Post on 13-Oct-2015
description
5/24/2018 OTC 25405 MS TNTurner (Submitted)
1/15
OTC 25405-MS
Analyt ical Development of Seastate Limits Associated to the AbandonmentOperations of a Structurally Compromised Subsea WellT. N. Turner and N. Murdoch, Atkins Energy - Oil & Gas, 920 Memorial City Way, Suite 700, Houston, TX 77024.
Copyright 2014, Offshore Technology Conference
This paper was prepared for presentation at the Offshore Technology Conference held in Houston, Texas, USA, 58 May 2014.
This paper was selected for presentation by an OTC program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not beenreviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect any position of the Offshore Technology Conference, itsofficers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Offshore Technology Conference is prohibited. Permission toreproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of OTC copyright.
Abstract
When drilling a subsea well, a loss of structural integrity of the conductor can lead to a substantial reduction in the well's
ability to resist the axial and lateral loads imposed by the drilling riser and BOP. This can then lead to hazardous conditionsand increased risks to the asset performing the drilling as well as the personnel onboard. This may leave the operator with no
choice but to abandon the well. In this situation, the hazards associated with the loss of integrity must also be considered
when planning the abandonment operation itself. This paper presents a case study that details the analysis methods used to
develop seastate limits for the abandonment of a subsea well with compromised structural integrity.
The well's high pressure and low pressure housings were observed by ROV inspection to have become un-latched during
drilling. The inspection also revealed substantial conductor subsidence which suggested a failure of the 30 conductor and
20 surface casing cement jobs. Initial investigations indicated the well lacked the required levels of structural integrity to
resist the loads involved with drilling to total depth. As a result, the decision was taken by the operator to suspend and then
permanently abandon the well.
This paper discusses the technical approach used to develop the seastate limits imposed during the abandonment operation.The methodology utilized advanced non-linear finite element analysis techniques in conjunction with a novel statistical
treatment of previously recorded in-situ BOP motion data.
Introduction
An operator began drilling a subsea well in a water depth of approximately 105m using a traditional semi-submersible
drilling rig. While running the 13-3/8" casing string, a remotely operated vehicle (ROV) observed movement in the BOP
bulls-eye during a routine survey. Further inspection identified that the high pressure (HP) and low pressure (LP) housingshad become un-latched and were separated by approximately 0.6m. The separation also suggested a failure in both the 30"
conductor and 20" surface casing cement jobs. The conductor had insufficient axial support given it had subsided downward
into the seabed after becoming unlatched from the HP housing. The conductor continued to subside, coming to restapproximately 1.5m below the original mudline elevation.
Following the discovery of the wellhead separation, Atkins was commissioned to undertake a series of structural analyses to
establish the consequences of the wellhead separation and provide guidance as to whether it was appropriate to continuedrilling. This culminated in the well being temporarily suspended citing concerns over the ability of the 20" surface casing to
support the weight of the freestanding BOP should there be a need to disconnect the riser and the lower marine riser package
(LMRP).
In the months following the suspension, the operator commissioned Atkins and the wellhead supplier to perform a number ofstudies to investigate the feasibility of recovering the well and to evaluate possible recovery schemes. Although initial
conceptual studies indicated that a recovery scheme might have been feasible from a technical standpoint, the operator
advised Atkins that they had taken the decision to permanently abandon the well.
5/24/2018 OTC 25405 MS TNTurner (Submitted)
2/15
2 OTC 25405-MS
To prepare for the abandonment, Atkins was contracted to perform further structural analysis of the well in order to define the
seastate conditions (i.e. permissible wave height and wave period combinations) where it would be safe to re-latch the BOPand drilling riser to the well to facilitate the abandonment. The seastate limits would assist the operator with managing the
risk associated with the abandonment and ensure that the structural and pressure containment integrity of the well did not
become further compromised.
System Description and General Behavior
While drilling the well, a typical top tensioned riser system was employed. It consisted of a tensioner system, an upper flexjoint (UFJ), and a telescopic joint, a series of slick and pup joints, a lower flex joint (LFJ) and a subsea LMRP/BOP stack.
The space-out of the riser system is presented as a simplified schematic in Figure 1a. A schematic of the subsea equipment
stack latched to the well is shown in Figure 1b. This shows the top of the LP housing 1.5m below the mudline elevation
corresponding to a wellhead separation of approximately 3.8m. This is the separation that was recorded prior to suspension of
the well.
(a) (b)Figure 1 Riser Space-out (a) and Post-separation Well Space-out (b)
The conductor consisted of a 30x1.5 X52 LP housing extension joint followed by three 30x1 X52 conductor joints
followed by a 30x1 X52 shoe joint. The conductor was originally drilled and set in a 36 pre-drilled hole. The surface
casing string comprised a 21" HP housing extension followed by a series of 40ft long joints of 20"x0.625 X56 casing. All
joints were connected using threaded connectors. The stick-up to the top of the HP housing was approximately 3.2m. Thewell had a 13-3/8" P-110 intermediate casing string landed on the wellhead just before it was suspended. There was a degree
of uncertainty in both the magnitude and the orientation of the conductors initial inclination though available evidence
suggested that it was unlikely to be much more than 0.5.
