Semi-Submersible Design - A New Generation of Offshore Accommodation

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/329962883
Semi-submersible Design - a New Generation of Offshore Accommodation

Vessels
Article in Ocean Engineering · January 2019

DOI: 10.1016/j.oceaneng.2018.11.053
CITATIONS READS
4 866
3 authors, including:
James R MacGregor
JRMA Ltd
23 PUBLICATIONS 30 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Historic Ships & Boats View project
Floating Production Vessels View project
All content following this page was uploaded by James R MacGregor on 22 December 2020.
The user has requested enhancement of the downloaded file.

Ocean Engineering 172 (2019) 759–787
Contents lists available at ScienceDirect
Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng
Semi-submersible design – A new generation of offshore accommodation T

vessels
James MacGregor∗, Savan Mayekar, Duncan Watson
Prosafe Offshore, Greenwell Road, Aberdeen, AB12 3AX, UK
ARTICLE INFO ABSTRACT
Keywords: This paper explores the naval architectural design considerations involved in a new fleet of offshore accom-
Semi-submersible modation semi-submersibles constructed by different owners in the period from 2008 to 2018. The principal
Floatel design features are compared with data from the previous generation of floatels constructed in the 1970s/80s.
Naval Architecture All fourteen vessels are compared at a high level, and in some areas detailed data from six of the new vessels
Newbuild
(three different design classes) is analysed. Information is provided on subjects such as weight estimation and
Accommodation
Weight
weight control, motions and station-keeping behaviour, resistance and propulsion. Key arrangement con-
siderations are discussed together with practical design data on internal volume and deck area.
1. Introduction Tamaki et al., 1984), and demonstrated an ability to remain safely

moored close alongside the client platform. The personnel bridges
1.1. History connecting the floatel with the platform were further developed at this
time, and resulted in a hydraulic telescoping gangway which (together
When the oil and gas industry embarked on the construction of large with the good motions of the semi-submersible design) allowed con-
offshore production platforms in the harsh environment of the North nectivity to be maintained under rough weather conditions. In extreme
Sea during the 1970s, there was a pressing need for temporary ac- weather, the gangway would be lifted and the floatel winched away to a
commodation vessels to support the large numbers of workers engaged stand-off distance until the weather moderated.
in “hooking up” and interconnecting the individual topsides modules Following the decline of large scale North Sea platform construction
which had been constructed onshore and lifted onto the platform by and hook up campaigns, some of these North Sea floatels were con-
crane barges. verted in the 1990s into drilling units (Ren et al., 2001) or floating
At that time, the available offshore crane capacity was limited, production vessels, or migrated to accommodation duty in the Gulf of
which kept the size of individual platform modules small, meaning that Mexico. Brazil also developed as a market for floating offshore ac-
the required offshore hook up labour force was large, and the hook up commodation solutions. No new floatels were built for nearly two
durations long. Initial construction operations managed by using decades.
available monohull barges, but the difficulties were considerable.
During the late 1970s and early 1980s a substantial fleet of semi-sub- 1.2. New designs
mersible vessels was built to service this demand.
That generation of floatels derived from semi-submersibles designed In the period 2003–2008, some vessel owners began to consider the
in the 1970s for midwater drilling operations in harsh environments construction of a new generation of semi-submersible accommodation
such as the North Sea, using spread mooring for station-keeping. The vessels. Reasons included the advancing age of the survivors from the
fleet was composed primarily of Aker H3 and Friede & Goldman original fleet (mostly aged 20–30 years by that time) and the perceived
Pacesetter designs featuring 6 or 8 slender columns, and multiple bra- need for support of ongoing maintenance campaigns on the (also
cings. Notable names included the Treasure Finder, Safe Supporter, Port ageing) platform infrastructure in the North Sea and elsewhere (Lamas
Royal and Polyconfidence. These units had operating displacements in and Carral, 2011).
the region of 20,000 tonnes. Fourteen of the vessels constructed as a result of this investment are
These vessels delivered excellent motion behaviour (Burnell, 1975; listed in Table 1.
∗
Corresponding author. Tel.: +44 7967 227 016.
E-mail address: [email protected] (J. MacGregor).
https://doi.org/10.1016/j.oceaneng.2018.11.053
Received 2 September 2018; Received in revised form 25 October 2018; Accepted 25 November 2018
0029-8018/ © 2018 Elsevier Ltd. All rights reserved.
J. MacGregor et al. Ocean Engineering 172 (2019) 759–787
Table 1
New semi-submersible accommodation vessels constructed from 2008 to 2018.
Vessel Floatel Floatel Superior, Floatel Victory, Safe Boreas, Neptuno, Safe Notos, Safe Safe Nova, OOS Tiradentes
Reliance Floatel Endurance Floatel Triumph Safe Zephyrus Atlantis Eurus Safe Vega
Year of Build 2010 2010 2013 2015 2015 2016 2017 2018
2015 2015 2015 2016 2017
Ordered by Floatel Floatel Floatel Prosafe Cotemar Prosafe Axis C. Helios (OOS)
Design Gusto & KFELS Gusto & KFELS Gusto & KFELS GVA GustoMSC GustoMSC GM 500A CR600
SSAU 3600 DSS20/NS SSAU 3000E Ocean500 Ocean500
400NG
Yard Keppel Keppel Keppel Jurong COSCO Qidong COSCO Qidong COSCO CIMC Raffles
Qidong
No. of Beds 500 440 560 500 450 750 500 456 600
Regulatory RoW Norway & RoW RoW & UK Norway & RoW RoW (and UK & RoW Aim at RoW (and
compliance possibly UK) Norway & possibly UK)
RoW
The table characterises the new vessels by year of build, accom- As such, they represent a considerable departure in terms of both
modation capacity, design type and ability to operate in the UK, size and style compared to the previous generation. This paper explores
Norway, or the rest of the world (RoW). The latter is an important the naval architectural design considerations involved in this new fleet
differentiator, as vessels aimed at the North Sea market must be de- of vessels, compared with the features of the preceding generation. All
signed for the most onerous metocean environment, and thus require fourteen vessels are compared at a high level, and in some areas a
more strength, air gap, mooring capacity and motion performance than deeper evaluation is carried out using data from six of the vessels (three
vessels designed for the rest of the world. classes, illustrated in Figs. 1–4). These are designs/projects with which
Furthermore, in Norway and the UK the vessels are considered to be the present writer(s) were more closely involved.
offshore installations, and must therefore comply with the same set of It is hoped that this information may provide guidance for future
shelf state laws and regulations as the hydrocarbon production and designers and operators and supplement the sparse public domain lit-
drilling installations. This adds a layer of certification, technical cap- erature on the preliminary design of semi-submersibles (Penney and
ability and cost beyond that that required for conventional marine Riiser, 1985; Gallala, 2013).
vessels engaged in the oil and gas industry. Norwegian requirements are
highest (freefall lifeboats, all single cabins etc), and a vessel certified for 1.3. Origins of new semi-submersible designs
work in Norway (subject to PSA and NMA regulations) will generally be
capable of certification in any other jurisdiction. A vessel capable of The Floatel Reliance is very similar to the floatel Safe Concordia
operating on the UK continental shelf (regulated by UK Health & Safety which was built by Keppel in 2005. These vessels are targeted for floatel
Executive, Energy Division) should be capable of operating everywhere duty in moderate environments such as Brazil and the Gulf of Mexico.
else, but not necessarily in Norway. The Concordia and Reliance SSAU 3600 design was derived from a
The designs of many of the modern generation of floatels are de- Keppel design (SSDT 3600) for a series of tender assist drilling vessels,
rived from semi-submersible designs intended for drilling operations in originally developed for operations in moderate sea states around SE
deep water and moderate environments, where dynamic positioning Asia. Some signs of these origins can be seen in the modest air gap,
(rather than mooring) and a high variable load capacity are prominent relatively small galley, DP2 rather than DP3 station-keeping, and a
design requirements. The new floatel designs typically have operating relatively high length/beam ratio compared to the more “square”
displacements around 30,000 tonnes, four (sometimes six) large col- layout of other semi-submersibles (tender assist rigs of this type require
umns with minimal bracing structures, and an enclosed deck box longitudinal stability for crane transfer of the drilling package to the
structure (one or two decks high) capable of providing the necessary adjacent client platform).
strength and accommodating the larger power generation systems. The Floatel Victory and the Floatel Triumph are in turn developed
Fig. 1. Safe Zephyrus – GVA 3000E design, constructed by Jurong Shipyard, Fig. 2. Safe Notos – GustoMSC Ocean 500 design, constructed by COSCO
Singapore. Qidong, China.
760
COSCO for the drilling/intervention vessel Island Innovator (delivered in

2012), and also used at CIMC Raffles as the basis for the COSL Pioneer
series of drilling rigs (ordered in 2005).
The OOS Tiradentes appears to be an in-house design from CIMC
Raffles, probably heavily based on their experience gained with the
Global Maritime and Agility design work for the COSL drill rigs.
1.4. Other competitor vessels
Not discussed in this paper are four other new vessels also engaged
in the offshore accommodation industry and competing with the “pure”
accommodation semi-submersibles. These are the OOS Gretha and OOS
Prometheus, built by CIMC Raffles, and the POSH Arcadia and POSH
Xanadu.
Although the OOS vessels do have twin pontoon semi-submersible
configuration, their designs are heavily influenced by the large con-
struction cranes onboard. The OOS Gretha is a 55,700t DP vessel with
no bracings but an asymmetric pontoon arrangement designed to sup-
Fig. 3. Axis Vega (now Safe Vega) – GM500A design, constructed by COSCO port two 1800t cranes. The OOS Prometheus is a smaller anchored vessel
Qidong, China (note alternative gangway pedestal at corner). with one 1100t crane. The two 47,000t displacement POSH vessels do
not have the typical twin pontoon configuration (the lower submerged
hull is a single “barge” type structure) and were originally designed
(and partly built) for purposes other than accommodation.
Also not discussed are the other types of vessel with which the semi-
submersibles must now compete. In the last 10 years a number of
monohulls have been built or converted to use dynamic positioning for
floatel operations (see Table 3b of Lamas and Carral, 2011). These have
been used offshore Australia, Brazil and Mexico, but the North Sea
weather seems to preclude their year round use in that market.
On the other hand, the dedicated accommodation semi-sub-
mersibles in the North Sea have seen a threat emerge from so called
“walk to work” DP monohulls which provide offshore accommodation
and transfer via a gangway but avoid regulation as offshore installa-
tions by virtue of not remaining permanently connected to the client
platform.
Also, several twin hull units, based on the SWATH concept and
marketed as Compact Semi-Submersibles (CSS) have managed to obtain
floatel contracts in Brazil, operating on DP (see description in Lamas
and Carral, 2011). These are much smaller than the conventional semi-
Fig. 4. Atlantis at work in Gulf of Mexico.
submersibles (84 m length, 32 m beam, 20.8 m depth, with displace-
ment around 12,000 tonnes) and consequently less expensive to build
from the Floatel Reliance, and intended to be capable of operations on and operate. Other competitors in certain markets include monohull
the UK continental shelf, at least for part of the year. The basic six barges and the jack-up type of accommodation unit, as described in
column hull configuration is retained, but the displacement and air gap Table 2 of Lamas and Carral (2011).
is increased. Their ability to deal with harsher environments is further
enhanced by introduction of 10 point chain mooring systems, and a 2. Size and main dimensions
move from 4 to 6 thrusters.
The Floatel Superior/Endurance are totally different to the Reliance/ 2.1. Weights, volumes - draft and air gap
Victory/Triumph, and are a DSS20™ design from Keppel/GustoMSC
closely related to deep-water semi-submersible drilling rigs previously Semi-submersibles are effectively constant draft designs, with a
built by Keppel such as the DSS20 Maersk Explorer (2003), the DSS21 small number of approved drafts at which they may be operated.
Maersk Developer series (2008 onwards) and the DSS38 design Gold Star Table 2a lists the displacements at operating and survival draft for the
(2008). These two vessels are designed to meet Norwegian require- new designs, together with the Tonnage according to the 1969 IMO
ments. Convention. The latter is a measure of the internal volume of the vessel.
The Ocean 500 hull design used for the Neptuno/Atlantis and Safe Table 2b gives corresponding data for a sample of three older vessels.
Notos/Eurus was developed by GustoMSC. It is related to designs which With operating displacements in excess of 30,000 tonnes, most of
GustoMSC had previously developed for the deep-water drilling market, the new generation of vessels are some 50% heavier than the older
and has similarities to the Floatel Superior/Endurance designs. generation.
The Safe Boreas basic design was prepared by GVA several years A critical feature for any semi-submersible is the air gap between
before the present building sequence commenced. This GVA3000E de- the still waterline, and the underside of the deck structure. There is a
sign is an enhancement of the previous GVA3000 Regalia vessel (de- close relationship between this dimension (especially for survival draft)
livered 1985), which became employed as a floatel after an early career and the environment in which the vessel can be safely operated when
as a dynamically positioned diving support and offshore construction moored (when dynamically positioned, a vessel has the option to seek
vessel. shelter in extreme conditions). Fig. 5 illustrates the contours of the 100
The Safe Nova and Vega were ordered by Axis at COSCO and make year return wave height for the North Sea region. It can be seen that the
use of a Global Maritime basic hull configuration previously used by northern regions have significant wave heights which can exceed 15 m,
761
Table 2a
Main dimensions, displacements and internal volume (new floatels).
Length overalla (on pontoons) (m) Breadth overalla (at pontoons) Displacement (tonnes) Tonnage ITC, 1969
(m)
Operating draft Survival draft Gross Net
Floatel Reliance 109.0 (93.9) 68.0 (45.2) 17,920 15,240 18,246 5,527
Floatel Superior 97.3 (90.0) 94.5 (64.5) 29,180 25,600 27,920 8,377
Floatel Endurance 105.0 (93.5) 103.0 (64.5) 30,860c 26,850c 30,803 9,241
33,594d 29,030d
Floatel Victory, Floatel Triumph 119.6 (98.1) 78.0 (53.4) 26,800b 22,700b 24,594 7,482
27,211 8,163
Safe Boreas, Safe Zephyrus 104.2 (97.5) 91.25 (71.25) 33,930 31,630 34,076 10,412
34,264 10,280
Neptuno, Atlantis 98.0 (95.0) 92.5 (67.0) 33,400 28,400 28,808 8,643
Safe Notos, Safe Eurus 98.0 (95.0) 92.5 (67.0) 33,400 28,400 30,278 9,084
Safe Nova, Safe Vega 112.2 (104.5) 100.6 (66.3) 37,700 36,020 37,140 11,140
OOS Tiradentes 106.45 (100.8) c. 95 (68.9) 35,494 – 38,434 11,530
a
Includes lifeboats and helideck.
b
Effect of Floatel Triumph column blisters not included.
c
Includes longer pontoons compared to Floatel Superior.
d
Includes longer pontoons and column blisters.
Table 2b
Main dimensions, displacements and internal volume (old floatels).
Length overall (on pontoons) (m) Breadth overall (at pontoons) Displacement (tonnes) Tonnage ITC, 1969
(m)
Operating draft Survival draft Gross Net
Regalia (1985) 95.4 (87.6) 91.6 (60.83) 21,030 19,260 17,624 –

