OTC-24720-MS

Ultra Deepwater Managed Pressure Drilling in Challenging Formations

Roar Malt, AGR Enhanced Drilling
Copyright 2014, Offshore Technology Conference

This paper was prepared for presentation at the Offshore Technology Conference Asia held in Kuala Lumpur, Malaysia, 2528 March 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 been
reviewed 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, its
officers, 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 to
reproduce 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
Ultra deepwater drilling is challenging. When easily fractured carbonates are encountered, the challenge increases further still.
High hydrostatic pressures, as well as the narrow gap between pore and fracture pressures, add to the complexity. Performing
successful cementing operations is equally challenging. This paper describes the drilling of a well at 2180m in the eastern
GoM. The formations were carbonates with potential for high-to-total losses. The rig was a new Sixth Generation DP MDU.
To manage the BHP and prevent losses, the rig had been equipped with a new type of DGD/MPD system; a pumped riser
system, with the pump mounted around 400m below the bell nipple. The pump reduces the riser level in a precisely controlled
manner to manipulate the BHP as required. Due to a very weak 20shoe the 17 ! section had a drilling window of only
0.3ppg. The pumped riser system was used to drill the section with nearly constant BHP, compensating for the ECD by
manipulating the hydrostatic column level while drilling and bringing the level back up for connections. No losses were seen.
The same applies for cementing of the 13 3/8 casing. In the 121/4 section the system was rested due to robust formations and
a generous drilling window but was used again successfully to ensure a good 9 5/8 cement job. In the 8.5section the pumped
riser system was again necessary due to potential severe losses in weak reef-type carbonates. It soon became evident that this
formation could not have been drilled conventionally as weak zones were penetrated at the start of the section. Loss pressure
was 8.4 ppg. Drilling with returns was still required in order to get the required geological information from the well, which
was achieved. As drilling proceeded a normal-pressured zone was encountered. The pumped riser system was used to balance
the resulting pressures downhole with required returns to the surface. The system was also used to perform a good Plug and
Abandonment operation after TD was called. This paper will describe the planning and drilling operation of the mentioned
ultra deepwater well and present the operational results.

Introduction
The MPD/DGD equipment was built by AGR Enhanced Drilling in 2011 for placement on the new sixth generation modular
drilling unit. The rig was hired for a 3 well campaign in ultra deep waters in the Gulf of Mexico. The MPD/DGD equipment
is a pumped riser system, where a subsea pump is mounted on the riser, in this case, 400m below sea surface, allowing
precisely controlled changes to the riser level and thereby to the Bottom hole pressure during operations.
The well described here is the third well in the campaign, and severe challenges were expected in the 8! section was due to a
very weak reef zone in the top of this section. It turned out that the 17 ! section was equally challenging.
The objectives for using the DGD/MPD system in this well were:
Assist the operator in drilling to the planned TD with minimum losses or formation related trouble time.
Prove that ECD can be eliminated and BHP can be manipulated as required
Ensure good hole cleaning in all sections of the well
Achieve returns if severe or total losses
Help ensure cementing of casings /liners could be done without losses.

The MPD/DGD equipment
The system consist of a pump module attached to the riser and necessary equipment to run and control this pump as it sucks
mud through an outlet in the riser and returns it to surface via a separate mud return line.
2 OTC-24720-MS


Figure 1 above shows the general MPD/DGD setup

The equipment consists of a Sub sea Pump Module (SPM) (6) containing 3 electrical pumps capable of pumping mud and
slurry at high rates.
The SPM is connected to a designated modified riser joint (8)where the return flow is led into a separate Mud Return Line
(MRL) (7). The MRL is secured in the Hose Handling Platform (4) from where it continues into the rigs flow line.
An umbilical line, mounted on a winch (3) is used to deliver the required 900 kW and to communicate with the surface control
system. The Electrical power and communication is handled by a separate control container (2). The operators and PCs
controlling the system is situated in a separate office/tool container (1) just by. In addition there is always an operator on the
drillfloor during operations.
OTC-24720-MS 3


Figure 2, picture showing the pump module (SPM) mounted
on the riser subsea

Figure 3, picture of the SPM as it is being deployed

The speed of the SPM pumps is increased or decreased to modify the mud level in the riser. If the SPM pump work faster than
the rig pumps, the riser level decreases, and if it works slower than the rig pumps, the riser level increases.
The riser level is controlled by entering a pressure setpoint in the MPD/DGD control panel. 4 very accurate positive pressure
sensors installed in the designated riser joint are used to supervise the pressure at this point. Changes in pressure leads the
SPM pump to speed up or slow down as required to keep the setpoint.
The changes in the riser level cause the hydrostatic column, and thereby the fluid pressure in the well bore, to change.
Changes, such as added ECD or increased hydrostatic pressure due to displacement of heavy fluids can thus be accounted for.

