Reliability of DownholeEquipment

A One Day Course on Understanding WellEquipment and Workover Failures and

A One Day Course on Understanding WellEquipment and Workover Failures andImproving Well Production Reliability.

George E. KingFebruary 2010

Safety MomentDon’t park within the rig guy wires.

Time Between FailuresReferring to the figure,the MTBF is the sum ofthe operational periodsdivided by the numberof observed failures.

Wikipedia

Misleading Numbers? Does it mean thatthe component will last 50+ years?

What is the Definition of MTTF and MTBF?

Mean time between failures

• Mean time between failures (MTBF) is the predicted elapsed timebetween inherent failures of a system during operation. MTBF can becalculated as the arithmetic mean or average time between failures of asystem.

• Mean time to failure (MTTF) measures average time between failureswith the modeling assumption that the failed system is not repaired.

1. Definition of MTBF depends on definition of what is considered a systemfailure. For complex, repairable systems, failures are considered to bethose out of design conditions that place the system out of service andinto a state for repair.

2. Failures that are left or maintained in an unrepaired condition, and donot place the system out of service, are not considered failures underthis definition. Malfunctions?

Wikipedia

Mean, Median and MTTFMean time to failure or MTTF, (a mean life function) is widelyused as the measurement of a product's reliability/performance.Calculated by dividing total unit(s) operating time by total numberof failures. Valid only when data is exponentially distributed; apoor assumption that implies failure rate is constant.

The mean or average is the mathematical average of the data in aset. A set of numbers of 1, 2, 3, 4, 100, would have a mean of1+2+3+4+100 = 110/5 = 22

The median is the value that splits the data is half. For the 1, 2, 3,4, 100 data set, the median is 3.

Should Reliability be Expressed as a Function ofTime or Cycles at Specific Conditions?

• Associated with time, the mean time to failurebecomes a measure of reliability, e.g. :– for cyclic equipment; the reliability at 50,000

cycles should be > 50%.

for gauges, the reliability at 200oF should be 90%.– for gauges, the reliability at 200oF should be 90%.

• Reliability of a product should be specified asa percentage value with an associated time.Ideally, a confidence level should also beassociated, which allows for consideration ofvariability of data being compared to thespecification.

Common MTBF Misconceptions

A battery may have a useful life of 4 hours anda MTBF of 100,000 hrs!

What this really indicates is that for a

Wikipedia

What this really indicates is that for apopulation of 100,000 batteries, there will beapproximately one battery failure every hourduring a single battery’s four-hour life span.

Failure Data BasesExamples…

String Item

ServiceTime,

years**No. of

FailuresMTTF,years*

Blast Joint 414 0 414

Downhole Packer/Hgr 837 7 119

Expansion Joint 3198 1 3198

Fow coupling 34794 0 34795

Gravel Pack Packer 1089 3 363

Gravel Pack Screen 1561 2 780

Landing Nipple 17324 1 17324

Millout Extension 6099 0 6099

Nipple for WR-ScSSV 4372 4 1093

Perm. Gauge Mandrel 1113 111 10

Prepacked Screen 1020 1 1020

The best way to use theMTTF is as a trend. Veryhigh numbers indicatethe part is very reliable.

Prepacked Screen 1020 1 1020

Production Packer 10268 17 604

Pup Joint 72078 4 18019

Screen 800 8 100Seal Assembly -Conventional 3101 13 238

Shear Out Safety Joint 584 0 584

Sliding Sleeve 2181 2 1090

Tubing 50485 76 664

Tubing Anchor 6569 2 3284

Tubing Hanger 8818 15 588

X-Over 16711 1 16710

Very Early and Incomplete Datafrom Sintef – 1995

the part is very reliable.

** Service times of lessthan about 250 years maynot be sufficientlypopulated to yield validnumbers.

Failures – Over Time

Your Warranty Period

Hazard AssessmentRisk = frequency xconsequence

Semi Quantitative Assessments

Technical Risk Identification andManagement

• Most risk analysis are focused on safety. Thispresentation focuses on managing technical riskduring wellwork.

– Technical risk management starts after the safetyassessment is complete, understood and approved.assessment is complete, understood and approved.

– Technical risk uses success and failure histories – both inthe company and in the asset.

– Understanding the well flowing system is the single,biggest goal of the operations in Petroleum engineering.

Technical Risk – Logic Sequence

• Action – each action involves sequences of possibilities – bothgood and bad.

• Potential outcomes – identify events that may occur duringexecution of the operation.

• Signposts - By study of analogous cases, identify the earlywarning signs that predict failures.Signposts - By study of analogous cases, identify the earlywarning signs that predict failures.

• Probability– assess likelihood of occurrence of unscheduledevents.

• Contingency plans – remedial action to recover from nonscheduled events.

• Construct a learnings loop, with a a single point of accountabilityand make it work.

Adapted from SPE 52968

Managing Technical Risk Decision Tree:Outcomes and Probabilities

Actions

OutcomesProbabilities

Reactions

This becomesextraordinarycomplex veryquickly.

Actions

SPE 52968

Technical Risk Ranking

• High – could result in losing the well orrequiring extensive rig work to repair.

• Medium – will significantly extend the work orsharply increase the complexity.sharply increase the complexity.

• Low – events requiring a management ofchange process decision, but not endangeringthe well.

Inspections – Do they help?

• Are the right items inspected?

– What creates a failure – in this part or another?

• When is it inspected?

• Does the inspection stop failures?• Does the inspection stop failures?

• What is done about out-of-spec parts?

– Rebuilt, recyled or just resold?

• What changes as a result of inspection?

RBI and Materials Operating Envelopes (MOE’s)

• Risk-Based Inspection (RBI) help focus resources &increase reliability.

• RBI studies rely on assumptions of past and futureoperating conditions.

• RBI focuses much more heavily on inspection activities

21

• RBI focuses much more heavily on inspection activitiesthan controlling operations & monitoring activities.

• Knowledge / control of unit’s operating envelope helps RBIplan.

• Materials operating envelope (MOE) studies, althoughcomplements RBI. Creating a MOE typically easier &more efficient in conjunction with RBI.

Buchheim, et.al., Equity Eng. Grp, 2006

Materials Operating Envelopes (MOE’s) DefineLimits

• MOE defines limits for each part of operatingparameters in a “unit” (pH, flow rate, temps,chemical or water inj. rates, acceptable levelsof corrosive constituents, etc.).

22

of corrosive constituents, etc.).

• If limits are not exceeded, degradationshould be predictable and reasonably low.

• If limits are exceeded, excessive equipmentdegradation due to corrosion could occur.

Buchheim, et.al., Equity Eng. Grp, 2006

Assessing Potential and Specific DamageMechanisms

• Material (general and specific information including heattreatment, chemistry, strength level, etc.)

• Service exposure (general and specific), normal andupset (trace amounts of corrosives, concentration,cycles, & particularly human factors.)

• How often and how quickly does damage occur?

23

• How often and how quickly does damage occur?• Mitigating factors (coking, crack closure, residual

stresses, coatings, chemical additives, water wash)• Any monitoring data or other warning systems (probes)• Previous inspections and effectiveness at targeting

particular mechanisms• Morphology of the damage

Buchheim, et.al., Equity Eng. Grp, 2006

In-Service Damage Types

• General Corrosion

• Localized Corrosion

• Pitting, Crevice, and Grooving Corrosion

• Planar Cracks

24

• Planar Cracks

• Branched Cracks

• Metallurgical Changes & Hydrogen Effects

• Distortion

Buchheim, et.al., Equity Eng. Grp, 2006

In-Service Damage TypesGeneral Corrosion

• High Temp Corrosion and H2/H2S Corrosion

• Moderate Velocity Sour Water

25

• Oxidation

• Atmospheric Corrosion

• Some Hydrofluoric Acid (HF) Corrosion

Buchheim, et.al., Equity Eng. Grp, 2006

In-Service Damage TypesLocalized Corrosion

• Low or High Velocity Sour Water

• Dilute Acid Corrosion

• Galvanic Corrosion

26

• Galvanic Corrosion

• Corrosion Under Insulation (CUI)

• Erosion/Corrosion

• Injection Point and Dead-leg Corrosion

Buchheim, et.al., Equity Eng. Grp, 2006

In-Service Damage TypesPitting, Crevice, and Grooving Corrosion

• Under deposit corrosion (scale, microbe,wax)

• Weld area attack

Water handling

27

• Water handling

• Amine salts

• Stainless steel attach from chlorides

Buchheim, et.al., Equity Eng. Grp, 2006

In-Service Damage TypesDistortion

• Bulging

• Blistering (H2)

28

• Creep

Buchheim, et.al., Equity Eng. Grp, 2006

Monitoring Methods

• Corrosion probes

• Hydrogen probes

• Coupons and physical probes

29

Coupons and physical probes

• Various measurements and scanning

• Stream samples

• Process variable monitoring

Buchheim, et.al., Equity Eng. Grp, 2006

Mitigation Methods

• Physically modify the process– Change temperature and/or velocity– Removal of stream fractions

• Chemically modify the process– Water washing

Injection of chemicals to change pH or tie up constituents

30

– Injection of chemicals to change pH or tie up constituentsor to form a film barrier

• Isolate the environment from the material– Organic coatings and thermal spray coatings– Metallic linings– Weld overlay

• Upgrade the materials

Buchheim, et.al., Equity Eng. Grp, 2006

Sources of Outcome, Probability, andContingencies

1. Review:• Standard industry operation prediction vs. asset

experience. Pick best-in-class operators andstudy their approach.

• How do best practices reduce problems.• How do best practices reduce problems.

2. Historical data bases.

3. Probability distribution function. Careful –lumping data together can be misleadingunless you understand individual conditions.

Adapted from SPE 52968

Where to begin?

• Historical performance – what has worked and whathas not?

• Are the metrics correct? Do they help or mislead?

• What early warning “flags” were evident in post• What early warning “flags” were evident in postappraisals of failures?

• What best (and worst) practices were observed?

Let’s Start

• First: Workover/Intervention evaluation providesthe most clues about failures.

– What is broken

– Why– Why

– What can be repaired

– What can be prevented

• Second: Production operations provides the mostclues about best practices. However, we seldomstudy things when they are working correctly.

Scorpion Plot• A scorpion plot is a plot by rate of production change

vs. economic return for workover/intervention jobs .

Cu

mu

lati

ve

Exp

en

datu

re

Uneconomic

Cumulative Productivity Gain

Cu

mu

lati

ve

Exp

en

datu

re

HighlyEconomic

ModeratelyEconomic

BorderlineEconomic

SPE 30649

Look at the type of job in each section of a

field-specific scorpion plot.

Highly Economic

11%

8%

22%

41%

7%8%

Acid Jobs

Cleanout

Cleanout & Stimulation

Conversion of GL to ESP

ESP Replacement

Integrity Repair

Moderately Economic

9%6%

11%

6%23%

16% Acid Jobs

Cleanout

Cleanout & Stimulation

Conversion of GL to ESP

ESP Replacement

Integrity Repair

Others

Lift Optimization

3%

41%Others

Borderline Economic0%

4%

8%

0%

17%

31%

40%

Acid Jobs

Cleanout

Cleanout & Stimulation

Conversion of GL to ESP

ESP Replacement

Integrity Repair

Others

29%Others

Uneconomic

11%

11%

7%

7%0%0%0%

4%4%

19%

4%

4%

29%

Acid Jobs

Cleanout

Cleanout & Stimulation

Conversion of GL to ESP

ESP Replacement

Integrity Repair

Others

Conversion of BL to ESP

Convesion of ESP to GL

Conversion to Water Inj.

ESP Installation

Recompletion

Water Shut-Off

SPE 88025

!

The number of jobs performed may holdsome clues to success.

Highly Economic (# jobs, % of total)

23, 11%

16, 8%

45, 22%

85, 41%

15, 7%

17, 8%Acid Jobs

Cleanout

Cleanout & Stimulation

Conversion of GL to ESP

ESP Replacement

Integrity Repair

The more the field doesa job, the better theyget at it if learnings areincorporated as the jobsprogress – for example,lift optimization.

6, 3%

85, 41%Integrity Repair

Others

Uneconomic

3, 11%

3, 11%

2, 7%

2, 7%

1, 4%

1, 4%5, 19%

1, 4%

1, 4%

8, 29%

Acid Jobs

Cleanout

Cleanout & Stimulation

Conversion of GL to ESP

Conversion of BL to ESP

Convesion of ESP to GL

Conversion to Water Inj.

ESP Installation

Recompletion

Water Shut-Off

lift optimization.

Some jobs, like watercontrol, remain verylow success ratebecause the symptomis treated, instead ofthe problem.

Some jobs have a high amount of risk -Fishing Limits

• Continued fishing past some economic limit is failure,even if the fish is retrieved and that portion of thehole is saved. (SPE 9102)

• Changes of recovering a fish decrease sharply withtime.time.

• Cost of the fishing operations (per day) must bematched against the sunk cost of the well.– Time and cost for a single round trip – historical time

– Knowledge about the fish and how to retrieve – BestPractice is drawings of BHA’s, photographs, camera work,etc.

What can be done to improve knowledge?

Approximate Drilling Fishing Success vs. Time

Successful Fishing

6

7

Why does success decrease with time? What jobs are highly impossible? Who makes thedecisions? When do you quit?

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Time (hrs)

Fre

qu

en

cy

SPE 9102

Cased hole fishing success with time.

This is a compilation of several jobs,but it indicates that cased holefishing is usually quickly successful.

If the fish is not quickly recovered,there are usually severe wellborethere are usually severe wellborecomplications that will prevent fishrecovery. What are they?

Fishing Time in Days

Going beyond equipment - Formation perm is NOTconstant – Effect of Declining Reservoir Press.

What are the consequences of reduced permeability – can it be avoided?

Dominance of Large Pores on Permeability – If thelarge pores are damaged, productivity drops!

20 m

Large pores and naturalfractures dominatepermeability.

20 m Not every instance is alike –pore throats, wetting layers,condensate phases, clays,minerals and fines influence theactual perm.

How can this potential problem be avoided? What has to change in the completion &workover approaches?

Risk of a Blowout – general data from 1960to 1996 – SPE 39354

Area Blowouts Number ofwells drilled

Blowouts per100 wells

The outcomes, probability and even the warning signs are obtainable from casestudies.

