Thesis PDF
Thesis PDF
Thesis PDF
MASTER’S THESIS
Study program/ Specialization:
Open
…………………………………………
(Writer’s signature)
Credits (ECTS): 30
Key words:
- Semi-submersible Pages: 8 + 56
- Air gap
+ Appendix: 8
- Hull dimension
- Stability + CD
- Heave motions
- Cost
Stavanger, 15.06.15
Abstract
One of the processes early in the development of offshore oil and gas fields is the concept selection.
Because semi-submersible and other alternative designs is very complex, a simplification in the
early design phase could lead to time and cost savings. The objective of this thesis is to develop a
simple spreadsheet, which estimates the hull dimensions and cost based on a few input parameters.
The estimated hull must have sufficient stability, heave motions and air gap.
All the important hull parameters have been evaluated and data from relevant semis have been
gathered. Based on these evaluations and data, the input parameters are selected and estimation
equations for the parameters are developed. The input parameters for the spreadsheet are topside
weight, weather condition and air gap margin. The estimations are mainly based on linear
relationships and are shown to give reasonable values when compared to existing semi-
submersibles. Especially the normal draft semis give very good estimations.
A parametric sensitivity study was conducted to look on the possibility to reduce heave motions.
The analysis of different hull configurations showed that draft is the only parameter giving a
noteworthy reduction. The necessity of reduced heave motions is the possibility to use steel
catenary risers. The necessary reduction was chosen to be roughly 50 %. The study showed that a
draft of 44 meters was necessary for this reduction, independent on semi weight. This deep draft
alternative is included in the estimation spreadsheet.
To secure sufficient air gap, an air gap analysis was conducted for five semis generated with the
estimation spreadsheet. The semis was analyzed for four different weather situations, which
represent the most common locations for offshore oil and gas production. One of the locations is
also analyzed in more detail. This detailed analysis show that the most critical weather situations
are for periods below TP and the additional response can result in an insufficient air gap. In addition
to estimate the required air gap, the spreadsheet accommodates the possibility to define the wanted
air gap margin. The analysis also show that the estimated semis have good stability and motions.
With only three simple input parameters, the estimation spreadsheet combines the presentation of
hull dimensions, weight, stability and cost in a good and user-friendly way.
i
Acknowledgments
The work with this thesis was done during the spring semester of 2015 at the University of
Stavanger. The thesis is the final requirement to the master’s degree in Offshore Technology. I
would like to thank the University of Stavanger for the knowledge and facilities it has provided me
during the education and the work with the master’s thesis.
I would also like to thank my faculty supervisor Jan Inge Dalane for the guidance and support
during the work on this thesis. The thesis objective developed by him, has been very interesting.
ii
Table of contents
Abstract .......................................................................................................................................................... i
Acknowledgments ..........................................................................................................................................ii
Table of contents ........................................................................................................................................... iii
List of figures ................................................................................................................................................. vi
List of tables ................................................................................................................................................. vii
Abbreviations .............................................................................................................................................. viii
1. Introduction ........................................................................................................................................... 1
2. The semi-submersible ........................................................................................................................... 2
3. Stability and motion theory................................................................................................................... 3
3.1 Stability of semi-submersibles............................................................................................................. 3
3.1.1 Initial stability ............................................................................................................................... 3
3.1.2 Stability terms............................................................................................................................... 4
3.2 Motion characteristics for semi-submersibles .................................................................................... 6
3.2.1 Heave ............................................................................................................................................ 6
3.2.2 Roll and pitch ................................................................................................................................ 9
4. Analysis theory .................................................................................................................................... 10
4.1 Semi-submersible Air Gap ................................................................................................................. 10
4.2 HydroD analysis ................................................................................................................................. 12
4.2.1 Wave directions, periods and location ....................................................................................... 12
4.2.2 Hydro and mass model ............................................................................................................... 13
4.2.3 Offbody points ............................................................................................................................ 13
4.3 Postresp analysis ............................................................................................................................... 14
4.3.1 Specific points............................................................................................................................. 14
4.3.2 Wave spectrums ......................................................................................................................... 15
4.3.3 Combined transfer functions...................................................................................................... 16
4.3.4 Statistical method....................................................................................................................... 16
4.3.5 Detailed analysis ......................................................................................................................... 16
5. Model Parameters ............................................................................................................................... 18
5.1 Topside layout and weight ................................................................................................................ 18
5.2 Air gap................................................................................................................................................ 18
5.3 Center of gravity ................................................................................................................................ 19
iii
5.4 Water plane area ............................................................................................................................... 19
5.5 Displacement ..................................................................................................................................... 20
5.6 Draft................................................................................................................................................... 20
5.7 Column dimension and shape ........................................................................................................... 21
5.8 Pontoon dimension and shape .......................................................................................................... 21
5.9 Distance between columns ............................................................................................................... 21
5.10 Natural heave period....................................................................................................................... 22
5.11 Natural roll and pitch period ........................................................................................................... 22
5.12 Summary of parameters .................................................................................................................. 23
6. The Semi-submersible Model .............................................................................................................. 24
6.1 Assumptions for the semi .................................................................................................................. 24
6.2 Default model dimensions ................................................................................................................ 25
7. Estimation Spreadsheet – Normal Draft ............................................................................................. 26
7.1 Air gap................................................................................................................................................ 27
7.2 Draft................................................................................................................................................... 27
7.3 Displacement and total weight ......................................................................................................... 28
7.4 Column width, pontoon width and pontoon height ......................................................................... 29
7.5 Column height and c-c distance ........................................................................................................ 31
7.6 Examples............................................................................................................................................ 32
7.7 Stability .............................................................................................................................................. 33
8. Sensitivity study for deep draft hull configurations ............................................................................ 35
8.1 Parametric study of alternative hull designs ..................................................................................... 35
8.2 Sensitivity study of alternative draft depths ..................................................................................... 39
8.2.1 Default model topside weight .................................................................................................... 39
8.2.2 20 000 tonnes topside weight .................................................................................................... 41
8.3 Summary............................................................................................................................................ 42
9. Estimation spreadsheet – deep draft .................................................................................................. 43
9.1 Displacement and total weight ......................................................................................................... 43
9.2 Column width, pontoon width and pontoon height ......................................................................... 43
9.3 Column height and c-c ....................................................................................................................... 44
10. Air gap analysis ................................................................................................................................ 46
10.1 Storm scenarios ............................................................................................................................... 46
10.2 Detailed analysis .............................................................................................................................. 47
iv
10.3 Results ............................................................................................................................................. 47
10.3.1 West Africa (4m, 16s) ............................................................................................................... 48
10.3.2 Mid-Atlantic (8m, 14s) .............................................................................................................. 48
10.3.3 Gulf of Mexico (12m, 15s) ........................................................................................................ 48
10.3.4 Norwegian Sea (16m, 18s)........................................................................................................ 49
10.4 Summary.......................................................................................................................................... 49
11. Cost Estimation................................................................................................................................ 50
12. The Spreadsheet .............................................................................................................................. 51
12.1 Input data ........................................................................................................................................ 51
12.2 Variable parameters ........................................................................................................................ 51
12.3 Fixed parameters ............................................................................................................................. 51
12.4 Output hull data .............................................................................................................................. 51
13. Discussion ........................................................................................................................................ 52
13.1 The models and assumptions .......................................................................................................... 52
13.2 Estimations and example semis ...................................................................................................... 53
14. Conclusion ....................................................................................................................................... 55
15. References ....................................................................................................................................... 56
Appendix A .................................................................................................................................................... A
Amplitude response for case 1 - 10........................................................................................................... A
Appendix B .................................................................................................................................................... C
Detailed air gap analysis ............................................................................................................................ C
Appendix C..................................................................................................................................................... E
Air gap analysis – normal draft .................................................................................................................. E
Air gap analysis – deep draft ..................................................................................................................... G
Appendix D .................................................................................................................................................... H
Spreadsheet user panel ............................................................................................................................. H
v
List of figures
vi
List of tables
vii
Abbreviations
viii
1. Introduction
One of the processes early in the development of offshore oil and gas fields, are the concept
selection. The concept alternatives must be detailed enough to estimate functionality and cost.