Preparation for drilling
This well was the third well in the drilling campaign so the MPD crew and the drill crew were familiar with the use of the
Pumped riser system and the interfaces to the rig and the drilling operations. To the operator however, the system was new
and the drilling programs had to be reworked to allow for use of the pumped riser system. In addition extra meetings were
held between the drilling management, MPD and other third party vendors to ensure that everybody were on line before
starting the operations.
Before drilling out the cement in the 17 ! section, a capacity test with 9.0 ppg mud was performed. This was both to verify
that everything worked as planned and to familiarise everyone with the using the system:


At 06:00 a tool box talk was held and the standard pre start check lists were filled out
Riser pressure (before altering riser level) and BOP pressure were recorded on a separate log sheet as 628.59psi and
3280psi respectively.
06:30, driller staged up mud pumps to 500 gpm.
06:50, MPD had reduced riser level to 100 m (a reduction of 149.38psi).
While waiting for mud loggers to confirm stable system, driller increased flow rate to test for the maximum amount
the shakers would be able to handle (using 9.0ppg mud), rate was confirmed as 1400gpm.
Once the system was confirmed stable, MPD continued to reduce riser level to 150 m (235 psi below full riser), as
riser level was decreased, the driller staged up pump rate to 1000gpm.
07:00, riser level was stable at 148m (222.05psi reduction) with the SPM running at approx. 57.3% power utilisation.
07:15, riser level had been decreased to 200m, PID controller was reacting well with a stable level reduction of
295.58psi, SPM power output being 67%. From this point on, surface Control Container temperatures were
monitored closely.
07:35, riser level had reached 250m (378.69psi), with SPM output at 72.5% and all temperatures well below alarm
tripping point. Flow rate was increased at this point to 1200gpm (07:38) and then onwards a further 100gpm to
1300gpm (07:40). Level was held for 5 minutes at this flow rate before beginning the final step to either 300m riser
level or a 90% power output on the SPM.
4 OTC-24720-MS
08:00, riser level had been reduced to 300m (456.29psi pressure reduction), SPM achieving this without reaching
90% output.
The fluid level was then held at this level in the riser for 10 minutes while all systems were monitored.
Riser level was brought back to bell nipple with flow rate at 1000gpm, and system was shut down at approximately
08:30, concluding the test.

Through the course of the test, the mudloggers monitored flow through the mud system in order to calibrate their EKDS when
using the pumped riser system and try to fingerprint the flow response when the riser level was manipulated. The graph
below displays recorded data from both MPD and Mudloggers for the testing period.

Figure 4, Plot from modlogger and MPD operator data during test with mud



The MPD graph above shows the riser level (blue curve) the corresponding riser pressure at the MPD sensors (violet curve)
and the RPM of the MPD pump (red curve). The step from 0-100 m and from 100 150 m level reduction was performed
rapidly whereas the change from 150 300m was done more gradually. This is due to the combined effect of increased
mudpump flow rate and reduced riser level. Both are forcing the SPM to work harder but also increasing the risk of overflow
at the shakers.
Note that The EKD flow is high when reducing the riser level. This is because the mud taken from the riser is added to the
flow out from the mud pumps and subsequently added to the active volume.
Also note that the SPM pump is controlled by the pressure setpoint and responds to the pressure sensors in the modified riser
joint. The riser level in meters is just a calculation based on MW and riser pressure and has no effect on the control of the
SPM.
Drilling 17 ! section
The expected LOT before drilling the 17! section was 10.2 ppg. However only 9.43ppg was achieved and this necessitated
active use of the pumped riser system throughout the section as the expected pore pressure at planned section TD was 9.3 ppg.
The initial plan had been to drill with 9.3 ppg mud and let mud weight rise to 9.4-9.5 at the end of the section while controlling
the ECD using the pumped riser system. Due to the low LOT it became necessary to reduce the riser level more to keep the
ECD below the fracture pressure at the shoe at all times.