Thought and investigation are needed to understand the data.

wells drilled 100 wells

Louisiana 123 29,000 0.42

Mississippi 20 11,000 0.18

OCS 245 180,000 0.14

Texas 450 310,000 0.15

Totals 832 380,000 0.16

Although average numbers present one view, numbers for a specific area (Louisiana)indicate the risk may be higher in that area. Why?

Blowouts vs. Type of Operations During Completions

Operation BO Activity BO

Installing equip. 25 Nipple down BOP 5

WOC 5

Casing running 3

Cementing casing 2

Fishing 1

Stuck pipe 1

LOT 1LOT 1

Set well plugs 1

Well Tests 10 WOC 5

Cementing Casing 2

Tripping In 1

Tripping out 1

Squeeze cementing 1

Circulation 5 Killing 2

Perforating 1

Cleaning the well 1

Gas Lifting 1

SPE 39354

Blowouts vs. Type of Operations During Workover

Operation BO Activity BO

Pulling Well Equipment 37 Pull Tubing 15

Stuck Pipe 4

Pull/Drill plugs 3

Pull WL 2

Logging 2

Perforating 1

Cleaning well 1

Snubbing out 1Snubbing out 1

Installing Equipment 17 Run Tubing 5

Install BOP 3

Run WL 2

Nipple Down BOP 1

Set Well Plugs 1

Acidizing 1

Abandon Well 16 Pull Tubing 8

Set Well plugs 4

Killing 2

Nipple down tree 1

Pull/Drill plugs 1

SPE 39354

Distribution (%) of Operation Phase Failures

Primary Barrier Exp Drlg Dev Drlg Complete Prod W/O WL

Swabbing 25 40 4 0 22 3

Low mud wt. 30 30 10 2 18 4

Drlg break 62 31 0 0 2 0

Form break dn 42 42 0 5 8 0

Wellhead failed 12 8 3 45 18 0

Trapped gas 18 28 13 0 35 3

Distribution of percent of specific failed barrier in blow out.

Gas cut mud 40 23 13 0 20 3

X-mas tree fail 0 0 0 71 29 0

Cement setting 16 10 79 0 5 0

SecondaryBarrier

BOP fail - close 34 44 4 1 24 4

BOP fail after 34 27 13 2 10 3

BOP not instal 18 13 30 0 43 0

Frac csg shoe 15 65 8 4 8 0

Fail to stab val. 26 24 3 3 47 0

Csg leaks 42 17 4 33 3 0SPE 39354

But, wait…..

• To have a failure, several steps in our designand application methods must fail.

• To prevent the failure, you must understandthe reasons it happened and how the systemthe reasons it happened and how the systemworks.

• Take the case of a well blowout.……..

Well Blowout Event Pyramid

110 kicks taken

1blowout

Would 1 blowout per 733wells be a good number forrisk evaluation?

Could you use surveillance tosharpen it?

Actually, the blowoutnumber averaged0.15 blowouts per100 wells, but ranged

733 wells drilled

In the late 60’s and mid 90’s, the blowout rate was 1 blowout per 2500 wells.

In high volume years (early 80’s) the blowout rate jumped to 1 blowout per 366wells. Why? Inexperienced crews and faster pace are factors, but new drilling areasand increasing depth also had an effect.

100 wells, but rangedfrom 0.05 to over 0.4in specific areas.

Raw data from SPE 39354

First - Kicks Are Well Activity Specific

Drilling Kick Stats by Operation in

Progress

1000

1200

Num

berofkic

ks

0

200

400

600

800

1000

Drillin

g

Tripping

Out

Tripping

In

Cas

ing

Circ

ulating

Testing

Other

Operation in Progress

Num

berofkic

ks

SPE 19914

But, for a kick to become a blowout, a barrier must fail

Primary Barrier BO Secondary Barrier BO

Swabbing 158 BOP failed to close 78

Too low mud weight 50 Rams not seated 14

Drilling breakthrough / unexp. high pressure 45 Unloaded too quickly 13

Formation breakdown – lost circulation 43 DC / Kelly / TJ / WL in BOP 5

Wellhead failure 40 BOP failed after closure 66

Trapped / expanding gas 40 BOP not in place 43

Gas cut mud 33 Fracture at casing shoe 38

X-mas tree failure 23 Failed to stab valve / Kelly / TIW 34

While cement setting 20 Casing Leakage 23

Unknown 19 Diverter – no problem 21

Poor cement 16 String safety valve failed 19

Tubing leak 15 Diverter failed after closure 17

Improper fill-up 13 Form. breakdown / lost circ. 15

Tubing burst 10 String failure 13

Tubing plug failurel 9 Casing valve failed 11

Packer leakage 6 Wellhead seal failed 10

Annular losses 6 Failed to operate diverter 7

Uncertain reservoir depth / pressure 6 X-mas tree failed 7

SPE 39354

What warning signs were apparent? When are they seen?What are the contingencies?

UKCS Kick Data – All Wells1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Total

orAvg.

NumberUKCS Wells

344 338 348 330 298 272 301 343 378 355 361 3668

Wells w/ kicks 34 32 45 41 34 27 39 38 45 39 36 410

% Wellsw/kicks

10 9 13 12 11 10 13 11 12 11 10 11%

SPE 56921

Kick frequency and the percentage of kicks that turn into blowoutsvary sharply throughout the world.

The percent of wells taking kicks doesn’t vary that much.

UKCS Kick Data - Workovers1995 1996 1997 1998 Total

orAvg

Estimated number ofworkovers

370 450 530 430 1780

Number of kicks during 5 10 3 2 20Number of kicks duringworkovers

5 10 3 2 20

Number of gas releases duringworkovers

1 2 1 6 10

% Workovers with kicks orreleases

2 3 1 2 2%

SPE 56921

UKCS Kick Data - Workovers1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Total

Cased Hole /WorkoverKicks

7 10 11 8 5 9 5 8 20 6 5 94

?

SPE 56921

The number of kicks during workovers often has to be viewedover time to gain consistency. Low well population can skewthe numbers.

?

Understand the Data! – Two ways oflooking at drilling kick rate.

Alberta Kick Rate by Depth of Kick

20

25

Kic

kR

ate

,k

ick

sp

er

10

0w

ells

Alberta Kick Rate by Depth of Well

40

50

60

Kic

kR

ate

,k

ick

sp

er

10

0

0

5

10

15

0 to1000 1001 to 2000 2001 to 3000 3001 to 4000 GT 4000

Depth of Kick

Kic

kR

ate

,k

ick

sp

er

10

0w

ells

SPE 19914

0

10

20

30

0 to1000 1001 to

2000

2001 to

3000

3001 to

4000

GT 4000

Well Depth, m

Kic

kR

ate

,k

ick

sp

er

10

0

we

lls

Kicks may vary with activity, experience and costdrivers. Alberta Kick Study

• 10 year study (1979-1988), 62632 wells, 2457 kicks.

• Average kicks = 3.9 kicks per 100 wells.– Exploratory – 5.7 kicks/100 wells

– Development – 3.2 kicks/100 wells

• By Year:• By Year:– 1979, 4.0 kicks/100 wells

– 1982, 2.6 kicks/100 wells

– 1986, 4.8 kicks/100 wells (lowest well count)

– 1988, 4.1 kicks/100 wells

SPE 19914

Lowest profitmargin, poormaintenance,low personnelcounts,

Technical Risk Management ThroughSurveillance

• A routine surveillance plan is critical toefficient reservoir management.

• Surveillance gathered on a healthy well, is thebasis for diagnosing problems.basis for diagnosing problems.

• All of it fits under the depletion plan.

Intervention Sequence

1. Management understanding of the need and the rewards of a strong basemanagement program.

2. Continuous investigation (surveillance) of specific data

3. Assessment of data => understanding of problem / potential.

4. Creation of a fit-for-purpose solution => candidate selection refinement!

5. Consistent ranking of all candidates:5. Consistent ranking of all candidates:

– Possibility of success

– Return on investment

– Grouping to allow learnings-related workover program

– Availability of people, equipment, money and weather.

6. Post job review, feedback into understanding loop.

7. Results tracking, learnings sharing, management updates.

Asset Value Opportunity

Accelerated, high percentage reserverecovery is the primary driver behindeconomic operation.

We know the least about a reservoir whenwe locate, design, drill and complete thewells.

Surveillance during production will change

SPE 52968

Surveillance during production will changethe way we understand the reservoir – buthow do can we capitalize on it?

Asset Value Opportunity – BaseManagement

SPE 52968

What makes an exceptional wellworkcampaign?

• Success of well intervention campaignsdepend on:

– Quantity / quality of candidates.

– Knowledge and experience of field application– Knowledge and experience of field application

– Ability to monitor and learn from the wellwork.

Expendatures in Successful Well Work Program

Successful well work programs have all requiredactions listed and funded

Repair, 30%

Enhancement,

24%

Integrity, 27%

Surveillance,

13% Other, 6%Repair

Enhancement

Integrity

Surveillance

Other

SPE 88025

The Value of Well Work is Often UnnoticedUntil It is Done.

Production with & without Well Work

Pro

du

cti

on

Rate

Enhancement

Repair / Optimization

Base w/o Well Work12 % Decline

Time

Pro

du

cti

on

Rate Base w/o Well Work

20% Decline

Wellwork can accelerate recovery andadd reserves.

Production with & without Well Work

Pro

du

cti

on

Rate

Enhancement

Repair / Optimization

Base w/o Well Work12 % Decline

Time

Pro

du

cti

on

Rate Base w/o Well Work

20% Decline

Economic Limit

Added Reserves Recovered

25%

65%

40%

60%

80%

Norwegian Study of Reserve Recovery vs. Tree

Location

Data from Gulfax and Statfiord Fields in Norwegian North Sea, data provided by Statoil

0%

20%

40%

Wet Tree Dry Tree

Recovery

Reliability – what will be discussed

• Downhole and Surface Equipment

• Completion Types

• Workover Techniques

• Production Operations

• We will bounce back and forth because all are tiedtogether – the way a well is drilled and completedaffects its life span, but in a sometimes differentmanner than production operations.

Downhole Equipment Importance

Tracking Trends – two operators

Trends – Moving Averages

Optimization Effect on Failures

How?

What?

Continued

Failure Frequency - FPY

Completions and Equipment

• Completion design, age and specificconditions influence failure rate.

• Equipment improvements can be majorchange elements.change elements.

Equipment - Overall ScSSV Reliability -Sintef

Overall ScSSV Reliability w/ Time

20

25

MT

TF

(years

)

WR = wireline retrievable

TR = tubing retrievable

0

5

10

15

20

WR

1977-

82

WR

1982-

86

WR

1987-

89

WR

1989-

92

TR

1977-

82

TR

1982-

86

TR

1987-

89

TR

1989-

92

MT

TF

(years

)

SPE 26721

TR = tubing retrievable

ESS and OHGP’s havedramatically improved inreliability for some operatorsreliability for some operatorsover time.

Sand Control – the concept…..

• Sand control may be sand exclusion (no sandproduced) to sand management (where somesand is produced).

– They have sharply different failure rates– They have sharply different failure rates

– They have sharply different failure places

– They have sharply different failure costs

Very good slide from Shell Brunei – What range can you operate in?

Sand management – drill cheap wells and take care of sand at surface – but manage erosion.

!

Note that in the Shelldata that sand declineswith time, but bursts ofsand will be producedwhen system flowingparameters change –KEEP IT STEADY!

Offshore and Onshore Trends

• Failure and Intervention Frequency is verydifferent for on-shore, platform & sub-sea wells

Sub-Sea Intervention Estimates(mix of GOM, North Sea, Brazil and West Africa)

307 7 6

1 1

Hydrate Removal

SS Tree, Choke and Pod

ScSSV System

Gathering and Flow Lines

30

12128

8

8

7 7Paraffin & Asphaltenes

Recompletion & Water Control

Sand Control Repair

Fill Removal

Scale

Lift & Unloading

Annular Pressure Damage

Selected Interventions – Completion and Production Issues - Omits Abandonments.

Cautions

1. Failure data and failure perception is often an artifact of thedata source. A person’s memory is selective and biased.Trust the overall well file data, but be cautious of specificdata points.

2. Failure definition & explanation are frequently disputed.

3. Short life zones skew the data towards longer perceived3. Short life zones skew the data towards longer perceivedreliability.

4. New field/well data (<5 yrs) skew the data towards longerperceived reliability.

5. Limited well numbers skew the data all over the place.

6. Failure frequency is NOT intervention frequency. Failuresoften stumble along for years at reduced rates.

Failure Definition

• Which is failure?1. Sub 40 micron fines produced at full flow.

– No – if design was not to plug.

2. Sand produced at full well flow.– Yes – economics based on full flow

3. Well production possible but choked back– Still a failure – anytime production is limited.

4. Well intervention required– Depends – water control? open new zone?

Classifying Failures

• Design failures – poor design or data frac

• Application – pumping or Q/C problem

• Infant failure – within 30 days

• Production failure – during producing life• Production failure – during producing life

• Failure by subsidence – a disqualifier

• Fines production? – depends on size

Calculated Pressure Drop Through Frac Pack

in Campos Basin

2000

2500

3000

Dra

wd

ow

n,

psi

0

500

1000

1500

0 10 20 30 40 50 60

Wells

Dra

wd

ow

n,

psi

SPE 73722

1 Darcy

250 md

20,000 bpd, 100 ft screen,

calculated flux = 200 bpd/ft PL measuredflux = 700 bpd/ft

Flux as a production limit? Yes, but measure,don’t calculate it.

100 md

50 md

Localized hotspots fromlayered flow can destroyscreens

Weave Damage on an Eroded Screen

01/28/98

Hot spot plugging and subsequent erosion can occur anywherewhere a hot spot develops.

BP - Trinidad

Fracturing may help link layers and avoid “hotspots” near the wellbore.

Does it work the same in horizontal wells?

Problems for SC Completions

• Too thin or too thick gravel between screen andcasing in OHGP/CHGP. Min > ¾”, Max <2.5”?

• Access to wellbore in high rate wells.

• Not getting perforations clean before packing (acidprepacks worked pretty well)prepacks worked pretty well)

• Unpacked perforations in Frac packs and HRWP?

• Mobile fines with any screen or “tight” gravel(Saucier = 6)

• Abrasion, hot spots and plugging

• Shales > 10% of interval

Failure Causes – Sand Control

Design21%Flux

17%

Sand Control Failure Root Cause

Compaction12%

sand control infantmortality

12%

Sand control failureunclassified

38%

Sand Control - Some Background

• Much of the oil and gas production currentlybeing discovered and developed offshore willrequire sand management approaches.