LMRP
BOP
15.73m
10.57m
17.33m
101.15m
96.58m
90.48m
78.29m
3.2m
0.0m
119.41m
129.37m
131.40m
128.23m
105.0m
Mudline
HPHousingDatum
TopofBOPLFJCentreofRotationTopofRiserAdapter
DrillFloorUFJCentreofRotationTopofTJ InnerBarrelTopofTJOuterBarrel
MSL
Topof15'PupJointTopof20'PupJointTopof40'PupJoint
Topof4x50'SlickJoints0.0m
1.50m
14.54m
26.73m
38.92m
51.11m
63.30m
3.2m
10.56m
15.72m
17.33m
LMRP
BOP
MudlineLPHousing
30"RL4
30" RL4
30" RL4
30" RL4
30"Shoe
WHDatum
TopofBOP
LFJC.o.R
TopofAdapter
1069m20"Shoe
230m20"Bondedto
Formation
3.36m20"RL4S
35.00m20" TOC
5/24/2018 OTC 25405 MS TNTurner (Submitted)
3/15
OTC 25405-MS 3
The soil in the region was known to consist of very loose, silty sand at the surface overlying strata of stiffer clay and sand.
This was based on the findings of a borehole. The support provided by the soil in the immediate vicinity of the well was
however subject to some uncertainty given the subsidence of the conductor. However, the dynamic response of the well andthe stresses generated in the various components were found to be governed by the movement of the surface casing within the
conductor. As such, the degree of lateral support provided by the soil was not believed to fundamentally influence the
response of the well in the subsided configuration.
The postulated behavior of the well under lateral riser loading is described in Figure 2. This figure illustrates the tendency forthe HP housing extension to bear against the inside of the LP housing forming a pivot point. Depending on the severity of the
response, secondary contact may develop further down inside the conductor. At any given instant, the BOP inclination ismade up of a dynamic component superimposed on top of a mean static component. The dynamic component is generated by
wave induced vessel motion transmitted to the BOP via the riser.
Initially, the BOP and un-cemented 20" casing freestand witha very low lateral stiffness. This means a small lateral load atthe LFJ causes the BOP to lean over and take up a stableposition with the outside of the HP housing extension incontact with the inside of the LP housing.
Once the HP housing extension makes contact with the insideof the LP housing, the BOP effectively pivots about the point ofcontact. The relatively low lateral stiffness of the BOP and 20"casing means that the BOP angle increases at a similar rate tothe lower riser angle. The BOP angle remains approximately0.1 higher than the lower riser angle due to the initial BOPangle of 0.2 developed prior to the HP housing extensioncoming into contact with the inside of the LP housing.
Secondary contact occurs approximately 12m below mudlineresulting in an increase in the lateral stiffness of the system.The rate of increase in BOP angle is much reduced beyondthis point such that the lower riser angle becomes greaterthan the BOP angle.
Figure 2 General Behavior of Well and Subsea Equipment Stack under Lateral Riser Loading
One of the inputs that was thoroughly investigated in earlier assessments was the top of cement (TOC) level in theconductor/surface casing annulus. A distinction was made between the elevation where the cement was considered to be
sufficiently competent that it bonded the conductor and surface casing together (or bonded the surface casing to the
formation) and the elevation where there was degraded cement in the annulus acting as a spacer.
The elevation where cement was considered to be sufficiently competent to bond the conductor and surface casing together
was estimated to be 230.0m below mudline. This was based on the downward movement of the 20" surface casing recorded
Mudline
TOC
Riser
Tension
Contact
Gravity
Riser anglelessthan
BOPangle
Mudline
TOC
Riser
Tension
Contact
Gravity
Riser angle
similarto
BOPangle
Secondarycontact
Mudline
TOC
Riser
Tension
Contact
Gravity
Riser anglegreaterthan
BOPangle
5/24/2018 OTC 25405 MS TNTurner (Submitted)
4/15
4
c
t
a
i
r
f
i
a
t
a
hen the 13-3/asing was con
he elevation
o be 35.0m t
elow. For the
nnulus acting
ecorded Da
efore drilling
arameters rel
tention wasowever, follo
ecovering the
timeline of
5th
25th
26
th
5th
8th
plot showin
arked increas
elieved to hainal resting pl
itial drilling
bserved durin
essel motion
eave) were re
eriod (Tz) der
eviation of ve
comparison
nd sway moti
ote is that theat the record
t the well cenas making to
ecordings fro
0.00
0.10
0.20
0.30
0.40
0.50
0.600.70
0.80
0.90
1.00
05/04
StandardDeviatio
n(deg)
8" casing wassidered to be e
here there w45.0m belo
purpose of th
as a spacer bel
ta Available
the well, th
ting to the r
to use the rewing the well
well. The data
ajor events d
pril
April
Aprilay
ay
the hourly s
e in the inclin
e first separatace, 1.5m bel
operation, the
these five da
(heave, surge
corded by se
ived from the
ssel surge and
between Figu
ons were elev
vessel motiod surge moti
er was somethe recorded s
strain gauge
06/04
07/04
08/04
09/04
10/04
landed. As thffectively bon
s cement in t mudline bas
e abandonmen
ow 35.0m bel
From Origin
operator de
sponse of th
corded data tead failure, t
was then used
ring initial dri
: BOP l
: Point
: ROV i: Condu
: BOP
andard deviat
tions is evide
d. Between thw mudline.
behavior of t
ys.