OTC-24720-MS 5

Figuere 5 below shows the expected geopressures. The black square shows the
achieved LOT and the clear blue curve shows the initially expected fracture
pressure.


With the current LOT and pressure profiles a mw of
9.3 ppg would fracture the shoe immediately. At the
same time a lower mudweight would result in a 13
3/8 shoe considerably shallower than planned.
Having the pumped riser system onboard made it
possible to drill to the planned TD without fracturing
the shoe. This was done by starting to drill the 17,5
section with a MW of 9.0 ppg. Riser level was kept
at 56 m to keep the ECD at 9.2 ppg. The Pressure
While Drilling sub was used to monitor the ECD at
all times during drilling. On connections the riser
level was increased to around 15m to compensate
for the loss of annular friction.











Figure 6 below shows how the riser level (black curve) was adjusted at the start of the section. The frequency (red curve) shows how much the SPM
pumps must work to maintain the various pressures.










When approaching the estimated pore
|pressure ramp at around 3000 m as can be
seen from figure 5 above, the MW was
increased to 9.3 ppg and the riser level
reduced to 91m. This gave an ECD of
9.21ppg and the possibility to increase riser
level more when encountering the pressure
ramp.

6 OTC-24720-MS




Figure7 below show riser level reduction during drilling and connections 2850 3010 m maintaining an ECD as close to 9.2 ppg as possible



At 3000m drilling depth the drilling flow rate was increased to a total of 1450 gpm including booster pumps. This was done to
improve hole cleaning. As can be seen from the plot above, the Bottom Hole Pressure increased to nearly 9.3 ppg with 9.3 ppg
mud. This was too close to the fracture pressure for comfortable drilling and the riser level was adjusted to 70-80m and later to
91 m to keep the desired 9.2 ppg BHP. On connections the level was raised to 33- 35 m.

Figure 8 below shows how the riser level was used to keep the BHP around 9,2 ppg all the time after increasing the flow rate..



To achieve a slightly better window for the cementing, the section TD was called when pore pressure estimates showed 9.1
ppg. Section TD was therefore somewhat shallower than initially planned.
OTC-24720-MS 7

Running and cementing the 13 3/8 casing
Due to the very narrow drilling window in this section there was a real danger of fracturing the well both while running the
casing and while cementing. Reducing the MW could induce a kick or a well collapse, so that was no option.

The simple, and yet effective solutions was running the riser while keeping the riser level at 38 m which corresponds to a
pressure reduction of 60 psi. This way the edge was taken of the surge maximum preventing fracturing of the well and still
staying above the porepressure when movement stopped.

Figure 9 below shows that formation will fracture while running casing in conventional mode, but in Pumped Riser mode it will stay
good.

2
0
0
0
.
0
2
5
0
0
.
0
3
0
0
0
.
0
3
5
0
0
.
0
m
Well
8 9 10 11 12
lbm/gal
Tripping RIH
Hydrostatic
Frac
Pore
RIH Press.
W
A
R
N
I
N
G

:

L
O
S
S
E
S
No Dual Gradient,
Mud 9.30 ppg
!P = - 65 psi
8 9 10 11 12
lbm/gal
Tripping RIH
Hydrostatic
Frac
Pore
RIH Press.
!"#$ &'#()*+,
#--$)*(.
/"( 0123 --4
56 7 89: -;)

Once the 13 3/8 casing was successfully run to total depth without losses, circulation and mud conditioning of mud was
achieved again with the manipulation of the hydrostatic pressure through the MPD System. The challenge of a cementing
operation with manipulation of the hydrostatic pressure through a MPD system is the feasibility to decrease such pressure
without risking well control during placement.
In this specific scenario, there was no possible solution to achieve the desired TOC without creating losses with the minimum
mud weight of 9.3 ppg allowed by formation at total depth. As analysis show no current solution would be available with the
conventional cementing technique. The gap between pore and frac pressure was so narrow, that it was not possible to reduce
the mudweight without an underbalance situation and a potential influx at the bottom hole.
The only possible solution to perform the job without losses or any influx due lack of well control, was achieved by:
1. Manipulating the hydrostatic pressure with the MPD System to keep it above the pore pressure gradient during the
beginning of the job;
2. Decrease the hydrostatic pressure in steps, to compensate for the pressure increase created by the heavier fluids
pumped during the cement job and placement in the annular space;

Analyses showed that to stay above the porepressure during the cement job would initiate fracturing of the previous casing
shoe when displacing the cement.
8 OTC-24720-MS

Figure 10 below shows that in a conventional case with MW at the minimum 9.3ppg, the entire job would be performed on the balance point
between fracture at previous shoe and influx at TD until the cement would start moving up the annulus. Than the 20 shoe would be fractured
immediately.