Safe Caledonia (1982) 91.3 (79.56) 69.7 (61.38) 20,470 17,985 19,230 5,770
Safe Scandinavia (1984) 106.3 (92.45) 93.9 (68.4) 28,315 25,777 25,383 7,615
and vessels will need air gaps of similar dimensions for safe operations. increase over the older designs. Even more marked is the relative lack of
Table 3 contains the key air gap parameters for the new vessels. increase in depth to underside of deck box. The new vessels display
The air gap at operating draft is less than at survival draft, and figures in the range 24 m–29 m, compared to a range of 25 m–33 m for
classification society approval documentation and the vessel Operations the older vessels. This arises because the new deck box structures are
Manual will contain instructions as to the weather conditions in which generally deeper than the old ones, and the increase has been added in
a vessel must deballast from the operating draft to survival draft. This a downwards direction rather than upwards. This reduces air gap and/
will typically stipulate wave heights in the region of 5–6 m Hs. The or draft.
stability regulations dictate the windspeeds which must be considered The drafts tabulated are those to underside of the pontoon. When
at survival (51.5 m/s) and at operating draft (36 m/s). azimuth thruster heights are considered, actual drafts increase by some
Provision of a high air gap is expensive in design terms, as it raises 4.0–5.5 m. Harbour drafts (with vessel deballasted to pontoon draft)
the centre of gravity of the deck box and any variable load to be sup- can be in the range of 13 m–15 m, which can severely restrict the lo-
ported above the upper deck. There is therefore pressure to minimise cations where the larger units can be berthed. There are very few lo-
the air gap, especially for designs with large and heavy deck boxes. This cations on the UK side of the North Sea where such semi-submersibles
is generally not a problem for deep-water drilling semi-submersibles can be brought alongside without having thrusters removed.
which mostly work in moderate environments, but can be an issue for
harsh environment vessels. 2.2. Dimensions of primary structural components
The “tallest” of the new floatels have survival draft air gaps in the
region of 14 m. This compares with values of more than 15.5 m for A twin pontoon semi-submersible structure is effectively “tied to-
older generation vessels such as the Safe Caledonia and Safe Scandinavia. gether” to resist wave splitting forces and racking moments by a com-
The design assumption for the earlier units was that sufficient air gap bination of the deck structure and the tubular bracings connecting the
would be provided so that no wave contact with the deck structure lower hulls. Redundancy in the structural design is required so that
would occur. The modern generation has been designed to allow some progressive collapse will not occur following loss of effectiveness of any
wave impact, provided that the structures are reinforced locally to deal one of these structural members. Such a failure mode was instrumental
with the associated slamming loads. Extensive model testing and nu- in the 1980 capsizing of the floatel Alexander Kielland.
merical analysis is required to examine phenomena such as wave run- None of the new generation employs vertical bracings to support the
up at the columns and other slamming effects. In late 2015 a fatality deck structure. The upper structure of many of the previous vessels was
occurred on a modern design drilling semi-submersible offshore based on a network of transverse and longitudinal girders, simply
Norway (PSA, 2016) when a wave hit the face of the deck box and plated over. The new vessels employ an enclosed deck box where the
several windows failed. This called the recent industry design practice shell plate and decks (together with internal bulkheads) play a sig-
for air gap into question, and new guidance has been published (Pessoa nificant role in strength, and the enclosed volume provides reserve
and Moe, 2017; Szalewski et al., 2017). However, the full implications buoyancy for stability. These features are summarised in Tables 4 and 5.
of this are likely to take several years to resolve. The modern generation of floatel semi-submersibles utilises fewer
Despite the overall growth in size of the new vessels, the typical bracings than 1970s designs such as the Aker H3, Sedco 700 and F&G
depth to main deck of around 35 m does not represent a significant Pacesetter. In the case of the four column designs, two transverse
762
Table 3a
Draft and air gap (new floatels).
Draft – operating Draft – survival Depth to underside Depth to main Air gap – operating Air gap – survival draft
(m) (m) deck box (m) deck (m) draft (m) (m)
Floatel Reliance 12.2 7.9 18.9 24.69 6.7 11.0

Floatel Superior, Floatel Endurance 18.0 13.0 27.5 35.2 9.5 14.5
Floatel Victory, Floatel Triumph 16.8 11.6 23.8 – 7.0 12.2
Safe Boreas, Safe Zephyrus 18.6 15.1 29.6 36.1 11.0 14.5
Neptuno, Atlantis 17.5 to 20.0 13.5 27.5 35.7 7.5–10.0 14.0
Safe Notos, Safe Eurus 17.5 to 20.0 13.5 27.5 35.7 7.5–10.0 14.0
Safe Nova, Safe Vega 17.75 15.75 29.25 37.45 11.5 13.5
OOS Tiradentes 17.75 to 20.0 14.25 to 15.75 28.25 36.25 10.5 to 8.25 14.0 to 12.5
bracings are used at each column pair, except for the Safe Boreas design
where only one bracing is used at each column pair. In that case, a high
degree of attention was paid to the stress and fatigue analyses and to
the construction quality and inspection of the connections.
Bracings are generally located above the pontoon tops to avoid drag
and slamming during transit, but should be sufficiently low so that
emergence and re-entry slamming does not occur when the unit is at
survival draft in large waves. The Floatel Reliance, Victory and Triumph
designs have transverse bracings which are close to the pontoon tops,
which is good for structural efficiency, but not ideal for transit. Also,
due to the shorter columns, most of the new vessels transverse bracings
are much closer to the survival draft (only 2.0–3.5 m submergence)
than the older generation, where submergences of 5.3 m and 6.0 m (e.g.
Pacesetter design Safe Caledonia and Aker H3.2 Safe Scandinavia) were
common.
With the exception of the Safe Boreas design, the larger examples of
the new generation have two working decks in the deck box. This gives
a length/depth ratio for these structures of around 8, compared to more
than 12 for the much more slender Safe Boreas deck concept. Despite
this, care is required with regard to the placing of openings in the
transverse and longitudinal bulkheads in these deck boxes to avoid
unnecessary stress concentrations.
The deck boxes have a double bottom of around 1.5 m height. In the
DP3 vessels, this space is used to arrange segregated routes for dis- Fig. 5. 100 Year extreme (winter) significant wave heights in North Sea (taken
tribution of cables and cooling water pipes between the three ma- from UK HSE report RR392).
chinery spaces in the deck box and the three pairs of thruster spaces in
the pontoons (ref. Fig. 22).
include the Floatel Reliance and the four units based on an Ocean 500
Note that in the case of the Floatel Reliance/Victory/Triumph, the
design (Neptuno, Atlantis, Safe Notos, Safe Eurus).
main deck is officially (i.e. for freeboard and stability calculation pur-
The changes may be summarised as follows;
poses) taken to be at the top of the double bottom rather than the up-
permost deck. For practical comparison in Table 5 these deck boxes are
considered to extend vertically to the upper deck. Furthermore, the • Safe Boreas and Safe Zephyrus had pontoon height increased by 9%
during design phase (prior to construction), which increased oper-
deckhouse erected on top of this double bottom extends transversely
ating displacement by some 3000t (9.7%)
only to the column centrelines, giving an aspect ratio of more than 2,
compared to the approximately square plan view of the other designs. • Floatel Triumph had blisters added to outboard sides of 2 midships
columns prior to delivery (not present on predecessor Floatel Victory)
• Floatel Endurance was designed with longer pontoons than pre-
decessor Floatel Superior, and then had a triangular (plan view)
2.3. Hullform modifications
blister added to each column prior to delivery (these are not present
on Floatel Superior), which further increased displacement by some
Nearly half of the new vessels have experienced some form of hull
2700t (8.8%) at operating draft
form modification during the design/building programme. Exceptions
Table 3b
Draft and air gap (old floatels).
Draft – operating (m) Draft – survival Depth to underside deck box Depth to main deck Air gap – operating draft Air gap – survival draft
(m) (m) (m) (m) (m)
Regalia (1985) 17.5 14.0 25.5 31.0 8.0 11.5

Safe Caledonia (1982) 18.29 13.72 29.33 33.83 11.04 15.61
Safe Scandinavia (1984) 22.0 17.5 34.5 (33.3 to girders) 37.72 11.3 15.8
763
Table 4
Numbers of columns and structural bracings.
No. of columns No. of transverse bracings Transverse bracing centre height - relation to survival draft
Floatel Reliance 6 3 6.09 m (pontoon top), 1.83m below survival draft

Floatel Superior, Floatel Endurance 4 4 11.11 m, 1.89 m below survival draft
Floatel Victory, Floatel Triumph 6 3 Just above pontoon top
Safe Boreas, Safe Zephyrus 4 2 12.0 m, 3.1 m below survival draft
Neptuno, Atlantis 4 4 11.5 m, 2.0 m below survival draft
Safe Notos, Safe Eurus 4 4 11.5 m, 2.0 m below survival draft
Safe Nova, Safe Vega 4 4 12.25 m, 3.5 m below survival draft
OOS Tiradentes 4 4 Not available
• Safe Nova and Safe Vega had sponsons added on inboard side of the new designs are typically 5 decks high, with a navigating bridge or
pontoons and blisters fitted to fore and aft column faces during control room located above.
construction, adding c. 5000t (15.6%) to operating displacement On the four Ocean 500 based designs, the control room is located at
the side of the vessel near to the gangway, but also with a view towards
Many newbuild drilling semi-submersibles have also experienced the sailing direction. On all the other designs, the navigating bridge and
similar problems during the design and build process. These modifica- DP bridge is located at the vessel centreline.
tions are generally implemented in order to improve the carrying ca-
pacity or stability of the units (e.g. to ensure that pontoon tops are not 3.3. Location of gangway relative to exhaust and living quarters
constantly submerged). They generally have unfavourable effects on
station-keeping, motions and transit performance. The personnel gangway connecting the vessel to the client platform
is the primary operational tool of a floatel, and the selected position
3. General arrangement dictates many aspects of the floatel interaction with the client platform
and also the internal arrangements of the floatel.
3.1. Overview Table 6 summarises the locations of the gangway on the new floa-
tels, with specific reference to other arrangement features such as the
All of the new designs share the following arrangement features; engine exhaust location and the living quarters. It can be seen that no
common theme has emerged amongst the new designs.
• Two slender lower hulls (pontoons), containing thruster rooms, Most of the new designs typically have the heavy generator sets
pump rooms, tanks for ballast, fuel and water and access trunk (one located at one end of the vessel. This is likely due to the parent drilling
example is illustrated in Fig. 6) designs, where this weight would be balanced by heavy mud pumps and
• Four or six vertical columns, containing access routes (elevators and mud pits at the opposite end. In the floatels the mud system weight is
stairways), auxiliary machinery spaces and smaller tanks and voids replaced by the weight from the large living quarters. The exhausts
(example is illustrated in Fig. 7) from these engines are generally arranged to discharge downwards, and
• Deck box (upper hull) containing one or two decks of machinery the fumes can cause unpleasant problems when drawn into the HVAC
spaces, workshops and store rooms, and elements of the living system of the client platform if that is nearby. Unfortunately, many of
quarters (examples are given in Fig. 8a and b) the floatel designs have chosen to place the gangway (and thus the
• Main/upper deck – of which approximately 50% is consumed by the client platform) at the vessel end near the engines and not near the LQ.
living quarters block and the other half is dedicated to workshops This is likely because the alternative location at the end near the LQ has
and stores and open deck space available for laydown or other been dominated by lifeboat arrangement considerations.
purposes (example is shown in Fig. 9) An end location of the gangway typically means that a projecting
• Lifeboat stations, cranes and personnel transfer gangway arranged part of the floatel (i.e. the pontoon end) will be closer to contact with
around the periphery of the unit, in various configurations the client platform than the floatel deck structures. To maintain the
target separation distance, the gangway base is often cantilevered some
3.2. Deckhouses distance away from the main deck. In addition, the critical (governed
by the stroke limit) longitudinal axis of the gangway is subjected to
Above the uppermost deck, two substantial deckhouses are erected; potentially large vessel motions in the surge axis. These issues do not
one for the living quarters at one end, and stores and one for work- exist with a side mounted gangway.
shops/stores/auxiliary spaces at the other. The living quarters blocks in The Floatel Reliance/Victory/Triumph designs have their engines
Table 5
Upper hull deck box characteristics.
Length (m) Breadth (m) LxB (m2) No. decks Deck Height Double Bottom Deck Box Height Deck box volume
(m) Height (m) (m) (m3)
Floatel Reliance 88.7 36.07 3200 1 4.57 1.22 5.79 18,524