Because the semi-submersible and other alternatives are very complex designs, a simplification in
the early phase is necessary to avoid unwanted planning time and costs. The objective of this thesis
is to simplify the process of hull dimensioning for semi-submersibles in the early phase of a project
by developing a simplified excel spreadsheet. This spreadsheet will give an estimate of hull design
and cost based on a few simple parameters. The estimated semi-submersible should have good
stability and motion characteristics. This will reduce engineering hours in the early phase of a
project and will make experimenting with different semi-submersible designs easier.
The development of the spreadsheet consists of the following stages: evaluation of the relevant
parameters and their influence on the model, document relevant theory for the design and analysis,
design relevant semi-submersible hull configurations and perform hydrodynamic analysis on them,
use the collected analysis data and theory to develop the spreadsheet.
1
2. The semi-submersible
The first semi-submersible platform in the oil and gas industry arrived in 1961 and used for drilling.
After recognizing the good motion and stability characteristics of the semi-submersibles, it has
since become one of the standard vessels for drilling, intervention and production offshore. Some
of the advantages of semi-submersibles are large deck area, large topside load capacity, possible
re-use of rig, less capex than fixed/spar, good motion and stability characteristics. The semi-
submersible rigs are used for operations in intermediate (300 m) to ultra-deep water (4000 m).
The hull of a semi-submersible mainly consists of a pontoon and columns, shown in figure 1. The
pontoon is the main contributor to the buoyancy and can be either a ring pontoon or two elongated
pontoons. As a larger body, the ring pontoon gives more buoyancy and therefore can support larger
topside loads. It also gives a symmetrical hull shape and therefore gives better motions for a fixed
vessel. This is favourable for production semis, which holds large and heavy topside processing
equipment and will stay at one location. The two elongated pontoons give a more streamlined
shape, which give better speed and stability
during transit between locations. This is
favourable for drilling/intervention semis that
will relocate to new areas regularly.
2
3. Stability and motion theory
3
3.1.2 Stability terms
The general symbol for stability is the GM term. It describes the distance from the metacenter (MC)
to the center of gravity (CG). MC is the point where the centerline of the vessel intersects with the
vertical line through the center of buoyancy (CB). Dependent on the value of GM, the vessel has
the stability behavior described earlier.
GM > 0 Stable equilibrium
GM = 0 Indifferent equilibrium
GM < 0 Unstable equilibrium
Even though a positive GM is wanted, the value should not be too large. Too large GM will affect
the roll and pitch motions negatively by generating larger roll and pitch responses. This result in
short roll and pitch motions that is very uncomfortable for personnel onboard the vessel. As long
as the vessel is in an upright position, the CG and CB should be in the same vertical line. If not,
the difference will create a moment around the axis and destabilize the vessel. When the vessel is
heeling, CG and CB will be off axis and work in opposite directions. This is caused by the effect
of CB moving towards the heeling and away from the vessel centerline. This will generate righting
moments by CG and CB which will work towards returning the vessel to its upright position. These
moments will increase in force as the heel is increased.
The equation for GM is,
𝐺𝑀 = 𝐾𝐵 + 𝐵𝑀 − 𝐾𝐺 (1)
4
The KB term describing the buoyancy is given by equation (2) and (3).
𝐾𝐵𝑃 ∙ ∇𝑃 + 𝐾𝐵𝐶 ∙ ∇𝐶
𝐾𝐵 = (2)
∇
The BM term describing the buoyancy is given in equation (4) and (5) [Gallala, 2013]. Equation
(5) can describe the second moment of area for all directions because of the platform symmetry.
𝐼
𝐵𝑀 = (3)
∇
1 3 𝑑𝑐 2
𝐼 = 4 ∙ [( ) ∙ 𝑤𝑐 ∙ 𝑙𝑐 + 𝑤𝑐 ∙ 𝑙𝑐 ∙ ( ) ] (4)
12 2
5
3.2 Motion characteristics for semi-submersibles
Motions of a floating vessel are a six degree of freedom system. Three of the motions are rotational
and three are translational. Rotation means that the vessel rotates around an axis and translation
means that the vessel moves along an axis. The rotational motions are roll, pitch and yaw. The
translational motions are heave, sway and surge.
The motions sway, surge and yaw are not important for describing the motion characteristics of the
semi-submersibles in this thesis. Production semis will always be fixed on location by anchors or
DP, which restrict these motions. The WADAM analysis in HydroD calculates the heave, pitch
and roll motions for the semi-submersible models. Because of the hull symmetry, the pitch and roll
motions will always be equal for equal wave directions. The natural period equations in this chapter
is based on the assumption of uncoupled and undampened motions.
3.2.1 Heave
The heave motion represented by the RAO chart in Postresp are shown in figure 4. The amplitude
is calculated based on transfer functions from the WADAM analysis and show vertical
displacement for wave height (m/m). There are two important sections of the RAO chart: The
natural period in heave, which is the peak at around 26 seconds and the smaller elongated peak at
around 14 - 20 seconds. The natural period in heave is calculated by equation (6) [DNV, 2011] and
represent one of the critical responses for the semi.
6
𝑚 𝑚𝑣 + 𝐴33
𝑇ℎ𝑒𝑎𝑣𝑒 = 2𝜋 ∙ √ = 2𝜋 ∙ √ (6)
𝑘 𝜌𝑤 ∙ 𝑔 ∙ 𝐴𝑤
The mass term consist of the total mass of the semi and the added mass. Added mass is a term
describing the additional mass that the semi drags along its vertical motion. The water particles
closely adjacent to the submerged hull of a semi will always follow the hull surface as it moves,
and the longer away from the hull the water particles are, the smaller the effect will be. Added mass
is calculated by equation (7) [DNV, 2011]. This equation uses strip theory to calculate the mass.
This means that the added mass is calculated for one vertical section of the pontoon with the length
of 1 meter. The mass for this section is then multiplied by the total length of the pontoon.
𝑤𝑝 2
𝐴33 = 𝐶𝐴 ∙ 𝜌𝑤 ∙ 𝜋 ∙ ∙ 𝑙𝑝 (7)
2
The added mass coefficient depend on the width to height ratio of the pontoon. It is based on
empirical data and can be found for example in the standard DNV RP-H103.
7
The natural period must be far enough away from the wave periods with high energy, if not the
response can be severe. The period with high energy waves depend on the location, but a maximum
high-energy period of around 16 seconds is normal for severe weather locations. Therefore, a
general minimum natural period in heave for a semi is around 24 seconds. The elongated peak at
around 14 – 20 seconds is the area that is important for the heave motions of a semi. This is the
periods where the waves have high energy and the semi must tolerate these wave energies without
causing to large heave motions. The requirement for heave motions in this area depend on the riser
system used. Especially steel catenary risers can have fatigue problems in this area.
8
3.2.2 Roll and pitch
The roll and pitch motions for semi-submersibles are generally in a range and amplitude that make
them non-problematic for the design of the hull. As with the heave motions, the motions are
calculated in the WADAM analysis and represented in the roll and pitch RAO. The natural period
in roll and pitch are calculated from equation (8) and (9) [DNV, 2011].
𝑚𝑣 ∙ 𝑟44 2 + 𝐴44
𝑇𝑟𝑜𝑙𝑙 = 2𝜋 ∙ √ (8)
𝜌𝑤 ∙ 𝑔 ∙ ∇ ∙ 𝐺𝑀𝑡
𝑚𝑣 ∙ 𝑟55 2 + 𝐴55
𝑇𝑝𝑖𝑡𝑐ℎ = 2𝜋 ∙ √ (9)
𝜌𝑤 ∙ 𝑔 ∙ ∇ ∙ 𝐺𝑀𝑙
9
4. Analysis theory
The vertical gap between sea surface and topside structure is a function of the wave elevation and
the platform motions. The wave elevation is described by a non-linear irregular wave function.