The future applied now: Incorporating a state of the art Dual Gradient System
Having this pumped riser MPD equipment available made it feasible to plan for a cement operation without losses. The
operator and the 3
rd
parties involved went through the cementing and casing running program, calculating the riser level
changes required to keeping the well stable and avoid losses during the operation. As this operation was the first of its type
not only on the rig but in the world, several meetings with the crews involved in the operation were needed to keep everybody
informed about the procedure. A separate a hazard analysis and risk control plan was performed before the cementing
operation could commence.
Based on the simulations, in close cooperation between all parties involved the necessary adjustments to the cement program
were made. Each cementing step was identified and the required riser level for each step was noted. The effect this status
would have on the operation and possible actions if something did not go to plan was identified. The advances of cementing
software allowed for every single detail of the cementing job, including U-Tube effect during the job produced due the
difference in hydrostatic columns both inside casing and in the annular space.

The pumping table that was used is shown below.












OTC-24720-MS 9


Table 2 below shows the planned pumping schedule and corresponding pressure reductions (riser level reductions)
STEP
Total
Volume
pumped
(bbl)
Required
TOTAL
underbalance
(psi)
Required
Underb.
(bar)
Hydrostatic
Column
length
(m)
Comments
WAIT ALWAYS FOR CLIENT/SLB
ENGINEER CONFIRMATION
STEP#1: Start
Cement job
with -150 psi
0.00 -150 10.3 94.2
Set hydrostatic level in order to have -
150 psi since the beginning of the job
(Est. Pressure @ EC-Drill System: 634
150 = 484 psi aprox)
STEP#2:
Decrease the
hydrostatic
pressure by -
200 psi
850 -200 13.8 125.6
Decrease the hydrostatic pressure by -
200 psi after 368 bbls displacement
with Mud.
ECD @ 13 3/8 Shoe = 9.30 PPG
ECD @ 2,861 m = 9.19 PPG
STEP#3:
Decrease the
hydrostatic
pressure by -
300 psi
1000 -300 20.7 188.4
Reduce the hydrostatic pressure by -
300 psi after 518 bbls dispaement with
Mud.
ECD @ 13 3/8 Shoe = 9.46 PPG
ECD @ 2,861 m = 9.09 PPG
STEP#4:
Decrease the
hydrostatic
pressure by -
350 psi
1150 -350 24.1 219.8
Reduce the hydrostatic pressure by -
350 psi after 668 bbls displacement
with Mud.
ECD @ 13 3/8 Shoe = 9.46 PPG
ECD @ 2,861 m = 8.86 PPG
STEP#5:
Decrease the
hydrostatic
pressure by -
400 psi
1190 -400 27.6 251.2
Reduce the hydrostatic pressure by -
400 psi after 708 bbls Mud have been
displaced);
ECD @ 13 3/8 Shoe = 9.41 PPG
ECD @ 2,861 m = 8.84 PPG
"


In the figure below simulation results showing the expected pressures at the 13 3/8" casing shoe. The EC drill modified curve
show that the hydrostatic pressure is kept well above the pore pressure while circulating and pumping the cement. This is also
the case when reducing the riser level to avoid fracturing the 20 shoe when displacing the cement.

Figure 11, simulated ECD at 13 3/8 shoe while cementing


The plot below shows that the shoe would fracture when displacing the cement conventionally, but as the Dual Gradient
System modified curve shows, the pressures will stay below fracture pressure if the riser level is reduced as planned.
10 OTC-24720-MS



Figure 12, simulated ECD at 20 shoe while cementing


The plot below shows the cement job. In this plot riser level, represented by the brown curve, is called called RAG (Riser air
gap). Pink curve shows the active mud pit volume

Figure 13, pressures, riser level and volumes during the cement job

There were no losses while running casing, circulating well nor during the cementing job.