• Some sand control methods have as little as -1 to• Some sand control methods have as little as -1 to+2 skins – Others generate +10 to +15 skins.

• Failure rate is NOT the same as intervention orrepair rate. Many fail, few are repaired.

• A failure is defined by inability to operate the wellat designed rates because of sand production.

Screen Failures – two main types

• Installation failures – damaged screen duringrunning.

• Production failure

– Usually most of the screen plugged with fines and– Usually most of the screen plugged with fines andthen flow erodes a hole.

– Less often – subsidence creates compression andscreen distortion.

A few things that cause screens to fail………

• Running screens– drag, sharp turns, windows, dope, shale,

• Pumping past screens– erosion, pressure, screenouts

• Pumping through screens• Pumping through screens– Fines in packing fluid, rate, volume

• Producing through screens - plugging– Fines in drill-in fluid, mobile fines, pressure drop

• Compaction loads

Top of screen 1 to 2 joints abovetop perfs or top of pay in openhole

Multi position gravel pack packerwith large crossover port forhigher rates (120% of tubingarea).

90 to 120 ft of blank pipe above thescreen – serves as a gravel reserve(along with the screen above thetop perf)

Annular clearance 1” to 3”between screen and casing oropen hole.

Slurry flow path pickled toremove dope, mill scale, mudand rust.

Undamaged Screen placed incorrect position – centralized.

holeopen hole.

Sump packer 5 to 10 ft frombottom perfs

Gravel displacementoutside perfs at least 45lb/ft – more can bebetter.

Washpipe inside screen 80% ofscreen ID

Perfs – 12 to 27 spf, DP or big holeand CLEAN!

Clean, low debris proppant sizedfor formation sand retention andmax permeability

Minimum blanks in screen meanminimum voids in pack.

Liquidreturn tosurface

slurryFailure points in the flow path duringfrac or gravel packing:

1. Crossover port

2. Casing oppositecrossover port.

3. The annular area between screen and casingwall.

a. Erosion from high velocity linear flow –a. Erosion from high velocity linear flow –minimal problem

b. Erosion from high velocity flow as the slurryenters a perforation.

c. Pressure drop in this area during high rateflow (fracs) can collapse screens – problemsare very rare, but watch clearances.

Flow Capacity of Clean Screen

6

8

10

12

Pre

ss

ure

Dro

p,p

si

2-3/8"

2-7/8"

0

2

4

6

0 5000 10000

Flow Rate, BPD/ft

Pre

ss

ure

Dro

p,p

si

2-7/8"

3-1/2"

Sand Control Reliability - problems

• Skin damage– Reservoir-to-wellbore limits

– Invasion of fines into gravel

– Crushing/breaking of gravel– Crushing/breaking of gravel

• Physical Damage– Screen Running Damage

– Erosion during production

– Corrosion – from produced and injected fluids

Primary Erosion Locations

• Directly opposite perforations

• sharp turns in the flow path

• where gas velocity is maximum

• eddy current and similar patterns• eddy current and similar patterns

• constrictions in the flow path

WirewrapWirewrap Screen Erosion w/ AirScreen Erosion w/ Air

Erosion tests by Baker - erosion failure can happento any screen – any design. Correct Application isabsolutely critical.

Damage created by running and pulling screen can result in immediate failure.

Failed screen during an Angolawater injection DST. Load setdown on screen caused amechanical burst.

Design Learnings

1. Crossover ports should have an area at least as large as the tubing area,preferably 130%.

2. Crossover ports should be shaped to assist the slurry direction changeand minimize turbulence.

3. The area of the screen/casing annulus should be 20% larger than thetubing – keep pressure drop below about 500 psi/100 ft.tubing – keep pressure drop below about 500 psi/100 ft.

4. Zones with very high permeability streaks may bridge the annulus withdehydrated sand plugs – special design is required.

5. Never stand screens in the derrick.

6. Use a screen table to run.

7. Adequate make-up room needed at joints, but minimize blanks wherevoids may occur.

Some Database Learnings

• Screen type as a failure factor in frac packing isovershadowed by:

– Running damage to the screen

– Screen-to-casing clearances– Screen-to-casing clearances

– And, “maybe” how pack is placed after frac.

Classic Failure Rate

25

30

35

40

45

50

#F

ailu

res

Early Failure –usually < 30 days.

Wearing out– period ofaccelerated failures

0

5

10

15

20

25

0.1 1 10 100

Time

#F

ailu

res

Long term stability

Completion Failure Rate by Age and Completion

Type

10

12

14

Co

mp

leti

on

sin

tha

tA

ge

Gro

up

Th

at

Fa

il(S

an

dC

on

tro

l

SOC=14/124CHGP=3/26

SOC=10/28

Sand Control - Production Failures

0

2

4

6

8

10

0 2 4 6 8 10 12

Completion Age, yrs

Co

mp

leti

on

sin

tha

tA

ge

Gro

up

Th

at

Fa

il(S

an

dC

on

tro

lF

ailu

res

) SOC

CHGP

OHGP

FP

CHGP =2/45

FP = 0/153

OHGP=0/18

FP=3/766

9/144

4/81

FP=0/27

FP = 0/70

OHGP = 0/2

CHGP 4/304

Snap Shot of Sand Control Failures - 2 yrs oldType ofCompletion

#Wells

Wellyears

DesignFailure

ApplicationsFailure

EarlyFailure

ProductionFailure

% Attempts % Attempts % Attempts Failures /well / yr

Screen Only 183 783 0.6 0.0 0.6 0.06

Cased HoleGravel Pack

369 1514 0 2.2 0.8 0.011Gravel Pack

Open HoleGravel Pack

175 507 0 9.7* 0.6 0.016*(<0.01?)

High RateWater Pack

187 544 0 0.5 0.5 0.009

Frac Packs 844 3369 1.7 2.4 0.2 0.004

Total Wells 1758 1617

* Skewed by early learning failures, probably cut at leastin half for better operators

Tubular Failures

• Mechanical

– Running problems

– Performance over time

– Collapse and Burst from operational forces– Collapse and Burst from operational forces

• Trapped annular pressuring

• Lift problems

• Corrosion

Well Failure Statistics

20%

25%

30%

35%

'94-2000'

'90-'93

We

llsD

eve

lop

ing

Leak

s

0%

5%

10%

15%

20%

0.0 5.0 10. 15. 20. 25. 30. 35. 40.

'90-'93

'80-'89

'70-'79

'58-'70

Vastar Putnam

We

llsD

eve

lop

ing

Leak

s

Well Life Prior to Casing Leak, Years

Note that well life, time period of completion and operator makessignificant differences in failure rate.

0.20

0.30

0.40

0.50

0.60

0.70

Failure

Rate

Casing Field Failure History

1990's

1980's

Con

nect

ions

Collaps

e

Wea

r

Brittle

unkn

own

1990's

0.00

0.10

0.20

Failure Mode

Time

Period

Source – Brian Schwind - PPI

Some failures have been decreased with time because of knowledge growth. Somefailures are sharply increased as new areas (deep water and increasing depth) areentered.

What happened?Corrosion?Change of operating conditions wasa major factor.

Deep Water Well Pipe Collapse

A simple mistake like forgetting to fill the pipe whilerunning (closed-ended) into a mud-filled riser cancontribute to pipe collapse.

Trapped Pressure Also a ProblemTrapped annular pressure from leaks or outside charging can bea factor. The problems include inner pipe collapse and outerpipe burst. Straight API strength ratings may be altered bycorrosion (inside and outside), pressure support from otherzones (increasing burst and/or collapse resistance), wear,tectonic loads, etc.)

Downhole Camera Inspection

Mechanical Caliper

What happens when the tubing isdropped in the well?

Mechanical Caliper

Abrasion Damage

Generally a problem inolder wells, deviated wellsand wells with a history ofmilling or fishing. It is verydifficult to spot and tofactor effect on pipestrength or equipmentstrength or equipmentreliability.

Erosion – by Formation Sand

A by-product of sand production in many cases. Thisone started with a small leak. Erosion is one of thefastest failure causes and can work to acceleratecorrosion with even very small amounts of solids in thefluids.fluids.

Control Lines

Running and protecting control lines is achallenge. Fishing them is even morechallenging.

Ice DamageIcing in wells – can be from weather or Joule-Thompson coolingof gas expansion.

Monobore:mixed grades,same weight

Mixed gradesand weights

Mixedweights,samegrade

Casing Design Options – think about running and setting packers.

Small diametersSmall diametersat the top of thewell may prevententry by somepackers.

5/26/2010 126

Allowing Tubular Movement

• Usually incorporate a PBR - polished borereceptacle, for a “stinger” or seal assembly toslide through.

• Shoulder out on the PBR - if it can move, it will• Shoulder out on the PBR - if it can move, it willeventually leak.

• Seals must match operating extremes as wellas general conditions.

5/26/2010 127

Wear Patterns in a Well. Why?

0.55

0.6

Re

ma

inin

gW

all

Th

ick

ne

ss

,in

.

Distribution of Wear in Recovered 9-5/8", 53.5 lb/ft, Q-125 Casingwith Initial 0.545" wall

Crescent Shaped Wear

0.3

0.35

0.4

0.45

0.5

0 5 10 15 20 25 30 35 40 45

Re

ma

inin

gW

all

Th

ick

ne

ss

,in

.

Joint Number

Uniform Wear

Crescent Shaped Wear

Corrosion

• Corrosion Basics– General corrosion theory

– Corrosion examples

• Specialty Problems• Specialty Problems– CO2 and H2S

– O2 in sea water injection

– Acid Treatment

– Packer Fluids

Major Causes of Corrosion

• Salt water (excellent electrolyte, chloride source)

• H2S (acid gas with iron sulfide the by-product)

• CO2 (Major cause of produced gas corrosion)

• O• O2 (key player, reduce any way possible)

• Bacteria (by products, acid produced)

Other Factors

• pH

• Chlorides (influences corrosion inhibitor solubility)

• Temperature (Increase usually increases corrosion)

• Pressures (CO2 and H2S more soluble in H20)Pressures (CO2 and H2S more soluble in H20)

• Velocity - important in stripping films, even forsweet systems

• Wear/Abrasion (accelerates corrosion)

• Solids – strips film and erodes metal

Chemical Corrosion

• H2S– weak acid, source of H+

– very corrosive, especially at low pressure

– different regions of corrosion w/temp.

• CO2• CO2– weak acid, (must hydrate to become acid)

– leads to pitting damage

• Strong acids - HCl, HCl/HF, acetic, formic

• Brines - chlorides and zinc are worst

O2 Corrosion

Dissolved Gas Effect on Corrosion

Overa

llC

orr

osio

nR

ate

of

Carb

on

Ste

el

There is no corrosion mechanism more damaging on a concentration basis than oxygen – smallamounts of oxygen, water and chlorides can ruin a chrome tubing completion in a few months.Injection wells are the most severely affected – minimise oxygen and don’t use chrome pipe ininjectors.

20 ppb O2 limit for seawater incarbon steel injection tubulars.

0

5

10

15

20

25

0 1 2 3 4 5 6 7 8

Overa

llC

orr

osio

nR

ate

of

Carb

on

Ste

el

O2

CO2

H2S

Dissolved Gas Concentration in Water Phase, ppm

0 1 2 3 4 5 6 7 8

0 100 200 300 400 500 600 700 800

0 50 100 150 200 250 300 350 400

O2

H2S

CO2

carbon steel injection tubulars.13Cr is CO2 resistant but verysusceptible to pitting corrosionin aerated brines. 5 ppb O2 issuggested as a limit, but eventhese levels have not beenconfirmed.

5/26/2010 133George E. King, Engineering

GEKEngineering.com

REMOVAL OF “PROTECTIVE” FILM

Corrosion - Best Practices

• Adopt a corrosion managementstrategy.

• Be aware of corrosion and erosioncauses.

• Completion planning must reflectcorrosion potential over well’s life.

corrosion in tubing exacerbated byabrasion from wireline operators.

corrosion potential over well’s life.

• Develop maintenance programs,measure corrosion.

• Know the corrosion specialists.

• Ensure inhibitors are compatible withmaterials and the reservoir!

• If tubing corrosion is suspected, DONOT bullhead fluids in the formation.

1970’s Industry Study of Failures

Method % of Failures

Corrosion (all types) 33%

Fatigue 18%Fatigue 18%

Brittle Fracture 9%

Mechanical Damage 14%

Fab./Welding Defects 16%

Other 10%

Same factors still a problem -

Causes of Petroleum Related Failures(1970’s study)

Cause % of Failures

CO2 Corrosion 28%

H2S Corrosion 18%

Corrosion at the weld 18%

Pitting 12%

Erosion Corrosion 9%

Galvanic 6%

Crevice 3%

Impingement 3%

Stress Corrosion 3%

Corrosion Types

• Galvanic – a potential difference between dissimilar metals in contactcreates a current flow. This may also occur in some metals at the grainboundaries.

• Crevice Corrosion – Intensive localized electrochemical corrosionoccurs within crevices when in contact with a corrosive fluid. Willaccelerate after start.accelerate after start.

• Pitting – Extremely localized attack that results in holes in the metal.Will accelerate after start.

• Stress Corrosion – Occurs in metal that is subject to both stress anda corrosive environment. May start at a “stress riser” like a wrench markor packer slip mark.

Corrosion Types

• Erosion Corrosion – Passage of fluid at high velocity may remove

the thin, protective oxide film that protects exposed metal surface.

• Hydrogen Sulfide Corrosion – H2S gas a water creates an acid

gas environment resulting in FeSx and hydrogen.

• Hydrogen Embrittlement – Atomic hydrogen diffuses into the• Hydrogen Embrittlement – Atomic hydrogen diffuses into the

grain boundary of the metal, generating trapped larger molecules ofhydrogen molecules, resulting in metal embrittlement.

• Hydrogen Corrosion – Hydrogen blistering, hydrogen

embrittlement, decarburization and hydrogen attack..

CO2 Partial Pressure

• Severity is a function of the partial pressure

– 0-3 psi - very low – non chrome use possible

– 3-7 psi – marginal for chrome use

– 7-10 psi – medium to serious problem– 7-10 psi – medium to serious problem

– >10 psi – severe problem, requires CRA even forshort term application.

Partial pressure is the mole fraction of the specific gas times the total pressure. If the CO2 moleconcentration is 1% and the pressure is 200 psi, the partial pressure is 0.01 x 200 = 2 psi.