Floatel Superior, Floatel Endurance 63.5 64.5 4096 2 3.25 1.2 7.7 31,537
Floatel Victory, Floatel Triumph – – – 1 – 1.5 – –
Safe Boreas, Safe Zephyrus 82.5 71.25 5900 1 5.0 1.5 6.5 38,208
Neptuno, Atlantis 66.0 67.0 4420 2 3.3 1.6 8.2 36,260
Safe Notos, Safe Eurus 66.0 67.0 4420 2 3.3 1.6 8.2 36,260
Safe Nova, Safe Vega 80.7 66.3 5350 2 3.3 1.6 8.2 43,870
OOS Tiradentes 83.6 67.6 4900 2 3.3 1.4 8.0 45,210
764
Fig. 6. Typical pontoon arrangement (Safe Boreas).
Fig. 7. Typical Column Arrangements (Safe Nova, forward starboard column).
765
Fig. 8a. Typical deck box arrangement (lower deck – Safe Notos).
amidships (a legacy of the parent tender assist drilling rig design). This direct visual contact from the control room (bridge)
allows for high exhaust stacks arranged port and starboard, which • Gangway location is reasonably far from engine exhausts
avoids fumes near the end corner mounted gangway, but does mean • Hydrodynamic forces pushing vessel onto the platform act in sway
that the living quarters must be insulated from noise and vibration from direction, and for the new hull designs this produces less motions
the nearby engines. than forces in surge direction, meaning better gangway operability
Some designs have been delivered with two pedestals for placement • Gangway does not interfere with normal crane operations, but can
of the gangway, but normally only one gangway is installed. still be reached by crane for gangway pedestal height adjustment
The Safe Notos/Eurus arrangement with a gangway mounted at the operations
side amidships (Figs. 2 and 9) reflects that of the older Safe Caledonia
(originally built as Safe Holmia in 1982), and has the following ad- With the benefit of hindsight, it is possible that the best possible
vantages; arrangement has not been provided on any of the new vessels. An ar-
rangement with the gangway located at a vessel corner near the LQ (not
• Gangway is near to the LQ, giving efficient access, and also allowing at the end of the vessel where the engines and exhausts are located)
766
Fig. 8b. Typical deck box arrangement (Tween deck – Safe Notos).
may provide the best balance of advantages and disadvantages. main deck are given in Fig. 8a and b and 9.
The corner position would allow the floatel to choose whether to In the upper hull, the primary considerations of layout involve the
operate beam-on or end-on to the platform, thus maximising operability following;
under varying metocean conditions. Proximity to the LQ would address
two concerns; ease of personnel access to living spaces, and avoiding • Easy access from gangway stair tower to client office areas
engine exhaust fumes interfering with the client platform and the • Easy access from gangway stair tower to workforce locker rooms
gangway. One disadvantage with such an arrangement could be diffi- and changing rooms
culty in arranging coverage by a deck crane for assistance in gangway • Access from locker/changing rooms to cabin areas
height adjustment operations. • Access from cabin areas to eating/recreation areas
3.4. Internal arrangements - accommodation The deck area allocated to client offices, change rooms and re-
creation rooms on some prior floatels was analysed prior to specifying
Typical arrangements of the decks within the upper hull and on and contracting some of the new vessels described in this paper. The
767
Fig. 9. Typical main deck arrangement (safe Notos).
findings are summarised in Tables 7 and 8, where they are expressed in • Easy routes from stores area to galley, and from galley to serving
relation to the total personnel capacity of the vessel. In general, the area
more modern units have more office area, reflecting the increasing • Good personnel flow from living/access areas through serving area
proportion of the offshore workforce engaged in desktop tasks. The and into mess room
previous vessels had large communal recreation areas such as cinemas, • Good personnel flow from mess room past plate return and waste
which may not be so necessary in modern vessels where more privacy is disposal areas and back to living/access spaces
available in the cabins, and each cabin is enabled with internet and
personal TV facilities. The cinema can serve as a meeting point for Table 9 shows the results of an analysis of deck areas allocated to
management addresses, and as a muster station in emergency. stores, mess room and galley on prior floatels, compared with similar
With regard to the galley, mess room and food stores, primary ob- ratios for modern offshore construction monohulls. The metrics for the
jectives in design should be as follows; mess room and galley are similar for both vessel types. The area for
food stores is relatively smaller on the floatels, probably because the
• Provision of sufficient storage and handling area, near to main deck floatels are resupplied regularly, whereas the construction vessels must
hatch or stores laydown areas (preferably allowing a food container be more self-sufficient, often working in remote areas or making long
to be placed at same level as the food stores) ocean transits.
768
Table 7
Two alternative pedestal locations – operating with gangway at

Note change made to Endurance exhausts – also moved to P&S
Deck box is strengthened to permit gangway pedestal at corner
Two alternative pedestal locations – completed with gangway

Working areas - changing rooms, offices.
Vessel may operate parallel to or perpendicular to platform

Range (m2 per PoB) Average (m2 per PoB)
Offices for client 0.66 to 1.18 0.85

Changing, Locker Rooms 0.52 to 1.33 0.90
Laundry & Linen Rooms 0.20 to 0.49 0.30
Exhausts fitted with water spray system

Table 8
Recreation areas.
corners away from gangway
side mid-length position Range (m2 per PoB) Average (m2 per PoB)
on end near centreline

Recreation rooms 1.09 to 2.27 1.75
The numbers of cabins and beds on the different vessels are sum-
marised in Table 10. The Norwegian rated vessels have the greatest
Notes
number of cabins due to the single occupancy requirement. In most of

the vessels the cabins are arranged along long conventional alleyways,
Exhaust direction
but in the Safe Boreas/Zephyrus they are arranged around two large
atriums, which provide natural light to the mess rooms and the inner
cabins. The presence of the atriums is probably one of the reasons why
Down
Down
Down
Down
Down
Down
the Gross Tonnage figures (Table 2a) for the Boreas/Zephyrus are more
Up
Up
Up
than 10% greater than those for the generally comparable Floatel Su-
Vessel end – away from LQ – near
perior/Endurance.
Vessel end – away from LQ and
Amidships P&S – away from
Amidships P&S – away from
Vessel end – away from LQ
3.5. Machinery spaces

Engine exhaust location
All the larger vessels have three engine rooms arranged side by side
at one end of the deck box. In most of the vessels, personnel access to
these rooms is in the fore and aft direction from the adjacent switch-
gangway
gangway
gangway
gangway
gangway
gangway
gangway
board rooms, or transversely through watertight doors. In the Safe

Boreas design, a full width transverse alleyway is provided between the
engine rooms and the switchboard rooms, giving direct access to each
Normal gangway orientation relative to
engine room. This alleyway was not possible in the Ocean 500 designs
Selectable (if corner location used)
Selectable (if corner location used)
(and presumably the DSS20 designs) because of high stresses in the

longitudinal bulkheads at that location (high shear force in deck box
coinciding with location of column shell structure).
Table 11 summarises the engine room sizes and clearances for three
Parallel to surge axis
Parallel to surge axis
Parallel to sway axis
different designs with approx. 30 MW installed power. The engine room

areas given exclude fuel service and settling tanks, which may be lo-
floatel motion
cated on the other side of engine room bulkheads. The figures show
Selectable
Selectable
Selectable
there are some tight clearances around the engines in all cases, but
demonstrate that the inline engine configuration allows considerably
more space between the engines than the vee arrangement.
Another large consumer of deck area in passenger vessels is the
Side - mid length (near LQ) OR corner
End centre-line OR corner (both away
space required for the air conditioning equipment. This is often un-
End – centreline – away from LQ
End – centreline – away from LQ
derestimated in the basic design process, and consequently problems

arise when additional space has to be found during the detailed design
Side – mid length – near LQ
phase when the actual sizes of the necessary equipment becomes

Corner – away from LQ
known. A typical HVAC installation for the LQ spaces alone of a North

Gangway location
Sea unit could include 17 air handling units, 105 actuated fire/gas
(away from LQ)
dampers, 21/22 inlet/outlet louvres and 21 local control units.

Prior to the specification and contracting of some of the new floa-
from LQ)
tels, the deck area allocated to HVAC rooms (excluding chiller plant
rooms) on some prior vessels was analysed (Table 12). The results were
not significantly different to those found for other vessels types, with
approx. 5–6% of the total air conditioned deck area requiring to be
Floatel Victory, Floatel Triumph
devoted to space for the HVAC equipment rooms. When this ratio was
Safe Boreas, Safe Zephyrus
provided at the basic design stage it was found to be sufficient.

Floatel Superior, Floatel
Safe Notos, Safe Eurus
Safe Nova, Safe Vega

Gangway locations.
Neptuno, Atlantis
3.6. Watertight subdivision

Floatel Reliance
OOS Tiradentes
Endurance
In order to maintain compartments sufficiently small for damaged

Table 6
stability reasons, and at the same time comply with the practical re-
quirements for access, the vessels require extensive (and expensive)
769
Table 9
Mess room, galley and food stores.
Semi-submersible accommodation floatels Monohull offshore construction vessels
2 2
Range (m per PoB) Average (m per PoB) Range (m2 per PoB) Average (m2 per PoB)
Mess Room 0.44 to 1.28 0.87 0.46 to 1.20 0.79

Galley 0.41 to 0.56 0.43 0.35 to 0.70 0.45
Dry food stores 0.13 to 0.21 0.16 0.47 to 0.91 (dry & cold 0.70 (dry & cold)
Cold/freeze stores 0.11 to 0.31 0.21
Table 10
Cabin arrangements.
No. cabins No. single cabins 2 man cabins 4 man cabins Bed capacity Safety certificate
Floatel Reliance 171 – – – 500 500

Floatel Superior, Floatel Endurance 440 440 0 0 440 440
Floatel Victory, Floatel Triumph 300 40 260 0 560 560
300 100 200 0 500 500
Safe Boreas, Safe Zephyrus 450 450 0 0 450 450 (can be 490)
Neptuno, Atlantis 229 36 29 164 750 750 or more
Safe Notos, Safe Eurus 264 28 234 0 500 500
Safe Nova, Safe Vega 274 92 182 0 456 500
OOS Tiradentes – – – – – 600
installations of watertight (WT) sliding doors. For example, the Safe

Notos (GustoMSC design) has 40 sliding WT doors and 8 WT hatches Table 12
indicated on the control panel. The Safe Boreas (GVA 3000E design) has HVAC room deck area as a Percentage of air conditioned deck area.
50 sliding WT doors (18 in each pontoon, 4 in the deck box, others in Range Average
the columns), and 14 WT hatches.
Semi-submersible accommodation floatels 2.2–6.1% 4.7%
Other types of diesel electric offshore vessel 3.7–8.0% 6.0%
Cruise ships (Kanerva et al., 2000) 3.6–10.3% 5.8%
3.7. Tank arrangements
A major consumer of internal volume in the pontoons and columns

are the tanks for fuel, ballast and potable water. Although a typical 4. Weights and stability
number seems to be 36, the number of ballast tanks can vary sig-
nificantly between designs – with two examples having 32 and 62 4.1. Lightship weight and steel weight
tanks, respectively.
Table 13 lists the designed tank capacities of the new vessels. While The larger examples of the new generation of vessels (lightship
potable water capacities do not vary much (typically storage is pro- weights around 22,000 tonnes) are some 50% heavier than the older
vided for one week's consumption), there is a significant range in the generation which had lightship weights in the region of 14,000 tonnes.
fuel and ballast tank capacities as designed. Ballast capacity is typically From the inclining tests carried out on 7 of the new floatels, the
more than 50% of the displaced volume. As Fig. 6 shows, it is good lightship is typically around 63% of the operating displacement, as
practice to ensure the fuel tanks are located on the inboard sides of shown in Table 14. A more accurate basis from which to make an early
pontoons, with water ballast tanks outboard and at the ends of the estimate of the lightship weight is the Gross Registered Tonnage, with
pontoons. which a very good correlation is displayed (averaging 0.664 tonnes per
For a typical vessel with 2100 m3 fuel capacity (1785 tonnes) the GRT). Unfortunately, this figure is rarely calculated early in a design
endurance on DP will be in the region of 40 days if thruster power process, but could be estimated quickly from the intended geometry.
consumption averages 30% of MCR, giving total power demand of For the 7 new floatels, the actual vertical centre of gravity of the
around 9000 kW and 44 tonnes/day consumption. lightship is in the range 68%–76% of the depth to the main/upper deck.
Table 11
Engine rooms.
Engine numbers and type Engine room area (m2) Clearance engine skid to aft Clearance engine skid to forward Space between engines closest components
bulkhead (m) bulkhead (m) (m)
2 x 9L32 160 1.86 2.3 1.76

2 x 16V25 233a 1.7 1.3 0.82
2 x 16V26 164 (outboard) 1.0 3.0 1.05
184 (central)
a
Area includes aux. machinery space.
770
Table 13
Tank capacities.
Fuel storage Potable water Ballast water
3 3
Capacity (m ) Capacity per installed MWe Capacity (m ) Capacity per Days at 0.25t Capacity (m3) Related to Operating
generator power (m3) PoB (m3) per man-day Displacement (m3 per tonne)
Floatel Reliance 1600 102.6 960 1.92 7.7 7,600 0.424

