Numerical Modelling of Installation Aids for Platform Installation

Numerical Modelling of Installation Aids for Platform Installation

Dr. Peter S. K. Lai, Xavier Chevalley Saipem UK Limited

ABSTRACT

The present paper details the numerical modelling of installation aids for platform installation

and demonstrates the use of the techniques to evaluate the corresponding dynamic loading

during installation. These aids include the bumper and guide system, supports, pin and bucket

docking system, fender system, leg mating unit and desk supporting unit. These installation

aids are mainly for topside deck module installations by lift and floatover operations. These

operations have been simulated in time domains and results are also presented in the present

paper. The usage of animation in presenting the simulated operation is also discussed.

1 INTRODUCTION In installing topside deck module onto an offshore platform, a number of different installation

aids are used to install the new module into the exact location whilst reducing the impact

loads during installation and to protect the equipment which is already on the platform located

around the new module. These installation aids include Bumper and Guide system, Docking

system (Pin and Bucket), Fendering and Leg Mating Unit and Desk Supporting Unit.

The installation operation is analysed with numerical simulation in time domain to finalise the

installation methodology and design of the structure and installation aids, and to define the

operational limits. Simplified assumptions are used in the numerical model to represent the

installation aids in most advanced analytical software available in the industry. However, the

corresponding analysis may not represent the actual marine operation.

The numerical modelling of these installation aids is investigated in detail in the present

study. The function of these installation aids is described and corresponding methodologies in

representing the aids in the numerical model, including the numerical equations, are presented

in this paper.

This paper is concluded with some analytical results from the assessed installation operations

together with discussion about the use of animation.

The following two “Real Life” cases are used.

1) Module installation by lifting: - module installed by Saipem’s semi-submersible crane

vessel, S7000, onto a semi-submersible floating production unit with bumper and guide,

docking pin and bucket and module landing systems.

2) Module installation by floatover: - module installed by floatover operation using the

Saipem’s cargo barge S45 with fendering system, leg mating unit and deck supporting unit.

Although the presented installation is for a fix jacket, the same technique can be applied onto

the floating platform.

2 INSTALLATION AIDS NUMERICAL MODELS

2.1 Lift Operation Heavy deck modules over 11,000t can be installed by heavy lift crane vessel, such as the

semi-submersible crane vessel, Saipem S7000, as shown in Figure 2.1-1. This installation is

mainly for new platform construction including Spar Buoy, Floating Production Unit and

jacket platforms.

Figure 2.1-1 Lift Operation

In order to extend the life time and function of existing floating and fix platforms, lighter

modules are removed and new module are installed onto existing platforms mainly through

lift operations as well.

In these operations, bumper and guide system, as shown in Figure 2.1-2, are used to protect

the existing structure and equipment from damage during the installation. This system also

guides the module into the final position. A docking pin and bucket system, as shown in the

Figure 2.1-3, is used for the final touch down and locks the module into the exact location on

the platform within strict tolerances.

Figure 2.1-2 Bumper and Guide System Figure 2.1-3 Pin and Bucket System

The pin and bucket system is also widely used for deck module supports stabbing in the jacket

legs and to dock a jacket onto pre-installed piles.

Finally, the deck module is supported vertically at the supporting structure.

2.1.1 Bumper and Guide system

The bumper and guide system is represented by two lines (AB and CD), as shown in the

Figure 2.1-4.

Figure 2.1-4 Bumper and Guide System Figure 2.1-5 Support System

At a specified time instant of the time domain simulation, the global locations of the bumper

(CD) and guide (AB) extremities are calculated. The line is presented as a vector with the

known position of its extremities. The coordinates of a point P on the guide post is defined by

the following equation:

( ) ( ) ABABAAAPPP LnZYXZYX ∗+= ,,,,

where ABn is the unit vector: ( ) BAZZYYXXn BAABABAB /,, −−−= and LAB is the length

of segment between A and P.

Similarly, the bumper can be represented by a vector (CD) and the coordinate of a point Q on

the bumper is defined as follows.

