1. 2003 SPE/IADC Drilling Conference
Well Control Procedures for Dual Gradient Drilling as Compared to Conventional Riser Drilling
February 21, 2003

2. 21.1
Well Control Procedures for Dual Gradient Drilling as Compared to Conventional Riser Drilling
Dr. Jerome J. Schubert
Dr. Hans C. Juvkam-Wold
Texas A&M University and
Dr. Jonggeun Choe
Seoul National University

3. Overview Introduction to Dual Gradient Drilling
Goal of the SMD Well Control Team
Comparison of Well Control for DGD and Conventional Riser Drilling
Conclusions

4. What is Dual Gradient Drilling?
Novel drilling system where the annulus pressure at the seafloor is reduced to near seawater HSP.
Results in a pressure gradient from the rig to the seafloor near that of seawater HSP, and mud gradient from the seafloor to the bottom of the hole

5. Dual Gradient Concept
Here we see a comparison of the pressure profile for conventional deepwater drilling and dual gradient drilling. The pressure at the seafloor is reduced to near seawater hydrostatic pressure, while a mud gradient is exerted in the wellbore from the seafloor down.

6. How is the dual gradient achieved?
Seafloor pumps and an external return line
Shell
DeepVision
SubSea MudLift Drilling
Injecting hollow glass spheres near the seafloor
Maurer Technology
There have been at least four projects that have studied dual gradient drilling. The first three use a pump on the seafloor to pump returns up a relatively small diameter (compared to the marine riser) from the seafloor to the surface. Maurer Technology has studied injecting hollow glass spheres in the mud return path.

7. Goal of the SMD Well Control Team
Develop Well Control Procedures for the SMD JIP that were at least as safe if not safer than conventional floating drilling operations.
The authors feel that these procedures are applicable for most DGD methods.
I was fortunate to have worked on the SMD project. Our goal was to develop well control procedures for this project.

8. How was the goal met?
We had to study the state of the art in conventional deepwater drilling
Determine what had to be modified or re-written for the SMD project.
New procedures were written and re-written as the project progressed.

9. How was the goal met? Perform risk analysis in the form of HAZOP
Modify or re-write procedures based on HAZOP
If the procedure was re-written, a new HAZOP had to be performed

10. How was the goal met?
Finally, most of these well control procedures were proven on a DGD test well.

11. Measurement of KCP
KCP is measured identically for DGD and Conventional
No DSV – rate must be greater than the freefall rate of the mud
W/DSV – must also measure the DSV opening pressure

12. Kick Detection Kick indicators Drilling break Flow increase Pit gain
Decrease in circulating pressure
Increase in pump speed
Well flow with pumps off
Increase in torque, drag, fill

13. Flow Increase MLP Increase Kick Begins MLP Inlet P Constant
In the SMD project a kick would first be detected by an increase in the MLP rate. When this is detected, the MLP would be set to operate at a constant rate equal to the pre-kick rate. This would result in an increase in seafloor annular pressure, bottom hole pressure and rig pump pressure.. Once these pressures stabilized, the influx has stopped. The increase in rig pump pressure from the pre-kick pressure plus the annular friction pressure is the standard Shut-in drillpipe pressure.
DPP Decrease

14. Well Flow w/ Pumps Off No DSV U-tube makes this much more difficult
Trend analysis is needed
700
gpm
600
500
400
Normal U-tube
tube Rate,
300
U-tube + kick
200 -
If there is no DSV to arrest the u-tube whenever the rig pumps are shut down, the seafloor pump will continue to pump until the u-tube has stabilized. In order to determine if the continued flow is from the u-tube or a kick, trend analysis will have to be used. If there is a DSV in place, any time the rig pumps are shut down, the MLP will cease, unless a kick has occurred. This is identical to the well flowing with the pumps off during conventional well control. U
100 15 20 25
Time, min

15. Well Flow w/ Pumps Off With DSV Shut down Rig Pumps
Continued operation of the Sea Floor Pump will indicate well flow.