Floatel Superior, Floatel 1800 81.5 800 1.81 7.3 16,360 0.561
Endurance 2450 86.9 800 – –
Floatel Victory, Floatel 1800 76.9 950 1.70 6.8 9,610 0.358
Triumph 1840 78.6 950 1.90 7.6 9,873 0.368
Safe Boreas, Safe Zephyrus 2100 69.7 920 2.04 8.2 15,045 0.443
Neptuno, Atlantis 3150 109.4 1000 1.33 5.3 14,475 0.433
Safe Notos, Safe Eurus 3150 109.4 900 1.80 7.2 15,350 0.460
Safe Nova, Safe Vega 2880 101 1190 2.61 10.4 15,750 0.418
OOS Tiradentes 3670 112 1550 2.58 10.3 14,570 –
The greatest contributor (about 60%) to the lightship weight is the

Table 15
structural steel weight. This is similar to the relationship observed for
Structural steel weight estimating relationships.
modern drilling semi-submersibles, where a sample of designs yields a
percentage in the range 55%–68% (average 62%). 7 of new generation floatels
For the larger vessels of the new generation, the structural steel
Range Average
weights are in the range of 12,000 to 17,000 tonnes. Table 15 sum-
marises norms derived from detailed estimates of steel weight and some Weight divided by Operating Displacement 0.34 to 0.45 0.382
other basis parameter. Again, the best basis for accurate early predic- Weight divided by Gross Tonnage 0.38 to 0.40 0.387
Weight divided by Estimated Lightship Weight 0.56 to 0.66 0.593
tion is the total enclosed volume, for which GRT is a proxy.
Weight divided by Actual Lightship Weight 0.57 to 0.68 0.609
Other weights vary depending on the scope of equipment and out-
fitting installed, and to some extent on the practices of the individual 1. Steel weight is structural steel weight including margin.
shipyards. Some ranges are given in Table 16 for the larger of the new 2. Estimated lightship weight is including margins.
generation vessels.
Major influences on the piping weight include choice of flanged or
sleeved pipe joints. For outfit steel, the result is influenced by the
shipyard design standards applied for ladders, gratings etc (tanker or Table 16
“offshore” standard may be too heavy). Estimated weights – including margin (tonnes) of Three designs.
Within the living quarters, a potential risk for weight and VCG Design 1 Design 2 Design 3
growth lies in the floor screed applied on top of the steel decks. Due to
Mechanical Equipment 2385 2605 3135
the large area and high elevation it is important to aim for a high
Piping 970 1560 1099
standard of flatness in these decks and control the pouring of the screed. Electrical Equipment, Bulks (cable, tray etc) 1046 1006 920
The cabins themselves may be fabricated in a “stick build” method HVAC Equipment and Ducting 591 500 613
using individual bulkhead and ceiling panels and associated toilet Living Quarters and Architectural Outfit 2920 1745 1780
module (wet unit) erected within the required footprint. Alternatively, Steel Outfit 2263 1853 1014
the cabin may be delivered as a single self-contained unit, installed in a

“plug and play” fashion with only a few connections to piping, cable
and HVAC services required. The weight of this should not be ignored,
Table 17
as single cabin modules of this type may weigh in the region of 1.2
Inclining experiment results compared to estimated weights and VCG for 4
tonnes each.
projects.
Parameter Source of Project 1 Project 2 Project 3 Project 4
4.2. Weight control during build Data
Lightship weight Lead ship 2.18% −4.37% −1.71% −1.81%

Due to their small waterplane area, semi-submersible projects are result Follow ship 1.56% −4.15% −1.74% –
notoriously sensitive to weight growth during design and build. Such VCG result Lead ship 0.29% 0.33% 3.21% 1.64%
adverse developments cannot be compensated for simply by operating Follow ship 1.33% 2.00% 2.63% –
LWT Vertical moment Lead ship 2.48% −4.06% 1.45% −0.20%
at a slightly deeper draft. Many newbuild projects leave the yards dis-
result Follow ship 2.91% −2.24% 0.84% –
playing visible blisters on the columns, often added at a late stage.
Table 14
Lightship weight and centre of gravity estimating relationships.
7 of new generation floatels Other semi-submersibles – pre 2012
Range Average Range Average
Lightship Weight Weight divided by Operating Displacement 0.58 to 0.67 0.63 0.49 to 0.68 0.533
Weight divided by Gross Tonnage 0.66 to 0.67 0.664 – –
Lightship Centre of Gravity VCG divided by Height to Main Deck 0.68 to 0.76 0.71 0.61 to 0.83 0.75
771
Fig. 10. Project/Vessel 1 - Monthly variation in predicted lightweight (net and with margin).
The recent series of floatel projects gives an opportunity to compare The actual vertical moment of lightship was higher than predicted
the accuracy of the weight estimating and control methods employed. for 4 vessels, and lower than predicted for the other 3. The worst error
All the projects produced regular weight control reports, tracking the was on Project 1, where the inclined vertical moment on the lead ship
estimated weights in each discipline, and monitoring the trends in was underestimated by 2.5%.
weight, centre of gravity and margin. Finally, the inclining tests carried These deviations may be compared with the estimating margins
out for each vessel produced hopefully accurate measurements of the which were in place at the time of contract, which were in the region of
actual weight and centre of gravity. 7% on lightship weight. In most cases this seems to have been sufficient,
Due to the deep draft required to fit the thrusters, the inclining tests but it appears that a separate margin on VCG would be prudent.
for many of these vessels were carried out away from the shipyard In passing, it may be worth recording that the most voluminous
quays, sometimes in relatively exposed locations, which were less than weight control reports proved to be the least accurate. The report for
ideal for such tests. the lead ship on Project 1 comprised 205 pages, while Projects 2 to 4
Table 17 lists the results of the inclining experiments for the lead used reports about 85 pages in length. There is a danger that excessive
ship (and follow ship where applicable) of four different classes (7 detail can be misleading if not supported by higher level sense checks.
vessels), compared with the formally predicted weight and centre of In the case of Project 1, informal deadweight surveys performed with
gravity prior to the test. The difference is expressed in percentage re- the vessel afloat some months prior to completion forecast a weight far
lative to the estimated weight and centre of gravity. closer to the actual result than the formal weight report.
Lightship weight at inclining was less than the predicted value (a The trends displayed in the regular weight reporting also reveal the
desirable result) in 3 of the 4 projects, but Project 1 failed (under- design development of the different projects.
estimated) by more than 2% on the lead ship. The error was reduced Fig. 10 shows the predicted net (without margin) and gross lightship
somewhat for the second vessel. weight for the lead ship of Project 1. This project experienced a sudden
For all four projects, the actual inclined VCG was always higher weight adjustment of more than 500 tonnes early in the project, which
(bad result) than predicted. In 2 out of 3 cases, the follow ship VCG was in fact required redesign of some of the buoyancy elements. Thereafter,
higher than the lead ship VCG, sometimes by a considerable margin. the figure shows how the predicted gross weight continued to increase,
The reasons for this worsening trend are not fully understood. although at a lesser rate than the net weight (due to reducing margin
Fig. 11. Project/Vessel 1 - Monthly variation in predicted Vertical Centre of Gravity.
772
Fig. 12. Project/Vessel 2 – Two Monthly reports of predicted lightship weight.
with increasing accuracy in the weight inputs). The estimate had not carried at the survival draft as the overall allowable vertical centre of
stabilised by the end of the reporting period and in fact this project gravity at the survival draft and the operating draft are linked by the so
proved to be heavier than these predictions. This caused a significant called “deballasting curve”. For transit conditions, it is desirable to
loss of variable deck load capacity compared to target values (which carry large quantities of fuel but typically smaller deckloads without
fortunately had been over-specified in anticipation of such problems). submerging the pontoon tops.
Fig. 11 shows the predicted vertical centre of gravity for the same The variable deck load requirements of floatels are generally much
vessel. This VCG for this project increased steadily throughout, but was smaller than those of semi-submersible drilling rigs, and this is reflected
showing signs of stabilisation near the end. The estimate turned out to in the actual capacities of the recently delivered units. While some have
be within 0.29% of the inclined value. advertised variable load capacities of 1500–2500 tonnes, many have a
In contrast, Figs. 12 and 13 show the history of a stable project true deckload capacity considerably lower than this. Often, a deckload
which did not experience surprises during design, nor dangerous errors capacity is quoted which actually includes variable loads distributed
in prediction. Both the weight estimate and the vertical centre of below the upper deck, such as personnel and effects, small tanks and
gravity prediction were stable for many months prior to the inclining stores which can amount to 200 or 300 tonnes. Also, the useful payload
test, which revealed a lightship lower than the predicted values, and a which can be carried depends on the presence (or not) of snow and ice
VCG slightly higher (0.33%) than predicted. in the stability calculations. In addition, the operating mode of the unit
Finally, Fig. 14 shows the history of a project which experienced can have a substantial effect, as the moored mode can give a lower
continuous difficulties and changes during design and build, expressed overall centre of gravity than DP mode due to wire paid off the high up
by the almost continuous increase in net (no margin) and gross (with winches, and mooring tension acting on fairleaders positioned low.
margin) predicted lightweight. Taking these factors into account, most of the new floatels will have
true deckloads (positioned 1 m above the main deck) in the region of
4.3. Variable deck load and typical loading conditions 1000 tonnes ignoring snow and ice, and around 700 tonnes accounting
for 200–400 tonnes of snow and ice. In practice, floatel working
The outcome of the design and build process should yield a vessel deckloads are usually smaller than this – often less than 200 tonnes.
with a useful carrying capacity in all its operating conditions. The vessel Table 18 compares some typical loading conditions for two older
deckload capacity at operating draft must also be capable of being floatels (one DP, one moored) with those of 3 new vessels. In these
Fig. 13. Project/Vessel 2 –Two-Monthly reports of predicted VCG.
773
Fig. 14. Project/Vessel 3 – Two-Monthly reports of predicted lightship weight (with and without margin).
conditions the older vessels achieve GM values around 3.5 m with true conditions given in the table. Even if snow and ice are ignored and a
variable deckloads of only 160 and 210 tonnes, with comfortable lighter, more realistic deckload of 326 tonnes is considered, the re-
margins on allowable vertical centre of gravity. No snow or ice is sulting GM in DP mode of 2.68 m is still not large, despite the 0.89 m
considered. margin on VCG. This is not a very different result from that for Design 3
The newer vessels are presented in conditions which include more in DP mode with 500 tonnes of deckload and snow/ice.
deckload (326–726 tonnes), with snow and ice also considered in some A very different result is obtained with Design 2 which has relatively
cases. It can be seen that Design 1 (with relatively slender columns) large and square columns and achieves higher GM values of around 6 m
may carry 726 tonnes together with snow and ice when using virtually in DP or moored mode when carrying around 700 tonnes of deckload or
all of the allowable centre of gravity. This produces a GM (transverse) snow/ice. These vessels also have a substantial margin (c. 2.5 m) on
of only 1.87 m, which is noticeably lower than the other vessel VCG remaining, indicating that higher deckloads are possible.
Table 18
Comparison of typical operating conditions at operating draft.
Vessel actual ID Older vessels New vessels
A B 1 2 3
Mode Moored or DP DP Moor Moor DP Moor DP DP

True Deck load Weight (t) 210 160 726 326 670 400 500
VCG (m) 32.45 36.1 37.1 37.1 36.7 36.7 37.95
Loads for operation Snow & ice (t) 0 0 204 0 0 367 204
Snow, ice VCG (m) 0 0 31.16 0 0 41.1 31.2
Mooring tension (t) 0 159 728 0 787 0 0
Tension VCG (m) 0 15.0 16.92 0 16.2 0 0
Small tanks, stores (t) 133 282 307.5 307.5 200 200 307.5
Small tanks VCG (m) 11.3 30.76 29.76 29.76 – – 29.76
Other variable loads Fuel 930 594 1950 1950 2948 2948 2000
Fuel VCG (m) 3.31 1.66 6.64 6.64 5.5 5.5 6.64
Fresh water (t) 446 760 700 700 750 750 500
Fresh VCG (m) 15.43 10.32 4.2 4.2 4.41 4.41 4.78
Other (t) 378 58 221 221 – – 50
Subtotal (t) 1754 1413 2871 2871 – – 2550
Subtotal VCG (m) 11.69 7.62 9.91 9.91 – – 10.75
Ballast water Weight (t) 4932 6823 6421 7753 8249 8310 9110
VCG (m) 3.87 4.04 3.477 3.477 – – 3.23
Overall vertical centre of gravity VCG solid (m) 17.332 16.219 20.946 20.12 17.44 17.28 19.56
FSC transverse (m) 0.08 0.252 0.166 0.166 0.16 0.18 0.144
FSC longitudinal (m) 0.18 0.681 0.194 0.194 0.40 0.44 0.353
VCG transverse (m) 17.416 16.471 21.11 20.29 17.61 17.45 19.70
VCG longitudinal (m) 17.513 16.900 21.14 20.31 17.84 17.71 19.91
Overall KG margin, KM and GM Allowable KG (m) 18.272 17.185 21.20 21.20 20.30 20.30 20.764
Margin on KG (m) 0.76 0.285 0.06 0.893 2.46 2.587 0.86
KMt (m) 20.72 20.12 22.98 22.98 23.58 23.58 22.22
KMl (m) 20.67 20.59 22.95 22.95 23.24 23.24 24.94
GMt (m) 3.30 3.65 1.87 2.68 5.97 6.12 2.52
GMl (m) 3.16 3.69 1.81 2.64 5.40 5.53 5.02
774
Table 19
Examples of weight growth through life – two existing semi-submersible floatels.
Vessel Years duration considered As built condition Later life condition Averaged annual increase
X 23 Lightship 12,485t Lightship 14,261t 77 Tonnes (0.62%)