Bumper Bumper

Guide

Pin

Bucket

A

B

C

D AB

CDP

Q LAB

LCD

PQ

P

O B

AD

( ) ( ) CDCDCCCQQQ LnZYXZYX ∗+= ,,,,

where CDn is the unit vector: ( ) DCZZYYXXn CDCDCDCD /,, −−−=

When the vector between PQ is perpendicular to vector AB and CD, the distance between P

and Q is the shortest distance between the guide post (AB) and the bumper (CD). This can be

represented by two governing conditions as follows

1) Vector AB has to be perpendicular to Vector PQ:

0=• PQAB ,

2) Vector CD has to be perpendicular to Vector PQ:

0=• PQCD

We have two equations here with two unknowns, LAB and LCD. Therefore the coordinates of

point P and Q and, hence, the distance between the centrelines of the bumper and guide post

(PQ) can be found. If the distance between the centrelines is less than the sum of their

external radius, an impact occurs between the bumper and guide and the difference is the

deflection. Hence, the impact force of the specified time instant can be calculated based on the

given stiffness of the system and the calculated deflection. The impact force will be acting on

points P and Q of two bodies.

The stiffness is linear based on the elastic behaviour of the guide post which is part of

permanent structural member. However, the guide post can be sacrificial member in some

cases which will be removed after the installation. Plastic deformation is acceptable in these

cases. The stiffness is non-linear with local indentation considered. The non-linear load and

deflection relationship is curve-fitted to a polynomial equation in order to increase

computational efficiency.

2.1.2 Supports

In general, the support only provides vertical support with no horizontal restrictions. The

footing of the deck module can slide along the supporting deck within the horizontal tolerance

from bumper and guide and/or pin and bucket systems.

It is represented by a point to plane impact model, as shown in the Figure 2.1-5. The point is

the support footing of the module and the plane represents the landing area.

The plane can be defined by two vectors with three points (O, A and B) on the supporting

deck structure and is described by the following equation:

01 =+++ cZbYaX

The coefficients a, b and c can be found as followings:

BBB

AAA

ooo

ZYXZYXZYX

W =

Where

BB

AA

oo

ZYZYZY

Wa

111

1−=

BB

AA

oo

ZXZXZX

Wb

111

1−= 111

1

BB

AA

oo

YXYXYX

Wc −=

The shortest distance (D) between the point P (the footing) and the plane (supporting deck

structure) can be found by using the following equation:

222

1

cba

cZbYaXD PPP

++

+++= eq. (2.6)

The normal vector of the plane and the distance D can be used to determine whether the point

P is above or below the plane and whether D is a gap or a deflection (with impact). Once a

deflection is calculated, the corresponding impact force can be found using the specified

structural linear stiffness based upon the elastic behaviour of the supporting structure.

2.1.3 Docking Pin and Bucket System

The docking pin and bucket system is one of the installation aids widely used for module

installation. The main purpose of this system is to facilitate the installation of the module into

the exact location and heading within strict tolerances. This system is used mainly for the

final installation stage before touch down. This system is also widely used in jacket

installation.

The system will typically consist of a docking pin with tapered end on one body and a

receptacle cone (bucket) on the other, as shown in Figure 2.1-6.

When the docking pin lowers down and enter the receptacle cone, the engagement can be

separated into the following three different stages, as shown in Figure 2.1-7.

1) The tip of the pin is within the receptacle cone. 2) The tip of the pin passes the bottom of the cone. 3) The parallel section of the pin enters the parallel section of the bucket.

Figure 2.1-6 Pin and Bucket System Figure 2.1-7 Stages of Engagement

2.1.3.1 Stage 1

The bottom of the pin is located between the top and bottom of the receptacle cone. When the

relative horizontal movement at the specified vertical position is bigger than the gap at that

vertical position, an impact is obtained, as shown in Figure 2.1-8. The resultant of the impact

force has to be normal to the slope surface of the cone.

Figure 2.1-8 Stage 1 Engagement

The slope (θ) is at the side of the cone. The gap between the bottom of the pin (P) and the

side of the cone is calculated based on the location of the pin and bucket.