When investigating air gap points near the columns or above the pontoon, it is important to take
into account that the wave will build up around the columns and on top of the pontoon. Because of
this effect, the most critical areas to investigate the air gap are in front of the columns, with respect
to the wave direction, and over the pontoon. Wave impact on the columns will cause water jets up
along the column. The wave elevation from this effect is hard to calculate and should be based on
empirical data. A specified point on the platform will move depending on heave, roll, pitch, sway,
surge and yaw motions. The available air gap at a specified time and point on the platform can be
written [DNV, 2012],
𝑎(𝑡) = 𝑎0 − [𝜂𝑁𝐸𝑇 (𝑡) − 𝛿(𝑡)] = 𝑎0 − 𝑟𝑡 (10)
10
The vertical displacement of a point on the platform, δ(t), is dependent on the heave, roll and pitch
motions at that specified point and can be written [DNV, 2012],
𝛿(𝑡) = 𝜉3 (𝑡) + 𝑦 ∙ 𝜉4 (𝑡) + 𝑥 ∙ 𝜉5 (𝑡) (11)
The vertical displacement of the waves around the columns, ηNET, depend on both first- and second
order wave motions for the incident, diffracted and radiated waves, and can be written [DNV,
2014],
𝜂𝑁𝐸𝑇 = 𝜂𝑖 (1) (𝑡) + 𝜂𝑑,𝑟 (1) (𝑡) + 𝜂𝑖 (2) (𝑡) + 𝜂𝑑,𝑟 (2) (𝑡) (12)
11
The slowly varying response cannot be found from the Postresp analysis and is therefore estimated
by empirical data, which gives a rough factor of 0.1 of the wave frequency response [DNV, 2014].
The slowly varying response of the platform is added to the wave frequency response after the
Postresp analysis. The total response is then given by the equation,
𝑟𝑡 = 𝑟𝑓 + 𝑟𝑠 = 𝑟𝑓 + [𝑟𝑓 ∙ 0.1] (13)
12
4.2.2 Hydro and mass model
The model that is analyzed is defined in this step. The hull model generated in GeniE is loaded into
the wizard and specified at a correct draft. The next step is to let WADAM calculate the total weight
and center of gravity for the platform. It is necessary to define the distance from CG to CB
manually. The calculated total weight, displacement and CG is now shown in the wizard, so this is
a good time to check the WADAM calculations against your own. WADAM calculates the radius
of gyration for the semi, but it is necessary to do these calculations manually to implement the
wanted topside weight in the equation.
In order to investigate the most critical areas around the platform, the air gap will be analyzed in
two points for each wave direction. For a wave direction perpendicular (0o) on the platform, the
points are located 5 m in front of the first column and 5 m in front of the second column. For a
wave direction diagonally (45o) on the platform, the points are located 5 m in front of the corner of
the first column and 5 m in front of the corner of the second column as seen in figure 6. It will not
be necessary to investigate other wave directions because of the symmetry of the hull. The specific
points are generated for the sea surface in WADAM, z-axis = 0, and for the semi in Postresp, z-
axis = 22.
13
Figure 7 - Air gap analysis points.
14
4.3.2 Wave spectrums
The wave spectrums investigated are single point JONSWAP spectrums, with the variables Hs, TZ,
γ and σ. The JONSWAP spectrum, “JOint North Sea WAve Project”, was a joint research program
in the North Sea, which was established to take care of the peak enhancement phenomena [Michael,
1999]. To describe the sea states on the Norwegian Continental shelf, the JONSWAP spectrum is
a good approximation. This spectrum includes a peakedness factor γ, which can describe more
extreme responses than for example the Pierson-Moskowitz spectrum. The Pierson-Moskowitz
spectrum is equal to the JONSWAP spectrum when the peakedness factor γ = 1.0. The sigma values
describe the width on each side of the peak. A-values for the low frequency side and B-values for
the high frequency side. In this thesis we have chosen the peakedness factor γ = 3.0 and the sigma
values σA = 0.07 and σB 0.09 which is reasonable values for the Norwegian Continental Shelf.
The variable TZ is the zero up-crossing period. While TP describes the period with maximum energy
density, TZ is a measure for the average period between zero up-crossings in a sea state. The zero
up-crossing period is found by the relation in equation (14) [DNV, 2011].
𝑇𝑍 (𝑇𝑃 , 𝛾) = 𝑇𝑃 (0,6673 + 0,05037𝛾 − 0,00623𝛾 2 + 0,0003341𝛾 2 ) (14)
15
4.3.3 Combined transfer functions
The response variables are transfer functions that can be used individually or multiple combined.
To analyze the air gap, it is created a response variable for each point specified on the semi earlier.
This variable takes into account all the six degrees of freedom and give the absolute vertical motion.
The combined response variable combines the response variable from the specific point on the
semi-submersible and the elevation of the sea surface. The output is the sum of these vertical
displacements, which describes both the negative and positive displacement in the z-direction.
The elevation of the sea surface has a factor of -1.2 and is computed by the off body points in
WADAM. The reason for it to have an additional factor of 0.2 is to account for the asymmetry of
the wave heights [DNV, 2014]. This transfer function takes into account the column placement and
the underlying pontoon for the second column. The rig motion at the specific point has a factor of
1.0.
16
√6,5 𝐻100 ≤ 𝑇 ≤ √11 𝐻100 (15)
The 100-year maximum wave height for a short-term 3hr sea state is approximated by multiplying
the HS with a factor of 1.9. It is also necessary to check the steepness criteria for the selected HS
and TP. The average wave steepness for a short-term irregular sea is [NORSOK, 2007],
2𝜋𝐻𝑠
𝑆𝑝 = (16)
𝑔 𝑇𝑝 2
SP Wave steepness
HS Significant wave height (m)
g Gravitational acceleration (m/s2)
TP Peak period (s)
17
5. Model Parameters
The design of a semi-submersible hull is a complex task and depend on many design parameters.
This complexity is why it would be convenient with a simplified design tool in the early design
phase to estimate the hull design. To simplify the model, some parameters are simplified by keeping
them constant and others are linearized. This chapter explains the most critical parameters, their
influence on the model and how they are treated in the hull estimation.
The topside weight is one of the parameters that is known when concept alternatives are developed.
Since it is a known value, it must be one of the input parameters for the spreadsheet.
18
platform. In the analyses, it is assumed that the air gap will have a fixed distance based on weather
conditions for all the cases. The air gap distance will affect the VCG of both the hull and topside,
therefore also the KG. This means that the hull design must counter the influence KG has on
stability (GM). Stability increasing adjustments can be done by: lower KG by increasing the hull
mass, increase BM by increasing the water plane area or/and the column center-center distance,
increase KB by increasing the draught (column displacement).
The air gap calculation is based on response analyses for the different models in relevant storm
scenarios. The response will be fitted for different designs based on displacement and weather
condition.
19
stability is slightly increased by the increased BM term, shown in equation (3) and (4). The overall
philosophy for water plane area design, is to keep it as small as possible.
5.5 Displacement
Displacement is the volume displaced by the pontoon and columns below the mean sea level. This
volume is what creates the upwards force (buoyancy) and its design volume is dependent on the
total weight from the hull and topside. In the estimations, the displacement is directly related to the
topside weight as the total weight is estimated by the topside weight. For deep draft semis, the
displacement is largely increased to displace the extra hull weight. The ratio between the
displacement in the columns and pontoon often lays in a certain area. For normal draft semis, the
columns displace about 25 % and the pontoon about 75 %. For deep draft semis, the columns and
pontoon displace about 50 % each. This has not been used to estimate the hull configuration, but
is evident for different designs in the spreadsheet.