Figure
sway, roll, pi
sors onboard
water surfacesway in the 0.
e 4 and Figur
ted and times
sensor was lns provided a
hat distorteday, but doub
s located on t
11/04
12/04
13/04
14/04
15/04
Pre Sep
depth of 230.ded to the for
e annulus actid on a revie
t analysis, it
w mudline.
al Drilling
ided to use i
drilling vess
o calibrate ee data was us
as input to th
lling of the we
nded on HP
here housing
nspection firstctor comes to
nlatched
ion of BOP i
t on the 25th
e 3rdof May as the same dr
e riser and w
Hourly Stan
tch and yaw)
the rig. Time
elevation reco
5s to 20s band
e 5 indicates t
where the sig
cated far forgood reflectio
y the vessels over the reli
e lower riser
16/04
17/04
18/04
19/04
20/04
ration
0m is below tation and rest
ng as a spacer of the recor
as assumed t
peration
t as an oppo
el, riser, and
isting riser aed as input to
abandonmen
ll is presented
ousing
s are first beli
observes seprest with top o
clination in t
f April, whic
nd the 8thofilling rig was
ell system du
ard Deviation
and water sur
histories of s
rdings are pre
width are pres
here is a gene
nificant wave
ard but onlyn of the surge
aw. Attemptsbility of the r
ere used to d
21/04
22/04
23/04
24/04
25/04
Date
e conductorrained in the a
was subject tded BOP and
at there was
tunity to imp
BOP stack.
alysis methothe analysis u
study.
below for ref
ved to have s
ration betweef LP housing
o orthogonal
corresponds t
ay, the top oto be used fo
ing abandon
of BOP Inclinat
ace elevation
ignificant wa
sented in Figu
ented in Figur
ral correlation
height was in
lightly starboat the well ce
were made toecorded yaw
velop a time
26/04
27/04
28/04
29/04
30
04
Sepa
hoe, this is thxial direction.
some uncertlower riser i
ement in the
lement a cam
etocean data
dologies andndertaken to a
rence.
parated
housings.5m below m
directions is
o the instance
the LP housithe abandon
ent was expe
ion
(compensated
e height (Hs)
re 4. Time his
5.
between time
excess of 1.0
rd of the wellnter, but the r
calculate theotions meant
istory of the l
01/05
02/05
03/05
04/05
ration
OT
depth where
inty. This waclination data
conductor/sur
paign to reco
was also rec
improve theirscertain the fe
dline
resented in F
where the hou
g was withinent as was u
ted to be sim
for the effect
and mean ze
tories of hourl
s where the v
. An importa
s centerline.presentation
ontribution thhis was not s
wer riser tens
06/05
07/05
08/05
09/05
RX
RY Un
25405-MS
he surface
estimateddescribed
ace casing
rd various
rded. The
accuracy.asibility of
igure 3. A
sings were
0.5m of itssed for the
ilar to that
s of vessel
o crossing
y standard
essel surge
nt point to
his meantf the sway
at the yawccessful.
ion.
11/05
12/05
13/05
BOPlatched
5/24/2018 OTC 25405 MS TNTurner (Submitted)
5/15
I
c
s
TC 25405-MS
evelopmen
he objectiveperation coul
t was initially
ave response
essels excursenchmark the
omplex and d
lthough thereeant the use
tatistical treatroposed inste
1) Use tinclin
2) Establvesselthe L
aband
3) Establcrossi
the 274) Use t
and di
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
14/03
SignificantWaveHeight,Hs
(m)
0.00
0.05
0.10
0.15
0.20
0.25
0.30
14/03
StandardDeviation(m)
of Analysis
f the analysibe carried ou
intended that
simulations t
ion from wellresponse of t
namically se
was confidenof FE modeli
ment of thed. The metho
e riser and wetion where ke
ish relationshi
surge and swhousing was
onment).
ish relationsh
g period andthof May.e above relati
rection combi
18/03
20/03
22/03
24/03
18/03
20/03
22/03
24/03
Figure 4 Si
Figure 5
Methodolo
was to establt without furth
beam type fi
o facilitate th
center line) ae FE model a
sitive nature o
ce that the FEng alone to e
eastate, vesseology propos
ll system beay well compo
ps between th
y. This step cbetween 1.0
ps between t
he wave direc
onships to est
ations.
28/03
30/03
01/04
03/04
05
04
28/03
30/03
01/04
03/04
nificant Wave
ourly Standard
y
ish, with reaser compromisi
ite element (
e developmen
d seastate chagainst the BO
f the BOP res
model couldtablish dyna
l motion, andd was as follo
FE model alents become
e BOP dyna
nsidered dataand 1.5m be
e vessel sur
tion. This step
blish the max
07/04
09/04
11/04
13/04
15
04
07/04
09/04
11/04
13/04
15
04
Height and Me
Deviation of V
onable confidng the structu
E) model of t
t of envelope
racteristics. Dresponse rec
onse was not
eventually beic seastate li
BOP motionws:
eady developverstressed.
ic inclination
recorded durilow the mudli
e and sway
considered th
imum expecte
17/04
19/04
21/04
23/04
25/04
Date
17/04
19/04
21/04
23/04
25
04
Date
n Zero Crossin
essel Surge an
nce, the seastal integrity of
he riser and w
describing s
ring an earlyordings descri
easily capture
calibrated, thits was aban
data recorde
d to perform
in the fore/af
g the periodne (i.e. simila
otion and th
e much larger
d BOP inclina
27/04
29/04
01/05
03/05
05/05
27/04
29/04
01/05
03/05
05
05
Housing Separat
and BOP Latche
Housing Separa
and BOP Latch
g Period
d Sway
ate conditionsthe well.
ell system be
fe combinati
phase of the sbed above, it
in the FE mo
limited naturdoned and a
d during the
imple static a
t and port/star
etween the 3rd
r to the positi
e significant
data set betwe
tion for a ran
07/05
09/05
11/05
13/05
15/05
07/05
09/05
11/05
13/05
15/05
ed
d
ed
ed
where the ab
sed to perfor
ns of vessel
udy, while attecame appar
del.