( ) ( )C

CLCU

LDD

⋅−

=2

tan θ ( )

22PLCL

C

CLCU DDL

DDhGap

−+

⋅−

⋅= ( )pcC zzLh −−=

If the relative horizontal movement between the pin and bucket (∆L) is greater than the Gap,

then impact occurs. The horizontal deflection is (∆H), which is contributed by the horizontal

DCU

DCL

LC

DPL

DPU

LP

h

Gap

∆L

C

P

compression (∆H’) due to horizontal stiffness and horizontal deflection induced by the

vertical compression (∆V’) with the vertical stiffness

( ) ( )22PCPC yyxxL −+−=∆

( )C

CLCU

LDD

VHGapLH×−

∆+∆=−∆=∆2

''

The Horizontal Impact Force can be calculated based on the horizontal deflection (∆H’)

together with the stiffness in X and Y directions, as follow. The Vertical Impact Force FV can

also be calculated accordingly with ∆V’.

xyPC

yPC

xH KHL

yyKL

xxKHF ⋅∆=

∆−

⋅+

∆−

⋅⋅∆= ''22

'VKF zV ∆⋅=

The resultant of the impact force will be normal to the surface of the cone and

( ) ( )C

CLCU

H

V

LDD

FF

⋅−

==2

tan θ

Once we have the global locations of point C (xC, yC, zC) on the cone and point P (xP, yP, zP)

on the pin, we have only two unknowns, (∆H’) and (∆V’). The vertical deflection can be

found by substitution and is listed as

( )

( )θ

θ

2tan

tan'

+

⋅∆=∆

xy

z

KK

HV '' VHH ∆−∆=∆

The corresponding impact force will be

Lxx

KHF PCxx ∆

−⋅⋅∆= '

Lyy

KHF PCyy ∆

−⋅⋅∆= ' zVZ KVFF ⋅∆== '

The impact forces are applied at the point P on the pin. Due to the fact that LC is short in

comparison, the impact forces are applied at the point C on the receptacle cone.

2.1.3.2 Stage 2

At this stage, the bottom of the pin has passed the bottom of the receptacle cone. However,

the bottom of the receptacle cone is in between the top and bottom of the tapered section of

the pin. When the relative horizontal movement at the specified vertical position is bigger

than the gap, an impact is obtained. The resultant of the impact force has to be normal to the

slope surface of the pin, as shown in Figure 2.1-9.

Figure 2.1-9 Stage 2 Engagement Figure 2.1-10 Stage 3 Engagement

The slope (θ) is the side of the pin. The gap between the bottom of the pin and the side of the

cone is calculated as follows

( ) ( )P

PLPU

LDD

⋅−

=2

tan θ ( )

p

pLpUPLCL

LDD

hDDGap⋅−

⋅−−

=22

( )pB zzh −=

Similar to the Stage 1, once we have the global locations of point B (xB, yB, zB) on the bottom

of the cone and point P (xP, yP, zP) on the pin, we have only two unknowns, (∆H’) and (∆V’).

The vertical deflection can be found by substitution and the corresponding impact force will

be L

xxKHF PBxx ∆

−⋅⋅∆= '

LyyKHF PB

yy ∆−

⋅⋅∆= ' zVZ KVFF ⋅∆== '

The impact forces are applied at the point P’ on the pin and point B on the receptacle cone.

2.1.3.3 Stage 3

At this stage, the top of the tapered section of the pin has passed the bottom of the receptacle

cone. When the relative horizontal movement is bigger than the gap, an impact is obtained.

There will only be horizontal impact force.

The gap between the bottom of the pin and the side of the cone is

2PUCL DD

Gap−

= ( ) ( )2'

2' PBPB yyxxL −+−=∆ GapLH −∆=∆

The horizontal impact force can be calculated based on the horizontal deflection (∆H)

together with the stiffness in X and Y directions, as follows.

h

Gap

∆L

B

P

P'

Gap

∆L

BP'

P

xypB

yPB

xH KHLyy

KLxxKHF ⋅∆=

∆−

⋅+

∆−

⋅⋅∆=2

'2

'

The corresponding impact force will be

LxxKHF PB

xx ∆−

⋅⋅∆= ' LyyKHF PB

yy ∆−

⋅⋅∆= ' 0.0== Vz FF

The impact forces are applied at the point P’ on the pin and point B on the receptacle cone.