VCG 21.3m VCG 23.30m 8.7 cm (0.41%)
Y 30 Lightship 14,192t Lightship 15,912t 57 Tonnes (0.40%)
VCG 22.92m VCG 26.67m 12.5 cm (0.54%)
Fig. 15. Relationship of waterplane area to displacement – old and new generation semi-submersibles.
It is worth noting that the governing stability direction for all these 5. Motion performance
vessels is usually longitudinal rather than transverse, even though the
hydrostatic stiffness (KM) is sometimes greater when inclined long- 5.1. First order motions
itudinally (e.g. if columns have blisters fitted). This is because the
pontoon tanks produce a greater free surface effect in that axis. Reduced movement in the open ocean environment is the primary
The new designs have KM values some 2–3 m higher than the older purpose of the semi-submersible configuration, and this objective is
generation. As a general design rule, pontoon volume should be re- achieved by giving priority to the following features;
duced as far as possible to maintain a high KM value.
• Resonant frequencies of motion in heave, pitch and roll should occur
at long periods, beyond those of the waves commonly found in the
4.4. Weight growth during life intended operating areas
• Distribution of buoyancy arranged so that wave excitation in com-
It is of course impossible to predict the changes which these vessels monly occurring wave periods is minimised (by arranging for
will encounter during their working lifetime, but experience shows that “cancellation” of up/down vertical forces and moments, and by
change is a common feature. Indeed, less than one year after delivery cancelling of pressure and inertia effects)
the Safe Boreas was modified by the placement of a temporary ac-
commodation module for 40 persons on the main deck. This should be The first of these features is obtained by having low stiffness, which
considered when setting the target deckload capacity of newbuild is achieved by a using a small waterplane area relative to the displaced
floatel, in addition to the expected normal operating deckload. mass. The second feature requires the submerged depths of the pontoon
Reference to previous vessels can provide some guidance, and upper and lower surfaces to be considered, as well as the “shielding” of
Table 19 contains data extracted from widely separated inclining tests the pontoon top by the columns placed thereon.
of two randomly selected older floatels. It can be seen that there is Fig. 15 indicates that the new vessels generally have a larger water
substantial growth in both weight and vertical centre of gravity. These plane area relative to their displacement than the older generation.
two vessels had at that time not undergone radical transformation or Despite this, the new vessels generally achieve heave periods no less
conversion, but had been subject to some upgrades and enhancements than 20 s, and resonant roll and pitch periods in the region of 40–50 s.
including additional accommodation modules, upgraded mooring These are not much shorter than the older vessels, and well beyond the
winches etc. average wave periods to be expected in most operating areas of the
Table 20
Periods of wave and swell spectra.
Sea Area West Africa Gulf of Mexico Brazil (Campos) West of Shetland (all year) Northern North Sea (all year)
Annual average Tp (s) 12.0 5.95 9.1 9.6 7.9

Percent of time when Tp exceeds 15s 12% 0% 5% 1.5% 0%
775
Table 21
Comparison of pontoon submergence of old and new designs.
Design Type Vessel Operating draft Pontoon depth Pontoon as percent of Pontoon top submergence
(m) (m) operating draft (m)
Modern 4 column type – 1980s Safe Boreas (as built) 18.6 9.75 52.4% 8.85
onwards Safe Boreas prior pontoon height 18.6 8.96 48.2% 9.64
change
Floatel Endurance/Superior 18.0 8.625 47.9% 9.375
Regalia (1985) 17.5 7.85 44.8% 9.65
Safe Notos/Eurus 20.0 9.0 45.0% 11.0
6 or 8 column type – 1970s basic Safe Caledonia (Pacesetter design) 18.29 6.13 33.5% 12.16
design Safe Scandinavia (Aker H3.2 design) 22.0 7.2 32.7% 14.8
Fig. 16. Heave transfer function for four designs of semi-submersibles.
world. The probability of encountering sea spectra with a peak period 5.2. Second order motions
beyond 15 s is rather low, as Table 20 indicates.
It is more difficult to compare the old and new vessels in terms of In addition to the first order motions (reactions to the frequency of
the second objective of reducing wave excitation. However, Table 21 the incident waves), floating vessels are also susceptible to so called
does compare the pontoon top submergence of a number of vessels. The second order effects, including steady and slowly varying wave drift
pontoon now consumes nearly half of the draught, compared to one forces. There is significant low frequency wave group energy at periods
third in the older designs. This means that the modern vessels have over 40 s, and this can be close to the semi-submersible natural periods
pontoon tops significantly closer to the still water line (c. 9 m) than the of pitch and roll. This may have consequences for station-keeping and
older designs (12–14 m), despite their larger size. They are basically also for motions such as pitch and roll.
less “submersible” in form than the older vessels. This is likely to make These second order effects are difficult to observe and predict op-
the pontoon top more susceptible to first and second order wave action, erationally, as they occur at such long periods. They can cause occa-
especially if the vessel is trimmed or heeled slightly, thus bringing the sional unexpected motions with amplitudes in excess of the underlying
pontoon top even closer to the surface. trend. Surge is a highly damped motion with low frequency peaks fol-
The overall effect of these differences is illustrated in Fig. 16, which lowing a strongly non-Gaussian process. One consequence is that the
compares the heave transfer function for two of the new four column extreme peaks are much larger than for a Rayleigh distributed process
floatel designs (GVA 3000E and DSS20NS) and two more traditional 8 with the same standard deviation.
column designs from Aker. The heave transfer functions are plotted During the 1980s, one aspect of these second order effects had been
together with the probability distribution of the wave periods expected observed in model tests and calculations (Atlar, 1987) whereby a semi-
in the northern North Sea (Haltenbanken seas states shown). submersible can take on a steady tilt angle in regular head seas. It is
The Aker H3 is an old design, representative of some of the first only relatively recently (Voogt et al., 2002; Voogt and Soles, 2007) that
generation floatels, and can be seen to have a very long natural heave the existence of these effects has been observed at full scale.
period (about 25 s) and low response in the commonly encountered A series of model tests found that this steady heel angle increased
wave periods from 5 to 10 s. The Aker H4.2 was considered a large for shallower vessel drafts (i.e. pontoon top near the surface – ref.
design in the 1980s (around 39,000 tonnes displacement) but now is Table 21) and increased current speeds. Increasing the vessel static
not much larger than some of the new floatels considered in this paper. stability was found to reduce or eliminate the steady list angle.
It has a generally similar response to the old Aker H3 (although with Furthermore, vertical forces on the pontoons can cause roll motions
slightly shorter heave natural period and higher responses), and has low in head seas (Voogt and Soles, 2007) which is counter intuitive to crew
heave response in the common wave periods and beyond. expectations. Increasing static stability can reduce such motions but not
The two new floatel designs show higher heave response in the eliminate them. Voogt and Soles (2007) report model test results and
range from 8 s upwards, around 100% more than the 8 column Aker real-world experience of low frequency roll motions of a drilling semi-
H4.2 design in 10 s waves. From about 12 to 17 s the performance is no submersible occurring in short period head seas combined with current.
worse than the old Aker H3 design. The motions had a period near the vessel natural roll period, suggesting
776
Fig. 17. Quadratic Transfer Function for Surge Wave Drift Force – Aker H3.2 old design and new design.
Table 22
Number and capacity of mooring lines.
Mooring type No. of mooring lines Line size Nominal Line MBL (t) No. x MBL (t)
Floatel Reliance N/A (DP only) 2 (harbour use) – – –

Floatel Superior, Floatel Endurance Wire 8 76 mm 490 3920
Wire 10 83 mm 580 5800
Floatel Victory, Floatel Triumph Chain 10 70 mm 510 5100
76 mm 610 6100
Safe Boreas, Safe Zephyrus Wire 12 90 mm 640 7680
Neptuno, Atlantis N/A 4 (harbour) – – –
Safe Notos, Safe Eurus Chain 10 76 mm 612 6120
Safe Nova, Safe Vega Wire 10 96 mm 750 7500
OOS Tiradentes N/A (DP only) None – – –
they are due to low frequency forcing. It was found that by changing the These can excite the natural period of the mooring or thruster system.
vessel heading 30–45° these roll motions could be completely removed. Resonant motions and hence large amplitudes can occur as a result of
Roll and pitch motions from these second order effects can be ad- modest input forces.
versely affected by low static stability characteristics (ref. Table 18). The modern generation designs have four columns, typically square
One report of model test investigations of the phenomena (Voogt et al., in cross section (with rounded corners), having lengths and breadths in
2002) states that a “disadvantage of very low initial stability is increased the range 13 m–14 m. This may be compared with the circular section
sensitivity to the energy in wave groups at a frequency close to the natural columns of the older designs, which had diameters in the range
period in pitch and roll. This means that although the roll and pitch natural 8 m–10 m for the corner columns, and smaller for the intermediate
periods are far removed from normal wave periods and therefore roll and columns.
pitch resulting from normal waves will be small, roll or pitch may be severely Fig. 17 compares the quadratic transfer function for horizontal wave
increased due to the natural period of wave groups. Such an increase was drift in the surge direction, for an old and a new design semi-sub-
observed in the model tests.” mersible. It can be seen that the new design has a very much larger
The details of one sudden motion incident involving a modern de- response in wave periods around 8 s. These wave periods are common
sign drilling semi-submersible offshore Norway have been published in the North Sea and elsewhere.
(Browne, 2013). It has been found with some vessels that the vessel's When analysed together with a wave energy spectrum, the mean
thrusters and dynamic positioning system can be used to reduce roll and surge wave drift force is considerably higher (by a factor of around 3)
pitch motions (Jenssen, 2010). for the new vessels than for the older slender column designs. One study
Robust initial stability is one solution to some of the above problems in waves of 7 m Hs shows horizontal forces of 50 tonnes compared to 15
and this can be obtained by large column dimensions. However, this tonnes (at 12s Tp) and 80 tonnes compared to 27 tonnes (at 8s Tp).
introduces issues related to second order effects on station-keeping, so The difference between the old and new designs is less severe in the
that careful compromise is required. sway direction for understandable reasons of column numbers and hull
geometry (the older designs presented a larger number of columns in
5.3. Effect of second order forces on station-keeping the sway axis). Nevertheless, in a 7 m Hs seastate the newer designs can
attract horizontal forces around 33–50 tonnes (increasing with reducing
The new semi-submersible designs also appear to be more suscep- wave period), some 50% higher than the older slender column designs
tible to horizontal excursion caused by second order wave drift forces. (22–35 tonnes).
777
Fig. 18. Configuration of 3 mooring windlasses on single drive shaft.
These characteristics have implications for the selection of the best operations than chain moorings, and this may be one reason why wire
location of the gangway during design, and the selection of the most has commonly been used for accommodation units in areas where their
appropriate heading (when possible) during operation, in order to drilling counterparts use chain. New DnVGL standards apply a fairlead
minimise horizontal excursions which consume the gangway stroke. bending reduction factor for wire ropes which generally reduces the
Some of the new designs appear to move less in a transverse direction, strength that can be considered in a mooring analysis by about 15%
and so are best positioned beam on to the waves, and the gangway is from the nominal/catalogue value.
probably also best positioned in a transverse rather than longitudinal Wire moorings result in a high centre of gravity for the mooring
direction. lines when stored on the winches. This has been mitigated in some
designs (e.g. GustoMSC Ocean 500) by fitting the winches in recesses
6. Mooring below the main deck.
Chain moorings allow a smaller and lighter winch/windlass (no
6.1. Design basis need to carry multiple wraps of wire), and also the weight of the
mooring line is carried low in the vessel. However, chain mooring in-
Most of the new vessels have their origin in dynamically positioned troduces downflooding points near the deck edges of the vessel, and the
designs. Mooring has been something of a secondary consideration in effect of a flooded chain locker on the GZ curve can prove problematic
some of the projects, and some have no offshore mooring capability when considering damaged stability.
installed. Indeed, without thruster assistance some of the new mooring
designs would not be adequate for harsh environment operation when 6.3. Chain lockers
subjected to the latest mooring analysis codes.
Table 22 shows the number and strength of mooring lines fitted on For those rigs opting to use chain instead of wire, the chain locker
each of the vessels. capacity previously represented a substantial demand on space within
The North Sea rigs generally have to comply with the requirements the pontoons and columns.
of ISO code 19001-07 Annex B2 Table B2 (ISO, 2013). This requires These volumetric demands have declined as mooring practices have
consideration of two line broken cases (in 10 year return weather), changed. Formerly, moored rigs carried their own anchors and a full
whereas previously only single line failure was generally considered in length of chain or wire (sometimes 1.5 km per leg). In the North Sea,
mooring design. In addition, modern analyses are conducted with modern mooring practice is for the anchors and most of the mooring
generally increased line factors of safety, longer metocean return per- line length to be laid in the field prior to the rig arrival. Anchor
iods, and the methods account more fully for relevant effects such as handling vessels are then used to retrieve and connect the rig lines to
line dynamics and 2nd order motions which might have been neglected the pre-laid lines. Around 600 m of chain or wire is sufficient for the
in the past. These requirements pose increased challenges for the handover and connection operations. This is one of the few areas where
mooring system design, and 8 point systems will be unsuitable for the modern design requirement is less onerous than before.
mooring close alongside fixed platforms. In practice, this has been the For example, the Safe Notos design has 12 chain lockers, with a total
case since at least the late 1980s, when NPD (Odland, 1990) and DNV volume of about 450 m3. The older Safe Scandinavia was originally
(Christiansen et al., 1994) required different factors of safety for such designed for the same size of chain with 12 lockers, but required a
moored operations, compared to open water drilling operations. At that volume of 1640 m3.
time, the more simplistic quasi-static analysis method was permitted.
6.4. Mooring winches
6.2. Choice of chain or wire mooring
Most of the previous generation of semi-submersibles used hy-
Wire moorings are generally quieter during retrieval/payout draulically powered winches or windlasses. With the new generation,
778
electrically powered winches have become more common. Table 23