16. Pit Gain W/DSV there is no difference
No DSV – No difference in kick detection. However pit gain after shut-in is equal to the pit gain after complete u-tube less the theoretical u-tube volume.

17. Shut-in on kick With DSV, SI is very similar to conventional
Shut down rig pumps,
Check for flow
If flowing, shut down MLP
Close BOP
With No DSV, preventing additional influx is difficult during u-tube.

18. MLP inlet P & DPP Increase
Shut-In on Kick
Kick Detected
Slow MLP
Rig Pumps Constant
MLP inlet P & DPP Increase
Here is the procedure for kick detection, stopping the influx, and measuring SIDPP using the dynamic shut-in procedure.

19. Shut-in Procedures
After the MLP and Rig pumps are returned to the pre-kick rates:
Allow the DPP and MLP Inlet P to stabilize
Record stabilized pressures and rates
Continue to circulate at constant Rig Pump Rate and Pressure until kick fluids are circulated out.
DPP is maintained by adjusting MLP Rate

20. SIDPP SIDPP is somewhat different.
W/DSV very similar to measurement of SIDPP with a float and is the
Post kick DSV opening pressure less the Pre kick DSV opening pressure.

21. SIDPP – No DSV Upon kick detection, slow MLP to pre-kick rate
Record the Stabilized DPP

22. Calculation of KWM Conventional Dual Gradient
To calculate KWM for dual gradient drilling the second equation in this slide is used.

23. DPP Pressure Decline Schedule
Calculating ICP is no different
FCP Conventional
FCP=KCP x KWM / OWM

24. FCP DGD

25. Driller’s Kill & Wait & Weight
Essentially the same for DGD and Conventional except for the differences noted earlier in measurement of SIDPP and shut-in.
MLP is used as the adjustable choke

26. Other Kills Volumetric Lubrication Stripping
Procedure have been developed but are not included in this paper.

27. Conclusions
The u-tubing that is expected in DGD causes some difficulties in many aspects of well control – none of them are show stoppers
The use of a DSV eliminates the problems associated with the u-tube phenomenon, but creates some of it’s own

28. Conclusions
The complications from the DSV are outweighed by the benefits
DSV makes well control seem more conventional, but it is not absolutely necessary.

29. Conclusions
Well control for DGD has been developed to a point where it is at least as safe if not safer than conventional riser drilling.
A well control training program for DGD will be essential for safe and efficient operations.

30. IADC/SPE 79880
The End

31. DGD with Seafloor Pumps

32. Recent Advances in Ultra-deepwater Drilling Calls for New Blowout Intervention Methods
Speaker: Ray Tommy Oskarsen Co-authors: Jerome Schubert Serguei Jourine
IADC Workshop Galveston “Deepwater Drilling: Where are we headed?”
Good morning ladies and gentlemen.
As a continuation of the SMD project a study on study on blowouts occurring in deepwater was undertaken at A&M.

33. Sponsors and Participants
Phase 1
Texas A&M University
Cherokee Offshore Engineering
Global Petroleum Research Institute
Offshore Technology Research Center
Minerals Management Service
The project was initiated by Texas A&M University and Cherokee Offshore Engineering.
MMS and OTRC are funding phase 1 of this project. We are looking for industry participation for phase 2. Several operators and service companies have shown interests. Phase 2 will be a JIP.

34. Drilling in ultra-deep water
Window between pore pressure and fracture pressure gets narrower
High pore pressures and low fracture pressures lead to more casing strings
More casing strings leads to more time spent on location
This leads to larger wellheads, even larger and heavier risers, and finally to bigger and more expensive rigs
With a standard BOP and many casing strings, you may not reach target.
Well control is more difficult - because of the pore pressure / fracture pressure proximity, and long choke lines with high frictional pressure drops
Industry believes major reserves are present in deepwater and ultra-deep water.
There are many problems with drilling in ultra-deep water, which most people are aware of.
The main problems are; you need large rigs, it is expensive, and well control is difficult.