2.2 Floatover Heavy deck modules can be installed by floatover method for semi-submersible, TLP and

fixed jacket structure. At the moment, the availability of heavy lift vessel with crane lifting

capability exceeding 10,000 t is very restricted in some areas, such as south-east Asia, Sea of

Okhotsk and Caspian Sea etc. Floatover would be the only installation method for heavy deck

module installation. The deck module will be transported by a flat top cargo barge and sit on

high supporting frame on the deck of the cargo barge, as shown in Figure 2.2-1.

Figure 2.2-1 Floatover Operations Figure 2.2-2 Installation Aids of Floatover Operations

The barge is towed into the gap between the legs of semi-submersible or jacket. After the

barge moves in, the barge is ballasted down and lowers the module onto the platform. In order

to minimise the impact load, special installation aids are used, such as Surge and Sway Fender

units, Leg Mating Unit (LMU) between the legs of platform and deck module and Deck

Supporting Unit (DSU) between the barge and deck module.

The usual arrangement of these units is presented in Figure 2.2-2. The surge fender will be

pressed on the leg of platform to restrict the barge surge motion during floatover operation.

There is a small gap between the sway fender and the leg of the platform. The LMU is usually

at the outmost support of the deck module. The DSU is at the inner supports on the barge

deck.

STER

N

A B

2

1

3

4

BOW

ALIGNMENT - MATING POSITION - LOAD TRANSFER

0.075m

TUG PULL 60t

SURGE FENDER IN CONTACT WITH JACKET LEGSLeg Mating Unit (LMU)

Deck Supporting Unit (DSU)

Surge Fender

SwayFender

STER

N

A B

2

1

3

4

BOW

ALIGNMENT - MATING POSITION - LOAD TRANSFER

0.075m

TUG PULL 60t

SURGE FENDER IN CONTACT WITH JACKET LEGSLeg Mating Unit (LMU)

Deck Supporting Unit (DSU)

Surge Fender

SwayFender

Leg Mating Unit (LMU)

Deck Supporting Unit (DSU)

Surge Fender

SwayFender

2.2.1 Sway Fender

When the barge moves in between the legs of the platform, impact occurs between the side of

the barge and the leg. Sway fenders are mounted on the sides of the barge to protect the barge

and the legs, as shown in Figure 2.2-3. The sway fender is typically wooden fender (with

higher stiffness) or rubber fender (with lower stiffness). It is common to use rubber fender to

reduce the impact but wooden fender is also used for benign sea area. The gap between the

fender and the leg normally is small (e.g. 75mm) in order to minimise impact load, especially

when the barge reaches the final mating position.

Figure 2.2-3 Sway Fender Figure 2.2-4 Surge Fender

The line to line impact numerical model used in bumper and guide system can be used here.

The fender and leg are modelled as two lines. As described in the section 2.1.1, the closest

distance between the leg and the surface of sway fender can be found at a specified time

instant. If the distance is less than the radius of the leg, impact occurs and the difference will

be the deflection. At the specified time instant, the impact force can be calculated based on

this deflection and the combined stiffness between the sway fender and the leg structure. The

load deflection curve of a non-linear system with rubber fender is curve-fitted into polynomial

equation in order to increase the computational efficiency.

2.2.2 Surge Fender

The action of surge fender unit is aligned parallel to the barge longitudinal axis to restrict the

surge motion. In the operation, the leading tug takes the barge into the opening between the

legs of the platform until the surge fender touches the leg. A mean tug pull will be used to

Sway Fender Surge Fender

apply a mean pre-compressed deflection on the surge fender to restrict the surge motion of the

barge.

The surge fender is usually a rubber fender with non-linear load and deflection relationship, as

shown in the Figure 2.2-5.

Figure 2.2-5 Load and Deflection curve for Rubber Figure 2.2-6 Surge Fender System

Surge Fender In the numerical model, the surge fender is simplified as a point (Q) on the cargo barge and

the leg of the platform will be represented by a line AB, as shown in Figure 2.2-6. At a

specified time instant, the global locations of the leg (AB) extremities and location of the

surge fender (Q) are calculated. The line is presented as a vector with the known position of

its extremities. The coordinates of a point P on the guide post is defined by the following

equation: ( ) ( ) APABAAAPPP LnZYXZYX ∗+= ,,,,

where ABn is the unit vector: ( ) BAZZYYXXn BAABABAB /,, −−−= and LAP is the length

of segment between A and P, which is the unknown.