The plug was not bumped during the operation. Cement inside casing was found @ 3238 m, 35 meters above the expected
depth corresponding to 12 barrels of displacement. Since the lack of pressure indication was not seen due to effect of the dual
OTC-24720-MS 11

gradient technique during the displacement, it was decided by all parties involved not to continue displacing (neglecting mud
compressibility), as the shoe could be washed. Good cement was tagged and drilled after wait on cement. Successful LOT was
performed obtaining a value of 11.46 ppg of equivalent mud circulating density, sufficient to continue with the drilling of the
next stage. The objectives of the job were clearly achieved and set a base and reference for future jobs performed with dual
gradient technique

Drilling the 12 1" section
The objective of the 12 # section was to drill through part of the Paleocene and set the 9 5/8 production casing at 3775 m
above any reefal growths or pore pressure reduction zones, in order to avoid communication of high pressure Palaocene zones
with low pressure fractured limestone zones, found in the same stratum. At the same time, setting the casing to shallow might
prevent effective mitigation of losses in the 8.5 section.
The use of ED-Drill had been incorporated into the drilling plans for 12 # section, but due to a leak discovered in an internal
seal in the 2
nd
pump stage during the 17 !" section it was decided to be on the safe side and rest the MPD system keeping it
ready in case of losses.
The 12 # was initiated and drilled conventionally with 9.6 ppg mud from 3340m to 3685 m MD/TVD without problems

12 OTC-24720-MS



Running and cementing the 9 5/8 casing

For this stage, the objectives of the cement job were:
Obtain TOC up to 3152 m (188 m above previous casing).
Provide competent casing shoe and FIT (Formation Integrity Test) in order to drill 8 !" hole to planned depth.

Prejob preparation

Bottom Hole Pressure Analysis

Given the expected low Frac Gradient, an analysis was made in order to evaluate the Dynamic pressure exerted by the cement
job. The simulations without the use of the dual gradient system were as follows:
Figure 14 Fracturing pressure and dynamic pressure while cementing
0 20 40 60 80 100 120 140
Time (min)
6
5
0
0
6
2
5
0
6
0
0
0
5
7
5
0
A
n
n
.

P
r
e
s
s
u
r
e

(
p
s
i
)

Frac
Pore
Hydrostatic
Dynamic
Depth = 3679 m

Figure 14 above shows the 300 psi extra BHP exerted due to the Cement Job. This pressure was meant to be lowered by using
the dual gradient system. However the pressure cannot be dropped from the beginning because of the formation Pore Pressure.
Doing so could induce an influx, being 100 psi below the Pore Pressure, as shown in the following graph:

Figure 15, Relation between pore and fracturing pressure at 300 psi BHP reduction
0 20 40 60 80 100 120 140
Time (min)
6
2
0
0
6
1
0
0
6
0
0
0
5
9
0
0
5
8
0
0
5
7
0
0
5
6
0
0
A
n
n
.

P
r
e
s
s
u
r
e

(
p
s
i
)

Frac
Pore
Hydrostatic
Dynamic
Depth = 3679 m

Given the scenario a step sequence had to be designed in order to accomplish an efficient bottom hole pressure management.
In an interdisciplinary effort between the Operator, the Managed Pressure System Company and the Cementing Company, it
300 psi
Above frac
100 psi Below
Pore
OTC-24720-MS 13

was found that the system had to be turned on throughout the whole pumping, and therefore a minimum volume had to be
constantly pumped. 12 m of riser volume was agreed, meaning 20 psi would be lifted off the BHP (Bottom Hole Pressure)
from the beginning and later the stepped sequence could be initiated. To achieve the following BHP behavior:

Fig 16, Pore pressure, fracture pressure and BHP when adjusting riser level actively
0 20 40 60 80 100 120 140
Time (min)
6
2
0
0
6
1
0
0
6
0
0
0
5
9
0
0
5
8
0
0
5
7
0
0
A
n
n
.

P
r
e
s
s
u
r
e

(
p
s
i
)

Frac
Pore
Hydrostatic
Dynamic
Depth = 3679 m

By following these parameters, a Riser Pump Sequence was designed so that the dynamic pressure was kept at the desired
level. This operation was meant to keep the pressures in between the formation limits as shown in the past graph.