e of the projeethodology
initial drilling
alysis to deri
board directio
and the 8thofon it would b
ave height,
en the 14th of
e of wave hei
17/05
19/05
21/05
23/05
25/05
Hs
Tz
17/05
19/05
21/05
23/05
25/05
Sway
Surge
5
ndonment
irregular
offset (the
empting tont that the
t scheduleased on a
operation
e the BOP
ns and the
May whenin during
mean zero
March and
ht, period
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
27/05
MeanZe
roCrossingPeriod,
Tz
(s)
27/05
5/24/2018 OTC 25405 MS TNTurner (Submitted)
6/15
6 OTC 25405-MS
5) Based on knowledge of the BOP inclination where one of the well components first becomes overstressed (asderived from the FE analysis) the combinations of wave height, period and direction where the BOP inclination
would result in the stresses in the well components exceeding permissible limits could be determined.
The philosophy behind the above approach was that the operator would have weather forecast information available that
would define the upcoming significant wave height, period, and direction. By knowing whether such wave heights, periods,
and directions were likely to cause overstress of critical well components, informed decisions could be taken regarding
latching or disconnecting the BOP and LMRP from the well.
Derivation o f Permissible BOP Inclinations
An initial static analysis was executed to establish the variation in the structural utilization of the various components of the
well system as the BOP inclination increased. This enabled the BOP inclination where critical well components became
overstressed to be established.
The analysis was performed using a Flexcom (MSC Kenny, 2013) FE model that represented:
the floating vessel, characterized by prescribed displacements at the drill floor
the upper flex joint
the marine riser
the lower flex joint
the subsea drilling equipment the HP housing, HP housing extension and the surface casing modeled to the depth of the conductor shoe
the conductor and its interaction with the surface casing (captured using pipe-in-pipe elements)
non-linear p-y and t-z soil springs used to represent the axial and lateral restraint provided by the soil
The model was subjected to an incrementally increasing vessel offset in order to generate an incrementally increasing lateral
load at the LFJ and hence an incrementally increasing BOP inclination. At each increment, axial forces and bending momentswere extracted and used to calculate combined axial and bending utilizations of the various well components. Checks
incorporated allowable stress factors from API-RP2A-WSD.
The results of the analyses were used to produce a chart showing the relationship between BOP inclination and the utilization
of each component as shown in Figure 6. This indicated that the maximum permissible BOP inclination was 1.43 prior tocutting and pulling the 13-3/8 casing and 1.96 following removal of the 13-3/8 casing. These limiting inclinations
incorporate a 0.5 allowance for the initial inclination of the conductor.
Figure 6 Variation in Component Utilization wit h BOP Inclination
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Utilisation
BOP Inclination (deg)
21" HP Housing Extension
20"x0.625" X56 Surface Casing
30"x1.5" X52 LP Housing Extension
30"x1" X52 Conductor
30"x1.5" RL-4 Connector (95ksi)
30"x1" RL-4 Connector (70ksi)
21" HP Housing Extension (No13-3/8" Casing)
20"x0.625" X56 Surface Casing (No13-3/8" Casing)
30"x1.5" X52 LP Housing Extension (No13-3/8" Casing)
30"x1" X52 Conductor (No13-3/8" Casing)
30"x1.5" RL-4 Connector (95ksi) (No13-3/8" Casing)
30"x1" RL-4 Connector (70ksi) (No13-3/8" Casing)
5/24/2018 OTC 25405 MS TNTurner (Submitted)
7/15
OTC 25405-MS 7
It should be noted that the permissible BOP inclinations derived in this part of the assessment were also provided in the
operating guidance document presented to the operator at the end of the study. Although it was recognized that permissible
BOP inclinations were of less practical use in planning operations, their derivation from simple static analysis meant theywere considered to be reasonably robust and therefore served as support to the more uncertain seastate limits.
Another item derived from the Flexcom analysis was the BOP inclination where the 20x0.625 X56 surface casing reached
its full plastic capacity and starts to form a plastic hinge. This was established on a relatively simple basis by defining an
equivalent ultimate utilization as:
UC
where yis the yield stress, ais the axial stress, bis the bending stress and SFis the shape factor (the ratio between theplastic and elastic section moduli). The above unity check equation was required as the Flexcom model did not include
material plasticity. The checks identified that the surface casing would reach its ultimate capacity at BOP inclinations of
2.72 and 3.23 before and after removal of the 13-3/8 casing respectively. These values again include an allowance of 0.5
for initial out of verticality of the conductor.
It was recognized that the above approach was simplistic and would not capture the effects of local buckling. For this reason,
a detailed nonlinear finite element analysis (using 3D shell elements) was performed to better substantiate the BOP
inclinations where the ultimate capacity of the system would be reached. This is further described in the following section.
Further iterations of the Flexcom analysis were performed to establish the sensitivity of the maximum permissible BOP
inclination to variations in the riser top tension. This was aimed at identifying whether there were any further practicalmeasures that could be taken to increase the permissible BOP inclinations and, hence, extend the permissible seastate limits.
However, it was found that the permissible BOP inclinations were not particularly sensitive to riser top tension.
Derivation of Seastate Limits
Correlation between BOP Motion and Vessel Motion
General
The first part of the procedure used to develop the seastate limits involved deriving a relationship between the vessel motionand the BOP dynamic motion during the period between the 3 rdand the 8thof May (the period when the LP housing was
between 1.0m and 1.5m below mudline). This initially involved the derivation and comparison of a select number of vesselsurge/sway and BOP inclination spectra to establish if a correlation existed between the frequency content of the two signals.