In this numerical model, the vertical coordinates of the P and Q are the same (ZQ=ZP) to have

impact. By substituting ZQ into the previous equation, the LAP is found and hence the

horizontal distance between P and Q can be found. The distance PQ at the instant is compared

to the pre-compressed distance. The difference is added onto / minus from the pre-compressed

deflection. The impact force can be found based on the curve-fitted load and deflection

relationship shown in Figure 2.2-5. With the small angle assumption, the impact force will be

in X (fore and aft) direction applying through points P on the platform and Q on the barge.

2.2.3 Leg Mating Unit

When the barge reaches its final mating position, the longitudinal position is restricted by the

pre-compressed surge fender and the gap between the sway fender and the leg is small. The

0

20

40

60

80

100

120

140

160

180

0 100 200 300 400 500 600

De f l e c t i on ( mm)

Tug Pull

Pre-CompressedDeflection

A

B

PQ

AB

LAP

horizontal motion of the barge is limited. In this situation, the seafastening structures will be

cut and the barge will be ballasted down. The first contact between the deck module and the

leg of the platform is at the Leg Mating Unit, as shown in the Figure 2.2-7.

Figure 2.2-7 Leg Mating Unit

The Leg Mating Unit (LMU) is a rubber shock block mounted at the leg of the deck module.

The initial contact surface is a cone shape component and has a non-linear load and deflection

relationship as with rubber fenders. In high load transfer condition, the LMU is closed and

deck and platform legs will be directly in contact with high linear stiffness.

The slope surface in LMU does provide low stiffness during impact not only vertically but

also horizontally in the initial stage. Due to the cone profile, the horizontal movement will

induce the vertical compression of the LMU. Therefore the horizontal compression includes

the horizontal deflection of the structure and the deflection caused by the vertical

compression.

The resultant of the impact force has to be perpendicular to the slope surface of the cone.

These conditions formed two boundary condition and two equations. The two unknown

deflections can be solved and the unit can be numerically modelled.

In high load transfer, the LMU is closed and the horizontal and vertical stiffness will be high.

Three different phases are considered in the engagement.

2.2.3.1 Stage 1- No Load Transfer

In this stage, the LMU is over the cone on the platform leg. There is no load transfer and a

vertical gap exists between the LMU and the cone, as shown in Figure 2.2-8.

LMU

Cone

Compressed LMU

Leg of Deck Module

Leg of Platform

Leg Mating Unit

Figure 2.2-8 Stage 1- no load transfer Figure 2.2-9 Impact Forces

The global coordinates of the point P of the LMU and the point Q of the leg of the platform

are calculated at a specified time instant. A vertical gap can be identified and the

corresponding horizontal gap (GAPH) is calculated based on the slope of the mating cone (θ).

If the horizontal movement (∆L) is larger than GAPH, impact occurs. The horizontal

deflection (∆H) is calculated.

)tan(θV

HGAPGAP = ( ) ( )22

QPQP YYXXL −+−=∆ )tan(

''θ

VHGAPLH H∆+∆=−∆=∆

This horizontal deflection (∆H) is a combination of horizontal structural deflection (∆H’) and

the horizontal deflection due to the vertical compression (∆V’) of the LMU. This equation

forms the first condition and the two deflections are unknown.

In general, the horizontal structural stiffness is linear contributed by the combined lateral

stiffness of the legs of the platform and the deck module. However, the vertical stiffness is

non-linear and is a combination of the non-linear shock absorber and linear vertical structural

stiffness at the legs of platform and deck module. With the deflections, the horizontal (FH) and

vertical (FV) impact forces can be found. Since the impact occurs at the slope of the cone, the

resultant is normal to the surface of the cone, ( )V

H

FF

=θtan , as shown in Figure 2.2-9. This is

the second condition.

This case creates two unknowns with two equations. Owing to the non-linear deflection and

load relationship of the LMU, the two unknowns, ∆H’ and ∆V’, are found by numerical

iteration and hence the FH and FV are calculated. Similar to the section 2.1.3.1, the Fx and Fy

can be found from the FH and Fz = FV.