35. Deepwater drilling projects
Dual Gradient Drilling
Casing Drilling
Expandable Casing
SX-riser
Many projects has emerged the last decade to overcome the previous mentioned challenges. These projects aim to drill faster, safer, and cheaper.
The picture shows the mudlift pump that was used during the Subsea Mudlif Drilling (SMD JIP) test well last year. It proves that dual gradient drilling is no longer a thing for the future.

36. Blowout Containment Procedures?
The most recent blowout containment procedures can be found in the “DEA – 63, Floating Vessel Blowout Control,” which was released September 1990.
DEA - 63 considered deep water up to 1500’
Envisioned “future work” in water as deep as 3500’
As seen many projects has been undertaken to guide us into the ultra-deep waters. What about blowout containment procedures….. Have they been keeping up with the current technology?
The last major work completed on deepwater blowout control was DEA 63, completed in 1990.
It was good work, but the work never considered drilling in water depths greater than 3500’.

37. Are We Ready? DEA-63 Cont. Focus on capping measures
No Dual Gradient Drilling
Concluded with recommendations for more work
DEA63 phase II was proposed in 1998 but didn’t get off the ground.
The concern was proper as the original DEA had understood and explained.
The objective of DEA63 was to develop innovative capping techniques. Today’s wells may not require these techniques as the downhole mechanisms are much better understood, particularly in view of the fragile formations which could not hold the resultant pressures from a well capping operation.
Other techniques are required and must be developed so that the problem can be solved in a logical, consistent manner.
Are We Ready?

38. Safety Pyramid 1 29 300 3000 Fatality LTA OSHA Recordable At-Risk
Do we need to study blow-outs? Should we be worried?
The slide is the graphic of the Safety Pyramid. This concept is familiar to the entire drilling industry and is used to present the idea of statistical probability of an incident. (It is simplified from its original version as presented by the International Loss Control Institute)
The phrase “work the bottom of the pyramid” is used to explain the methodology of reducing the number of qualified events by focusing attention by all levels on the potential severity, were it not for some factor, of a given incident.
Behavioral-based safety programs are becoming widely used in the industry with measurable success, but-even more importantly for this presentation-it’s a standard which an instant identification can be made to well control issues.
Using widely-accepted concept of the Safety Pyramid, these linkages exist among incidents of increasing severity
1 fatality = 29 LTAs (Lost Time Accidents)
=300 Recordable Incidents (Occupation Safety & Health Administration)
=3000 At-risk Behaviors
Sometimes called Near Misses
“Work the Bottom of the Pyramid”
3000
At-Risk
Behaviors
Albert H. Schultz - DuPont

39. Statistics Podio Study of OCS Blowouts, 1996
1 Blowout for every 285 wells drilled
2.7% of the wells studied deeper than 15,000 ft
These accounted for 8% of the blowouts
Wylie and Visram, 1990
1 Blowout for every 110 kicks
SINTEFF Deep Water, 2001
52 kicks for every 100 wells drilled
79% of kicks had significant problems
At least 21% of kicks resulted in loss of all or part of the well
1992 to 2001 we drilled 1015 wells in water >1500 feet deep
We can gather data from literature to create a safety pyramid for blowouts.
SPE Paper “Trends Extracted from 1200 Gulf Coast Blowouts During ” Also, noted in the conclusions: Blowouts continue to occur at approximately a constant rate.
Database has not been updated since that time.
Here we are trying to work the bottom of the pyramid and we don’t have much of a clue about the severity of the problem or how many near misses there are to address, or even kicks.
This is the reason we’ve requested the weekly report data from MMS which appear to have well control or other extraordinary events in them. Besides sifting this data, we have requested as-built drawings vs. APDs so that anamolies might be investigated to see whether a well control event caused the change, and also a review of Sundry Notices which should give a good clue as the condition of the well at any point.
This is a very significant gathering of data even with these relatively conservative sift points.