Riser Pump Stepped Sequence
In order to mitigate the losses expected during the cement job a Riser Pump Sequence was designed to lower the dynamic
pressure as follows:
Job Execution:
Table 3, pumping schedule and corresponding riser level changes for the 9 5/8 cement job
Execution
Volumes Density Rate
Riser
Level
Equivalent
Pressure
Design Execution Design Execution Design Execution
bbl bbl ppg ppg bpm bpm m psi
Spacer 50 60 11.0 11.0 6 6 12 20
Slurry 115 115 12.5 12.5 4 3-4 12 20
Spacer 10 10 11.0 11.0 6 6 12 20
Mud 160 160 9.6 9.6 7 6-7 12 20
Mud 20 20 9.6 9.6 3 2.5-3 12 20
Mud 220 220 9.6 9.6 7 6-7 12 20
Mud 50 50 9.6 9.6 7 5-7 61 100
Mud 20 20 9.6 9.6 7 5-7 98 160
Mud 20 20 9.6 9.6 4 4 164 270
Mud 20 20 9.6 9.6 4 4 177 290
Mud 16 16 9.6 9.6 3 3 202 330

Actual Bottom Hole Pressure during Cement Job
By following the previous design the BHP was lowered efficiently. However during the analysis of the data after the cement
job it was observed that the pressure did exceed the formation limit over 50 psi.





14 OTC-24720-MS


Figure 17, post job calculation of Dynamic pressures during cement job
0 25 50 75 100 125 150 175
Time (min)
6
2
5
0
6
0
0
0
5
7
5
0
A
n
n
.

P
r
e
s
s
u
r
e

(
p
s
i
)

Frac
Pore
Hydrostatic
Dynamic
Depth = 3679 m


This information was confirmed by the Mud Loggers since losses were detected during and after the cement job. Losses were
seen throughout the cement job. While pumping the cement 130 barrels were lost. While displacing the cement up the annulus
55 barrels were lost and finally 35 barrels were lost after bumping the plug.
The hydrostatic pressure in the riser was however exactly as planned during the entire job as the plot below show. It is likely
that at least some of the losses were due to a leak in the MRL that was discovered shortly after resuming drilling.

Figure 18, Data from the MPD control system showing stable pressure readings

The final result of the cement job was successful. An FIT was carried out with no problem to increase pressure, meaning a
strong seal was left on the shoe.

Drilling the 8 !" section

The objective of this phase was to collect data from the Paleocene to below an unconformity ~4486 m, representing the area of
interest of the well, with a high risk of severe loss circulation below the unconformity.
The section commenced at 05:00 hrs 5th October 2012. Cement and shoetrack was drilled with the MPD riser valves closed
and the pump off. This is to prevent plugging of the MRL or SPM with parts from the float or collar. Also 10 m of new
formation was drilled from 3678m without the pumped riser system activated. Prior to start drilling a LOT of 11.3 ppg was
achieved. This is slightly above the estimates from the geopressure log. MW was 9.3 ppg.

BU was circulated and the EKD was calibrated before drilling resumed. At 3695m losses to the formation were seen, and at
3699m a flow check was made, revealing static losses of 16 bbls pr hour.

RP: 44 bar = 12 m
(cement displ.)
RP: 23 bar = 201 m
(bump plug)
50 psi
Oveburden
OTC-24720-MS 15

At this stage the MPD system was activated. The riser level was lowered in steps to 30, 60,100 and eventually 130 m. This
corresponds to a reduction in BHP of 47, 95, 158 and 206 psi respectively. The pumps were initially staged up but were
reduced again due to severe losses. The top fill pump was activated in the start of the process. Later it caused some concern
during flow check since it was not turned off causing misinterpretation of the flow check . After this the top fill pump was not
used on this section.

Figure 19, plot from mudloggers showing steady riser level and decreasing active volume



As the figure above shows, the losses did not improve when activating the MPD/DGS System Instead it seemed to get worse.
The Pumped Riser system was therefore isolated and troubleshooting performed. A Leak was discovered and the Mud Return
Line had to be recovered to repair the leak.

In the meantime the rig continued drilling without MPD, trying to cure the losses with LCM. At 3731m losses were extra
severe and riser level dropped to 340m static. This was found to correspond to a leak off (or SH min) pressure of 8.4ppg.