FV

FH

θθθθ

θθθθGAPvθ

GAPH

LMU

PlatformLeg

θPlatform

Leg

LMU

∆∆∆∆L

P

Q Q

P

2.2.3.2 Stage 2- Low Load Transfer

In this stage, the load is transferred from the barge onto the leg of the platform. There is no

horizontal gap and LMU is compressed, as shown in Figure 2.2-10.

Figure 2.2-10 Stage 2- Low Load Transfer

As with the previous section, the global coordinates of the point P of the LMU and the point

Q of the leg are known. From the vertical coordinates of points P and Q, a compression (∆V)

of the LMU can be identified and the corresponding impact load can be found based on the

non-linear load and deflection relationship.

Similar to the stage 1, the horizontal deflection (∆H) combines with horizontal structural

deflection (∆H’) and the deflection due to further vertical compression (∆V’) of the LMU.

( ) ( )22QPQP YYXXH −+−=∆

)tan(''

θVHH ∆+∆=∆

In this stage, the vertical impact force is the total load due to the compression (∆V) and

further compression (∆V’) due to horizontal deflection. Similar to stage 1, the two unknowns,

(∆H’) and (∆V’) are found and hence the impact forces, Fx, Fy and Fz, are calculated for the

specified time instant.

2.2.3.3 Stage 3- High Load Transfer

In this stage, the LMU is fully compressed with majority of the weight of deck module

transferred from the barge onto the platform. Usually, it is over 80% of the weight transferred

onto the legs of platform. The vertical and horizontal stiffness are linear and pure structural

stiffness due to the fact that the LMU is full compressed and assumed to be rigid.

At a specified time instant, the vertical coordinates of points P and Q can be calculated. The

vertical compression can be identified. The corresponding (FVc) is calculated from the load

and deflection relationship.

LMULMULMU

Vertical Impact ForceFVc

LMU Length - ∆∆∆∆V

LMU

∆∆∆∆H

∆V'

θθθθ θθθθQ

P P

Q

Similar to previous cases, the horizontal deflection (∆H) is a combined deflection from

horizontal (∆H’) and vertical (∆V’) deflections. The resultant of the impact force is normal to

the slope of the cone.

( ) ( )22QPQP YYXXH −+−=∆

)tan(''

θVHH ∆+∆=∆ ( )

V

H

FF

=θtan

The horizontal and vertical impact forces are

xyQP

yQP

xH KHH

yyK

Hxx

KHF ⋅∆=

∆−

⋅+

∆−

⋅⋅∆= ''22

'VKFF zVcV ∆⋅+=

This stage is also with two unknowns, (∆H’) and (∆V’), and two equations. The further

vertical deflection (∆V’) is calculated by substitution as follows and hence, horizontal

deflection (∆H’) can be found.

( )

( ) ( )θθ

θ

tantan

tan'

xyZ

Vcxy

KK

FHKV

+⋅

⋅−∆⋅=∆

Similar to previous stage, the impact forces, Fx, Fy and Fz, are calculated for the specified time

instant.

2.2.4 Deck Supporting Unit

The deck module sits on a supporting grillage on the deck of cargo barge. In high load

transfer condition, majority of weight is supported by the leg of the platform. The deck

module can be separated from the supporting grillage on the barge prematurely due to

dynamic response which can cause high re-impact load. In order to reduce the re-impact load,

Deck Supporting Unit (DSU) with rubber shock absorber block is used, as shown in Figure

2.2-11.

In low load transfer condition, the DSU is closed and fully compressed. The deck module is

supported at the point P on the grillage. The stiffness of the support is high and linear. In high

load transfer condition, as example, with less than 30% of weight is left on the barge, the

DSU is opened and the stiffness at these support points will be reduced significantly. Hence,

the re-impact load is reduced. A typical load and deflection curve of DSU is presented in

Figure 2.2-12 for reference.

Figure 2.2-11 Deck Supporting Unit Figure 2.2-12 Load and Deflection relationship of DSU

The global coordinates of the points P and Q are calculated for a specified time instant. If the

vertical location of point Q is greater than that of point P, the DSU unit is opened and

corresponding compression is the fully compressed deflection minus the difference between

vertical locations between point P and Q. The vertical impact force is calculated based on this

compression and the non-linear load and deflection relationship. The point to plane impact

method described in section 2.1.2 can also be used as more accurate model to cope with high

sliding situation in an unlikely event.