5.6 Draft
Draft, also called draught, is the distance from the keel of the pontoon (bottom of the hull) to the
mean sea level. This value has a strong impact on the heave motions of the platform in the periods
below and up to peak period. In figure 5, these periods are expressed by the smaller peak to the left
of the natural peak period. The heave motions is a result from the wave forces acting on the
pontoon. The longer away from the waves the pontoon is, the smaller the forces and heave motions
will be. The dynamic pressure generated by waves decreases exponentially with increased water
depth. This means that the ideal design of a semi-submersible has as large draft as possible, while
still keeping the platform stable.
To increase the draft, the pontoon displacement must decrease and column displacement increase.
There is two ways of achieving this: One way is relevant for the assumptions made in this thesis,
which says that the pontoon width must be the same as the column width. This means that a
decrease in pontoon width and height will also decrease the column width. The other solution is to
keep the column and pontoon width the same, but decrease the pontoon length, but this reduces the
allowable topside VCG. For both of these solutions, it is necessary to increase the submerged
column length to achieve correct displacement.
20
5.7 Column dimension and shape
The columns can be either circular or quadratic. With quadratic columns, the pontoon are often the
same width as the columns, which will allow the stresses in the structure to spread correctly. The
alternative is that the pontoon width is narrowed slightly towards the middle, to get the right
motions and displacement ratios. The column width determines the water plane area and when
correlated to the pontoon width, it is one of the main parameters for the design. It influences the
draught, natural period in heave, hull VCG, KB, BM and the displacement ratio between pontoon
and column. Generally, as small column as possible is wanted since this can increase draught and
have positive impact on stability and heave motions as mentioned earlier.
21
5.10 Natural heave period
The natural heave period is a very important factor when considering the motion characteristics of
a semi-submersible rig. The closer the wave period is to the semi-submersible’s natural heave
period, the larger the response motion. The worst case is when the natural heave period of the semi-
submersible coincides with the wave period, which results in resonance motions. Resonance
motions will increase until either, the acting force (wave period) changes, or the mass of the semi-
submersible changes. The natural heave period is described by equation (6), where the important
parameters are the total mass, added mass and the stiffness. Total mass and stiffness is known from
the geometry of the platform. The added mass is a more complex parameter, dependent on the
submerged horizontal hull planes and a mass coefficient. By collection empirical data from the
WADAM calculations, the added mass and thereby the natural heave period can be estimated in
the spreadsheet.
For semi-submersible production platforms, it is necessary to estimate the heave motions because
it determines the fatigue loads on the production risers. Especially for steel catenary risers (SCR),
where all the vertical motions is distributed only in the touch down point. This leads to stricter
heave motion requirements for SCR than flexible risers.
The natural roll and pitch frequencies for semi-submersibles are generally not an issue for the
design. Therefore, it will not be accounted for in this thesis.
22
5.12 Summary of parameters
Based on these evaluations and their significance for the stability and motion characteristics, some
parameters will have a fixed value and other will vary depending on linearization from model
analyses. These simplifications is necessary to be able to generate the estimation spreadsheet. Their
influence on the accuracy will be discussed in the chapter 13.
To achieve excellent heave motions, some parameters must differ from the normal standpoint.
Which parameter who influences the heave motions in a good way is found by a parametric study
for the parameters who is suspected to have a positive influence. This is done in chapter 8.
23
6. The Semi-submersible Model
The semi-submersible models used in the calculations and analyses are simplified models only
consisting of the hull structure. The hull model created in GeniE consists of the outer shell, which
will be subjected to water forces. Supporting structure, like braces and beams are neglected from
the design, as only the outer shell shape will influence the hydrodynamic analyses.
24
6.2 Default model dimensions
For creating the spreadsheet and initial hydrodynamic analyses, a default model design is provided
to set up the correct formulas and calculations. The base model is an already conducted project
with good results from the stability, motion and air gap analyses. The generated spreadsheet for
estimation of hull design will follow the configuration of the default model and give the default
model design output if the topside weight and weather location is equal.
25
7. Estimation Spreadsheet – Normal Draft
The spreadsheet estimates hull dimensions based on the two inputs, topside weight and 100-year
significant wave height. The outputs are hull geometry dimensions, weight and stability. The
outputs are based on the default model and alters this model up and down to achieve decent stability
and geometry for different input values. The estimation spreadsheet has a topside weight range
from 20 000 to 80 000 tonnes. This includes most of today’s production platforms.
The reasoning behind the estimations are based on existing platforms, experimentation and general
knowledge of decent hull design. Some of the achieved results are compared to relevant existing
platforms to show the relevance of the estimates.
Input Output
- Air gap
- Draft
- Topside weight - Column width
- 100-year Hs - Column height
- Air gap margin - Column c-c
- Pontoon width
- Pontoon height
- Displacement
- Total weight
- Stability
- Cost
26
7.1 Air gap
The air gap has been fixed at 22.0 meter for all the estimated models in this chapter and for the
deep draft estimations. To see the final air gap estimation, see chapter 10. The air gap of 22.0 meter
was selected because the dimension of the default model and it would be sufficient for all platforms
in all-weather scenarios. The air gap for some of the semis will be decreased from the analyzed
22.0 meter. This will result in a larger available VCG for the topside or the possibility to alter the
hull dimensions to improve motions or reduce volume (cost).
7.2 Draft
The draft of the default model is 24.0 meter, which is in the range of normal values for regular
platforms. Initially, the draft was a fixed value (24.0 m) for all the estimated platforms. However,
this led to some problems with the stability of the platforms, which further generated unreasonable
hull dimensions. Therefore, the draft varies from 22.0 – 26.0 meter respectively for small to large
platforms. Data from existing platforms shows that this range of draft is common. Comparing the
estimated draft with drafts of the example platforms would not give any reasonable knowledge
since the draft for these platforms may be selected based on wanted motion characteristics and can
vary in a greater way than the other dimensions. The draft of the estimated models should give
heave motions in an acceptable range.
d Draft (m)
Wtop Topside weight (kg)
As many of the estimation equations are build up in the same way, the draft equation will be
explained in more detail here. The equation generates a linear increase in draft from the minimum
value 22.0 meter. To assign the minimum value to a topside weight of 20 000 000 kg, 20 000 000
is subtracted from the topside weight in the numerator. The fraction will then be zero for this
topside weight. If the equation shall generate 24.0 meter for the default model, the numerator must
27
be twice the size of the denominator for this topside weight. The sum in the numerator is 50 661 050
– 20 000 000 which equals 30 661 050. Then the denominator is selected to be 15 330 525 (half)
to achieve the 24.0 meter draft. Many of the other equations are build up in the same way, a ratio
between the weights in the fraction, added with the minimum value. Some estimation equations
will not have a minimum value, but be restricted by the range of the topside weight.
The topside weight percentage varies based on empirical data from relevant platforms and the
default model. It also takes into account to achieve decent stability for the platform. The default
model is relatively a medium-large platform and has a topside weight percentage of 45.78 %. Large
platforms can achieve a topside weight percentage of about 50 % and the smallest platforms about
35 %. This relationship has been found when investigating existing semi designs. See section about
example semis for references to investigated data. The reason behind the variation in topside weight
percentage is mainly that additional displacement is achieved with less steel weight, than the
increase in topside weight. The internal equipment inside the hull also does not scale in the same
way as the hull. From stability experiments and relevant existing platforms, the topside weight
percentage is selected to vary from 38.0 % for the smallest platform to 53.22 % for the biggest.
𝑊𝑡𝑜𝑝 − 20 000 000 𝑘𝑔
𝑊%,𝑡𝑜𝑝 = 0.38 + (20)
394 100 899.7 𝑘𝑔
From the topside weight percentage, the total weigh together with the hull weight is calculated.
This also gives the required displacement of the hull. Shown below is the displacement of estimated
models based on the topside weight percentage compared with existing platforms. The chart shows
that the estimated displacement and total weight is reasonable relative to the topside weight.
28
Figure 11 - Topside weight versus displacement, with examples.