CO2 corrosionCO2 partial pressure vs. corrosion. Notethat mpy or mills per year of total wear maybe low but pitting damage can be very high.

CO2 localised attack in 7” productiontubing

Severe CO2 corrosion in tubing pulled from a well. One reason for the attack was that thetubing was laying against the casing, trapping water that was replenished with CO2 fromthe gas flow.

Thinned and embrittledtubing twisted apartwhen trying to breakconnection during atubing pull.

Corrosion weakened pipe – large areas can be affected.

This is the starting point – but whatcaused the break? Clues are foundin the well environment, the historyand the area surrounding the break.

CO2 CORROSION ISOPLOT

0.8

0.9

1.0

pH 5

0.9 - 1.0 mm/y 0.8 - 0.9 mm/y 0.7 - 0.8 mm/y 0.6 - 0.7 mm/y 0.5 - 0.6 mm/y 0.4 - 0.5 mm/y

30 40 50 60 70 80 90 100 110 120 130 140 150

-4

-3

-2

-1

0

1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Temperature, degC

Log(Pco2)

0.4 - 0.5 mm/y 0.3 - 0.4 mm/y 0.2 - 0.3 mm/y 0.1 - 0.2 mm/y 0.0 - 0.1 mm/y

07-01: What the liner looked like

• Well completed on 1/15/98

• 2-7/8” L-80 liner

Source – Jennifer Julian, Alaska BU

Liner Corrosion-7-01

• Lessons learned:– Smaller tubulars corrode/erode faster than larger

tubulars

– Before doing a RWO, strongly consider a linercalipercaliper

– Production response to corroded calipers can bevery gradual (i.e. they don’t look “broken”) withproblems occurring with steep changes in gasproduction.

– Run chrome liners.

Source – Jennifer Julian, Alaska BU

Mills/per year or mm/yr may not be a goodindicator when the metal loss is in pitting.

Trench corrosion common from CO2 attack.

Chloride Stress Cracking

• Starts at a pit, scratch or notch. Crackproceeds primarily along grain boundaries.The cracking process is accelerated by chlorideions and lower pH.ions and lower pH.

Stress Sulfide Corrosion

• Occurs when metal is in tension and exposedto H2S and water.

• Generates atomic hydrogen. Hydrogen movesbetween grains of the metal. Reduces metalbetween grains of the metal. Reduces metalductility.

Grading of pipe is useful, but doesn’t tell the whole story – e.g. N-80 and L-80.

Domain Diagram for C110

Domain Diagram for Super 13Cr

pH 4.5

5.5

ACCEPTABLE

0.03bara

pH 4.5

5.5

ACCEPTABLE

0.03bara

3.5

0.001 0.01 0.1 1.0

pH2S (bara)

Domain Diagram For The Sulphide Stress Cracking Limits

Of 95ksi Super 13Cr Alloys In High Chloride (120,000 ppm Cl-) Waters

UNACCEPTABLE

FURTHER ASSESSMENT REQUIRED

3.5

0.001 0.01 0.1 1.0

pH2S (bara)

Domain Diagram For The Sulphide Stress Cracking Limits

Of 95ksi Super 13Cr Alloys In High Chloride (120,000 ppm Cl-) Waters

UNACCEPTABLE

FURTHER ASSESSMENT REQUIRED

Hydrogen Sulfide Corrosion

• Fe + H2S + H20FeSx + H2 + H2O

• FeS - cathode to steel: accelerates corrosion

• FeS is a plugging solid

• Damage Results• Damage Results– Sulfide Stress Cracking

– Blistering

– Hydrogen induced cracking

– Hydrogen embrittlement

H2S corrosion is minimized by sweetening the gas (knocking the H2S out or raising pH.

SSC Failure of Downhole TubularString in Louisiana

Video

Crevice Corrosion• The physical nature of the crevice

formed by the tubing to couplingmetal-to-metal seal may produce alow pH aggressive environmentthat is different from the bulksolution chemistry – hence amaterial that looks fine when testedas a flat strip of metal can fail whenthe test sample (or actual tubing)the test sample (or actual tubing)includes a tight crevice.

• This damage can be very rapid inwater injection wells, wells thatproduce some brine or in wellswhere there is water alternating gas(WAG) sequencing – causingfailure at the metal-to-metal seals ina matter of months.

O2 Corrosion

Dissolved Gas Effect on Corrosion

Overa

llC

orr

osio

nR

ate

of

Carb

on

Ste

el

There is no corrosion mechanism more damaging on a concentration basis than oxygen – smallamounts of oxygen, water and chlorides can ruin a chrome tubing completion in a few months.Injection wells are the most severely affected – minimise oxygen and don’t use chrome pipe ininjectors.

20 ppb O2 limit forseawater in carbon steelDissolved Gas Effect on Corrosion

0

5

10

15

20

25

0 1 2 3 4 5 6 7 8

Overa

llC

orr

osio

nR

ate

of

Carb

on

Ste

el

O2

CO2

H2S

Dissolved Gas Concentration in Water Phase, ppm

0 1 2 3 4 5 6 7 8

0 100 200 300 400 500 600 700 800

0 50 100 150 200 250 300 350 400

O2

H2S

CO2

seawater in carbon steelinjection tubulars.13Cr is CO2 resistant butvery susceptible to pittingcorrosion in aeratedbrines. 5 ppb O2 issuggested as a limit, buteven these levels have notbeen confirmed.

A split in the side of 5-1/2”casing. Cause was unknown –mechanical damage (thinningby drill string abrasion) wassuspected.

Wear Damage

Abrasion by solids, gas bubbles or liquid droplets may significantly increase corrosion bycontinuously removing the protective oxide or other films that cover the surface following theinitial chemical reaction.

Most graphs do not show the effect of too low a velocity on the corrosion rate. When thesurface is not swept clean, biofilms can develop or the surface liquid layer may saturate withCO2 or other gas, increasing corrosion. Minimum rates are about 3.5 ft/sec for clean fluids.

Note the effect ofincreasing flowing fluiddensity on corrosionrate.

Also – presence of solidsAlso – presence of solidsin the flowing fluids verysignificantly lowers themaximum permissibleflow rate.

Corrosion increases after water cutreaches 10 to 20%. The cause isremoval of the protective oil film. In thethird phase, the pipe is completelywater coated and corrosion ratebecomes more constant.

Wetting of the surface bywater significantlyaccelerates corrosion.

Top, Left: Chrome pipe after acidizing with the proper inhibitorand inhibitor intensifier.

Bottom, Left: Chrome pipe after acidizing with a marginalinhibitor.

Bottom, Right: Chrome pipe after acidizing without aninhibitor.

15% HCl, 2 hour exposure

WeldsThe heating that occurs during the welding process will cause theweld metal and the heat affected zone around the weld to bephysically different from the surrounding, original metal.

An anode is created by this difference.

An anode can start here or here.

Heataffectedzone

Weld metal (added anddifferent from originalbase metal)

Base metal

Bacterial deposits on injection tubing. Pittingunder the bacterial colony can be severe.Anaerobic

SRB’s - sours the well/reservoirIron Fixers - slime and sludgeSlime Formers - formation damage

–

Erskine – Failure of 25Cr Duplex SS

Source – BP Corrosion – John Alkireand John JW Martin

Many of the super alloy failureshave been linked backed to thebrines used for completions.

Cracking initiated at a stressriser – impact or wrenchmark.

Bacterial CorrosionBacterial Corrosion

Co2 corrosion in box, just past gap in coupling.

Parted coupling – CO2 corrosion – note trenches and pits.

Severe O2 corrosion in a surface line, just downstream of a connection.

Sacrificial anode (magnesium) from an offshore platform. This was a round bar stock anode.

Sacrificial Anodes - Galvanic Series in SeaWater

1. Magnesium

2. Zinc

3 soft aluminum

4.cadmium

5. hard aluminum

6. steel

7. stainless steel (300 series)

8. lead

9. brass and bronze

10. Inconel

11. Hasteloy C 276

Controlling Corrosion

1. Maintain high pH

2. Control gas breakout

3. Use passive metals

4. Remove Oxygen

5. Control velocities

6. Lower chlorides

7. Bacteria control

8. Acid/brine use considerations and alternatives

9. Liquid removal

10. Inhibitor injection

11. coatings

Acid Inhibitor Mixing

• Set up for “rolling” the tank, not slowcirculation - upper layer must be mixed in

This Won’t Work!This Won’t Work!

pump

Unless youget the toplevel caughtin thecirculation

Acid Inhibitor Mixing

• Check the inhibitor layer is possible

This will workThis will work

pump

Inhibitorpulled in

Inhibitor Mixing

• What will work?

– Air sparging orrolling the tankwith gas

– paddle mixing that createsa vortex

mixer

– vigorous rolling

• Effect of oxygen in acid -– oxygen increases corrosion

– oxygen saturation in acid is 7 parts per billioneffect is limited

Chromium

• Increasing Cr content of the alloy increasesthe Cr content (and film resistivity) ofcorrosion layer. Above 10% Cr in alloy,composition of layer is constant. Why usecomposition of layer is constant. Why usemore???? - chemical resistance.

• For 13% and 22% Chrome tubulars, criticalerosion velocities are twice carbon steels inCO2.

Tubular Selection Criteria – someconsiderations

• Embrittlement– hydrogen

– chloride stress cracking

• Weight Loss Corrosion– H2S-CO2-H2O-NaCl systems– H2S-CO2-H2O-NaCl systems

– CO2-H2O-NaCl

• Localized Corrosion

• Acidizing

• Galvanic

• Strength

• Cost and availability

Crack in the casing immediately below the wellhead.Probably due to a minor defect in the tubular andperhaps compounded by wellhead stress.

Equipment Approx Failure Ratefailures/well/yr

Packer 0.004

Rough Downhole Failure Rates

0.009

Cement/lap 0.012

SS Pods 0.023

SS SSSV 0.014

SS SSSV issues 0.024

SSSV

Flow Assurance

• Paraffins and Asphaltenes

• Scales

• Emulsions

• Hydrates• Hydrates

Wax Deposition impacts operations

Reduces throughputIncreases pressureReduces revenueIncreases risks opening pig receivers

Muge Erdogmus - BP

Wax in Oilfield Fluids

In solution No immediate problem

In suspension Can lead to flow problems, i.e. viscosityIn suspension Can lead to flow problems, i.e. viscosity

DepositedBlock or restrict flow in wells/flowlines/export pipelines

Gel Formation Can lead to restart problems – worst case is

totally blocked pipeline $$$$

Muge Erdogmus - BP

HARD/ SOFT WAX DEPOSITS

SOFT DEPOSITS• Formed under low shear• Formed under low heat flux• Open waxy structure• Contain large amounts of

entrapped oil• Become harder with time in

HARD DEPOSITS• Formed under high shear

• Often contain large amountsof asphaltene

• Wax inhibitor itself mightcause hard deposit!!!• Become harder with time in

flowing environmentcause hard deposit!!!

Generally:Larger VolumeEasier to Remove

Generally:Lower VolumeHarder to Remove

Muge Erdogmus - BP

Frictional pressure loss

In single phase flow frictional pressure drop isgiven by:

f L r u2

DPf = –––– ––––––d 2

where:d 2

where:f = friction factor (Darcy-Weisbach)L = length of piped = pipe diameterr = fluid densityu = fluid velocity

f is proportional to Re (and hence viscosity

Muge Erdogmus - BP

Viscosity of a Non-Newtonian Crude

Muge Erdogmus - BP

Controlling Wax Deposition

StrategyAvoidance:

Prevent wax fromforming / sticking

Manage:

Tool KitThermal:

Insulation, heat, flushing

Chemical:

Inhibitors, solventsManage:

Control rate andclean at acceptablefrequency

Remediate:Intervene whennecessary

Inhibitors, solvents

Mechanical:

Pigging, scraping

Novel / Black Box:

Ultrasonics, magnets

Muge Erdogmus - BP

High Thermal Insulation Systems

Inner Pipe

Water Stop / Spacer

Annulus

Field Joint

Inner Pipe

Water Stop / Spacer

Annulus

Field Joint

Aim to maintain fluid arrival temperatures at surface above WAT

Pipe-in-pipe

Pipe bundles

Outer Pipe

AnnulusInsulation

Outer Pipe

AnnulusInsulation

Muge Erdogmus - BP

Wax Chemical Additives

• Crystal modifiers– Distort wax crystals, often through co-crystalisation, inhibiting

further growth– Pour point depressants– Wax inhibitors

• Dispersants– Penetrate wax deposit, disperse and mobilise particles– Penetrate wax deposit, disperse and mobilise particles

• Detergents– Render wax water wet and prevent agglomeration

• Bugs– Consume heavier fractions

Muge Erdogmus - BP

HeatingFor avoiding wax when system is cold or melting deposited wax

Heat Tracing

IPB - Integrated ProductionBundle

Skin Effect CurrentTracing

Integrated ProductionUmbilical

Direct ElectricHeating System

S

Electrical Heat TracingFlowline

Muge Erdogmus - BP

Hot Oiling

• Useful in flowlines or equipment

• An increase in the oil temperature of only 10°C will

quickly soften wax

• Achieving the Wax Melting temperature is not• Achieving the Wax Melting temperature is not

normally necessary in a flowing system

• Should not be used in wells unless plugs are set to

keep the wax saturated hot oil out of the lower well

(no more than 10 joints down).

Modified from Muge Erdogmus - BP

Pigging• Scraper pig is most common method of removing wax

deposits

• Pig is swept along by the oil flow and mechanicallyscrapes wax off the walls

• A small amount of oil bypasses the pig to disperse theremoved solids and prevent a wax slug from forming

• In a well, sucker rods and wireline scrapers do asimilar jobsimilar job

• Foam pigs may be used initially if thick deposits are– These deform to fill pipe, but can build up a slug in front

• Chemical pigs may be useful.

Modified from Muge Erdogmus - BP

Asphaltenes• Defined as Pentane insoluble

• Heaviest and largest molecules in the hydrocarbon mixture

• Characteristic black color

• Become unstable with significant changes in density,usually due to changes in pressure

• Oils likely to exhibit precipitation have high gravities, low• Oils likely to exhibit precipitation have high gravities, lowasphaltene contents, are highly undersaturated and haveunfavorable Resin to Asphaltene or (Sat. + Asp.) / (Aro. +Res.) ratios.