3 NUMERICAL SIMULATIONS In the engineering phase of the project, the installation operation is analysed using time

domain simulation. Saipem UK use the time domain simulation program, LIFSIM, reference

[3.1], from the Maritime Research Institute Netherlands (MARIN) The Netherlands.

LIFSIM handles up to three coupled bodies and calculates the motion responses by solving

the 18 coupled motion equations with fluid reactive forces described by convolution integrals.

Hydrodynamic coefficients of floating bodies are read into the program. The program allows

user to define additional external forces to each body. Users can write their own interface

subroutine which will be called at each time step in solving the coupled body motion

equations. The program provides 6 degrees of freedom motion responses of each body as

input to the user interface subroutine. The user can use the motion responses to convert into

additional external forces and moments, as described in previous sections, to the centre of

gravity of each body. These forces and moments are then considered in each integration time

step in solving the motion equations.

The described numerical models have been programmed in the time domain simulation using

the mentioned user interface facilities of LIFSIM.

Deck Support Unit

Deck Module

Supporting Grillage

Deck Support Unit

Deck Module

Supporting Grillage

Q

P

∆∆∆∆V

DSU Load and Deflection relationship

0.0

2000.0

4000.0

6000.0

8000.0

10000.0

12000.0

14000.0

0.000 0.050 0.100 0.150 0.200 0.250 0.300

Deflection (m)

Forc

e (k

N)

Full Compressed Design load

Des

ign

Com

pres

sion

∆∆∆∆V

3.1 Lift Operation An installation of a 400t deck module onto a moored semi-submersible Floating Production

Unit (FPU) is simulated, as shown in Figure 3.1-1. The installation is carried out by the semi-

submersible crane vessel (SSCV), S7000, with the module on the hook. In order to position

and protect the existing machinery and structure, front and side bumper and guide systems,

pin and bucket system and supports are fitted on the 400t deck module and FPU.

0 200 400 600 800 1 103×0

10

20

30

40

0

10

20

30

40

0

BG_front i

38.268

0

BG_sidei

10000 Ti

Figure 3.1-1 Deck Module Installation Figure 3.1-2 Front and Side Bumper &Guide Impact Loads

The simulation for quartering sea with a 2.0m Hs with 7 sec Tp using long crested

Torsethaugen wave spectrum is presented for reference. In the simulated operation, the SSCV

approaches the front guide post with a slow constant velocity (e.g. 1m/min) and establishes an

overboom to make the module lean on to the guide post. This will stabilise the deck module.

Then the module moves sideways with the same slow constant velocity. Similarly, an

overboom on the side bumper and guide system is also achieved. The module will then be

lowered down onto the target position by engaging the pin and bucket systems and landed on

the supporting deck structure.

The front and side bumper and guide impact loads are shown in Figure 3.1-2. Figure 3.1-3

shows the impact loads at the pin and bucket system. Similarly, the impact load at the support

is shown in Figure 3.1-4. The presented values are from one of the guide posts, pins and

supports for reference. As shown in Figure 3.1-2, the module approaches the front guide post

in the first 100 seconds and high impact load occurs in the initial phase. After the over boom

is established, the module leans onto the guide post and the impact load reduces. In case of

high slow drift motion, the module could separate from the guide post and re-contact with a

corresponding high impact load.

Bumper and Guide System

Pin and Bucket

Support

0 200 400 600 800 1 103×0

20

40

60

2−

0

2

4

6

855.122

0

PIN_Horz i

7.801

0.187−

PIN_Verti

10000 Ti

0 200 400 600 800 1 103×0

100

200

300273.404

0

SUPPORTi

10000 Ti

Figure 3.1-3 Pin and Bucket Impact Loads Figure 3.1-4 Impact Load at the Support

The module starts to move sideway after 500 seconds and high impact load occurs during the

initial contact. At about 700 seconds, separation occurs from side guide post and results in

high impact load. The lowering of module starts from 800 seconds. Impact occurs at the slope

section of the pin and bucket and vertical impact load occurs. When the pin goes through the

bucket, high horizontal load occurs in the initial phases. After the pins lock the module

horizontally, the horizontal impact load reduces due to the small gap between the parallel

section of the pin and bucket, as shown in Figure 3.1-3. Similarly, high impact load occurs

when the module support contacts the landing structure and impact load is significantly

reduced after the module settles down on the landing structures.