Small platforms has a column width of about 16.0 – 18.0 meter. The large platforms has a width
of about 23.0 – 24.0 meter. There are not many very large platforms to compare against, but as an
estimate, the column width for the largest estimated platform is set to 25.0 meter. The smallest
platform receives a width of 16 meter. This gives a smaller range of variation for the platforms
larger than the default model (2.0 m) than the smaller ones (7.0 m). This has to do with the
stabilization of the platforms. Generally, the larger the platform the larger must the deck area be.
Therefore, as the platform gets bigger and topside weight increases, the height of the platforms
does not change significantly. By reducing the increase in column width, the column c-c can
increase more and provide more stability for the wider topside.
29
To implement this into the spreadsheet, the calculation of column width is different based on
whether the topside weight is above or below the default model.
For topside weights below the default model, the calculation of column width follows this relation.
𝑊𝑡𝑜𝑝 − 20 000 000 𝑘𝑔
𝐶𝑤 = 16 𝑚 + (21)
4 380 150 𝑘𝑔
The estimation of pontoon height follows the same method as the column width and is based on
relevant existing platforms. The height for platforms smaller than the default varies in the range of
8.5 – 11.0 meter. For larger platforms, it varies from 11.0 – 12.0 meter.
30
Based on topside weight, these values are calculated. The chart below shows the estimated column
width compared with relevant examples. The available examples for the larger platforms are non-
existent also for this chart, but for the medium to small platforms, the estimations are a reasonable
good match with the examples.
The column height is the sum of the air gap and draft. Note that it includes the pontoon height. If
not stated else, the column starts at the top of the pontoon for the other parts of the thesis.
31
𝐶ℎ = 𝐴 + 𝑑 (27)
The chart below shows that the range for the column c-c is reasonable. Note that the chart shows
the total pontoon length, which includes the column width in addition.
7.6 Examples
Existing semi designs have been studied to be able to compare the estimations with existing data.
This has shown to be valuable when developing the equations as they work well as a guiding
parameter. Interesting relationships between topside weight and hull dimensions have been found
for many of the dimensions, which have been used in the estimation equations. The data is mainly
gathered from “worldwide survey of semi-FPSs and FPUs”, by Mustang Engineering. The data is
also quality check with manufacturers where it has been possible.
32
Table 2 - Semi-submersible production platforms [Wood Group Mustang, 2011], [GVAC.SE, 2013], [Aker Kværner, 2003],
[OFFSHORE-TECHNOLOGY.COM, 2015].
7.7 Stability
The estimated hull dimensions generates an estimate on the stability of the platform. The stability
is represented by the VCG for the hull, topside and total (KG). KG is calculated based on equation
5 with the assumption of a constant GM of 4.0. The VCG for the hull is more complicated to
calculate and depends on the arrangement of mooring, equipment, ballast tanks etc. Therefore, it is
very difficult to estimate a reasonable distribution of the weight in the hull. One easy approach
could be to assume a uniform distribution of the weight on the hull, but this would generate a too
high VCG, as the hull is generally heavier in the lower parts. This is mainly because of ballast and
storage tanks in the pontoons and lower columns.
The method chosen to estimate the hull VCG is to assume uniform distribution of the weight, but
add a correction factor for the pontoon. Because of uniform distribution of weight, volume can
replace the weight term. The default model has a VCG for the hull of 12.0 meter, but with uniform
distribution has 16.5 meter. To achieve 12.0 meter, the correction factor must be 2.35. Simulating
for other topside weights, this assumption gives reasonable results. The formula for estimating hull
VCG is shown below.
𝑃 𝐶
[𝑉𝑝 ∙ ( 2ℎ ) ∙ 2.35] + [𝑉𝑐 ∙ (𝑃ℎ + 2ℎ )]
𝑉𝐶𝐺ℎ𝑢𝑙𝑙 = (28)
𝑉𝑐 + (𝑉𝑝 ∙ 2.35)
33
The VCG for topside is calculated based on KG and hull VCG. The formula for estimating topside
VCG is shown below.
𝑊𝑡𝑜𝑡𝑎𝑙 ∙ 𝐾𝐺 − 𝑊ℎ𝑢𝑙𝑙 ∙ 𝑉𝐶𝐺ℎ𝑢𝑙𝑙
𝑉𝐶𝐺𝑡𝑜𝑝 = (29)
𝑊𝑡𝑜𝑝
34
8. Sensitivity study for deep draft hull configurations
The purpose of a deep draft semi is to minimize the heave motions to allow the use of for example
steel catenary risers. When developing the spreadsheet, the focus is first on how the other variables
influence the heave motions. Then, the results from those studies are used to study the effects of
different draft configurations. The study of the dimensions is done by generating different cases
and studying the heave and pitch RAO for those. A good estimation is achieved when the heave
and pitch motions are lowered by about 50 % from the normal draft semi. To achieve larger draft,
the hull size and thereby its weight must increase. Increased hull size also leads to increased costs.
Therefore, the hull size is increased minimally to achieve the wanted response decrease.
For heave RAO, the area around the first peak (about 12 – 18 s) is evaluated. For the pitch RAO,
the peak values are evaluated.
In the default case, the hull weight was governed by equation (20). In these deep draft cases (1 –
6) the first parameter, increased hull weight percentage, describes how much the original weight
must increase to accommodate the extra hull displacement. This parameter is wanted as low as
possible to minimize cost.
35
The objective of case 1 and 2 is to show the effect of increased column width, when keeping
all other parameters constant.
The objective of case 3 and 4 is to show how the semi responds to increasing and decreasing
the column c-c.
Case 5 illustrates the effect of decreasing the pontoon height.
Case 6 is an extreme case for all the positive configurations and only used for illustrational
purposes.
1 2 3 4 5 6 Default
4,0000
3,5000
3,0000
Amplitude (m/m)
2,5000
2,0000
1,5000
1,0000
0,5000
0,0000
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38
-0,5000
Period (s)
1 2 3 4 5 6 Default
0,8000
0,7000
Amplitude (m/m)
0,6000
0,5000
0,4000
0,3000
0,2000
10 11 12 13 14 15 16 17 18 19 20 21 22
Period (s)
36
The first thing to discuss from the studies is the problem with the natural peak period in the
amplitude for some of the cases. Normally the peak has an amplitude between 1.5 and 3.0, but here
an amplitude of e.g. 50 and 100 are reported. This is not a problem for the results, since it is not
this area of the graph that is evaluated. The problem with the peaks is probably caused by something
with the dampening of the system and the models does not include Morrison model, which would
contribute to damping. This problem should be analyzed if further work is done on the thesis.
1 2 3 4 5 6
0,2000
0,1800
0,1600
Amplitude (m/m)
0,1400
0,1200
0,1000
0,0800
0,0600
0,0400
0,0200
0,0000
4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38
Period (s)
1 2 3 4 5 6
0,2000
0,1900
0,1800
Amplitude (m/m)
0,1700
0,1600
0,1500
0,1400
0,1300
0,1200
0,1100
0,1000
10 12 14 16 18 20 22
Period (s)
37
The results from case 1 and 2 shows that the column width has no noteworthy influence on the
heave RAO. On the other hand, it is seen that pitch motions is reduced (7.8 %) when increasing
column width from 24.5 to 26.0 meters.
Case 3 and 4 evaluates the influence from the column c-c distance. Figure 15, show that a reduction
has a positive effect and an increase has a negative effect on the heave RAO. The same applies to
the pitch RAO. The effects are very small, when considering that the c-c distance is altered by
plus/minus 4 meters, which is a relatively large change. The positive effects for the heave and pitch
RAO compared to case 1 are respectively 3.5 % and 1.3 %.
Case 5 shows the effect of reducing the pontoon height and the result show a positive effect. The
reduction of both heave and pitch RAO compared to case 1 are respectively 7.0 % and 4.0 %.
Compared against the default model, all the 40-meter draft cases gives excellent motions. The worst
(case 4) and best (case 5) have respectively a reduction in heave RAO of 38 % and 44 %. The
results show that the main contributor to the motions is the draft, which was expected. The next
stage for generating the estimation spreadsheet is to configure the equations to increase the draft,
while altering the other dimensions as recommended from the study, until 50 % reduction in heave
motion is achieved.