• Precipitation can also occur due to commingling

• Problem Avoidance: Pressure maintenance in reservoir.Pressure/density maintenance during sample handling.Avoid volume changes.

Precipitated Solids Filter ExperimentsAsphaltene Phase Diagram

Saturation Pressure Curve

Upper Asphaltene OnsetSingle Phase Oil Region

No Asphaltene Precipitation

Asphaltene PrecipitationReservoir P&T

Pre

ssu

re

Temperature

Two Phase Liquid and Gas Region

Separator P&TSeparator P&T

Lower Asphaltene Boundary

Pedcor- Corelabs

Asphaltene Stability

• Maltenes and resins form the micelle

• Asphaltene is the small platelet (35A) held inthe middle of the micelle

• Dispersed platelets are not usually a problem• Dispersed platelets are not usually a problemalthough the oil may have a high viscosity

• When micelles are upset and broken, theplatelets coagulate and form a mass.

One form of asphaltene – hard balls of dry asphaltene platelets thatadhere to each other. Other forms are viscous tar-like masses

World Wide Crude Oil Chemical Compositions (SARA)Hydroc arbon Numbe r

Field Asphaltene Re s in Aro matic Saturated Total of

(Wt %) (Wt %) (Wt %) (Wt %) (Wt %) Sample s

Athabas ca 23.3 28.6 32.1 15.9 48.1 15

Wabas ca 21.6 30.6 32.1 15.6 47.7 7

Peac e Rive r 48.7 23.2 20.5 7.6 28.1 3

Cold Lake 20.6 28 30.5 20.9 51.4 7

E. Ve ne zue la 12.6 32.4 36.4 18.6 55 5

Average on 22.9 30.6 30.4 16.1 46.5 46

46 Heavy Oils46 Heavy Oils

PB HOT (EOA) 14.13 13.37 28.1 44.4 72.5

PB HOT (WOA) 10.38 20.42 28.23 40.97 69.2

W. Ven. (ne ar 13.2 12.9 38 35.9 73.9

HOT)

Co nve ntio nal 14.2 28.6 57.2 85.8 517

Normal Oils

PBU Normal 16.52 1.9 31.93 49.67 81.6

Oil 18.42

Schrader Bluff 4.9 29.0 24.7 41.5 66.2 15

Asphaltenes

• precipitated by:

– CO2

– acid

– pH– pH

– turbulence

– chemical shift that upsets micelle

Microscopic photos of asphaltene aggregationMicroscopic photos of asphaltene aggregation

Rice U. - P. Zhang

O - Nothing could be seen o - Fine particles

x - Tiny aggregation X - Large aggregation

Ultra low viscosity oils, usuallyasphaltic or paraffin based, present aspecial set of problems in damagerecognition and treatment.

Few methods are successful inprevention. Solvents are common buthave poor effectiveness at removal.

Asphaltic Sludges

• Form very viscous masses, often after contactwith spent acid; frequently catalyzed by iron

• Sludges are serious problems because theycannot be easily removed.cannot be easily removed.

• Test the oil with spent acid and 1000 ppm ironbefore acidizing any oil reservoir. Use anti-sludge and an effective iron control additive.

Scales

• Usually a precipitate from a brine thatbecomes saturated with a material due to achange in the conditions within a well.

• Scale precipitation may be driven by mixing• Scale precipitation may be driven by mixingincompatible waters, but can also be causedby out-gassing, shear, turbulence, andtemperature and pressure loss.

Scales

• calcium carbonate - upset driven

• calcium sulfate - mixing waters, upset, CO2

• barium sulfate - mixing waters, upset

• iron scales - corrosion, H S, low pH, O• iron scales - corrosion, H2S, low pH, O2

• rarer scales - heavy brines

Some scales form in layers, often driven by an “upset” in the flow dynamics of the system.These deposits can form almost anywhere, on any surface, but the deposits are usually justdownstream of the location of an upset in the flow system. The location of the scale is often anindicator of what is causing the precipitation.

Calcium carbonate scale deposit showing ability to “cleave” along the layer boundaries. Unusualshape was from deposition on a flapper of a SSSV.

Calcium sulfate scale from slow growth in a high water cut reservoir.

A rapidly formed deposit of calcium sulfate formed after an acid was mixed with a scaledissolver chemical that had removed a deposit of calcium sulfate scale at the perforations.

Calcium Sulfate scale that completely blocked a section ofdownhole tubing – this piece was from a connection.

Scale crystals formed slowly in a gathering line.

Scale Location

• at pressure drops - perfs, profiles

• water mixing points - leaks, flood breakthru

• outgassing points - hydrostatic sensitive

• shear points - pumps, perfs, chokes,• shear points - pumps, perfs, chokes,

• gravel pack - formation interface

Barium SulfateScale from a

Scale may appear at the surface first –or last. All depends on the scale and

flowing conditions.

Scale from aNorth SeaProductionSeparator

Scale at the gravelpack / formation

interface.

Calcium carbonate scale wasfound at the interface of thefound at the interface of thegravel pack interface with theformation. The precipitationtrigger was pressure drop,resulting in out-gassing of CO2with an accompanying pHincrease.

Scale Prediction

• Chemical models - require water analysis andwell conditions

• Predictions are usually a “worst case” - this iswhere the “upset” factor comes in.where the “upset” factor comes in.

– added shear - increased drawdown, chokechanges, etc.

– acidizing

– venting pressure

Scale Deposition – mostly in top of well

Very Even Accumulation

Beware of predicting deposition points until youunderstand unloading efficiency. Workover hydraulics

in wellbores are critical.

Scale deposition points or “slug recovery” of debris?slug recovery” of debris?

Putting it to use.

• Localized scale deposition may be from:

– Temperature or pressure crossing precipitationthreshold at a specific point.

• Could it be changed by altering rate, temp loss or holding• Could it be changed by altering rate, temp loss or holdinga bit more pressure?

– Might be affected by shear at a restriction

• SSSV, collapse, nipple profile, fluid mixing, etc.

• Consistent scale deposition along the lengthusually from basic incompatability.

Emulsions

• Multiple phases that do not separate quickly;usually requires an energy source.

• If oil and water do not separate quickly, thenlook for the stabilizing mechanismlook for the stabilizing mechanism

• Emulsions are frequently blamed for damage,however, most emulsions are formed in thetubing or lift system by gas breakout or addedenergy.

Types of Emulsions

• oil-in-water

• water-in-oil

• gas-in-water (foams and froths)

• solids-in-liquids (muds, etc.)• solids-in-liquids (muds, etc.)

• Over twenty different combinations that canbe called emulsions.

Energy Sources

• lift system

• gas breakout

• shear at any point in the well

• choke• choke

• gas expansion

Stabilizers

• surfactant (film stiffeners)

• solids (silt, rust, wax, scale, cuttings)

• emulsion or component viscosity (preventsparticle or droplet contact)particle or droplet contact)

Widely

Deformation

Changes in Fluid Viscosity with Change in Internal Phase of Dispersed or EmulsifiedFlow

Increasing internal fraction of the “emulsion”

52% 74% 96%

Viscosity

WidelyDispersed Contact

Inverted

Expansion of gas occurs as the gas rises from the bottom of the well. Theexpanding gas can entrain and carry liquid with it if the flow rate reaches criticalvelocity (the velocity necessary to lift liquid).

2500 ft

1075 psi

Remember – the volume of the gas bubble (andindirectly the velocity of the upward flowing fluid) iscontrolled by the pressure around it. This pressure isprovided by the formation pore pressure andcontrolled by the choke and other back pressure

5,000 ft

2150 psi

controlled by the choke and other back pressureresistances.

The type of flow pattern changes with the expansion of the gas. One or more of the flowpatterns may be present in different parts of the well. The flow patterns may explain differencesin lift, corrosion and unloading.

Mist Flow – external phase is gas with a small amount ofliquid

Channel or annular flow

Slug or churn flow

Piston flow

Bubble flow

Single phase liquid flow

Depth andPressure

Hydrates

• Crystalline or clathrate materials of frozenwater and gas. Can be formed at temps over32oF (0oC) and on pressure reduction.

Press

Temperature

hydrates

no hydrates

saltier water

Hydrates – clathrates of gas and water

Source – Muge Erdogmus

Hydrate Phase Behavior

Pre

ssu

reHydrocarbon

Liquid + Water

Hydrate +WaterHydrate

+ Ice

Temperature

Pre

ssu

re

+ Ice

Ice +Gas

Liquid Water +Hydrocarbon Gas

Hydrate FormationLocus

Solid polymer lumps or “Fisheyes” and microgels filtered from liquid HEC polymer.Microgels or “fisheyes” after straining a liquid HEC dispersion through a 200 mesh screen in ashear and filter operation for gravel pack fluid preparation.

FacilitiesPressure

Flow LineBack-pressure

Hydrostatic

Wellhead Pressure Chokes• Reservoir (e.g. coning)• Sand-control• Erosion constraints• Commercial e.g. Gas contract• Facility Constrained e.g Gas Handling

Exportlosses

The Physics !

P Reservoir - targetTubing FrictionPressurelosses

BHFP

Near-Wellbore Lossese.g. Skin and turbulence

P Draw-down[‘Good’ Pressure Loss]

P Reservoir - actual

losses

BP

Support Slides

• Pipe Banding, Grading and Re-use

Safety Moment – Chains help you go.They don’t keep you from skidding.

Some Basic “Formation Damage” Causes

Completion Stimulation ProductionProblems - 1

ProductionProblems - 2

DepletionEffects

Particles frommuds & brines

Poor selection orapplication

Wetting Hydrates Natural fracclosure

Fluid invasion offormation

Poor well-to-formation link

Water blocks Bacterial debris Matrix permreduction

Cement debrisand fluids

Mineral damageby fluids

Mineral Scales Turbulence athigh rates

Water influx,phase changes

Perfs – number,phasing, size,depth, cleanup

Emulsionsrelated to pH andgas use

Dew point andbubble point –relative perm.

Emulsions,foams, froths andsludges

Formationdisaggregation –sanding, fines

Limitedformation entry

Damage bypolymer &surfactant

Paraffin (wax)precipitation

Corrosion,erosion, andabrasion

Lift problemsfrom fluidchanges

Screen and G.P.restriction

Precipitates andsolids releasedby acids

Asphaltenes andTars

Deliquificationand lift

Subsidence

The Effect of Damage on Production

Rate = (DP x k x h) / (141.2 mo bo s)

Where:

DP = differential pressure (drawdown due to skin)DP = differential pressure (drawdown due to skin)

k = reservoir permeability, md

h = height of zone, ft

mo = viscosity, cp

bo = reservoir vol factor

s = skin factor

Root Causes of Residual Damage AfterClean-up Flow….

• High perm formations less affected?

– Major damage removers:

• Flow Rate per unit area,

• Flow Volume per unit area,• Flow Volume per unit area,

• Pressure pulse?

• Drawdown per unit area – a control?

First Problem

We don’t understand cleanup byflow…

It’s a matter of flow rate and volumethrough a given area.

For the same paythickness, a horizontalwell or a fractured wellmay contact 100’s oftimes more pay zone areathan a vertical well.

Clean-up flow is“diluted” by thelength of intervalopen at once forcleanout.

Now, think about the set drawdown – say500 psi - per unit area, the velocitygenerated and the total volume per area.

A 10 ft pay in avertical well w/ 6”diam. yields contactarea of 16 ft2

A 1000 ft pay in ahorizontal well w/ 6”diam. yields contactarea of 1600 ft2

Which has the potential of cleaningup faster and more completely?

Horiz. Well – assume 5x more than verticalflow (typical) – would generate 5000 bpd,but spread over 1600 ft2 – and release aclean-up flow of only 3 gal per hour per ft.

5000 bpd

Vertical well - 500 psidrawdown and an inflow of1000 bbl/day/ft spread outover just 16 ft2 will generate aclean-up flow of 110 gal perhour per ft.

clean-up flow of only 3 gal per hour per ft.The velocity may be too low to get goodcleanup.

1000 bpd

Second Problem

• We don’t understand damage.

– How it got there

– How it is removed.

– How to prevent it.– How to prevent it.

– What operations put the well’s productivity atrisk.

PI Change After Workover - Perfs Protected

20

30

40

50

%C

han

ge

inP

I

0.3 1.2 3.1 4.6 8.4

PI of wells

When perfs were protected, that was little risk of long term damage.

-40

-30

-20

-10

0

10

20

1 2 3 4 5 6 7 8 9 10 11

%C

han

ge

inP

I

Short Term PI Change

Long Term PI Change

SPE 26042

Damage in Fractured Wells with Unprotected

Perforations

-20

-10

0

1 2 3 4 5

%D

am

ag

e PIi = 7.55

When the perfs were not protected, the well was damaged.

-80

-70

-60

-50

-40

-30

%D

am

ag

e

PIi = 1.92

PIi = 6.21

PIi = 7.55

PIi = 7.55

PIi = 18.1

Effect of Scraping or Milling Adjacent to Open

Perforations

0

10

20 Perfs not protected by

LCM prior to scraping

One very detrimental action was running a scraper prior to packer setting. Thescraping and surging drives debris into unprotected perfs.

-60

-50

-40

-30

-20

-10

0

1 2

%C

ha

ng

ein

PI

Short Term PI Change

Long Term PI Change

Perfs protected by

LCM

SPE 26042

Kill Pills: Summary of Overall Effectiveness in

Non Fractured Wells

5

10

%C

han

ge

inP

I

Sized Borate Salts

(4)

Cellulose

Fibers (3) HEC Pills No Pills

Sized particulates, particularly those that can be removed, are much less damaging than mostpolymers.

-25

-20

-15

-10

-5

0

1 2 3 4 5 6 7

%C

han

ge

inP

I

Sized Sodium

Chloride (12)

No Near Perf Milling

(6)

Near Perf Milling or

Scraping (3)

No Near Perf

Milling (8)

SPE 26042

Near Perf Milling

or Scraping (10)

Third Problem

• We don’t know enough about timing of damageremoval.

– Variety of causes• Polymer dehydration

• Decomposition of materials• Decomposition of materials

• Adsorption, absorption and capillary effects

• Field data from Troika (100,000 md-ft) show initialflow improves PI, but later flow does not.

Note that PI remained improved, even after drawdown was reduced.

Effect of Drawdown on PI

Cleanup Lessons

• On initial cleanup, PI erratically increased asthe choke was opened. The typical responsewas a decrease, as if the well or part of theflow pathway were loading up, followed by asharp PI increase, seemingly when thesharp PI increase, seemingly when theobstruction was unloaded.