The upper plots in Figure 7 present the vessel surge and sway spectra and the water surface elevation spectra for two alternate
one hour periods on the 4thof May and the 6thof May. The lower plots present the BOP inclination spectra during the same
periods. The first point to note from Figure 7 is that in both cases the majority of the BOP response occurs in the 5.0s to 10.0s
bandwidth. The second point to note is that in both cases the vessel motion spectra also show a significant amount of contentwithin this bandwidth. However, it is apparent that for the case shown in right hand plots of Figure 7, there is a significant
amount of vessel motion content in the 10.0s to 15.0s bandwidth. This does not appear to translate into any significant BOP
motion.
Review of further spectra indicated that the majority of the BOP response was in the 5.0s to 10.0s bandwidth irrespective of
whether the vessel exhibited a second response peak in the 10.0s to 15.0s bandwidth. The reason for this trend was believed
to be due to the natural period of the BOP stack lay in the 5.0s to 10.0s range. This meant that it was prone to excitation bythe vessel motion in this bandwidth and relatively unaffected by the vessel motion in the 10.0s to 15.0s bandwidth.
Based on the above observation, it was postulated that a relationship between the vessel motion and the BOP motion could be
established by comparing the standard deviation of vessel motion in the 5.0s to 10.0s bandwidth and the standard deviation ofBOP motion over the full bandwidth.
5/24/2018 OTC 25405 MS TNTurner (Submitted)
8/15
8 OTC 25405-MS
Vessel/Wave Motion Vessel/Wave Motion
BOP Response Spectra BOP Response Spectra
Record for 00:00 - 01:00 4th May Record for 14:00 - 15:00 6th May
Figure 7 Vessel Motion and Seastate Spectra (top) and BOP Response Spectra (bottom) for May 4th and May 6th
Development of Relationship
Following the above observation, the standard deviations of vessel motion in the 5.0s to 10.0s bandwidth and the standard
deviations of BOP motion over the full bandwidth for each hour between the beginning of the 3rdof May and the end of the
7thof May were extracted.
The left plot in Figure 8 presents the standard deviation of BOP inclination in the fore/aft direction plotted against the
standard deviation of vessel surge. The right hand plot presents the standard deviation of BOP inclination in the
port/starboard direction plotted against the standard deviation of vessel sway.
The blue data points in Figure 8 are taken from instances when the mean BOP inclination was in a direction similar to theplane of the BOP motion. For example, the blue points in the left hand plot are from instances when the BOP had a
prominent mean inclination in the forward direction and the blue points in the right hand plot taken from instances when theBOP had a prominent mean inclination in the starboard direction. The red data points in both of the plots in Figure 8 aretaken from instances when the BOP did not have a prominent mean inclination in the plane of motion. The trend lines in
Figure 8 are associated only with the blue data points.
The first point to note from Figure 8 is that in both cases, although there is some scatter in the data, there is a reasonable (andrelatively linear) trend between the vessel motion and the BOP motion when the mean inclination is in a direction close to
that of the dynamic motion. The correlation is believed to be better during these instances because the HP housing extension
is expected to have been pressed up against the inside of the LP housing with the BOP/HP housing/surface casing string
pivoting about this point in a reasonably uniform fashion. At instances when the mean BOP inclination was in a directionperpendicular to the plane of the dynamic motion, it is possible that there would have been more movement of the HP
housing extension and surface casing within the conductor such that the motion may have been less uniform. This may be
why the correlation with the vessel motion is not so good during these instances.
Majority ofvessel
response is in5s to 10sbandwidth.
Vessel exhibitssignificant
response in both
the 5s to 10sbandwidth andthe 10s to 15s
bandwidth. The
latter does nottranslate intoBOP motion.
RiserNatural
Period
Majority ofBOP
response is in5s to 10s
bandwidth.
Majority ofBOP responseis in 5s to 10s
bandwidth with
minimal BOPresponse in 10s
to 15s
bandwidth.
RiserNatural
Period
5/24/2018 OTC 25405 MS TNTurner (Submitted)
9/15
OTC 25405-MS 9
The second point to note from Figure 8 is that for motion in the fore/aft direction, the BOP inclination is 9.9 times the vessel
surge but for the port/starboard direction, the BOP inclination is 4.6 times the vessel sway. This implied the BOP wasapproximately twice as sensitive to vessel motion in the fore/aft direction as it was to motion in the port/starboard direction.
The reason for this is not entirely clear though three possible explanations were identified.
The first possible explanation is the natural period of the BOP stack is different in the two orthogonal directions, such that the
level of excitation caused by vessel motion in the two directions is not consistent. However, in order for the natural period tobe different in the two directions it would require either the foundation stiffness or the lateral mass of the BOP to be different
in the two directions. Work performed to calibrate the foundation stiffness suggests there was not much difference in thefoundation stiffness in the two orthogonal directions. Furthermore, the only way differences in the lateral mass could be
significant is if the BOP was substantially asymmetric (i.e. causing different hydrodynamic added mass under motion in the
two orthogonal directions). In practice, inherent uncertainty exists when estimating the added mass for such a
hydrodynamically complex structure; however, available BOP drawings gave no indication that the added mass would be
particularly different in the two orthogonal directions.
