP

LEYGTH

too

250

(M)

Fig. 5. Extent of damage due to forward collision without rupture of the reinforced hull (a) v = 1 mkec. (b) v = 3 mkec. (c) v = 5 m/set, (d) v = 7 rdsec.

A.P.

71

(a)

SHIP

F.P. 4

LENGTH

(M:

7

VERTIC. HoRIzo.

-*---

3

SO

0

(bl

A.P.

101

F.P.

2c:

150 SHIP

LENGTH

250

DAMAGE OAWAGE

300 A.P.

(“:

A

----’

I

/ 0 (=I

I

i

50

150

r.p.

SHIP

2c7

i

1

250

300 A.P.

----.

VERTIC. HORIZO.

DAMAGE DAMAGE

pzzq

*

T.P.

SO

I

6

I 0 (6)

OAmAGE OAMAGE

(“1 ,

LENGTH

A

1:

VERTIC. HORIZO.

100

2””_I

150 SHIP

LENGTH

(“1

250

300 A.P.

Fig. 6. Extent of damage due to forward collision with rupture of the standard hull. (a) v = 1 m/set, (b) v = 3 mkec, (c) v = 5 mkec, (d) v = 7 mkec.

72

0

a.23

0.s

0.75

1

COLLISION

(a)

1.2s TIME

1.5

(b)

-

0.25

0.5

u.75

1

1.2s

COLLISION TIME

2

2.25

(SEC.)

3,

0

1.7s

1.5

VERTICAL

1.75

[SEC.)

Fig. 7. (a) Generated force and (b) extent of damage due to the side coffision of the standard hull during maneuver.

CAMAGE

2.5

73

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17. 18.

Structural Crashworthiness, N. Jones, “Structural Aspects of the Ship Collisions”, Jones and Wierzbicki, Butterworths, 1983, pp. 309-332 P.A. Frieze, R.C. McGregor and I.E. Winkle, “Marine and Offshore Safety”, Elsevier, 1984. W. Johnson and A.G. Mama& “Crashworthiness of Vehicles”, Mechanical Engineering Publications Ltd., London, 1978. A.N. Kinkead, “Some Collision and Grounding Considerations for Refrigerated Marine and Offshore Safety, Frieze, McGregor and Winkle, ElGas Carriers”, sevier, 1984, pp. 159-182. of Double Bottoms in Preventing Oil Outflow from J.C. Card, “Effectiveness Tanker Bottom Damage Incidents”, Marine Technology, 197.5, pp. 60-64. N. Jones, “Plastic Behaviour of Ship Structures”, Transactions of the Society of Naval Architects and Marine Engineers, Vol. 84, 1976, pp. 115-145. N. Jones, “A Literature Survey on the Collision and Grounding Protection of 1979, U.S. Coast Guard, Washington, D.C. Ships”, Report SSC-283, V.U. Minorsky, “An Analysis of Ship Collisions with Reference to Protection of Nuclear Power Plants”, Journal of Ship Research, Vol. 3, 1959, pp. 1-4. H. Vaughan, “Bending and Tearing of Plate with Application to Ship-Bottom Damage”. ‘The Naval Architect’, Journal of the Royal Institution of Naval Architects, May.1978, pp. 97-99. H. Vaughan, “Damage to Ships Due to Collision and Grounding”, Det norske Veritas. Oslo. Technical Renort 77-345. August 1977. of Structural Damage on Kockums K.R. Johnsen and H. Vaug’han, “Estimatio; 133000 m3 Liquid Natural Gas Carrier Due to Collision and Grounding”, Det norske Veritas, Oslo, Technical Report 77-275, July 1977. J. Poudret, M. Huther, P. Jean and H. Vaughan, “Grounding of a Membrane Tanker, Correlation Between Damage Predictions and Observations”, Symposium on Extreme Loads, Arlington, pp. 125-132, Society of Naval Architects and Marine Engineers, 1981. H. Vaughan, “The Tearing Strength of Mild Steel Plate”, Journal of Ship Research, Vol. 24, June 1980, pp. 96-100. N. Jones and W.S. Jouri, “A Study of Plate Tearing for Ship Collision and Grounding Damage” Journal of Ship Research, Vol. 31, 1987, pp. 253-268. of an B.M. Johansson, A.J. Keinonen and B. Mercer, “Technical Development Environmentally Safe Tanker”, SNAME, Proceedings of the STAR Symposium, June 1981. I.F. Glen, C.G. Daley and G. Tam, “Analysis of the Structure of the Proposed CCG Polar Class 8 Ice Breaker Under Extreme Ice Loads”, Transactions SNAME, 1985. H. Vaughan, “Flexural Response of Ships to Impact Loads”, Transactions of the Royal Institution of Naval Architects, Vol. 128, 1986, pp. 259-267. G.V. Asadi, “Dynamic Response of Ship Structures to Impact Loads”, Ph.D. Thesis, University of British Columbia, May 1989.