Figure 20, showing that losses stopped when reducing riser level to 280m.

Shortly after, based on new geological
information combined with the severe losses,
the mud was displaced to 8.8 ppg. During
displacement the riser level was adjusted to
minimize the losses.

Drilling resumed with the pumped riser
system active on the 8th October, and prior to
drilling the riser level was reduced from 180
to 280m. Simultaneously the pumps were
staged up to ensure sufficient hole cleaning
above the loss zone. The loss rate reduced
considerably with riser level at 280m as the
plot to the left shows.

Drilling continued with riser level at 280 m.
On connections the riser level was increased
to 180 m. Such riser level changes were
performed in co-operation between the MPD
operator and the driller and typically took 4
6 min, whereof 2-3 minutes was driller
staging up or down pumps. This can be seen
in the plot below.
Riser Level
(Black) 0 300m
Vol +/- (Black)
ACT Vol (red)
Riser Level
280m
No more
losses
16 OTC-24720-MS


Figure 21, Performing connection while partly compensating for loss of hydrodynamic friction





ECD as measured with the PWD sub was 8,6 ppg with riser level at 280 m. As drilling continued, the riser level was further
reduced to 300 m. This was to reduce losses further.
At 3802 m an influx took place. This is clearly visible on the plot below. The riser level was increased to 130m, and then the
well was flow checked. The flow check was negative with this riser level. The well was circulated but no gas was seen and
based on the returns it was concluded that the kick was water.

Figure 22, Showing the influx and subsequent filling of riser.


The plot in figure 23 shows how the kick looked at the pumped riser system. Note that the flow meter (red curve) picks up the
kick first, than the net kVA. The Saiv pressure (blue) follows the set point (pink) showing that the change in riser pressure is
due to a planned increase of riser level.





This shows it took 4,5 min
to increase riser level from
280m to 180m before the
connection.
Driller stages down pump in
2 minutes
It took a little more (6 min)
to decrease Riser level again
again.
An influx takes place
The riser level is increased
to 150m
OTC-24720-MS 17




Figure 23, Data from the MPD control system showing the influx and the subsequent raising of riser level



After regaining control drilling continued and riser level was initially reduced to 210 m. Drilling continued while trying to
establish stabile conditions with various losses and gains. Riser level was frequently adjusted but generally kept at 110 - 120 m
while bringing it up to 75 m at connection.

Based on the data from drilling it was found that The loss zone at 3730m has a loss pressure of 8.4 ppg whereas the influx
zone is hydrostatic at 8.6 ppg. Returns show that the influx is fresh water.

Figure 24, Plot of the formation pressures at the loss zone and below

Figure 24 above shows the situation. When staying above the pore pressure, the BHP is way above the fracture pressure at
3730m and losses occur. By adjusting the riser level (and thereby the BHP) the amount of fluid flow into the loss zone is
controlled. This way it is possible to drill with returns.
At 3878 m, the MWD failed and drilling continued without PWD. At the same time it became gradually more difficult to keep
the MW. The return MW was measured to 8.6 ppg.
Drilling continued with frequent fine adjustments to the riser level but generally with the level between 150 and 170m. LCM
and hi wis sweeps were pumped at regular intervals.
On the 11
th
October several cement plugs were pumped but no effect was seen.
Drilling was resumed with EC-Drill, maintaining level at 90-110 m (50-55 m on connections) while trying to keep minimal
losses to TD at 4250 m (20-50 bbl/h).
18 OTC-24720-MS

After trips, while washing down the riser level was changed to reduce the BHP of pumps on/off. This can be seen in the plot
below. Note the adjustments of riser level for each connection during pumping in.

Figure 25, Pumping into hole while adjusting riser level after a trip


The staging up and down of rig pumps in cooperation with MPD became more smooth and systematic during the operation,
without the connection taking more time.

Figure 26 riser level changes on connections or when stopping rig pumps



This is also clearly seen in the plot from the pumped riser control system below.