The second possible explanation is that the recorded vessel sway motions contained a yaw component, meaning these were
consistently higher than the sway at the drill center location. This would have resulted in the ratio between recorded sway and
BOP inclination in the port/starboard direction appearing artificially low in comparison to the ratio between vessel surge andthe BOP inclination in the fore/aft direction.
The third possible explanation is that the positioning of the choke and kill lines around the circumference of the riser could
have affected the load transmitted into the well in the two orthogonal directions.
Due to the complexity of the system response and the limited time frame that was available to complete the study, it was not
possible to definitively prove or disprove any of the above three theories. There remained a possibility that all three theories
could have contributed to the observed trends. Notwithstanding the inability to resolve the anomaly, the fact that the same,well configuration, BOP, riser system, vessel and vessel heading were to be used during the abandonment as were used
during initial drilling, meant that it was considered appropriate to adopt the directionally dependent relationships between
vessel motion and BOP motion in the assessment.
Figure 8 Relationsh ip between Standard Deviation of BOP Inclination wi th Vessel Surge (left) and Vessel Sway (right)
Correlation between Vessel Motion and Seastate
Initial Review
The second part of the procedure used to derive the seastate limits involved developing a relationship between the seastate
parameters (wave height, period and direction) and the vessel motions. This considered an extended data set covering the
period between the 14thof March and the 27th of May.
Plots showing the relationship between the standard deviation of surge and sway in the 5.0s to 10.0s bandwidth and thesignificant wave height are presented in Figure 9. Figure 9 shows there is a considerable amount of scatter in the data, which
suggests there were factors other than the significant wave height which were dictating the magnitude of the vessel motion.
5/24/2018 OTC 25405 MS TNTurner (Submitted)
10/15
10 OTC 25405-MS
The two most likely factors identified were the mean zero crossing period of the seastate and the direction of wave
propagation relative to the vessel.
Figure 9 Relationship between Wave Height and the Standard Deviation of Vessel Surge (left) and Vessel Sway (right)
Treatment of DirectionalityWave directionality recordings were not available. Therefore an attempt was made to establish the dominant wave direction
at each one hour instance using the ratio between the surge and the sway standard deviations in the 5.0s to 10.0s bandwidth.
The philosophy behind this approach was that in cases where the wave approached the rig in a head seas direction, the surge
motions would be significantly larger than the sway motions. Similarly, for cases where the waves approached the rig in a
beam seas direction, the sway motions would be significantly larger than the surge motions.
A plot showing the resultant motion in the 5.0s to 10.0s bandwidth normalized with respect to wave height plotted against the
dominant direction of the motion is presented in Figure 10. The first point to note from Figure 10 is that the data is
concentrated around a direction of 60 with none of the data points at an angle of less than 30. This could be interpreted tomean that the majority of the waves approached from a direction close to quartering. This was likely to be partly true because
the vessel was orientated to point North West and it was known from regional metocean data that a significant proportion of
the waves propagate from North to South and would therefore have approached the vessel from a quartering direction.However, it was thought the observation was also partly a result of wave spreading effects, which meant that even when the
dominant wave direction was in a predominantly head or beam seas direction, there was still an appreciable amount of vessel
motion in the direction perpendicular to the direction of wave approach.
A second point to note from Figure 10 is that the magnitude of the vessel motion is slightly higher when the dominant wavedirection tends towards 90. This suggested the vessel exhibited a more onerous response under beam seas waves than it did
under head or quartering seas waves. This was consistent with the response predicted by available analytically derived vessel
RAOs.
Based on the above, it became clear that it was not possible to conclusively define the direction of wave approach from the
vessel motions. It was, however, considered appropriate to split the data set into points that were consistent with a
head/quartering seas direction and points that were consistent with a beam seas direction.
Figure 11 and Figure 12 present plots showing the relationship between the standard deviation of surge and sway in the 5.0sto 10.0s bandwidth and the significant wave height for the head/quartering and beam seas dominated data points,
respectively. Although there is still considerable scatter in the data, it is appreciably less than was observed in Figure 10 prior
to splitting the seastates based on their directionality.
5/24/2018 OTC 25405 MS TNTurner (Submitted)
11/15
OTC 25405-MS 11
Figure 10 Resultant Vessel Motion w ith respect to Dominate Wave Direction
Figure 11 Relationshi p between Significant Wave Height in the Head/Quartering Sea Direction wi th Standard Deviation of VesselSurge (left) and w ith Standard Deviation of Vessel Sway (right)
Figure 12 Relationship between Significant Wave Height in the Beam Sea Direction with Standard Deviation of Vessel Surge (left)and with Standard Deviation of Vessel Sway (right)
Treatment of Wave Period
The effect of wave period was taken into account by further sub-dividing the directional data sets into 0.5s wide wave period
bands. For brevity, plots showing the relationship between the standard deviation of surge and sway in the 5.0s to 10.0sbandwidth and the significant wave height for alternate wave periods are not presented here.
Derivation of Seastate Limits
BOP Inclination Standard Deviations
The first activity involved with deriving the seastate limits was to establish the standard deviation of resultant BOP
inclination associated with a range of significant wave height and mean zero crossing period combinations for the twoalternate directional data sets (i.e. head/quartering and beam directions). This was achieved using the relationships between
BOP motion and vessel motion and the relationships between vessel motion and seastate described above. For each
Sway DominatedSurge Dominated
5/24/2018 OTC 25405 MS TNTurner (Submitted)
12/15
12 OTC 25405-MS
significant wave height and mean zero crossing period combination, the resultant standard deviation of BOP inclination,
BOP, was calculated as follows:
9.9 4.6
Where surgeis the standard deviation of vessel surge and swayis the standard deviation of vessel sway. The factors of 9.9 and4.6 were taken from the trend lines in the left and right plots in Figure 8, respectively. As noted earlier, the trend lines in
Figure 8 are associated with only the blue data points as these points are associated with the more uniform motion producedwhen the mean BOP inclination is in the same plane as the orientation of the dynamic motion. This is when the response was
expected to be most onerous such that neglecting the red points associated with less uniform motion was not considered
unreasonable.