The Safe Notos and Safe Eurus are fitted with four sets of electrically Results of DP Simulations assuming co-linear environment.
driven chain windlasses, one at each corner. Two of the sets include 3 Intact Loss of single Two thrusters
windlasses (Fig. 18) while the other two are installed with 2 windlasses, thruster unavailable
with space and other provisions made for future upgrade to 3, to give a
Wave height Hs (m) 11.6 9.6 8.6
12 point mooring system if required.
Wave period Tz (s) 9.0 9.0 9.0
The 10 windlasses are rated for 76 mm R4 chain with a MBL of 612 Wind speed (m/s) 38.5 32.1 29.2
tonnes, and have a parking brake capacity of 620 tonnes. Windlass Surface current speed (m/s) 0.46 0.46 0.46
pulling capacities include 270 tonnes at 13 m/min (high load, low
gear), 108 tonnes at 28 m/min (high speed, low gear), as well as 42
tonnes at 86 m/min (high speed, high gear).
On the Safe Notos each corner of the vessel is fitted with a 690 kW Removal of bolsters has also been encouraged due to safety concerns
motor driving up to 3 windlasses on a common shaft. The electrical regarding the risk of a stowed anchor coming loose and damaging the
system is arranged with frequency converters for variable speed, and bolster and then the watertight hull. This was less of a concern on the
braking resistors for absorbing reverse power during payout. The older permanently moored floatels (although such damage did occa-
system is designed to allow multiple windlasses to work simulta- sionally occur) as the anchors were usually off the bolster. With the new
neously. floatels operating on DP for much of the time, stowed anchors would be
The winches may be controlled from a console, located in the vessel exposed to this risk for a greater percentage of time. A serious incident
Central Control Room (Bridge), as well as from local controls placed in occurred to one such vessel in the Norwegian sector (Andersen et al.,
four winch control cabins (one at each corner of the unit). 2014) and this has acted as a catalyst to discourage anchor stowage on
Other features associated with winches on a floatel include a deluge bolsters on such vessels.
system located above each windlass and a seawater washing system for
the chain. The deluge system is intended to minimise the risk of spark 7. Dynamic station-keeping and propulsion
ignition during rapid pay-out of the chain during an emergency. Release
of the chains on one side of the floatel is a technique which may be used 7.1. Dynamic positioning
to move a moored vessel away from a fire/gas event on the client
platform by using the energy stored in the opposing chain catenaries. All of the new vessels have a dynamic positioning capability, with
Further separation distance may be obtained by powered winching. offshore spread mooring sometimes absent, or something of an after-
thought in some designs. This is in contrast to the previous generation
6.5. Fairleaders and anchor bolsters where mooring was the primary means of station-keeping and only a
very small number of vessels had a DP capability.
On the Safe Notos, each mooring line is provided with a seven The principles of dynamic positioning are well presented elsewhere
pocket self-aligning fairlead mounted on the column. The fairleads are and not repeated here. Most of the new designs have DP3 classification
self-lubricating and designed for submerged operation. The fairleads and three engine rooms. For these vessels, that means a worst-case
are capable of 180° travel, and arranged so that they may be locked in failure mode involving loss of two thrusters (due to loss of one engine
position parallel to the column face to avoid swinging and banging room or switchboard room). The weather conditions for design vary
when not in use during transit and DP operations. from case to case, but Table 23 presents the performance of one modern
Compared with the older designs, the new vessels have shorter design under various failure modes.
columns, due to generally smaller air gap (Table 3a) and deeper pon- In addition, this vessel can hold station with one thruster failed in
toons (which gives reduced pontoon submergence - see Table 21). conditions with high surface current (1.4 m/s), waves of 5.9 m H and
These shorter columns make it more difficult to arrange mooring fair- winds of 16 m/s (1 min average). These are requirements for floatels
leaders sufficiently low while at the same time allowing the chain or operating on DP offshore Brazil.
wire to depart at a sufficiently steep angle without hitting the anchor These limiting weather conditions are calculated using well estab-
bolsters (if fitted). lished methods. However, it should perhaps be noted that the modern
Some of the new vessels (e.g. the four Floatel vessels) have the drilling semi-submersible which experienced wave impact damage
fairleaders placed very close to the operating draft waterline (or even offshore Norway in late 2015 (PSA, 2016) was subjected to a sudden
above in the case of Floatel Victory/Triumph). On some others the fair- large excursion due to the horizontal impulse applied to the deck box.
lead was moved lower from the designer's initial position, to about 2 m Such events are not captured in DP simulation tools or onboard DP
below waterline. predictors. This therefore makes it critical that DP vessels are not op-
This subject is relevant when considering the operations of supply erated in situations where a loss of air gap may occur and cause such a
vessels alongside the floatel. These vessels are becoming larger and dangerous loss of position. Fortunately, the DP mode allows floatels to
deeper drafted, and mooring lines near the surface can restrict the room remove themselves from forecast severe weather conditions in a way
for manoeuvre. This is a concern for operating draft only, as supply which moored semi-submersibles cannot.
boats are not allowed alongside at survival draft due the risk of damage
in excess of those considered in the damaged stability regulations. 7.2. Thrusters
The effect of the shorter columns on the relationship between fair-
leaders and anchor bolsters on the newer designs has been described To achieve the desired dynamic positioning capacity, all the new
above. This geometrical problem is countered to a great extent by the vessels are fitted with thrusters of the fixed pitch type, driven by ver-
fact that many of the modern floatels are not fitted with anchor bolsters. tically mounted variable speed motors, supplied from frequency con-
The older vessels had anchor bolsters fitted below all fairleaders, and vertor drives. The previous generation of floatels employed variable
required to carry up to 3 anchors on the bolsters for insurance purposes pitch constant speed thrusters. The fixed pitch type of thruster certainly
when in transit as they had no independent means of propulsion. The reduces the risk of failure due to mechanical or hydraulic system de-
new designs are generally well endowed with redundant propulsion fects, but the associated frequency converter drives introduce com-
systems, and can thus argue to carry no anchor, or only one on a bolster plexity and potential new sources of failure.
when in transit. To maximise commonality of spares, most vessels have opted for all
779
Table 24
Station-keeping equipment installed.
DP class No. of thrusters Thruster rating (MW) Total thruster power (kW) Power – Operating Displacement (kW/t)
Floatel Reliance DP2 4 2.5 10,000 0.558

Floatel Superior, Floatel Endurance DP3 6 3.2 19,200 0.657
3.5 21,000 0.625
Floatel Victory, Floatel Triumph DP3 6 3.2 19,200 0.716
Safe Boreas, Safe Zephyrus DP3 6 4.0 24,000 0.707
Neptuno, Atlantis DP3 6 3.7 22,200 0.665
Safe Notos, Safe Eurus DP3 6 3.7 22,200 0.665
Safe Nova, Safe Vega DP3 6 3.8 22,800 0.605
OOS Tiradentes DP3 6 3.8 22,800 –
Fig. 19. Results of speed and power trials – three hull designs.
Fig. 20. Collation of Speed Trial Results – versus function of speed and dis-
placement.
thrusters sharing the same direction of rotation and the same propeller
blades. The thrusters are generally fitted with tilted nozzles to direct the
outflow away from the underside of the pontoons and so avoid loss of Although it is not easy to achieve on a semi-submersible design
thrust due to the Coanda effect. equipped with six thrusters and only four columns, a design which al-
The thruster powers installed are listed in Table 24, and related to lowed the thrusters to be retracted vertically into the hull would be
vessel size by means of a power/weight ratio. potentially attractive. It could allow vessel berthing in a wider range of
The smaller 2500 kW thrusters of a previous generation projected ports, and save considerable time and cost when carrying out repairs or
around 3.8 m beneath the pontoon, while the larger modern floatels planned overhauls. Some of the monohull DP drillships have employed
have thrusters projecting 5 m–5.4 m below the keel, and weighing retractable thruster arrangements for many years (MacGregor et al.,
43–64 tonnes. These thrusters are too heavy to be handled by the on- 2001).
board cranes, unless a lifting capacity of more than the typical 50
tonnes has been installed. 7.3. Speed performance
Even if a suitable crane is in place to assist in thruster changeout
operations, the procedure for removing and reinstalling an underwater Speed trials results for transit draft are presented in Fig. 19 for
demountable thruster on these vessels is time consuming and ex- Designs 1, 2 and 3 with different pontoon characteristics as shown in
pensive, requiring the use of divers and heavy lifting operations inside Table 25. For Design 1, results are given for the lead ship (1a) at 97.7%
the thruster room. Due to the deep draft, these operations frequently of pontoon draft and a displacement of 27,082 tonnes, and for the
need to be carried out far from shore where exposure to weather is follow ship (1b) at 97% of pontoon draft and a displacement of 26,792
increased. tonnes. A further set of results is given for 1b, trimmed 1° by the stern.
Table 25
Pontoon Proportions for Vessels subject to Speed Trials.
Length (m) Beam (m) L/B ratio Height (m) End shape - profile End shape – plan view Displaced volume (m3) Block coefficient
Design 1 (vessels 1a, 1b) 97.5 15.0 6.50 9.75 Flat top and bottom Elliptical 13,469 0.944
Design 2 95.0 14.0 6.78 9.00 Square with radius Square with radius corners 11,905 0.994
corners
Design 3 104.5 13.65a 7.66a 9.75 Square with radius Triangle 14,285 1.027a
17.55b 5.95b corners 0.799b
a
Excluding sponsons.
b
Including sponsons.
780
Table 26
Results of turning trials.
Draft as % of pontoon Speed Tactical Diameter Advance (m) Transfer (m) Time to turn Notes
depth (knots) (m) (seconds)
Ocean Confidence (Ren et al., 131% 6.0 370 252 155 195 for 90° All 8 thrusters turned 45°
2001) 6.0 147 200 65 147 for 90° All 8 thrusters turned 90°
4.1 293 278 120 230 for 90° 4 thrusters turned 45°
3.8 130 160 52 160 for 90° 4 thrusters turned 90°
Design 1 97% 10.1 573 417 255 – 2 thrusters steering
7.7 582 378 240 – 2 thrusters steering
Design 2 96.4% 4.5 – – – 925 for 360°
4.3 – – – 731 for 360°
7.8 – – – 445 for 360°
7.3 – – – 371 for 360°
Design 3 98.5% 9.5 – – – 580 for 360°
10.0 – – – 659 for 360°
Table 27
Crash stop trials results.
Vessel Draft as % of Speed (knots) Time to Stop Time to Full Astern Distance
pontoon depth (seconds) (sec) travelled (m)
Ocean Confidence (Ren et al., 2001) 131% 3.8 68 – 108a

5.6 70 – 141a
6.7 74 – 160a
3.3 68 – 81a
3.6 67 – 86a
4.3 80 – 11a
Design 1 97% 9.4 46 121 112b
10.1 54 131 138b
Design 2 96.4% 9.3 64 – –
4.5 100 – –
Design 3 98.5% 8.6 68 – 206c
a
Ahead reach.
b
Track length.
c
To stop.
For Design 2, the results relate to a condition at 96.4% of pontoon speed. The superior performance is more marked when one notes the
draft and a displacement of 23,526 tonnes. For Design 3, the results increased displacement which is being propelled for the lesser power.
relate to a condition at 98.5% of pontoon draft and a displacement of There does not appear to be much performance difference between
28,770 tonnes. Designs 1 and 2, although the following design differences are noted;
With pontoons having length from 95 to 104.5 m, the Froude
Number at 10 knots is between 0.16 and 0.17. This is well below the • Design 2 pontoons are slightly more slender than Design 1, despite
sharp increase in wave resistance which occurs at Froude numbers Design 1 having a more streamlined bow and stern shape
above 0.4. • Design 1 pontoons are deeper (more wetted surface) than Design 2
It can be seen that Design 3, with the more slender pontoon, re- • Design 2 is less heavy than Design 1 in these conditions.
quires considerably less power than the other designs for the same
For Designs 1a/b, it can be seen that the second vessel required
considerably more power than the first, for reasons which are not fully
understood. When Design 1b was tested trimmed 1° by the stern this
gave much better results than the level trim test.
Fig. 20 combines the results for all vessels, by plotting required
thruster power against a function of displacement raised to the power
0.6667 (intended to represented wetted surface area) and speed raised
to the power 2.7. The correlation is not particularly good, but a clear
trend is evident.
7.4. Manoeuvring
At the same time as the speed trials reported above, manoeuvring

and crash stop trials were conducted on the new floatels. The six
powerful azimuthing thrusters mean that turning and stopping perfor-
mance is excellent.
In Tables 26 and 27, the results of these tests are compared with
those from the Ocean Confidence (Ren et al., 2001), a heavily modified
Fig. 21. Typical arrangement of three thruster pairs. Aker H3.2 hullform (former floatel) fitted with eight 3000 kW
781
Fig. 22. Typical connections between three sets of power generation/distribution and thrusters.
Table 28
Main generators.
DP2 or No. of Cylinder bore Engine No. of Generator rating Total power Power – Operating
DP3 engines (cm) layout cylinders (kWe) (kWe) Displacement (kWe/t)
Floatel Reliance 2 4 26 12 V 48 3900 15,600 0.87