40. Blowout Pyramid 1 Blowout 20 Well Bore Losses
80 Significant Well Control
Problems
110 Kicks
When the pyramid is applied to deepwater well control, the relative severity of the incident can be classified much as other incidents.
For instance, a blowout is the most catastrophic event and can be viewed for these purposes as the top of the pyramid.
The second level would correspond roughly with an LTA and would consist of lost hole sections, lost wellbores, production and data.
The third level would consist of an official well control event which was handled by closing in the preventers without loss of any hole section.
The bottom level would correspond with the Near-Miss level of incident reporting and consist of well control events which did not require BOP shut-in to handle.
The safety pyramid is really an incident pyramid and is applicable to non-productive activity
1 Blowout for every 110 kicks
20% lost hole sections (SINTEF)
1 Kick = dozens At Risk Behaviors
High gas, ballooning, improper fill
? At Risk Operations
200+ Wells Drilled

41. Are wells in deep water likely to occur more frequent?
Higher pore pressure gradients
Difficulties in handling highly compressed gas
Increased exposure time
Longer open hole sections
More tripping time
Increased risk of lost circulation
Blowouts has been occurring regularly throughout the petroleum history. As shown by Podio: deepwater wells accounted for only 2% of all wells drilled, yet they account for 8% of the blowouts. Why do blowouts occur more frequently in deepwater?
It goes back to well control problems which arise as we move into deeper water.
No blowout has yet occurred in ultra-deep water (water depths of 5000ft or greater) but statistics show it is likely to happen. Are we ready to handle it?
Odds are not in our favor!

42. Deep Water Blowouts Proposed practical solutions: capping,
injecting solidified reactive fluids,
dynamic kill/momentum kill,
inducing bridging
There are four ways to gain control of a blowing well.
No blowout in deepwater has yet to be successfully capped. The stab has to be guided by rovs, which are not designed to be maneuvered against the streams of a blowing well. Industry are working on underwater belt-vehicles, much like the ones used to cap land wells, to be used for deep-water stabbing operations.
Gunk (diesel & gel) or a cement with instant setting time can be used to kill a blowout.
A dynamic kill or momentum kill can be attempted either from relief well, or from drill string in well if it is in place.
We want to look at a new technique; induced bridging.

43. Fastest and Least Expensive
Mode of Control Duration
39 % 9 11 5 19 3
15%
36%
14%
10% 4% 7%
Bridging
BOP
Cement
Depletion
Equipment
Mud
Relief Well
Missed
0-1 hour
1 hour-1 day
1-3 days
3 days-1 week
1 week- 1 month
> 1 month
Missed
Skalle : Bridging is the most common method of blowout control in OCS (39.6%).
Flak : Natural well bridging would shut off most blowouts.
Adams and Kuhlman : Formation bridging is responsible for stopping many shallow blowouts.
Literature has shown that the fastest and cheapest method of blowout control is bridging. Can we induce it? And more importantly; do we want to induce it. Will an induced bridge lead to an underground blowout? Will it leave us in worse condition?
FOR MORE INFO...
SPE 53974, IADC/SPE 19917, /references/ 02_Ultra-deepwater %20blowouts.htm

44. Bridging Scenarios
Sergei Jourine has completed pioneering work on the bridging problem. His model is still in development, but it has already shown promising results.
This slide shows the general concept of his model, and all the factors that need to occur for a bridge plug to kill the blow-out.

45. 1. Well is out of Control 3 1 2
Flow and Geomechanics Models
3. Stress-Strength Relationships
2. Stress and Pressure Distributions
1. Wellbore and Reservoir Performance Relationships
The first step to calculate if bridging is to occur is to calculate the pressures in the well. This slide show the the wellbore and reservoir performance curve.