Based on these results, a new parametric study is done. This time the draft is varied while the other
parameters are kept constant. Three cases are studied with the draft at 40.0, 42.0 and 44.0 meters.
The dimensions of the hull is selected based on the data from the previous study.
38
8.2 Sensitivity study of alternative draft depths
39
7 8 9 Default
5,0000
4,5000
4,0000
3,5000
Amplitude (m/m)
3,0000
2,5000
2,0000
1,5000
1,0000
0,5000
0,0000
4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38
-0,5000
Period (s)
7 8 9 Default
0,8000
0,7500
0,7000
0,6500
0,6000
Amplitude (m/m)
0,5500
0,5000
0,4500
0,4000
0,3500
0,3000
0,2500
0,2000
0,1500
0,1000
10 11 12 13 14 15 16 17 18 19 20 21 22
Period (s)
40
The charts show that the natural peak period for heave is reduced from 26 to 24 seconds for the
deep draft models. This is a compromise to achieve excellent motions for the semi, but is acceptable
as the wave energy is limited around these periods. The total reduction in heave RAO for the three
cases are respectively 43.1, 46.4 and 49.4 percent. The achieved heave RAO for case 9 is
considered a good result and will govern the estimation spreadsheet for the deep draft semis.
Normal Case 10
Topside weight (tonnes) 20 000 20 000
Hull weight (tonnes) 32 632 65 313
Total weight (tonnes) 52 632 85 313
Increased hull weight (%) 0 62.1
Topside VCG (m) 10.6 12.1
Column width (m) 16.0 17.0
Column c-c (m) 69.0 76.5
Pontoon width (m) 16.0 17.0
Pontoon height (m) 8.5 8.0
Air gap (m) 22.0 22.0
Draft (m) 22.0 44.0
41
Case 10 Normal draft
4,0000
3,5000
3,0000
Amplitude (m/m)
2,5000
2,0000
1,5000
1,0000
0,5000
0,0000
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38
-0,5000
Period (s)
The results from the alternative hull configuration in case 10 show a reduction of heave RAO of
49.6 %. This prove good motions and show that the results from case 1 to 9 can be used to generate
the estimation spreadsheet for all the variable topside weights.
8.3 Summary
The results show that the heave RAO is mainly dependent on the draft of the semi. A reduced
column c-c distance also showed small positive effects on the RAO, but will influence the natural
peak period negatively. When developing the estimation spreadsheet for deep draft semis with
excellent motions, the main parameter will be the draft. The other dimensions will only follow the
recommendations given in the earlier results to achieve good stability and natural peak period.
These results show the same recommendation as a sensitivity study conducted by a joint industry
project by Statoil, Hydro, Shell, BP and Marintek in 2006. Their study took into account more hull
configurations, but the conclusion was the same. “The sensitivity study has shown that from all the
geometric variations that have been tested, very few if any gave sufficient effect on the vertical
riser porch in fatigue sea states”[Arnesen, et al., 2006]. Based on these findings the verdict was
that a more drastic change to the design parameters must be implemented to reduce heave motions
for the riser porch, by increasing the draft.
42
9. Estimation spreadsheet – deep draft
The estimation spreadsheet for deep draft semis is based on the same principles as for the normal
draft. The spreadsheet will estimate hull dimensions from the inputs topside weight and 100-year
HS. The spreadsheet can estimate hull configurations for topside weights between 20 000 and
51 000 tonnes. To achieve the excellent vertical motions, the draft was found to be 44 meter for
both the default model and the 20 000 tonnes topside model. From this, it is assumed that the draft
is fixed at 44 meter for all the estimations. This chapter describes the estimation equations that
have been changed from the normal draft estimations.
𝑊𝑡𝑜𝑡𝑎𝑙 = ∇ ∙ 𝜌𝑤 (31)
43
the range 8.5 – 12.0 meter. For deep draft semis, the range is reduced to 8.0 – 9.0 meter for the
smallest and largest topside respectively.
The pontoon height varies linearly and is described by the equation below.
𝑊𝑡𝑜𝑝 − 20 000 000 𝑘𝑔
𝑃ℎ = 8 𝑚 + (32)
30661050 𝑘𝑔
Equation for calculating topside weight for 20 000 – 25 000 tonnes topside weight. The other ranges
are calculated in the same way but different fixed parameters.
𝑊𝑡𝑜𝑝 − 20 000 000 𝑘𝑔
𝐶𝑐−𝑐 = 76.5 𝑚 + (33)
2 000 000 𝑘𝑔
The column height is estimated with the same method as for normal draft.
44
Below is a chart showing the differences in displacement, column width and draft between the
normal and deep draft semi estimations.
45
10. Air gap analysis
The air gap is analyzed for normal and deep draft semis with 20 000, 50 000 and 80 000 tonnes
topside weight with the standard 22.0 meter air gap. The analysis is done in the Postresp software.
A single point Jonswap wave spectrum is generated for each storm scenario and together with the
air gap location and wave direction merged as a response spectrum The response spectrums are
analyzed with a storm duration of 3 hours (10 800 s) and a probability of exceedance of 0.1. After
the responses are generated, they are increased by 10 % to include slowly varying motions in the
rig. The detailed analysis investigates some of the periods just below the peak period for the given
HS, which is expected to give greater response.
It is assumed that the minimum design air gap will be 12.0 meter and increase in increments of 2.0
meters. The response is allowed to overrun the air gap by about 2.0 meters as described earlier.
The selected locations give an evenly increase in severity of the storm conditions. The locations
selected are Gulf of Mexico, Mid-Atlantic, Norwegian Sea and West Africa.
Table 7 – 100-year storm situations [Swail, Ceccacci, Cox, 2000], [NORSOK, 2007].
Significant wave height, HS (m) Peak period, TP (s) Zero-up crossing period, TZ (s)
West Africa 4.0 16.0 12.2
Mid-Atlantic 8.0 14.0 10.7
Gulf of Mexico 12.0 15.0 11.5
Norwegian Sea 16.0 18.0 13.8
46
10.2 Detailed analysis
To investigate the wave spectrum for the Norwegian Sea in more detail, it is necessary to check
the critical wave periods below TP and their wave steepness.
The most critical periods for a storm condition with HS = 16.0 and TP = 18.0 are found by using
equation (15). This result in the period range 14.05 ≤ T ≤ 18.28. Periods with 1s increments will
be used, which gives four additional periods to the original. The steepness for these periods and
correlated wave heights must then be checked against the NORSOK criteria with equation (16),
(17) and (18). The results are shown in table 8.
Table 8 – Wave steepness for detailed analysis.
The correlated wave height to the additional periods are found by extrapolating the contour diagram
for Norwegian Sea. As seen in table 8, HS for the two lowest periods was too steep and needed to
be corrected to fulfill the steepness criteria.
10.3 Results
The air gap response was investigated at four points around the semi for two wave directions. The
results show the maximum response from one of the points for each storm scenario. The available
air gap is described as the initial air gap minus the response.
Table 9 – Respons data for analyzed semis.
20’ - normal 20’ - deep 50’ - normal 50’ - deep 80’ - normal
West Africa 4,8 5,5 5,9 5,5 6.6
Mid Atlantic 11,4 12,0 14,1 13,3 15.7
Gulf of Mexico 15,5 17,1 19,3 18,9 21.4
Norwegian Sea 17,1 21,5 20,6 24,0 22.2
Norwegian Sea + 0.8 + 1.6 + 2.4
(Detailed Analysis) + 4.6 % + 7.7 % + 10.8 %
47
Detailed analysis for the Norwegian Sea conditions indicates that a detailed analysis becomes more
important as the semi becomes heavier. A response increase of up to about 10 % could make an air
gap insufficient for the selected location and result in unwanted damage to structure and equipment.
Table 9 show that the response is worsened for the deep draft semis. This is because the increased
draft reduces the vertical motions of the semi and thereby prevents the semi from following the
wave motions. The increased response for the heavier semis come from the same principle, but
here the weight reduces vertical motions.