• Very little partially broken polymer wasrecovered, but the early load water recoverymatched the increase in PI.

Cleanup Lessons

• The wells cleaned up steadily with increasingdrawdown in the period of time immediatelyfollowing start of backflow. Cleanup wasincreasing, measured by increasing PI, at theend of the first short cleanup periods (shut-inend of the first short cleanup periods (shut-inof the well).

Cleanup Lessons

• After initiation of production operations (thiswas after the first cleanup flow), no furthercleanup of damage was seen, regardless ofdrawdown. The reason is not known, but maybe due to polymer adhering or cooking out??.be due to polymer adhering or cooking out??.

• Lower skins were linked to both sand flowbefore the completion was run (sand surgeremoved damage), increased cleanup flowvolumes (and drawdown) on the initial cleanup,and more effective frac stimulations.

Fourth Problem

• We don’t understand how damage impactseconomic return.

Example Economics - Skin Sensitivity

140

150

160

170

PV

-10,

$m

m

25

30

35

40

PV10

100

110

120

130

0 5 10 15 20 25

Skin

PV

-10,

$m

m

10

15

20

25ROR

Type ofDamage

MostProbableLocation

Impact onProductivityif notRemoved

SurgePressureNeeded forRemoval?

Flow Vol.Or TimeNeeded forRemoval?

RemovalHampered byLimits onCleanup Vol?

AlternateMethods ofRemoval orPrevention

Mud Cakein Pay Zone

FormationFace

Moderate toSevere

Yes SpurtVolume

Yes Acids, Soaps,Enzymes

Mud Filt. <12” Minor - mod. No No

Damage Removal

Whole MudLoss

Fracturesand largevugs

Severe Possible,Can help infew cases.

Depend onCond. fewcasessimilar

Yes, flowcombined withsolventtreatment

Few successfulwhole mudremovals whenvol. > 200 bbls.

CementFiltrate

<12” intopay

Only if claydamage

No No

PerforationCrush Zone

½” aroundperf

Moderate toSevere

Yes 4 to 12gal/perf

Yes Surge smallzones intochamber, acids,pulses, fracs

Formationsand inperfs

perftunnels

Most severe Cleanupand reperf

Depends oninitial &lateractions

Develop goodperf andprepack actions

Type of Damage Skin range CommentsMud Cake in PayZone

+5 to +300,+15 is typical

Mud skin is usually shallow and has more impact whenturbulence and non-darcy skin problems are most severe. Mudcake is usually by-passed by perforating.

Mud Filtrate +3 to +30 Filtrate usually recovered by steady flow and time. Related torelative perm effects. This is usually a short lived problem (1 to

Damage Effects

relative perm effects. This is usually a short lived problem (1 to3 weeks)

Whole Mud Loss (inpay zone)

>+50 Options depend on mud volume lost. Enzymes, solvents andacids for small volumes (<10 bbls). Sidetrack if over 1000 bbls.Low solids mud can be removed by concentrating on viscosifierdestruction or dispersment.

Cement Filtrate +10 to +20 Very shallow clay problems. Perforate with deep penetratingcharges to get beyond. Use leakoff control on cement.

Perforation CrushZone

+10 to +20 Perf small intervals underbalanced. Isolation packer breaksown,explosive sleeve breakdown (very simple) - must beaccomplished prior to gravel packing.

Formation sand inperfs

>+50 Most severe typical damage - cleanout and recompletionrequired

Mud Damage

• Common problems

– fines in the mud - physical plugging

– wetting of formation by mud surfactants

– emulsions– emulsions

– reactions with the formation fluids

– reaction with the formation clays

Oil based mud cleanup is a special case, requiringdispersal of the OBM emulsifying agents andwetting of the particles to prevent damage.Contact with acid, as shown in the following slides,will produce some severe sludges that are verydifficult to break.

A 50-50 mix of 14 ppg OBM and 15% HCl.The resultant sludge formed immediatelyand was stable for months.

Polymer Damage

• From: muds, pills, frac, carriers

• Stable? - for years

• location - depends on form polymer was in

– dispersed properly - surface to deep in formation– dispersed properly - surface to deep in formation

– in pills and mass - right in perfs

Particles in the Fluid

• Solids from tanks,lines and fluids

• Severe problem, butoften ignoredoften ignored

Particulate Damage

• Unintended particulates - “dirty fluids”– filter fluids to 5 microns at Beta of 1000

– maximum NTU of 30, preferable is 20

– clean tanks, lines - how about tubulars?– clean tanks, lines - how about tubulars?

• Particles in fluid loss pills– mixed in proper range for perm encountered?

– Will it come off formation? Can it come back thrupack? Thru screen? What about removal?

Horizontal Well Formation DamageTheories

Zone of Invasion - Heterogeneous Case

50423010

Stress on the Formation – three packersSwell packer, 50 psiexerted pressure,7.5 ft elementcontact length

Inflatable packer,2000 psi exerted

In the bottom twopackers, the forceexerted on theformation issufficient to createa fracture at least2000 psi exerted

pressure, 10 ftelement contactlength

Mechanical packer,6000 psi exertedpressure, 0.5 ftelement contactlength

a fracture at least12” deep withoutadded fluidpressure.These fractures mayfocus frac initiationduring a frac job.

SPE 123589

Model of Breakdown Pressure

SPE 123589

Modeled Stress around the Wellbore

The finite elementmodel of stress arounda horizontal well showsthe lowest stress at thetop of the wellbore.

Fracture initiation mayFracture initiation maybe easiest at the topand bottom of thewellbore unlessdisturbed by formationheterogeneities.

Inflatable Bridge Plug (IBP), orInflatable Casing Packer (ICP)?

• Inflatable plugs –Reliability Factors

• # times inflated in a run

• Maximum Inflation

• Element compositionand length

• Set Point

• Casing or OH?

• Inflatable Packers –Reliability Factors

• Slide Damage

• Set Point

• Casing or OH?

• Permeability of zone?

• Inflation Sequence

• Shrinkage?• Casing or OH?

• Permeability of zone?

• Shrinkage?

• Time?

Learning? – caliper the set point first!

After initial setting, inflatable plugsdo not come back to initial

diameters. Allow about a 20% ODincrease in clearance calculations.

Maximum Flex Point

Inflatables rely on expansion of an inner rubber bag that pushes steel cables or slatsagainst the wall of the pipe or the open hole. The only gripping ability is generated by thefriction of the steel against the pipe or open hole. This is critically dependent on theinflation pressure and the exterior slat or cable design. For a permanent seal, placeseveral bailers of cement on top of the inflatable.

Baker5/26/2010 301

George E. King EngineeringGEKEngineering.com

Damage from over inflation can bepermanent. Know the ID of the set

points.

How much differential well pressurecan an ICP hold?

Depends on the clearances, the expansion percentage, whatinflates the IBP, the permeability of the formation where it isset, the length, the fluids on the outside and the temperature– and that’s just for starters.

Failures on Running Plugs (RBPs) andPackers

• Failure Causes:

– Running the plug too fast (fluid viscosity dependent)

– Deposits on the tubular wall – wax, scale, corrosion

– Deformed or corroded tubing – caliper?

– One heavy wall joint in the midst of a string – caliper?

– Setting tool performance at this depth, deviation, temp?

– Pipe body too hard (slips won’t grip in V-150?)

– Pressure differential out of plug capacity

– Temperature beyond either initial or long termtemperature capacity (composite plugs break down).

Halliburton Energy ServicesGeneral Guidelines For SealsHalliburton Energy ServicesGeneral Guidelines For Seals

(1)

PEEK(2), (4) Ryton Fluorel(3) Aflas(3) Chemraz(3) Viton(3) Neoprene(3) Nitrile(3) Kalrez(3) Teflon(3)

Filled Unfilled Unfilled Filled Unfilled Filled Filled Filled Filled Unfilled

350 350 450 350 325 300 275 450 400 325

(177) (177) (232) (177) (163) (149) (135) (232) (204) (163)

Above Below

15,000 10,000 15,000 5000 5000 5000 3000 15,000 15,000 5000

(103) (68.9) (103) (34.4) (34.4) (34.4) (20.7) (103) (103) (34.4)

A A A A A B B NR NR A A A

A A B B A B B C A A A A

A A A A A A A B B A A A

(2), (4)Compound

Service °F

(°C)

Pressure psi

(MPa)

Environments

H2S

CO2

CH4 (Methane)

Hydrocarbons

(2), (4)

A-Satisfactory B - Little or no effect C - Swells D - Attacks NR - Not recommended NT - Not tested

NOTE: (1) This information provides general guidelines for the selection of seal materials and is provided for informational purposes only. Seal Specialists with Halliburton Energy Services should be consulted for the actual selection of sealsfor use in specific applications. Halliburton Energy Services will not be liable for any damage resulting from the use of this information without consultation with Halliburton Seal Specialists.

(2) Contact Technical Services at Halliburton Energy Services - Dallas for service temperature and pressure.

(3) Back-Up Rings must be used.

(4) There could be a slight variation in both temperature and pressure rating depending on specific equipment and seal designs.

A A A A A A A B C A A A A

A A A C A A A NR NR A A A

A A C B A C C B A A A A

A A A A A A A NR NR A A A

A A NR A A NR NR NR B A A A

A A A A A A A C A B A A

A A NR A A NR NR NR NR NR B B

Diesel A A A NR A A A B B A A A

Hydrocarbons

(Sweet Crude)

Xylene

Alcohols

Zinc Bromide

Inhibitors

Salt Water

Steam

5/26/2010 305

Temperatures in the Well? Circulating orHigh Rate Injection?

0

2000

4000

6000

30 40 50 60 70 80 90 100 110 120 130

Tubing

Tbg Fluid

Casing 1

Undisturbed

0

2000

4000

6000

30 40 50 60 70 80 90 100 110 120 130

Undisturbed

Tbg Fluid

Tubing

Casing 1

8000

10000

12000

14000

16000

18000

Circulation pump rate = 8-BPM

BHST= 122*F

BHCT= 98*F

8000

10000

12000

14000

16000

18000

Frac job pump rate = 35-BPM

BHST= 125*F

BHTT= 86*F

5/26/2010 306

Low Pressure Applications

Model AD-1

Tension Set

•Mechanical Jay

•Waterflood

Model G

Compression Set

•Mechanical Jay

•Low Pressure•Waterflood

•Disposal

•Shallow Production

•Emergency Shear

Release

•Low Pressure

Production

•Emergency

Rotational Release

RETRIEVABLE CASING PACKERMODEL “R-3” DOUBLE GRIP

• Double grip, compression set packerwith unloader

• Medium performance, generalservice packer

• J slot mechanism for setting andreleasingreleasing

• Hydraulic, button type holddown

• Bypass design speeds equalizationand resists swab-off

• Differential lock helps keep bypassvalve closed

Medium ServiceModel A-3 Lok-Set

•Double grip, compression setpacker

•Medium performance,

general service packer

•Sets with right hand rotation•Sets with right hand rotation

and slack-off weight

•Can be landed in compression,

neutral or tension

•Releases with right hand

rotation and up strain

Heavy Duty ServiceHeavy Duty ServiceModel Hornet HDModel Hornet HD

Double grip, compression or tension setpacker

Medium performance, to 10,000 psi @ 350Fgeneral service packer

Sets with ¼ turn right hand rotation andslack-off weight

Can be landed in compression,

neutral or tension

Releases with ¼ turn right hand rotation

Packer Performance Envelope

• Failure Modes

• Pressure Ratings

• Limitations

Packer Performance Envelope

SIZE 194 "SAB-3" 47 X 38 PACKERT

EN

SIL

E

9-5/8" 53.5# Csg (8.535 I.D.) - 4140, 18-22 Rc BODY & PISTON COVER, CAST IRON CONES

Graphical Presentation of Rated Performance Limits

Represents Effects of Combined Loading on Components

PRESSURE

FO

RC

E

0

0

ABOVE BELOW

TE

NS

ILE

SE

T-D

OW

N

Performance EnvelopePacker Performance Envelope Summary

Above Pressure Below

Tensio

n

1 2

3

4

FO

RC

E

Packer Failure Modes

(as represented ingraph)

1 Tensile failure of LH

square thread

2 Tensile failure atbody/guide thread

FO

RC

E

4

Com

pre

ssio

n

3 5

7 6

Above PRESSURE Below

3 Body collapse

4 Packing elementsystem failure

5 Pin collapse atbody/guide thread

6 Bearing failure at sealassembly

7 Body lock ring failure

Effects of Material SelectionT

EN

SIO

N

0

PERMANENT PACKER RATING ENVELOPE

With Standard Materials

With NACE Materials

SE

TD

OW

N

PRESSURE ABOVE PRESSURE BELOW0

18-22 Rc Body

& Cast Iron Cones30-36 Rc Body & Cones

Performance Envelope / Tube Move

•5-1/2" 26# Tbg

TUBING HANGER

•Tubing Affected Areas

Top Joint Tension

Packer To Tubing

•4-1/2" 18.8# Tbg

7-5/8" OD 33.7# Csg

KC-22 ANCHOR

PACKER

Perforations

•Annulus

Packer To Tubing

Force

Compressive Loads

Differential Pressures

Packer Failure Causes

• Body Collapse– ID of seal bore contacts OD of seal assembly

– Caused by differential pressure above or belowthe packer, or by packer and tubing forces; or by acombination of the two.combination of the two.

– Virtually all of the pressure and forces exerted onthe packer are locked in by the slip system. Thebody remains collapsed even after the forces areremoved.

5/26/2010 316George E. King Engineering

GEKEngineering.com

Packer Failure Causes

• Body Collapse (continued)

– Since pressures and forces are dependent on the cross-sectional area of the packing element, they are dependenton the casing ID.

– Each size packer is used in a range of casing ID’s (weights)– Each size packer is used in a range of casing ID’s (weights)

– As casing ID increases, the expansion of the packingelement mustbe greater - adverselyaffects packer rating.