Ratio between Standard Deviation and Maximum Inclination
The second activity was to convert the hourly standard deviations of resultant BOP inclination into an hourly most probable
maximum BOP dynamic inclination.
In assessment of offshore structures, it is common practice to assume the peaks in the water surface elevation are Rayleighdistributed so the most probable maximum amplitude in 1000 cycles (a typical number of waves in a 3 hour seastate) is equal
to 3.72 times the standard deviation (1.86 times the significant wave amplitude). It is only appropriate to apply this ratio toresponse variables if the peaks in the signal are also Rayleigh distributed.
The non-linear and dynamically sensitive nature of the BOP response, coupled with one hour instances being studied, meant
there were concerns that it may not be appropriate to apply the factor of 3.72 in the case. The recorded inclinometer data was
processed further to establish the ratio between the hourly maximum and the hourly standard deviation of inclination
experienced in practice. The results of this are shown in Figure 13.
Figure 13 Ratio between Hourly Peak BOP Inclination and Hour ly Standard Deviation of BOP Inclination
The average of the ratio over the period of interest was found to be 4.43, which was rounded up to 4.5 for use in deriving theseastate limits. It is acknowledged that the above approach is relatively crude and possibly non-conservative. It is also
possible that further analysis of the available data would have enabled a more statistically representative value for the most
probable maximum response in a one hour period to be determined. However, such analysis would have been highly time
consuming and it was not clear that it would result in a ratio significantly different to 4.5.
Derivation of Acceptable Seastates
The resultant BOP inclination standard deviations under a range of significant wave height, mean zero crossing periodcombinations were converted into most probable maximum BOP inclination amplitudes using the ratio defined above. These
were then presented in the form of a scatter diagram type plot as shown in Figure 14. Figure 14 was derived from the
head/quartering seas data set with a similar plot derived for the beam seas seastates.
The shading in Figure 14 is based on the following:
Green : BOP inclination such that stresses in all well components are within permissible limits.
Red : BOP inclination such that API allowable stresses are exceeded (BOP inclination in excess of 0.93).
5/24/2018 OTC 25405 MS TNTurner (Submitted)
13/15
OTC 25405-MS 13
Black : BOP inclination such that the ultimate plastic capacity of the surface casing is exceeded (BOP
inclination in excess of 2.22).
The inclinations of 0.93 and 2.22 defined above are both 0.5 less than those derived from the static FE analysis described
earlier. This was to provide an additional 0.5 allowance for mean inclination of the BOP resulting from excursion of the rig
from the well center.
Within allowablelimits
Allowablestressesexceeded
Failure ofsurface casingexpected
Figure 14 Most Probable Maximum Peak Dynamic BOP Inclination Ampli tude (degrees) for comb inations of Hsand Tz(Head/Quartering Seas)
The plot in Figure 14 is applicable to the scenario where the 13-3/8" casing is still in place. Additional (less restrictive) plots
applicable to the case after the 13-3/8" casing had been cut and pulled were also produced.
Derivation of Ultimate Capacity Using Nonlinear Finite Element Analysis
Given that the seastate limits were highly restrictive, there was a need to better understand the risk associated with exceedingthem. For this reason, further non-linear, finite element analysis of the conductor and surface casing was performed to better
define the ultimate plastic capacity of the upper surface casing. This would allow a better understanding of the BOP
inclination at which the surface casing pipe would be expected to locally buckle and collapse. In particular, it was consideredimportant to know how soon after the plastic capacity is reached that local buckling would occur and how abrupt the failurewould be. The analysis would also provide confirmation that local buckling of the 21 HP housing extension (where it bears
against the inside of the LP housing) would not occur prior to failure of the 20x0.625 X56 surface casing.
Due to the limited available timeframe coupled with the inherent complexities involved with performing non-linear buckling
analysis, it was agreed with the operator that a relatively simple approach would be taken. The main simplification made wasto approximate the conductor and the soil as being rigid. It was considered acceptable to make this approximation, as it was
known from previous work that the stresses in the upper surface casing were governed by the deformation of the surface
casing and HP housing extension within the conductor. The additional, and relatively small, deflections of the conductorwithin the soil have a secondary effect on the surface casing stresses.
The finite element model presented in Figure 15 was developed for analysis in ABAQUS (Dassault Systmes Simulia Corp.,
2012) and represents the inner and outer casing components installed between an elevation of 35m below the mudline (which
was the assumed TOC) extending upwards to the top of the HP housing, 3.2m above the mudline. The model exploits thesymmetric nature of the well structure using a single symmetric boundary condition applied through the center of the well
geometry. The FE model did not include the 13-3/8 casing. This is because the 13-3/8 casing would be expected to provide
a degree of restraint to the surface casing, limiting the consequences of any local buckling. As such, the scenario where the13-3/8 casing has already been cut and pulled was expected to represent the worst case for analysis.