46. 2. Wellbore Instability 4 3 3. Stress-Strength Relationships
Unstable
Moderate
Stable 3
Wellbore Stability Model
4. Solid Production Potential
3. Stress-Strength Relationships
Depending on the formation surrounding the wellbore, the solid production potential may be determined.

47. 3. Solid Production 4 5 Blowout 4. Solid Production Potential
Stable Fluid Flow 5
Blowout 4
Massive Solid Production
Negligible Solid Production
Solid Production Model
5. Actual Solid Production
4. Solid Production Potential
Concentration
Time, sec
Distance, m
If the formation is highly unstable you may get massive solid production which is the building stone for a bridge plug.

48. 4a. Wellbore Collapse 5 6 5. Actual Solid Production
Flow and Geomechanics Models
6. Outflow Performance with Actual Solid Load
5. Actual Solid Production 5
Massive Solid Production
Negligible Solid Production
Total Wellbore Collapse 6
Wellbore
Bridging
As solids are being produced from the walls of the wellbore they are circulated out.

49. 4b. Bridge Formation 6 7 Blowout
Stable Fluid-Solid Flow 7
Blowout
Flow and Geomechanics Models
7. Bridge and Formation Stability
6. Outflow Performance with Actual Solid Load
Bridge
If the flow rate is sufficient to transport the solids out of the well the well will continue to flow.
However, if the flowrate is not sufficient to transport the solids out of the well bore a bridge plug will be set.

50. 5. Bridge Stability 7 Underground Blowout
Formation Failure
Underground
Bridge Failure
Wellbore
Bridging
Flow and Geomechanics Models
7. Bridge and Formation Stability
If a bridge plug is set high above the flowing formation, an underground blowout may occur.
Alternative, pressure build-up may break the bridge-plug and the well is flowing again.

51. Deep Water Tendency
From field data a most likely scenario can be constructed from our model. The blue dots illustrate the most likely path for deepwater and ultra-deep water wells.
This is in agreement with data from litterature.

52. Center for Tectonophysics, TAMU
Rock Properties
Center for Tectonophysics, TAMU
In Progress
Currently the model is being tested with cores from the deep sea drilling project.
The solid production model has other applications in production, such as sand control and cavity like completions.

53. Well, if it doesn’t bridge….
Present thinking: Relief well is the only option
MMS NTL 99-G01
Requires assurance that operator is capable of handling blowout operations such as relief well
There is a fatalistic mindset in the industry that a relief well, due to the unique geometry of a deepwater well, is the primary well-killing option for a deepwater blowout.

54. Dynamic Kill Simulator
BOP
SEAFLOOR
A dynamic kill can be done from the drillstring in the well or by drilling a relief well. The objective of a dynamic kill is to create enough frictional pressure in the blowing well to choke the influx of formation fluid. Note that the frictional pressure is directly proportional to the length from the injection point to the seafloor.
To successfully plan a dynamic kill a dynamic kill simulator should be used. It is important to determine the flowrate required to kill the influx. The flowrate determines what kind of surface equipment we need. Also, the optimum injection point needs to be determined. If the pressure in the well is high, multiple relief wells may be required.
We need a dynamic kill simulator for case studies and to verify procedures.

55. 0.052x20,000x16 = 16,640
0.052(10,000x ,000x23,4) = 16,640
10,000 ft
8.6 ppg
4,472 psi
23.4 ppg
16 ppg
What happens as we move into deeper water.
For a land well we would need a 16 ppg mud to control a psi pressure.
In ultra-deep water we need a 23.4 ppg mud to control a psi pressure. No muds available with this density.
The length of the blowing well is halved. The frictional pressure is proportional to this length, which means the circulation rate must be much higher for a deepwater well.
Same principle apply to a dual gradient well vs a conventional well, if the equipment above seafloor is still intact.
20,000 ft
16,640 psi

56. Dynamic Kill Comparison
20,000’ onshore well with 16 ppg
20,000’ deepwater well in 10,000’ of water with 16 ppg
10000’ of ’ of 23.4 ppg?
Friction pressures developed during dynamic kill could be much less in a deep water well
Can we choke it back at the mudline?
How?