3.2 Floatover Operation A time domain simulation has been carried out for installing a 14000t deck module onto a

jacket structure by floatover method in a quartering sea with 0.75m Hs and 7 sec Tz using

long crested Jonswap Spectrum. The deck module is on the Saipem cargo barge, S45.

Although the presented results are installing a module onto a fixed structure, the methodology

is similar for a floating structure such as semi-submersible FPU.

The cargo barge, S45, with the deck module is at the installation position, as shown in Figure

2.2-2. The lead tug applies a mean pull of 60 t onto the barge and is counter-acted by the

reaction of the rubber surge fender which reduces the barge surge motion. The gap between

the rubber sway fender and the leg is small with 75mm to reduce the corresponding impact

loads. A constant pump rate of 6.2 t/sec has been simulated to increase the draft of the barge.

Figures 3.2-1 and 3.2-2 shows the surge and sway fender impact loads. The horizontal and

vertical impact load of LMU is presented in Figure 3.2-3 and the DSU vertical load is shown

in Figure 3.2-4. The presented values are one of the surge and sway fenders and one of the

LMU and DSU for reference.

0 1 103× 2 103× 3 103× 4 103× 5 103×0

50

100

141.386

0

FSURGE_B2i

4.75 103×0 Ti 0 1 103× 2 103× 3 103× 4 103× 5 103×

0

100

200

300

388.18

0

FSWAY_A2i

4.75 103×0 Ti

Figure 3.2-1 Surge Fender Impact Loads Figure 3.2-2 Sway Fender Impact Loads

0 1 103× 2 103× 3 103× 4 103× 5 103×0

50

100

150

200

0

1 103×

2 103×

3 103×

200

0

FLMUH_A1 i

3.712 103×

0

FLMUZ_A1 i

4.75 103×0 Ti 0 1 103× 2 103× 3 103× 4 103×

0

1 103×

2 103×

3 103×

4 103×

4.063 103×

0

FDSUZ_A2 i

4.75 103×0 Ti

Figure 3.2-3 Horizontal and Vertical Impact Figure 3.2-4 DSU Impact Loads Loads of LMU The pre-compression load is 30t for each fender. Figure 3.2-1 shows the compression load is

zero at a number of time instants which means separation between the surge fender and the

leg of the platform. High horizontal impact loads occurs in the LMU in the initial contact

phase. When the LMU is continuously in contact with the cone of the leg (i.e. continuously

with non-zero vertical impact load), there is no horizontal gap between the LMU and the cone

on the leg of platform. Hence the horizontal impact load is significantly reduced. High

fluctuation of the vertical load in DSU occurs at about 4500 seconds when the Module start to

separate from the DSU.

4 VISUALISATION It is important to check these numerical models and simulation. Visualisation is a practical

mean to check the modelling by converting the numerical simulation into animated action.

Saipem UK used GLview Inova from Ceetron, Norway to convert LIFSIM time history

responses into animation. It provides a means to check the coupled body dynamic behaviour

during the impact. Animation has been created for the presented simulations and realistic

dynamic behaviours during impact have been found.

5 CONCLUSION The presented numerical modelling has been applied in engineering projects. Although no

detailed correlations have been carried out, analysis results are found to be practical and

match with our experience. In addition, the animation presents a realistic dynamic behaviour

which matches with our observation.

6 ACKNOWLEDGEMENT The authors would like to acknowledge the contribution from Mr. Dario Giudice, Naval

Architect Coordinator, Saipem Singapore Pte Ltd, and Mr. Briac Herve, Naval Architect,

Saipem UK Limited, in the numerical analysis. In addition, the authors also acknowledge the

support to this work from Mr. Richard Harrison, Engineering and Welding Manager, Saipem

UK Limited.

7 REFERENCE 3.1 “LIFSIM User Guide”, MARIN, The Netherlands