As can be seen in Figure 15, the FE model includes only the inner surface of the 30 conductor, the LP housing, the LPhousing extension, and the 30 connectors. These components are all represented with rigid shell elements. Modeling the
conductor as a rigid shell meant that a bit of conservatism would be introduced to the results. This is because the lateral
contact force transmitted by the HP housing extension upon contact with the LP housing would be dissipated wholly through
deformation of the inner string. However, in reality, this lateral contact force would also result in some local and globaldeformation of the conductor as well as the surrounding soil.
5/24/2018 OTC 25405 MS TNTurner (Submitted)
14/15
14 OTC 25405-MS
The geometry of the inner string included the 20" surface casing, the 21 HP housing extension and the 20" connectors.
These components were modeled using deformable four node general purpose non-linear shell elements. The HP housingwas represented with non-linear beam elements because this component is very stiff compared to the rest of the inner string.
This was also because contact surfaces were unnecessary for the HP housing considering its location above the top of the LP
housing. Connectivity between the HP housing and HP housing extension was achieved by means of kinematic couplings
between the interface nodes of these components.
Non-linear elastic-plastic material properties were assigned to the surface casing and HP housing extension. Surface based
contact pairs were used to simulate the interaction between the inner and outer strings.
Modeling the deformation of the surface casing as it would occur with increasing amounts of BOP inclination was achieved
in a similar manner to the static analysis completed in Flexcom; however, as the detailed FE model only represented the
portion of the well up to the top of the HP housing, the effect of incrementally increasing the vessel offset was simulated by
applying an equivalent, incrementally increasing, set of lateral (shear) and bending loads to the node at the top of the HPhousing. These loads were derived from the static analysis previously conducted in Flexcom. Local element shear, effective
tension, and bending were extracted from the appropriate element within the Flexcom model and applied as a follower force
to the top node of the HP Housing in the detailed FE model. Loads of a secondary nature (e.g. hydrostatic pressures and
buoyancy) were omitted from the FE model on the basis that these forces would have little effect on the results and wouldlikely have required disproportionate amount of modeling effort.
Figure 15 Detailed Finite Element Model of the Conductor and Surface Casing Strings
Initial inclination of the conductor was not modeled. Instead, this was accounted for by simply subtracting 0.5 from theangles at which certain levels of stress and deformation in the surface casing were observed (similar to that done in the
Flexcom analysis). Figure 16 presents the relationship between the bending moment at the top of the HP housing and the
plastic strains at the critical cross section in the surface casing against the BOP inclination.
HP Housing(Beam Elements)
TOC Fixity(-35m)
21 x HP Housing Extension
Inner Surface of LP Housing
(Rigid)
30 x 1-1/2 X52 LP HousingExtension
20 x 5/8 Connector
20 x 5/8 X56 Surface
Casing
Inner Surface of 30x1Connector (Rigid)
20 x 5/8 X56 SurfaceCasing
20 x 5/8 Connector
Inner Surface of 30 x 1 X52Conductor Pipe (Rigid)
5/24/2018 OTC 25405 MS TNTurner (Submitted)
15/15
OTC 25405-MS 15
Figure 16 HP Housing Bendi ng Moment & Plastic Strain at Sample Location Plotted With Respect to BOP Inclination
The following observations were made from the results of the detailed FE analysis:
The BOP inclination of 3.8 where the full plastic moment capacity of the surface casing is reached is reasonablywell represented by the Flexcom model and the simplified calculation procedure outlined previously.
The surface casing is capable of sustaining loads associated with a BOP inclination of up to approximately 3.8 (3.3once conductor initial inclination is allowed for) before inelastic material behavior leads to large permanent
deformation of the surface casing.
Analysis suggests that failure would ultimately come in the form of local buckling occurring in the surface casingbody. The buckle would likely develop in the surface casing a few inches underneath the top 20 connector.
The onset of the buckling failure would occur gradually and not in such a manner that results in a rapid, catastrophiccollapse of the casing. BOP inclinations in excess of 5.5 can be tolerated without collapse of the surface casingthough the well would have sustained significant permanent deformation by this stage.
Local buckling of the 21 HP housing extension (where it bears against the inside of the LP housing) is not expectedto occur prior to the surface casing forming a plastic hinge.
The FE analysis therefore provided some assurance that if down-hole conditions were such that disconnecting the BOP,
LMRP and riser would have introduced significant risks, remaining connected to the well in seastates where the API based
permissible BOP inclinations were slightly exceeded would be unlikely to result in a sudden, unstable collapse of the surfacecasing.
Conclusions
The loss of integrity of a subsea wells conductor can lead to a dramatic reduction in a wells ability to resist the axial andlateral loads imposed by the drilling riser and BOP. This can lead to the well having to be abandoned and force abandonment
operations to adhere to very restrictive seastate limits.
This paper has described how finite element analysis was used in combination with statistical processing of previously
recorded seastate, vessel motion and BOP motion data to define the permissible seastate conditions where safe abandonment
of a well with compromised structural integrity could be performed. Advanced non-linear finite element analysis was alsoused to understand the risk associated with exceeding the API allowable stress based seastate limits.
References
API. (2010). Recommended Practice for Planning, Designing and Constructing Fixed Offshore Platforms - Working Stress Design API-RP2A-WSD (21st ed.). Washington, D.C.
Barltrop, & Adams. (1991). Dynamics of Fixed Marine Structures (3rd ed.). Oxford: Butterworth-Heinemann Ltd.Dassault Systmes Simulia Corp. (2012). Abaqus v6.12 User Manuals. Providence, RI, USA.MSC Kenny. (2013). Flexcom User Manuals (Version 8 ed.). Galway, Ireland.
Plastic hingeformation
Onset ofLocal
Buckling