10.3.1 West Africa (4m, 16s)
For the West Africa conditions, the variations and response are relatively small for all the semis.
This leads to the air gap calculation for this condition and semis can follow the same estimation
without giving to large source for error. For this condition, the semis will have the minimum air
gap of 12.0 meter, which will allow enough air gap under all conditions.
48
10.3.4 Norwegian Sea (16m, 18s)
The semi response in Norwegian Sea locations is the most severe of the analyzed scenarios. The
variations is somewhat linearly and predictable with roughly a 2.0 meter increase between the
normal draft semis and about 4.0 meters between normal and deep draft semis. If assuming linearly
response increments between the semis, the distribution of the selected air gap is as follows for
normal and deep draft semis.
Table 12 – Air gap estimation for Norwegian Sea.
10.4 Summary
When considering the detailed analyzed responses, the additional response tilted the selected air
gap for the larger semis from 20.0 to 22.0 meter. This show the importance of analyzing the
significant wave height for periods below the peak period when doing detailed engineering.
As all the analyses in this thesis is done with an air gap of 22.0 meter, the lower recommended air
gap will for many of the semis give a possibility to alter the design and/or allow a higher VCG for
the topside.
49
11. Cost Estimation
The cost estimation is calculated from the hull volume. It uses the fixed parameters density of hull
(kg/m3) and hull cost (NOK/kg). These values are provided by the supervisor and are based on
earlier design cases. The reason for not using the estimated hull weight and calculate the weight
from the hull density parameter, is that the estimated hull weight takes into account all loads acting
on the hull. E.g., ballast will not add any production cost. The density of the hull takes into account
both steel weight and equipment weight. Cost estimation is shown in equation (34).
𝐶𝑜𝑠𝑡ℎ = 𝜌ℎ ∙ 𝑉ℎ ∙ 𝐶ℎ (34)
50
12. The Spreadsheet
Here follows a guideline for using the spreadsheet. The spreadsheet estimates hull dimensions for
topside weights between 20 000 and 80 000 tonnes. A figure of the spreadsheet user panel is
provided in Appendix D.
51
13. Discussion
One of the assumptions for the models is that rounding of column and pontoon corners are
neglected. For real design situations, the columns and pontoon will always be designed with
rounded corners to smoothen out flow around the hull and add to structural integrity. These are not
issues for the studies in this thesis, but rounded corners would reduce the displacement and water
plane area some. With a reasonable rounding of 3m radii on the columns of the default model, the
water plane area will reduce by 1.5 %. This is not enough to alter the behavior of the semi
noteworthy. The rounding radii of pontoons are generally smaller, e.g. 1m. The total displacement
effect with rounded columns and pontoon would therefore be negligible.
The topside weight is a fixed input parameter. During the lifetime of a production facility,
requirements can change, which results in modifications and maybe new modules in the process
area. This generates additional weight and movement of the total center of gravity. Because of this,
it is important to allow some margin in the topside VCG, which is implemented in the estimations
in the spreadsheet.
The hull weight estimated for the semi-submersibles includes live loads, riser loads, ballast,
equipment and steel weight. This generates some uncertainty about the distribution of weight in
the hull and the real hull weight. As stated in the thesis, the weight is assumed uniformly distributed
but with an additional factor for the pontoons. This assumption obviously generates a source for
error as the VCG of the hull is not only dependent on its volume, but will vary depending on the
design and production situation. Some variance in the height of the hull VCG will not alter the KG
52
severely enough that the allowable topside weight and VCG becomes unsuitable. The estimated
allowable VCG for the topside should have a large enough margin in all the cases to tolerate some
variance. When comparing with the default model, the estimated hull VCG with the pontoon added
factor is reasonable. The change in hull VCG for increased/decreased draft and hull size have been
evaluated with the supervisor and found to give reasonable values.
The total weight of the hull has some uncertainty. The weight includes e.g. riser loads, which is not
only dependent on hull size, but depend mostly on the selected riser material and water depth. The
number of production risers will influence the total riser weight and the number depend on the
reservoir size. The larger the reservoir/production, the more risers. A larger production leads to a
larger topside weight, which again leads to a larger hull, so this weight factor will follow the sizing
of the semi. The total weight of a riser system is generally in the area around 2000 tonnes, which
means it will often not have a significant impact on the total hull weight. As mentioned about the
margin for the topside VCG, altering ballast weight will not generate trouble with the allowable
weight and VCG.
The GM is fixed at 4.0 meters for all the models. In real final design studies, the GM is normally
in the area from 3.0 to 4.0 meter. Fixing the GM at 4.0 will not generate any problems with the
design, but it can hinder the possibility to exploit all the potential of a design. E.g. reducing GM
slightly could allow for a small reduction in hull size, resulting in reduced cost or the possibility to
increase the allowable topside weight and VCG.
Many of the equations are based on linear relationships with topside weight or on empirical data
from existing rigs. This will not always give the most fitting value for a dimension, but as estimated
values, they are shown to generate decent values. The purpose of the thesis is only the estimate the
53
dimensions to make early phase design comparison and decisions easier, not to give exact final
design dimensions. One way of getting more accurate estimations could be to use other forms of
equation fitting than linear.
Some of the dimension values for the existing semis have been difficult to quality check, and
therefore the comparison can be misleading. This is only for a small amount of the data and the
values should not be way off if they are wrong. The dimension comparisons include many example
semis to minimize this source of error.
One of the important sources of error in the analysis of different draft configurations is the lack of
damping in some of the models, which generates severe peaks around the natural period in heave.
The source of the problem is unknown. The important area for the study, which is the first elongated
peak, showed correct values for the default model and reasonable values for the estimated cases.
This means that the damping problem does not affect the conclusions from the study.
One of the objectives of the thesis was to estimate a hull with minimum weight while still having
good stability, motions and air gap. This objective is not implemented in the spreadsheet as a factor,
as it has not been enough time to do the required analyses. However, during experimentation of
different hull dimensions and estimation equations, the results giving the lowest hull weight have
been chosen.
The deep draft hull estimations assumes the necessity of 50 % heave RAO reduction. How large
reduction is necessary obviously depend on each individual case. Especially water depth is a factor.
Because the reduction effect of increased draft reduces as draft increases, the assumption of 44
meter draft could be many meters too deep, to achieve the required reduction. In what range the
required reduction percentage varies is not known, but it could be interesting to study the range
and correlation to other parameters.
54
14. Conclusion
The developed estimation equations for normal draft semis have been shown to give reasonable
results when compared to existing semi designs. The hull design estimated for 20 000 and 80 000
tonnes topside weight for the air gap analysis in chapter 10, show good stability and motion
characteristics. In an overall view, the results with the normal draft estimations are very good and
can possibly work as the intended objective of the thesis, a design decision support tool.
The estimation equations for deep draft semis showed the ability to reduce the heave motions
considerably while having good stability. With the lack of comparable data, it is difficult to know
if the estimations give recommended design solutions, but the objective of reduced heave motions
is fulfilled.
Recommended air gap for all the possible hull configurations is provided for the most common
weather situations. Depending on requirements and client requests, the wanted air gap margin can
be set. This function can save much time during early phase design decisions, because the air gap
analysis requires modelling, analyses and post-processing to get comparable values.
During the development work with the estimation equations, interesting findings for semi design
have been identified. E.g., topside weight percentage was shown to have a strong relationship to
the total semi weight. Many other design parameters also followed strong trends, which means that
estimation of the hull parameters can be reasonable precise.
The estimation spreadsheet combines the presentation of estimated hull dimensions, weight,
stability and cost in a good and user-friendly way. With only three simple input parameters, the
estimations give relatively accurate data and may indicate that the spreadsheet will be used.
55
15. References
Arnesen, G., Aker Kværner; Dalane, J. I., Statoil; Aramanadka, S. S. B., Shell; Herfjord, K.,
Norsk Hydro; Snell, R., BP; Stansberg, C. T., Marintek (2006). Integrated Semi and Steel
Catenary Risers (SCRs) in Deep Water and Harsh Environment Conditions. OTC 18259. OTC,
Texas, USA.