5/26/2010 317George E. King Engineering

GEKEngineering.com

Consequences of Body Collapse

• Not a catastrophic failure (sealing notcompromised)

– often cannot remove seals from packer bore

– Safe Operating Region - 3 on envelope

5/26/2010 318George E. King Engineering

GEKEngineering.com

Packer Failure Causes

• Packing Element System Failure

– failure of the element occurs when the elementextrudes through the back-up system

5/26/2010 319George E. King Engineering

GEKEngineering.com

Packer Failure Causes

• Causes of packing element failure

– temperature of seal material exceeded

– excess pressure on element causes extrusion

– chemical attack (breakdown or softening)

– gas permeation and sudden decompression causingblisters and seal ruptures

– backup system failures

– seal bore corrosion or erosion leaves sealing surface rough

5/26/2010 320George E. King Engineering

GEKEngineering.com

Consequences of Seal Failure

• Catastrophic failure (sealing compromised)

– seal is lost - leaks may cause other problems

– often cannot remove seals from packer bore(baked, hardened, fused, etc.)(baked, hardened, fused, etc.)

– Safe Operating Region - 4 on envelope

5/26/2010 321George E. King Engineering

GEKEngineering.com

Packer Failure Causes

• Pin collapse at the body/guide connection

– like a body collapse, collapse of the pin connection at thelower connection with maximum deflection at the middelof the lower thread.

– Differential pressure can occur when the seals are in a seal– Differential pressure can occur when the seals are in a sealbore extension below packer or if a wireline plug is set in anipple in the tailpipe.

– Consequences - non-catastrophic

– Region - 5 on envelope

5/26/2010 322George E. King Engineering

GEKEngineering.com

Using The Packer Envelope in WellDesign

• Plot data from tubing movement and stressprograms on a packer envelope to determineif the loads fall in the safe area during initialsetting, production, stimulation and killsetting, production, stimulation and killoperations.

5/26/2010 323George E. King Engineering

GEKEngineering.com

Temperature Extremes

• The extremes of temperature change (higherthan normal) are usually seen in operationsinvolving cyclic thermal processes.

• Lower than normal temperatures may be seen• Lower than normal temperatures may be seenin permafrost, sea floor penetrating and CO2

operations.

5/26/2010 324

Setting the Packer

• Chances of setting packers go up sharply whena casing scraper is run. (Remember the burrson the perforations?)

• The quantity of debris turned loose from thecasing wall is often severe! (Tens of poundsworth!) Watch the formation damage.

5/26/2010 325

Packer Set Point Requirements

• Avoid setting packer in thesame joint where previouspackers have been set.

• Avoid doglegs, fault locationsor high earth stress zones

• Adequate cement and bondrequired behind pipe atpacker set point

• Remove burrs from pipeabove packer set point

• Remove debris (dope, millscale, mud, cement, etc.) oncasing wall (fills slip teeth)

• Well pressures are withinrange of packer at set pointpacker set point

• Caliper casing above andthrough the packer set point

• Clearance between packerand casing at set point iswithin rated range of packer

• Avoid zones of high corrosion,either internal or external.

range of packer at set point

• Pipe alloy compatible withsetting slips (hardness ofcasing relative to packer slips)

• Slip design & contact areaacceptable for slip holding

• Weight applied to packer canbe transferred to formation

5/26/2010 326

Information Required Before SettingPacker or Plug

• Wellbore drawing with all diameters

• Last TD tag – rerun?

• Doglegs and deviations

• Viscosity of fluid in wellbore– Calculate running speed vs. surge/swab.– Calculate running speed vs. surge/swab.

• Copy of reference logs

• Where have other packers been set (avoid that joint)

• Set point requirements

• How can it be equalized if it has to be pulled?

5/26/2010 327

Job Checks

• Measurements from CCL to a packer referencepoint.

• Run in hole at about 100 fpm, slowing at IDrestrictions.restrictions.

• Using CCL/GR, log up and correlate depths

• Set packer – look for line weight reduction

• Disconnect and log up a few collars (may beslightly off depth after disconnecting).

5/26/2010 328

Job Checks

• Drop back and gently tag packer with settingtool to confirm depth.

• Log back up a few collars.

5/26/2010 329

Packer Setting Guidelines

• Drift

• Scraping

• Casing Support

5/26/2010 330

Drift the Casing

• Casing ID requirements above the set point

• Casing ID requirements below the set point

• Check the drift to deepest point with drift ofdiameter and length of packer.diameter and length of packer.

5/26/2010 331

Clean/Scrape The Casing?

• Removal of perforation burrs minimizes elastomerseal damage

• Removal of cement, mud, pipe dope and mill scaleminimize debris that can fill the slips.

• Scraping casing can increase packer setting success• Scraping casing can increase packer setting success

• Scraping casing can also produce some severeformation damage if perforations are not protected.

5/26/2010 332

Casing Scraper – Designed to knock offperforation burrs, lips in tubing pins,cement and mud sheaths, scale, etc.

It cleans the pipe before setting apacker or plug.

The debris it turns loose from the pipemay damage the formation unless thepay is protected by a LCM or plug.

5/26/2010 333

Effect of Scraping or Milling Adjacent to Open

Perforations

0

10

20 Perfs not protected by

LCM prior to scraping

One very detrimental action was running a scraper prior to packer setting. Thescraping and surging drives debris into unprotected perfs.

-60

-50

-40

-30

-20

-10

0

1 2

%C

ha

ng

ein

PI

Short Term PI Change

Long Term PI Change

Perfs protected by

LCM

SPE 26042

5/26/2010 334

Lift Systems

• Optimizing the lift system is generally one ofthe most economic well operations.

ESP – Root Failure Cause Considerations

1. Collect data from first few weeks afterinstallation and from last few weeks beforefailures.

a. Initial production data verifies design informationa. Initial production data verifies design information

b. Data before failure show changes in wellperformance or pump wear.

2. Maximum run times can only be achieved if theRoot Cause of Failure is identified.

SPE 68789

ESP Typical Run Times

• Extremely wide variance– Short times 50 to 100 days – where pumps handle

high gas content, solids or very hot wells in smallcasing

– Medium times – 600 to 800 days – where user– Medium times – 600 to 800 days – where userand supplier pull pumps at first sign of problemsand examine.

– Longest run times – 10 years plus – idealconditions (low temperature, low GOR, high fluidlevels, instrumented completions)

Run Times - Optimizing

Increasing Run Time -

Specific Need Data

• Tubing and casing pressure

• Pumped fluid volume

• Fluid composition

• Presence, amount, size and type of solids• Presence, amount, size and type of solids

• Drive output volts

• Amps

• Operating speed

• Fluid level, BHP, BHT

Review the root cause failure fortrends between wells.

– Power fluctuations

– Type of installations

– Water and oil types – compatibility with rubber

– Solids

– Gas content

– Temperatures– Temperatures

– Service technician

– Pump supplier

– Shroud type and design – cooling

– Casing size and clearances

– Pump setting location

• For a “complete” list of failures and repair solutions – see SPE68789 by Jim Lea and Mike Wells.

Review the Previous ESP Pull History

• Example – motor burnout resulting from insulationfatigue from overheating due to a plugged pump.

• Set points in deviated sections have very high failuresdue to bearing and shaft wear.due to bearing and shaft wear.------------------------------------------------------------

• Calculated well production from field data can bemisleading – accuracy is often poor.

• Get accurate production data.

Unusual problems

• Gassy wells – under-load (UL) settings werenot correct – did not allow pump shutdownwhen the pump became gas locked.

• Oversized motor in hot or low flow wells –• Oversized motor in hot or low flow wells –current draw was like idle load and UL did nottrip.

ESP Gas Lock Description

• “At the low intake pressures the density difference between theliquid and the gas phase can approach three orders of magnitude.The viscosity difference can be two to three orders of magnitude.The two fluids, when subjected to the forces in a centrifugalpump, can flow at different rates and sometimes in oppositedirections. The impeller accelerates the fluid and centrifugal forcedirections. The impeller accelerates the fluid and centrifugal forcemoves the fluids to the peripheral exit. The magnitude of theforce is related to the impeller rotational rate, the radial locationof the particle of fluid and its density. The gas, being less dense, isforced to the low side of the impeller vane. The gas begins toexpand, surge and start separating the liquid from the leadingedge of the vane, blocking the passage.”

B.L. Wilson, SPE Gulf Coast ESP Workshop, April 1998.

Failure Rates and Predictions

• Mostly mechanical

• The actual rates vary widely with conditionsand company

Deep Water Production - 2004

shallow casingfailure/ pressure

6%TRSCSSV & Inserts

14%

Flux11%

DWP Historical Well Failure Mechanism

Foot valve3%

Leaks packer & tbg6%

Horz. OH sandcontol failures

17%

Damaged OH screen3%

Compaction8%

sand control infantmortality/ design

6%

Sand control failureunclassified

26%

Failure Rate Variation Reflects CompaniesExpertise, age of wells & Area of Operation

Company Device Failure Rate1

failures/well/yrArea of Operations

Company A (wet tree) ScSSVs 0.011 GOM

Company B (wet tree) ScSSVs 0.002 GOM

Company C (dry & wet) ScSSVs 0.004 North Sea

Industry Overall (dry & wet) ScSSVs 0.1052 GOMIndustry Overall (dry & wet) ScSSVs 0.1052 GOM

1. About 60% of the ScSSV failures involve control line crushing, connectionproblem or plugging. More problems are noted at higher well pressures and inareas of flow assurance problems.

2. Reflects older wells in the shelf where ScSSV failure & malfunction is common3. Intervention rate is sharply lower on wet tree wells than dry tree wells.

Failure Rates for Dry Tree Well Components

Equipment TotalWells

Failure Ratefailure-per-well-per-year

Tubing 209 0.0083

Packer 209 0.0024

SSSV failures 209 0.0071

SSSV malfunctions 209 0.0047SSSV malfunctions 209 0.0047

Control Lines (crush, leak, plug) 209 0.0091

Well Head (Dry) 209 0.0005

Well Head Valve 209 0.0024

Choke 209 0.0024

Cement Lap Failure 209 0.0095

Other Failure 209 0.0047

Failure Rates for Wet Tree Well Components

Equipment TotalWells

Failure Ratefailure-per-well-per-year

Wet Tree Failure 72 0

Wet Tree Seal Failure 72 0.017

Wet Tree Pod Failure 72 0.023

Wet Tree Valve Failure 72 0.048Wet Tree Valve Failure 72 0.048

Wet Tree Choke Failure 72 0.005

ScSSV Failure 72 0.014

ScSSV Incidents 72 0.024

Flow Assurance Issues 72 0.01 to 0.15

Sand Control Failures 94 0.013

Sand Control Failures (circa 2004, SPE 84262)CompletionType

#Wells

TtlYrs

DesignFail, %

ApplicationFail, %

EarlyFail, %

Prod.Fail, %

Prod.Fail,f/w/y

SubsiFail,f/w/y

IntervenRatei/w/y

Injectors 30 85 0 10 0 31.3 0.070 0 0.059

ScreenlessFracs

26 107 0 27 0 7.7 0.019 0 0.808

Cased &Perf

61 336 0 1.6 0 41 0.074 0.003 0.024

Screen Only 194 756 0.5 0 1 21.6 0.056 0.0013 0.019Screen Only 194 756 0.5 0 1 21.6 0.056 0.0013 0.019

Expandable 197* 262 0.5 3.6* 0.5 3 0.023*

CHGP 387 1664 0 2.3 0.8 5.2 0.012 0.0006 0

OHGP 208* 613 0 7.7* 0.5 4.8 0.016* 0.0016 0.021

HRWP 187 556 0 0.5 0.5 2.7 0.009 0 0.002

Frac Pack 842 3351 1.5 2.4 0.2 2 0.005 0.0015 0.001

Total Wells 2132 7538

* Expandable screen and OHGPs were in early time at this point – better results now.

Producer Injector

Interventionfreq i/w/y

Interventionfreq in i/w/y Source and Comments

Root Failure major minor major minorSand control completion -stand alone screen

0.056 0.08 0.1 0 BP's 2000 well Sand Control FailuresDatabase (SPE 84262)

Sand control completion -gravel pack completion

0.011 0.025 0.05 0 BP's 2000 well Sand Control FailuresDatabase (SPE 84262)

Sand control completion -frac pack completion

0.005 0.007 BP's 2000 well Sand Control FailuresDatabase (SPE 84262)

Liner hanger 0.01 0.015 Data on hanger failures in GOM andlower 48 US

Intervention Rates – Land Wells(US & Canada, 2000 to 2004)

lower 48 USTubing, sweet gas, nonchrome, >7 psi CO2 partialpressure

0.2 0.05

General reliability data GoM and USOnshore

Tubing, sweet gas, nonchrome, >5<7 psi CO2 partialpressure

0.17 0.05

General reliability data GoM and USOnshore

Tubing, sweet gas, nonchrome, <5 psi CO2 partialpressure

0.1 0.05

General reliability data GoM and USOnshore

Tubing, Chrome13, CO2partial pressure <7

0.01

Tubing, water injector,chrome

0.3 0.1 General Reliability, Comments byCorrosion Group on injectors

Tubing, water injector, non-chrome

0.1 0.05 General Reliability, Comments byCorrosion Group on injectors

Producer InjectorInterventionfreq i/w/y

Interventionfreq in i/w/y Source and Comments

TT Patch / Repair 0.005 to 0.01 Estimate based on Alaskarepairs

Setting Plugs 0.1 Observed problems fromAlaska and Lower 48 USA,North Sea study on settingplugs.

Problems Encountered During Specific Operations, circa2000-2004

plugs.Tailpipe 0.02 0.02 Estimate from observations of

plug sticking and corrosionproblems.

Packer 0.0024 0.01 0.0024 0.005 Data from vendors (Baker)and direct experience andanalysis of failed packers inAlgeria and US.

PBR or similar 0.07 0.02 0.07 0.02 Stuck seal data fromHalliburton (Tom Rey) andstinger retrieval data fromfishing book.