Floatel Superior, Floatel Endurance 3 6 25 16 V 96 3680 22,080 0.76
26 16 V 96 4700 28,200 0.84
Floatel Victory, Floatel Triumph 3 6 26 12 V 72 3900 23,400 0.87
Safe Boreas, Safe Zephyrus 3 6 26 16 V 96 5020 30,120 0.89
Neptuno, Atlantis 3 6 32 9L 54 4800 28,800 0.86
Safe Notos, Safe Eurus 3 6 32 9L 54 4800 28,800 0.86
Safe Nova, Safe Vega 3 6 32 9L 54 4750 28,500 0.76
OOS Tiradentes 3 6 5445 32,670 N/A
controllable pitch thrusters. That vessel was tested at a deep transit remain available after the loss of any single space, without having to
draft of 10.5 m, with a displacement of 33,740 tonnes. rely on any complex cross connections in the electrical power and
cooling systems for the thrusters.
8. Power generation, electrical and automation systems The thrusters are typically grouped in pairs according to the prin-
ciple in Fig. 21, so that loss of one distribution system will result in loss
8.1. Overall system architecture of thrusters at opposite corners of the unit, thus minimising the re-
duction in station-keeping capacity. The cable routing and piping
For the vessels intended to achieve DP3 notation, the balance be- routings to achieve these groups can be complex, especially considering
tween minimising equipment installation and station-keeping perfor- that there are two or four pump rooms in the pontoons, but three engine
mance was achieved by dividing the propulsion system into three equal rooms in the deck box requiring a separate supply of cooling water.
families (each with two generator sets, one switchboard room and two Fig. 22 illustrates one scheme for cable routing.
widely separated thrusters). This allows 66% of the installed power to
782
Table 29
Comparison of electrical systems on three classes of floatel design.
Safe Boreas, Safe Zephyrus Safe Notos, Safe Eurus Safe Nova, Safe Vega
Concept Generating Voltage 11,000 V 6,600 V 11,000 V

Other voltages onboard 690 V, 450 V, 230 V 450 V, 230 V 690 V, 450 V, 230 V
No. MV switchboard rooms 3 3 3
No. of LV switchboard rooms 0 (part of MV rooms) 0 (separated part of MV rooms) 3
Outer Bus Tie for Main switchboard ? No Yes Yes
Main generators 6 x 5950 kVA 6 x 6000 kVA 6 x 5938 kVA
Emergency/harbour generator None – 2 main generators are rated for emergency 1500 kW at 450 V 1540 kW at 690 V
duty
Switch board Number of sections 3 3 (with splitter CB in each) 6 (3 pairs)
Number of consumer sections in main 11 16 (plus 3 spares) 12
board
Thruster transformers 6 x 5000 kVA 6 x 4550 kVA 6 x 4500 kVA
11,000/690 V 6600/2200 V 11,000/690 V
Ship Service transformers 3 x 3500 kVA 6 x 3000 kVA 6 x 2800 kVA
11,000/690 V 6600/450 V 11,000/690 V
Other transformers 2 x 1000 kVA 4 x 1000 kVA None
11,000/690 V for winch VFDs 6600/710 V for winch VFDs
Spare capacity None 3 × 1250A CBs 3 cubicles
Notes Cranes fed from 450 V boards Winches fed from 690 V board
8.2. Main engines reduction in fuel consumption from c. 190 g/kW-hr to 176 g/kW-hr
under ISO conditions.
The power to achieve the required station-keeping performance is
considerable, and this dominates the overall system load balance and 8.3. Electrical and automation systems
engine selection. Installed powers range from 20 to 30 MW, as shown in
Table 28. Table 29 illustrates the range of choices made in the configuration
One problem with such high installed powers is that normal DP of the main switchboard, and other choices such as selection of gen-
loads are much lower than the governing design case. In DP3 mode with erating voltage, powering of mooring winches etc. Despite common
three pairs of engines, some clients will insist on 6 engines constantly features such as six thruster transformers, quite substantial differences
running to avoid temporary blackouts. In such cases the engines are may be observed in the number of sections and cubicles in the main
subjected to prolonged operations at low load. switchboard, and the number and capacity of the supply transformers to
The new floatels have selected a range of different medium speed the lower voltage systems.
engines. The smaller bore 25 cm and 26 cm vee configuration engines The onboard marine equipment is monitored and controlled by in-
with 900 or 1000 rpm used in the earliest new vessels are a legacy from tegrated automation systems, with distributed I/O cabinets around the
the previous generation of floatels. These characteristics allowed en- vessel, a number of processing stations and distributed operator inter-
gines of higher power/weight ratio and a lower height, which were faces (VDU based).
important for the previous smaller semi-submersibles. On most of the vessels, the vessel automation system takes care of
For example, a 5000 kWe generator set based on a vee configuration the power management of the main generators and consumers, and is
26 cm engine would weigh 70 tonnes and be 3.7 m high, compared with supplied in an integrated package together with the DP control system.
84 tonnes and 4.0 m for a 4800kWe inline set from the same maker but Usually the fire and gas detection and emergency shutdown systems
based on a 32 cm and 720 rpm design. This is a difference of nearly 90 (ESD) are supplied as part of the same integrated package. The HVAC,
tonnes on a six generator installation. crane and mooring winch installations are controlled by their own
However, these benefits came with a penalty in the form of higher dedicated systems, with interfaces to the ship's control network.
fuel consumption, and (in the new high power installations) a large The DP sensors do not include the hydro-acoustic or taut wire sys-
number of cylinders to maintain. The later vessels in the current gen- tems normally fitted on construction vessels (although these may be
eration opted for 32 cm bore inline engines, as they had sufficient retrofitted), but focus on relative positioning systems including
weight carrying capacity and height in the deck box to accommodate gangway stroke sensor, Cyscan (laser referencing) and Radius (radar
them. This allowed a reduction in cylinder count from 96 to 54, and a scanning), as well as DGPS.
Table 30
Installed lifeboat capacity.
Lifeboat type Layout of boat No. boats Boat Total persons in Philosophy Capacity with one boat Capacity with one side
stations capacity boats out of service out of service
Floatel Reliance Davit 3 sides (2, 4, 4) 10 200%

Floatel Superior, Floatel Freefall 3 sides (4, 2, 2) 8 70 560 Alternative N+1 490 280
Endurance
Floatel Victory, Floatel Davit 3 sides (4, 4, 2) 10 100 1000 200% 900 600
Triumph
Safe Boreas, Safe Zephyrus Freefall 3 sides (2, 3, 3) 8 70 560 Alternative N+1 490 420
Neptuno, Atlantis Davit 3 sides (4, 4, 4) 12 125 1500 200% 1375 1000
Safe Notos, Safe Eurus Davit 2 sides (4, 4) 8 125 1000 2 × 100% 875 500
Safe Nova, Safe Vega Freefall 3 sides (2, 3, 3) 8 70 560 Alternative N+1 490 350
OOS Tiradentes Davit 2 sides (5, 5) 10 120 1200 2 × 100% 1080 600
783
Table 31
Cranes fitted to Floatels.
Crane 1 Crane 2
Main hoist Whip hoist Main hoist Whip hoist
Floatel Reliance 120t x 15 m 16t x 49 m 25t –

30t x 47 m
Floatel Superior 50t x 30 m (20t x 55 m) 15t x 60 m 50t x 30 m (15t x 60 m) 15t x 60 m
Floatel Endurance 100t x 16 m 15t x 60 m 100t x 16 m 15t x 60 m
Floatel Victory, Floatel Triumph 120t (45.8 m boom length) – 64t (50.4 m boom length) –
Safe Boreas, Safe Zephyrus 50t x 26 m (48 m boom length) 14t x 56 m 50t x 16 m (56m boom length) 14t x 56 m
Neptuno, Atlantis 300t x 25 m 25t x 71 m 75t x 18 m 15t x 56 m
Safe Notos, Safe Eurus 300t x 12 m radius 15t (60 m boom length) 50t (55m boom length) 15t (60 m boom length)
92t x 45 m radius (55 m boom length)
Safe Nova, Safe Vega 70t x 35 m 15t x 50 m 40 t x 20 m (knuckle boom) 20t x 45 m
OOS Tiradentes 60t x 7.5 m (10t x 40.5 m) – 60t x 7.5 m (10t x 40.5 m) –
9. Other equipment advantageous if a similar solution could be created for the semi-sub-
mersibles.
9.1. Gangway
9.2. Lifeboats
The hydraulic telescopic gangway is the central operational piece of
equipment installed on these floatels. Most of the vessels have been Clearly, with large numbers of personnel onboard, the provision of
fitted with units having a mean length of 38 m and a stroke of ± 7.5 m. suitable lifesaving facilities is an important aspect in any floatel. The
The weather conditions sufficient to permit landing of the gangway standard used to calculate the number of lifeboat seats required varies
on the platform are typically a significant wave height of 3.5 m or less, from vessel to vessel, as shown in Table 30. Some vessels have managed
on a declining wind speed trend. to obtain exemptions so that the total number of lifeboat seats is not in
After landing, the gangway stroke is monitored and the telescopic accordance with the IMO MODU Code (which requires 100% capacity
motions trigger a two stage alarm system, and eventually an auto-lift to remain available if all the lifeboats in any one location are lost or
function whereby the gangway is lifted and slewed away from the rendered unusable), but is in line with the so-called N+1 philosophy.
landing platform in order to avoid damage. Stored energy in hydraulic The latter simply provides one lifeboat in excess of the number required
accumulators is used to accomplish these operations at high speed. to accommodate all the personnel onboard. Such exemptions typically
The first alarm is typically set at stroke of +5 m and -4 m and require full compliance with IMO MODU Code when in transit (i.e.
triggers an audio-visual alarm on the navigation bridge (central control restricted PoB), and a formal evacuation analysis for situations when
room). The second stage alarm is typically set at +6.5 m and −5.5 m the full complement is onboard. On these new vessels, the freefall
(or ± 6 m) and triggers the audio-visual alarm on the bridge and in the lifeboats are supplemented by escape chute and raft systems.
local gangway control cabin. An audible alarm and red flashing light is
activated at each end of the gangway to discourage further use. The 9.3. Cranes
auto-lift is typically set at +7.5 m and −6.5 m (or ± 7 m).
In addition to the automated disconnect function, the gangway is All the floatels are fitted with two deck cranes, usually of a modest
subject to management intervention depending on the frequency of size, capable of handling stores to and from supply vessels. All of the
occurrence of certain motions, based loosely on a Rayleigh distribution cranes are offshore type fitted with fast auxiliary/whip hoist to aid
of extreme response. The gangway would be disconnected if two these operations, and should be rated with a dynamic amplification
movements exceeding 5 m occurred in 10 min, or on account of any suitable for ship-to-ship service. Some of the floatels have one crane of a
single stroke of 6 m or more. Similar criteria exist for the vertical an- larger load rating, to increase the possibility of assisting the client
gular motions. platform owners with construction activities. These cranes are typically
On all of the semi-submersible floatels in operation so far, the height not rated for subsea operations, and do not have heave compensation.
of the gangway pedestal can only be adjusted by dismantling and in- Table 31 summaries the size and rating of cranes fitted on the new
serting/removing sections. This is frequently required when the vessel vessels, which do not differ significantly in performance from those of
moves from one client platform to another and can be a time consuming the previous generation.
and expensive operation. In the North Sea, some platforms are designed Almost all of the cranes selected are electro-hydraulic lattice boom
for a large wave clearance and can be very high, whereas some FPSOs in wire luffing type in order to save weight, respecting the weight sensi-
Brazil may have main deck heights similar to the deck of the floatel tivity of the semi-submersible concept. However, on Safe Nova/Vega
which vary depending on cargo loading. On the Safe Notos, the one of the cranes is a box section ram luffing type. The heavier weight
gangway pedestal and associated stair tower are designed in sections so of this type was accepted in order to gain the improved dexterity and
that the gangway can be landed at heights from 15.7 m to 32.2 m above controllability for deck load handling provided by the knuckle boom
the operating draft waterline. Some monohull floatels (e.g. Edda Fides) feature included.
have telescoping pedestals to avoid this problem, and it would be In addition to these main cranes, some of the vessels are fitted with
Table 32
Analysis of Installed Potable Water System components – various vessels.
Potable water pumps Hot water calorifiers Hot water circulating pumps
Typical installation sparing concepts 2 × 100% or 4 × 50% 2 × 100% or 4 × 50% 2 × 100%