DNV (2011). Modelling and Analysis of Marine Operations. Recommended practice DNV-RP-
H103. DNV-GL, Oslo, Norway.
DNV (2012). DeepC - Improved Confidence in Deep Water Concepts. DNV Software Seminar,
pp. 5. DNV-GL, Oslo, Norway.
Gudmestad, O. T. (2014). Lecture material – Stability. OFF580 Marine Technology and Design.
University of Stavanger, Stavanger, Norway.
Michael, W. H. (1999). Sea Spectra Revisited. Marine Technology, Vol. 36, No. 4, pp 211-227.
56
NORSOK (2007). Actions and Action Effects. NORSOK Standard N-003, Rev. 2. Norsk
Standard, Lysaker, Norway.
Nygård, M., Aker Kværner (2003). Column Structure Design. OGP Marine Risks Workshop
Proceedings, No. 344, pp. 59 - 79. IOGP, London, United Kingdom.
Swail, V. R.; Ceccacci, E. A.; Cox, A. T. (2000). The AES40 North Atlantic Wave Reanalysis:
Validation and Climate Assessment. 6th International Workshop on Wave Hindcasting and
Forecasting. Monterey, California, USA.
Wood Group Mustang (2011). Worldwide survey of Semi-FPSs and FPUs. Link:
http://www.mustangeng.com/NewsandIndustryEvents/Publications/Publications/2011%20WW%
20Semi%20Production%20Survey.pdf. Link verified 9th June 2015.
57
Appendix A
A
Table 14 – Amplitude response for case 7 – 10.
B
Appendix B
C
Table 16 – Detailed response analysis for default topside weight.
D
Appendix C
Air gap analysis – normal draft
Table 18 – Response analysis for 80 000 tonnes topside weight.
Wave spectrum Standard Dev. Duration Number of Response Level Cor. Response Level
(m) (s) frequencies (m) (m)
Gulf of Mexico 3,9 10800 1144 16,7 18,4
Mid-Atlantic 2,9 10800 1165 12,4 13,7
Norwegian Sea 4,0 10800 1085 17,0 18,7
West Africa 1,2 10800 1129 5,1 5,6
Gulf of Mexico 4,5 10800 1136 19,4 21,4
Mid-Atlantic 3,3 10800 1161 14,3 15,7
Norwegian Sea 4,7 10800 1059 20,4 22,4
West Africa 1,4 10800 1114 6,0 6,6
Gulf of Mexico 4,5 10800 1008 19,2 21,1
Mid-Atlantic 3,3 10800 1025 14,1 15,5
Norwegian Sea 4,6 10800 972 19,7 21,6
West Africa 1,4 10800 997 5,9 6,5
Gulf of Mexico 3,5 10800 941 15,1 16,6
Mid-Atlantic 2,6 10800 964 10,9 12,0
Norwegian Sea 3,9 10800 871 16,5 18,2
West Africa 1,1 10800 921 4,7 5,2
E
Table 19 - Response analysis for default topside weight.
Wave spectrum Standard Dev. Duration Number of Response Level Cor. Response Level
(m) (s) frequencies (m) (m)
Gulf of Mexico 3,4 10800 1084 14,6 16,0
Mid-Atlantic 2,5 10800 1101 10,8 11,9
Norwegian Sea 3,7 10800 994 15,7 17,3
West Africa 1,0 10800 1065 4,5 4,9
Gulf of Mexico 3,9 10800 1090 16,9 18,5
Mid-Atlantic 2,9 10800 1118 12,3 13,5
Norwegian Sea 4,4 10800 976 18,7 20,6
West Africa 1,2 10800 1062 5,2 5,8
Gulf of Mexico 4,1 10800 1015 17,5 19,3
Mid-Atlantic 3,0 10800 1029 12,9 14,1
Norwegian Sea 4,3 10800 959 18,4 20,3
West Africa 1,3 10800 1002 5,4 5,9
Gulf of Mexico 3,8 10800 1066 16,3 17,9
Mid-Atlantic 2,8 10800 1092 11,9 13,1
Norwegian Sea 4,3 10800 956 18,2 20,0
West Africa 1,2 10800 1040 5,1 5,6
Wave spectrum Standard Dev. Duration Number of Response Level Cor. Response Level
(m) (s) frequencies (m) (m)
Gulf of Mexico 3,0 10800 1236 12,9 14,2
Mid-Atlantic 2,2 10800 1270 9,4 10,4
Norwegian Sea 3,3 10800 1102 14,1 15,5
West Africa 0,9 10800 1202 4,0 4,4
Gulf of Mexico 3,2 10800 1191 13,8 15,2
Mid-Atlantic 2,3 10800 1234 10,0 11,0
Norwegian Sea 3,6 10800 1043 15,6 17,1
West Africa 1,0 10800 1150 4,3 4,8
Gulf of Mexico 3,3 10800 1159 14,1 15,5
Mid-Atlantic 2,4 10800 1183 10,4 11,4
Norwegian Sea 3,5 10800 1063 15,1 16,6
West Africa 1,0 10800 1136 4,3 4,8
Gulf of Mexico 2,9 10800 1059 12,4 13,7
Mid-Atlantic 2,1 10800 1097 8,9 9,8
Norwegian Sea 3,4 10800 927 14,4 15,9
West Africa 0,9 10800 1022 3,9 4,3
F
Air gap analysis – deep draft
Table 21 – Response analysis for deep draft default tonnes topside weight.
Wave spectrum Standard Dev. Duration Number of Response Level Cor. Response Level
(m) (s) frequencies (m) (m)
Gulf of Mexico 3,8 10800 1056 16,4 18,1
Mid-Atlantic 2,8 10800 1089 11,9 13,1
Norwegian Sea 4,7 10800 889 20,0 22,0
West Africa 1,2 10800 1018 5,2 5,7
Gulf of Mexico 4,0 10800 996 17,2 18,9
Mid-Atlantic 2,8 10800 1045 12,1 13,3
Norwegian Sea 5,2 10800 825 21,8 24,0
West Africa 1,3 10800 951 5,5 6,1
Gulf of Mexico 3,8 10800 970 16,0 17,6
Mid-Atlantic 2,7 10800 1000 11,5 12,7
Norwegian Sea 4,6 10800 833 19,6 21,5
West Africa 1,2 10800 939 5,1 5,6
Gulf of Mexico 4,0 10800 993 17,0 18,7
Mid-Atlantic 2,8 10800 1041 12,0 13,1
Norwegian Sea 5,1 10800 822 21,6 23,8
West Africa 1,3 10800 949 5,5 6,0
Table 22 - Response analysis for deep draft 20 000 tonnes topside weight.
Wave spectrum Standard Dev. Duration Number of Response Level Cor. Response Level
(m) (s) frequencies (m) (m)
Gulf of Mexico 3,5 10800 1090 15,0 16,4
Mid-Atlantic 2,5 10800 1140 10,6 11,7
Norwegian Sea 4,3 10800 906 18,5 20,3
West Africa 1,1 10800 1041 4,8 5,3
Gulf of Mexico 3,6 10800 1040 15,6 17,1
Mid-Atlantic 2,5 10800 1096 10,9 12,0
Norwegian Sea 4,6 10800 862 19,5 21,5
West Africa 1,2 10800 991 5,0 5,5
Gulf of Mexico 3,5 10800 1048 14,9 16,4
Mid-Atlantic 2,5 10800 1095 10,6 11,6
Norwegian Sea 4,3 10800 881 18,3 20,1
West Africa 1,1 10800 1005 4,7 5,2
Gulf of Mexico 3,6 10800 1018 15,2 16,8
Mid-Atlantic 2,5 10800 1072 10,7 11,7
Norwegian Sea 4,5 10800 844 19,3 21,2
West Africa 1,2 10800 970 4,9 5,4
G
Appendix D