57. Dynamic Kill Simulator
Static Part:
Common User Input
Static Data During Simulation
Dynamic Part :
Data that Changes with Time
Transient Effects
Computational Part:
Pressure Calculations for Given Moment in Time
The dynamic kill simulator will be separated in 3 parts.

58. Static Part Reservoir properties Formation fluid Well geometry:
#Relief Wells?
Injection Points?
Reservoir properties
Formation fluid
Well geometry:
Number of relief wells
Blowing well geometry
Inflow and outflow of kill fluid
Dual Gradient?
Hanging Drillstring?
BOP?

59. Deep Water Blowouts
4 deepwater sustained underground blowouts controlled by Boots & Coots
3 broached mud line gas flows (20” casing set BOPs installed)
1 BOP Failure Gas Blowout
No oil blowout has reported to date
The dynamic kill simulator should be applicable to any type of uncontrolled flow.
FOR MORE INFO...
Flak L.: “Control of Well Issues”, “Marine Insurance – Facing the Changed World”, International Union of Marine Insurance-NEW YORK – 2002, on-line org/Flak.htm

60. Static Part: Determine Uncontrolled Flow
Below Casing Seat:

61. Deliverables for Dynamic Kill Simulator
Fully Three Phase Transient Multiphase Flow Model
Any Possible Well Configuration
All Possible Leakage Points
Dual Gradient Drilling Option
Multiple Influx Zones
Lost Circulation at Weak Zones
Newtonian and Non-Newtonian Kill Fluid
Bridging Prediction
Simulator Written in Java Code
What we need….

62. Comparison of Dynamic Kill Simulators Available
There are many dynamic kill simulators on the market. None of them fits our need. They are either lacking functionality and/or are not available for us to use.
It is important to note that our dynamic kill simulator is not meant to compete with simulators on the market.
Dynamic kill simulator
will be a tool for us to
develop kill procedures.

63. Questions we need to answer:
Can a well be dynamically killed when half the well bore is gone?
How do you dynamically kill a well when half the well is full of sea water?
How do you model the kill operation?
Will it bridge?
Can you induce bridging?
Do you want it to bridge?
Conclusion….. these are the questions that needs to be answered in the future.

64. Question Cont. With our high reliance on bridging
Should we not understand the mechanisms of bridging better than we do now?
Should we gain an understanding of the factors that contribute to bridging?
Are there ways that we can promote bridging?
Should we not have a mechanism where we can predict where the bridge is likely located?
In long open hole sections, do we really want the well to bridge?
With our high reliance on bridging, should we not understand the mechanisms of bridging better than we do now?
Should we not gain a better understanding of the factores that contribute to bridging?
Are there ways that we can promote bridging?

65. Questions Cont. Only 1 DGD well has been drilled to date
Little thought has been given as to how a blowout on a Dual Gradient well will be killed.
Can we expect to be able to use “conventional” blowout containment methods?
Only 1 DGD well has been drilled to date and it was in less than 1000’ of water.
No one has given any thought (and reported it) as to how a blowout on a Dual Gradient well will be killed.
Can we expect to be able to use “conventional” blowout containment methods?
Vertical intervention is not likely to work if the drillstring or BOP’s are not intact.
It is likely that a relief well will have to be drilled with DGD technology.

66. Deliverables A best practice guide for blowout procedures.
A study to determine the likelihood of of a well bridging.
Ways to induce bridging.
The consequences of undesirable bridging.
A dynamic kill simulator for conventional and dual density wells
Blowout control methods for dual density wells.
Cost estimate for deepwater intervention.
A final report in electronic format.
The objectives of this study will be……