Capacity of 100% system (average) 122 L per person per hour 10.3 L per person 22.5 L per person per hour
Total installed capacity (average) 290 L per person per hour 21.2 L per person 44 L per person per hour
784
one or two small pennant handling cranes, to assist mooring operations offshore. The laundry rooms occupy reasonably large areas in the ac-
in deck areas where the main cranes cannot reach. This may be on commodation, as indicated in Table 7. A common basis for selecting the
account of mooring winches being located underneath the helideck or number and capacity of washing machines is to calculate the total re-
other reasons. A typical specification for such a pennant handling crane quirement based on 2.5 kg per person per day. By then assuming that
would be an outreach of 18–20 m in order to reach the anchor handling each machine can deliver 9 or 10 wash cycles per day (allowing 80 min
vessel. The required capacity of the crane at these radii is rather per cycle) the number of machines required can be determined. Some
modest, as the weight of a pennant wire should not exceed 2.5 tonnes, spare capacity is usually installed in case of breakdown. The number
but a suitable dynamic amplification factor should be included. and capacity of the dryers should match the washing machines.
Such cranes do not require a cabin, and are usually arranged with a
telescopic or knuckle boom or both in order to minimise storage issues.
The main hoist should be single fall only (as is the case for the whip 10. Project execution
hoist on the main cranes), and equipped with an automatic overload
protection system (AOPS). Also similar to the main crane whip hoists, 10.1. Contracting
high speed operations are required; such as luffing in 60 s, hoisting with
an empty hook at 80 m/min and hoisting on load at 50 m/min. As indicated in Table 1, all the new vessels were constructed in
Singapore and China, with two yards (Keppel, COSCO Qidong) deli-
9.4. Fresh water generation, sewage treatment and laundry equipment vering 11 of the 14. None were built in Korea, probably because the
Korean yards were busy with other offshore projects during this period,
On vessels with such high personnel complements, substantial and and they judged (correctly) that few of the accommodation vessel
reliable fresh water generation and distribution systems are essential, projects would develop into a series of more than 2 units.
despite the fact that the units are regularly serviced by supply boats The period in which most of these vessels were constructed coin-
from shore. cided with a boom in construction of offshore vessels, and shipyard
Fresh water makers are typically of the reverse osmosis type, and capacity was severely strained. This meant that these floatel projects
installed with a sparing concept of 2 × 100%, 3 × 50% or 4 × 50%. had to compete for shipyard resources against other projects, often for
The Floatel Superior/Endurance can generate 225 tonnes per day, while owners with long series of more valuable projects (e.g. drilling units).
the Floatel Victory/Triumph may produce 250 tonnes per day. All fourteen of the vessels were constructed under turnkey design
On the Safe Notos, three fresh water makers are installed, each with and build contracts with the shipyards. In the case of Floatel
150 tonnes/day capacity. Four potable water transfer pumps can take International, and possibly the OOS Tiradentes, the shipyard had some
suction from the four storage tanks at a rate of 35 m3/h per pump (the shareholding in the vessel owning company.
pumps may operate in parallel to deliver 70 m3/hr). In the case of the five Floatel International vessels and the OOS
Hot water calorifiers and hot water circulating pumps are equally Tiradentes, the shipyard possessed the basic design capacity in-house or
critical. These must be sized for peak demand at shift changes and meal via an existing joint venture. For the eight other vessels, the owner first
times, rather than the average consumption per day. Table 32 contains developed a basic design together with an independent design com-
typical sizing data extracted from analysis of a number of offshore pany. The bidding shipyards were then requested to quote on the basis
vessels – not only floatels. of that design. After contract, the basic designer then worked under the
Sewage treatment plant is equally important, and adequate re- control of the shipyard, with a scope of work which varied from project
dundancy and capacity must be provided. A typical installation would to project.
include 3 × 50% or 2 × 100% plant, with 100% based on the combined
black and grey water generated by the personnel onboard.
As a minimum, the equipment must comply with the IMO MEPC 10.2. Scope
159/55 amendment. Care is needed when specifying the plant to not
simply specify the numbers of personnel, but the total hydraulic load Each of these vessels represents a substantial and complex project in
and the total organic load per day. There are no rules specifying how its own right. Some of the key project volume metrics are listed in
such load may be calculated, but useful guidance may be obtained from Table 33. Perhaps surprisingly, the cable and pipe statistics exceed
the recommendations of the Federal Maritime & Hydrographic Agency those of some drilling semi-submersibles built in the same yards during
(BSH) in Germany and some of the manufacturers. the same period. The large numbers of cabins on the floatels have a
The laundry is a similarly important facility on such vessels, al- direct multiplying effect on the quantities of IT and entertainment
though in Norway all laundry is sent ashore for processing and return cables and domestic piping systems.
Table 33
Comparison of Key Volume Metrics for two different vessel designs.
Vessel 1 Vessel 2
Electrical installation Cable length installed (m) 875,000 926,220

Cable terminations 44,959 54,302
Average cable length (m) 38.9 34.1
Piping systems construction Pipe spools 29,312 26,658
Pipe system hydrotests 722 324
Compartment completion LQ spaces 559 –
Technical rooms 217 –
Tanks 140 46 in columns
46 in pontoons
Completion and commissioning Control system I/O loop checks 9289 10,837
MC test packs 5447 2662
Commissioning test packs 314 270
Sea trials test memoranda – 117
785
Fig. 23. Durations to delivery from contract and cutting steel.
Table 34
Durations from contract and steel-cutting to delivery.
12 accommodation semi-submersibles built in Singapore and China 22 drilling semi-submersibles built in Singapore
Range Average Range Average
Contract to delivery (months) 25.8 to 38.9 33.6 33.2 to 50.9 40.9

Cut steel to delivery (months) 22.8 to 34.9 27.9 27.1 to 41.4 31.1
10.3. Construction method 10.4. Durations
All of the vessels were constructed by means of lifting the deck box Fig. 23 shows the times from contract to delivery and steel cutting to
in large megablock sections on top of the already assembled pontoons delivery for 12 of the vessels. The Nova and Vega are excluded from this
and columns. In the case of the projects in Singapore and COSCO comparison as their construction was significantly delayed for a number
Qidong these operations were carried out using large floating cranes, of reasons, including market conditions. Market conditions may also
and the deck structure was typically erected in 3 or 4 megablocks have resulted in smaller delays to one or two of the other vessels.
weighing 1000–1600 tonnes. In the case of the CIMC Raffles build, the Taken overall, there is no clear improvement between follow-on
20,000 tonne capacity Taisun crane could be used to erect the upper vessels and first of class. The times from contract to delivery are
hull with a single lift. sometimes longer for the follow-on vessels, because these were some-
With sufficient preparation it was sometimes possible to complete times contracted at the same time as the lead ship. Similarly, the times
the lifting of all four megablocks in four days. On at least one project from cutting steel to delivery for the follow-on vessels do not show a
the dimensional control in prefabrication was so good that it was pos- clear improvement compared to the performance for the lead ship. This
sible to call for the pre-welding fit-up inspection one day after the last can be due to capacity constraints and other priorities at the shipyards.
lift. Table 34 shows that in Singapore the typical time from cutting steel
At Jurong Shipyard, the pontoons were placed afloat (by traditional to delivery for a drilling semi-submersible was 31 months. This is 3
dynamic launching) and then placed in drydock where the bracings months longer than for the Singapore built floatels. From contract to
were installed and the columns erected. Keppel Shipyard used a variety delivery, the drilling rigs took 6 months longer than the floatels. These
of lower hull assembly methods, including a floating dock with erection differences are primarily due to the extra time needed to deliver, in-
carried out by a floating crane. Keppel was able to make use of hull tegrate and commission the drilling equipment, which is more complex
elements supplied by subsidiary yards in South East Asia, with complete than anything installed on the floatels.
pontoons towed from China and the Philippines, and other blocks de-
livered from these locations and Thailand. 11. Conclusions
At COSCO Qidong, the pontoons and columns were erected on land
and connected by the bracings, before the complete assembly was After a hiatus of more than 20 years, a major fleet renewal pro-
carefully skidded onto a submersible barge from which it was then gramme has been carried out in the offshore accommodation semi-
undocked, ready for erection of the deck box. submersible industry. Despite a variety of owners, designers and ship-
The thrusters were installed immediately prior to sea trials, with the yards being involved, the 14 new vessels display certain common
vessel spread moored at a sufficiently deep water location away from characteristics. The new designs offer greater comfort for the offshore
the shipyard. This meant that the vessel was practically unable to return workforce by means of single or two person cabins. While significantly
to the yard after completion of sea trials, so the construction and testing larger and heavier than their 1970s/80s predecessors, the new designs
had to be more complete at commencement of sea trials than might have relatively shorter columns and relatively greater waterplane area
otherwise be the case. to displacement ratios, which results in different hydrodynamic
786
response. Whereas the previous generation focused on station-keeping Christiansen, P.E., et al., 1994. Tender assisted drilling in the north Sea. In: OTC paper
by means of mooring, the new designs are all capable of dynamic po- 7458, Offshore Technology Conference, Houston.
Gallala, J.R., 2013. Hull Dimensions of a Semi-submersible Rig. MSc thesis. Norwegian
sitioning operations, with mooring as an option. The paper reviews University of Science and Technology, Trondheim.
principal design data on subjects such as internal volume and deck area, ISO Code 19901-7, 2013. Petroleum and Natural Gas Industries - Specific Requirements
weight estimation and stability, motions and station-keeping beha- for Offshore Structures – Part 7: Station-Keeping Systems for Floating Offshore
Structures and Mobile Offshore Units. International Standards Organisation.
viour, resistance and propulsion. Jenssen, N.A., 2010. Mitigating excessive pitch and roll motions on semi-submersibles. In:
Dynamic Positioning Conference. Marine Technology Society, Houston.
Acknowledgements Kanerva, M., et al., 2000. The Future of Ship Design by Deltamarin. MPI Group, UK
published by.
Lamas, M., Carral, L., 2011. Offshore and coastal floating hotels – flotels. Transactions
Thanks are due to Prosafe Offshore for permission to publish the 153 Royal Institution of Naval Architects.
paper. The views expressed are those of the authors alone and do not MacGregor, J.R., et al., 2001. Detailed Design and Construction of DP Drillships for
Deepwater Operations. Transactions, Royal Institution of Naval Architects.
represent those of any of the organisations mentioned in the paper. The
Odland, J., 1990. Mooring of the Veslefrikk B close to the wellhead platform. In: OTC
contributions of the many personnel from Prosafe, GVA, GustoMSC, paper 6305, Offshore Technology Conference, Houston.
Jurong Shipyard and COSCO Qidong who participated in the realisation Penney, P.W., Riiser, R.M., 1985. Preliminary Design of Semi-submersibles. Transactions,
of the projects which are the central topic of this paper are gratefully North East Coast Institution of Engineers & Shipbuilders, vol. 101.
Pessoa, J.M., Moe, A.M., 2017. Air gap on semi-submersible MODUs under DnVGL Class –
acknowledged. Current and future design practice. In: OTC paper 27693-MS, Offshore Technology
Conference, Houston.
References PSA, 2016. Investigation of an Incident with Fatal Consequences on COSL INNOVATOR
30th December 2015. Petroleumstilsynet, Norway.
Ren, H.H., et al., 2001. Vessel capability sea trials for dynamically positioned mobile
Andersen, T.L., et al., 2014. The FLOATEL SUPERIOR loose anchor incident and its sig- offshore drilling units. In: Dynamic Positioning Conference, Marine Technology
nificance for design and operation of semi-submersibles. In: OMAE Paper 2014- Society, Houston.
23974, ASME 33rd International Conference on Ocean. Offshore and Arctic Szalewski, P., et al., 2017. Horizontal wave impact loads on column stabilised semi-
Engineering, San Francisco. submersibles operating in harsh environment. In: OTC paper 27774-MS, Offshore
Atlar, A., 1987. Towards the understanding of steady tilt phenomenon in semi-sub- Technology Conference, Houston.
mersibles. In: 6th Offshore Mechanics and Arctic Engineering Symposium (OMAE Tamaki, I., et al., 1984. Full scale measurement tests on the new semi-submersible
1987), Houston, Texas, 1987. polycastle. In: OTC Paper 4731, Offshore Technology Conference, Houston.
Browne, V.C., 2013. Assessment of Low Frequency Roll Motions on the Semi-submersible Voogt, A.J., et al., 2002. Mean and low frequency roll for semi-submersibles in waves. In:
Drilling Rig COSL PIONEER. Thesis. Norwegian University of Science & Technology, Paper No. 2002-JSC-285, ISOPE Conference, Kitakyushu, Japan.
Trondheim. Voogt, A.J., Soles, J., 2007. Stability of deep-water drilling semi-submersibles. In: 10th
Burnell, A.J., 1975. Comparative performance of semi-submersibles in the North Sea. In: PRADS Symposium, Houston.
Society of Petroleum Engineers paper 5292.
787
View publication stats

Semi-Submersible Design - A New Generation of Offshore Accommodation

Uploaded by

Copyright:

Available Formats

Semi-Submersible Design - A New Generation of Offshore Accommodation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Semi-Submersible Design - A New Generation of Offshore Accommodation

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Semi-submersible Design - a New Generation of Offshore Accommodation

Article in Ocean Engineering · January 2019

Historic Ships & Boats View project

Floating Production Vessels View project

The user has requested enhancement of the downloaded file.

Contents lists available at ScienceDirect

Semi-submersible design – A new generation of offshore accommodation T

ARTICLE INFO ABSTRACT

1. Introduction Tamaki et al., 1984), and demonstrated an ability to remain safely

COSCO for the drilling/intervention vessel Island Innovator (delivered in

1.4. Other competitor vessels

Regalia (1985) 95.4 (87.6) 91.6 (60.83) 21,030 19,260 17,624 –

Floatel Reliance 12.2 7.9 18.9 24.69 6.7 11.0

Regalia (1985) 17.5 14.0 25.5 31.0 8.0 11.5

Floatel Reliance 6 3 6.09 m (pontoon top), 1.83m below survival draft

Floatel Reliance 88.7 36.07 3200 1 4.57 1.22 5.79 18,524

Fig. 6. Typical pontoon arrangement (Safe Boreas).

Fig. 7. Typical Column Arrangements (Safe Nova, forward starboard column).

Fig. 9. Typical main deck arrangement (safe Notos).

Two alternative pedestal locations – operating with gangway at

Deck box is strengthened to permit gangway pedestal at corner

Two alternative pedestal locations – completed with gangway

Vessel may operate parallel to or perpendicular to platform

Vessel may operate parallel to or perpendicular to platform

Vessel may operate parallel to or perpendicular to platform

Offices for client 0.66 to 1.18 0.85

Exhausts fitted with water spray system

on end near centreline

number of cabins due to the single occupancy requirement. In most of

Vessel end – away from LQ – near

Vessel end – away from LQ – near

Vessel end – away from LQ – near

Amidships P&S – away from

Vessel end – away from LQ

3.5. Machinery spaces

board rooms, or transversely through watertight doors. In the Safe

Selectable (if corner location used)

(and presumably the DSS20 designs) because of high stresses in the

Parallel to surge axis

Parallel to sway axis

different designs with approx. 30 MW installed power. The engine room

End centre-line OR corner (both away

End – centreline – away from LQ

derestimated in the basic design process, and consequently problems

phase when the actual sizes of the necessary equipment becomes

Corner – away from LQ

Corner – away from LQ

known. A typical HVAC installation for the LQ spaces alone of a North

dampers, 21/22 inlet/outlet louvres and 21 local control units.

provided at the basic design stage it was found to be sufficient.

Safe Notos, Safe Eurus

Safe Nova, Safe Vega

3.6. Watertight subdivision

In order to maintain compartments sufficiently small for damaged

Mess Room 0.44 to 1.28 0.87 0.46 to 1.20 0.79

Floatel Reliance 171 – – – 500 500

installations of watertight (WT) sliding doors. For example, the Safe

A major consumer of internal volume in the pontoons and columns