Figure 27 below shows a typical connection towards the end of the well. Pressure at riser pressure sensor during drilling is
31.6-31.8 psi (90 110m) and increased to 36.1 psi 50 55m on connections.
In this particular case the ramping time is as follows:
Ramp down (603 - 0 gpm) in 5 stages. Time 14:19:43 14:23:15 (3 min 32 sek)
Ramp up (0 603 gpm) in 5 stages. Time 14:28:35 14:32:47 (4 min 12 sek)


Staging down
Riser level
Staging up
Riser level
0-300m


OTC-24720-MS 19




Figure 27, plot from MPD control system showing a typical connection



The 8 !" section was concluded at 4250 m after drilling 570m in what would have been an undrillable total loss scenario
without the pumped riser system. Contractor was able to achieve the data and samples they were looking for despite the
challenging formations encountered. Prior to P&A the pumped riser system was used to maintain level control during both
wireline and wiper trip, with the riser level at about 85 m.

Well control measures
The well was drilled using the normal methods of influx detection (active pit levels, PVT readings, gas/salinity levels in
return flow) as the return flow was always to either the port or starboard flowlines, taking the conventional return route
thereafter (as per Scarabeo-9 setup).
During the drilling, there were one clear incident with influx into the wellbore whilst the MPD/DGS system was in use. In
addition several flow checks, both static and dynamic was performed due to suspected inflow.
For flow checks the most effective flow check was to close the Riser Isolation Valve and isolate the SPM (without turning it
off, remain on idle) rather than stop the SPM. The riser pressure sensors were then used to monitor riser level. String would
now u-tube back to riser but with an identifiable increase?? in volume that can be finger printed.

Conclusions
The pumped riser system turned out to be at least as important for the success of this well as anticipated. It allowed drilling
and isolation of the 17 !" section with a nearly undrillable window. It allowed the performance of a good 9 5/8 casing
cement as well as P&A plugs and subsequent inflow tests. And not the least it allowed the continued drilling of the 8 !"
section even after encountering a zone of total loss and it was possible to continue drilling after encountering a higher
pressured water zone.
The rubber mud return line was a weak point in the set up due to the potential for external wear and subsequent leaks. Equally
the internal seals in the pump could have been more robust as one of the started leaking causing concerns.
The total conclusion is still that the operation was an absolute success and that this equipment has proved its value in deep
water and narrow margin drilling. With the necessary improvements on seals and mud return conduit it may prove to be a step
change in this type of drilling environment.


Acknowledgement
The author would like to thank AGR Subsea and Schlumberger for their support during this project and for permission to
publish these results. Special thanks to Alexander Baez and Arevalo Marcano for their support and help for the planning and
execution of the operation, and to Rodrigo Giron Rojas for invaluable help on the cementing parts The author would also like
20 OTC-24720-MS

to thank the Scarabeo 9 rig crew and management for their assistance and co-operation throughout the operations.
We would also like to thank Tony Power and Rodger Littlechild for valuable input throughout the drilling operation.


Figure 28 The Sixth generation deepwater drilling unit Figure 29The SPM unit during installation on the drilling rig

























References
1 USRE43199 Arrangement and Method for regulating bottom hole pressures when drilling deepwater offshore wells

SPE OTC-24720-MS151100-MS- First Application of EC-Drill in Ultra-Deepwater, Proven Subsea Managed Pressure Drilling Method
M. Mir Rajabi, SPE, K. Toftevg, R.S. Stave, SPE, AGR AS, and R. Ziegler, SPE, PETRONAS

OTC 22889-MS A Step Change in Safety: Drilling Deepwater Wells with Riser Margin
Robert Ziegler, SPE, PETRONAS

SPE 130308-MS Deepwater Riserless Mud Return System for Dual Gradient Tophole Drilling
Dave Smith, SPE, AGR Subsea; Warren Winters, SPE, BP America; Brian Tarr, SPE and Robert Ziegler, SPE, Shell; Iskandar Riza and
Malik Faisal, PETRONAS

SPE 137319-MS - Deploying the World's First Commercial Dual Gradient Drilling System
J. David Dowell, SPE, Chevron
SPE/IADC 164561
Successful application of Deep Water Dual Gradient Drilling
R. Ziegler, P. Ashley, PETRONAS, R. Malt, K Toftevag and R. Stave AGR-EDS-ORS


OTC-24720-MS
First Successful Commercial Application of Dual Gradient Drilling in Ultra-deepwater GOM
Robert Ziegler, Mohd Saiful Anuar Sabri and M Ramdan B Idris, Petronas; Roar Malt and Roger Stave, AGR Enhanced Drilling