Gas Lift - Land Wellscirca 1995-2004

Interventionfreq i/w/y Source and Comments

Gas lift failures 0.01Data from Weatherford (Rick Seagraves),modified, and BP Alaska

Gas Lift Optimization 0.08Data from Weatherford (Rick Seagraves),modified, and BP Alaska

Chemical Inj Mandrel 0.01Data from Weatherford (Rick Seagraves),modified

Chem Inj Valve 0.05Data from Weatherford (Rick Seagraves),modified

North Sea Well Event(Intervention)

% ofTime

InterventionFrequencyfailures/well/yr

InterventionFrequencyfailures/well/yr

InterventionFrequencyfailures/well/yr

Tubing RetrievableSSSV 20% 0.067 0.067Gas Lift Optimization 30% 0.1Gas Lift Equipment

North Sea Well Intervention Datacirca 2000

Gas Lift EquipmentFailures 3% 0.01Screens 15% 0.05Recompletions 12% 0.03Tubing Leaks 6% 0.01 0.045 0.011Stimulation 6% 0.01Logging 4% 0.01 0.05Seal Failures 3% 0.01 0.01 0.01Fishing 1% 0.03 0.03Sanding Up 0.06 0.025Profile Modification 0.01

Perf Condition ApproximateOpen

Perforations

References

Overbalance in Mud – moderate pressure well,no surge afterward

20% Downhole camera pictures, general well behaviorof Mobile bay and Opon well perforating

Overbalance in Mud – high pressure well, surgeand flow after perforating

30% SPE 16894

Overbalance in brine – no surge 35% SPE 15816

Overbalance in brine, surged 40% 15816, 16212

Overbalanced in acid, surge or no surge 45% Amoco Canadian Experiments

Perforating PerformanceCirca 1980 to 2004

Overbalanced in acid, surge or no surge 45% Amoco Canadian Experiments

Extreme Overbalanced Perforating, k<1 md, P>1.4 psi/ft with pumping after firing

60% Marathon Experiments and TerraTek tests(Dees, et.al.)

Extreme Overbalanced Perforating, k>1 md, P>1.4 psi/ft with no pumping

45% Amoco Canada experiments, Arco experiments

Underbalanced perfs, no flow 40% Downhole camera work, 14321, 16212,

Underbalanced perf, surge and flow >4 gal/perf,k>1md, h<50ft

50% 14,321, 16212, GOM experience (Bonomo andYoung’s Amoco work)

Underbalanced perf, surge and flow >4 gal/perf,k>1md, h>50ft

40% Anschutz Ranch Experience, 14321, 16212,

Underbalanced perf, surged and washed 75% 16212, Bonomo and Young’s Amoco work

Sand Abrasive Methods 80% 55044, downhole camera work from Canada

Bullet perforating, brittle formations such ascoal – some water wells

45% Coal well work, Mounds experiments (Amoco),Water well performance data

Bullet perforating, all other formations 25% Downhole camera work, well performance data

Activity % failure onfirst run

% failure onsecond run

Comments

Cement Packerplacement bybullheading

50% 25% Problems with contaminationand leaks

18

Cement Packerplacement by CT

10 to 15% 5%

Various Intervention SuccessCirca 1991 - 2005

placement by CTDual Hydraulic PackerRecovery

15 to 50%

Acidizing, matrix 30% to 70% historical results fromIndustry survey by Arco, 1991.Carl Montgomery

Wax Removal 20% to 60% Statistical survey ofmechanical and chemicalremoval techniques applieddownhole.

Scale Removal 10% to 30% Heavily dependent on removalmethod, mechanical, jettingand chemical methods varywith application.

Est. Int. Operation Success Rate Percent success on the first run or attempt.1981-85 1986-90 1991-95 1996-00 2001- Source

One Trip GP Systems 50% 72% 82%Robert Stomp, Conoco, IPQC Best Prac Sand Control,July 30/31, 2002

TCP - less than 200 ft 70% 92%M. Cloud and R. Kirkpatrick, Amoco, Internal report,1988

E-Line Perforating 98%Marathon Explosive Safety Conference, January20/21 2002

Tubing Cutoff - tension pulled, tool>/= 80% tubing ID 75% Amoco Study of Tubing Cut-Off SuccessTubing Cutoff - tension not pulled, or

Estimate of Initial Operations Success RateCirca 1981 to 2006

Tubing Cutoff - tension not pulled, ortool < 80% of tubing ID. 25% Amoco Study of Tubing Cut-Off Success

Fracturing (Hard Rock) 90% 92% Failure Frequency Data Base, BP led effort,

Frac and Pack 91% 95% Failure Frequency Data Base, BP led effort,

Frac and Pack - TSO 80% approximation

Gravel Pack - cased hole 97% Failure Frequency Data Base, BP led effort,

Gravel Pack - open hole 91% Failure Frequency Data Base, BP led effort,

High Rate Water Pack 99% Failure Frequency Data Base, BP led effort,

Expandable Sand Screens 86% Failure Frequency Data Base, BP led effort,

Screen Only Completions 98% Failure Frequency Data Base, BP led effort,

Perm Pkr Setting - scraper used 95% Dual Well Completion Operation Report

Perm Pkr Setting - scraper not used 85% Dual Well Completion Operation Report

ScSSV, succcess init run & test 98% Failure Frequency Data Base, BP led effort,

Acidizing, general use 30% Carl Montgomery, CEA study, 1995

All Coiled Tubing ops independent ofnumber of runs, non train wrecks 98% Nowsco North Sea StudyCoiled Tubing ops, no problemsrequiring NPT 75%

Activity % failure onfirst run

% failure onsecond run

Comments

WL run to EOT in 2-3/8”tubing

14% improve if cool water circulated4

WL run to EOT in largertubing

<2%

WL Plug setting 5% Assumes low scale, low paraffin environment

WL Plug pulling 20% 15% Debris over plug is major problem

Problems Encountered During Wirelineand CT Operations

WL Plug pulling 20% 15% Debris over plug is major problem

CT Plug Setting 10 to 15% Problems in sensitivity and depth control

CT plug pulling 10 to 15%

WL Perforating 2%to 3% <1% detonator/conductivity problems, assumes tubing isopen to TD

32

CT Perforating 5% to 8% 3% detonator/gun-to-gun failure, assumes tubing is open toTD

32

Tubing Puncher Charge 5% Depends on magnetic decentralizer operation33

Tube cut off, below packer 75% 75% Incomplete cut without tension29

Tube cut off, above packer 20% 20% Insufficient overpull, coatings & heavy or alloy pipe29,30

Sliding Sleeve Operation 10 to 50% depends on age, corrosion and debris, improve with CTimpact tool on CT

Problem in a Dual Incidence of failure(% of wells or units

over well life)

Comments and References.

Communicationbetween strings, nosliding sleeve

1 to 10% Lower figure reflects optimum seal selectionand good design of tubular and wellhead

4

connections, increases with sourenvironment

Dual Completions Failures

environment

Communicationbetween strings, slidingsleeve present

10 to 50% Lower figure reflects low debrisenvironment, increases with sourenvironment or debris

4

Dual Packer Failure 7% to15% Various factors including unseating due tocool fluids and leaks of the elements

42,43

Expansion Joint Failure to 50% Leaks42

Blowout, risk of 1blowout in 20 year welllife (side/side dual)

0.0023 Assumes a gas lifted completion21

Blowout, risk of 1pressure controlincident in 20 year welllife, concentric

0.0032 Assumes a gas lifted completion21

Difference Between Failure Rate andActual Intervention Rate - Trinidad Wells

Teak 70’s# wells

Teak70’swell-yr

Teak70’sF.R.

Early FailRate (%)

Teak 80’swells

Teak 80’swell-yr

Teak80’sF.R.

Early FailRate (%)

Teak90’swells

Teak 90’swell-yrs

Teak90’s F.R.

Early FailRate (%)

Avg Sand Control FailRate (f/w/y)

55 134.3 0.28 16 19 86.6 0.08 11 8 17.2 0.12 25

Avg Sand Cntrl InterRate (i/w/y)

55 134.3 0.17 19 86.6 0.03 8 17.2 0.06

No Cntrl Failures -deep zones) f/w/y

15 41.3 0.15 7 6 19.6 0.05 17 2 5.9 0.17 50

No cntrl Interv (i/w/y) 15 41.3 0.02 6 19.6 0.05 2 5.9 0.17

Bare Screen Failuresf/w/y

7 17.6 0.28 14

Bare Screen Interv.i/w/y

7 17.6 0.23

CH GP Failures f/w/y 31 71.2 0.35 23 13 67 0.09 8 2 5.9 0.11 17

CH GP Interv. i/w/y 31 71.2 0.24 13 67 0.07 2 5.9 0

OH GP Failures f/w/y 2 4.3 0.47

OH GP Interv. i/w/y 2 4.3 0.24

Producer Injector

platform subsea platform subsea

Root Cause major minor major minor major minor major minor

Downhole sand control failure 0.1002 0 0.1002 0 0.0200 0 0.0200 0

Liner hanger / packer 0 0 0 0 0 0 0 0

Isolation devices 0 0.0021 0 0.0021 0 0.0021 0 0.0021

Tailpipe 0 0.0021 0 0.0021 0 0.0021 0 0.0021

Packer 0.0021 0 0.0021 0 0.0021 0 0.0021 0

PBR or similar 0 0 0 0 0 0 0 0

Tubing 0.0025 0 0.0025 0 0.0025 0 0.0025 0

Gas lift mandrel 0.0140 0.0140 0.0104 0.0104 0 0 0 0

DHPG 0 0 0 0 0 0 0 0

Chemical Injection Mandrel 0 0 0 0 0 0 0 0

DHSV 0.0173 0.0380 0.0173 0.0380 0.0035 0.0243 0.0035 0.0243

Tubing hanger 0.0033 0 0.0033 0 0.0033 0 0.0033 0

Bullhead Scale Squeezes 0 0.0113 0 0 0 0 0 0

Coiled Tubing Scale Squeezes 0 0.0873 0 0.0397 0 0 0 0

Scale Sidetracks 0.0086 0 0.0039 0 0 0 0 0

Scale Milling 0 0.0366 0 0.0166 0 0 0 0

Example of a WellIntervention Predictionin West AfricaDevelopment

Scale Milling 0 0.0366 0 0.0166 0 0 0 0

Wax 0 0.0021 0 0.0021 0 0 0 0

Hydrates 0 0.0056 0 0.0056 0 0 0 0

Asphaltenes 0 0 0 0 0 0 0 0

Sand clean-out 0 0.0042 0 0.0042 0 0.0042 0 0.0042

Production logging 0 0.1013 0 0.0193 0 0.1013 0 0.0486

Downhole fluid samples 0 0 0 0 0 0 0 0

Downhole memory gauges 0 0.0050 0 0 0 0 0 0

Downhole plugs 0 0 0 0 0 0 0 0

Reperforation 0 0 0 0 0 0 0 0

Water shut-off 0 0.0186 0 0.0186 0 0 0 0

Gas shut-off 0 0.0032 0 0.0032 0 0 0 0

Recomplete 0 0 0 0 0 0 0 0

Sidetrack 0.0085 0 0.0085 0 0.0085 0 0.0085 0

Stimulation / fracturing 0 0.0186 0 0.0186 0 0 0 0

Production / injection monitoring 0 0 0 0 0 0 0 0

Well integrity 0 0 0 0 0 0 0 0

X-mas tree 0 0 0.0152 0.0152 0 0 0.0139 0.0069

Controls 0 0 0 0.0214 0 0 0 0.0214

Flowline / pipeline 0 0 0 0.0145 0 0 0 0.0104

Subtotal 0.1565 0.3500 0.1634 0.2316 0.0399 0.1340 0.0538 0.1200

Total 0.5065 0.3950 0.1739 0.1738

Average WOW% by Vessel Type

Comparison of WOW Percentages by Vessel Type (NNS, 'M' Task Limit)

50.0%

60.0%

70.0%

DP Vessel

MODU

0.0%

10.0%

20.0%

30.0%

40.0%

jan feb mar apr may jun jul aug sep oct nov dec

Month

WO

W%

25Frank Ketelaars, Jardine &

Associates Ltd., Oct 19, 2000

Base Case Results – DP Vessel UtilisationDPVessel UtilisationBreakdown

Gas Lift Valve

5.2%Transit/Mob

12.9%

Zonal Isolatio

19.0%

ScaleRemoval

4.4%

SandRemoval

2.6%

Lower Completi

8.7%Reperforation

0.9%

ScaleSqueeze

12.2%

SCSSV

4.0%Well Logging

10.0%

Xmas Tree

3.8%

Stimulation

7.2%

Upper Completi

9.0%

AverageDPutilisationover 2000-2007periodis71daysper annum

6Frank Ketelaars, Jardine &

Associates Ltd., Oct 19, 2000

Definitions (pt 1)

• Intervention frequency – based ininterventions per well per year.

Total number of workoversTotal number of workoversfreq = ----------------------------------------------

(total well number) x (cum years operation)

Definitions (pt 2)

• Major intervention – requires a rig and killing well

• Minor intervention – wireline or CT, well not typ.killed (some include data gathering, some don’t) –data gathering not included in this report.data gathering not included in this report.

• Early failure – problem corrected while the rig is onthe well – not really a intervention frequencyconcern, but failure to repair make it a concern.

Value of Early Data?

• Not much value for early interventionfrequency numbers.

• Early data on subsea wells dominated byprototype and first series equipmentprototype and first series equipment

– Steep learning curve (design, installation,operation)

– Numerous failures

Intervention Frequency Variables

• Initial clean-up effectiveness*• Effectiveness of initial T.S.O. fracturing*• Equipment test tolerance• Dry or wet tree• How much damage will you put up with?• How much damage will you put up with?• Facility/Injectors at capacity?• Is data collection an intervention?

* Assumes the early problems were not corrected –effect on intervention frequency can be severe.

Intervention Frequency Ranges

• Clean-up effectiveness – 0.025 to 0.17• T.S.O. fracturing success – 0.1 to 0.4• Equipment test tolerance - 0.026 to 0.23• Wet or dry tree – 0.075 (wet) to 0.15 (dry)*• Wet or dry tree – 0.075 (wet) to 0.15 (dry)*• Tolerance for damage? 0.0 to 0.09+• Facility at capacity? – 0.0 to 0.09

* Dry tree interventions easier and lower risk. Drytree reserve recovery is 1.5 to 2.5 x that of wettree.

25%

65%

40%

60%

80%

Norwegian Study of Reserve Recovery vs. Tree

Location

Data from Gulfax and Statfiord Fields in Norwegian North Sea, data provided by Statoil

0%

20%

40%

Wet Tree Dry Tree

Recovery

Intervention Causes

• Pod Failures - 20 to 30% of total

• Flow assurance problems - 20% to 30%

• Subsurface safety valves -10 to 20%

• Profile modification – 10 to 15%• Profile modification – 10 to 15%

• Seal problems -10% to 15%

• Sand control failures -5% to 10%

Intervention Frequency in GOM

• Discount early data (e.g. Mars – 0.33 at onetime, Uropa: 0.4 (?) due to hydrates)

• Troika: 0.086 wells/year, but could be 0.26 iftwo wells with early failure problems aretwo wells with early failure problems areworked over soon.

• Pompano: 0.027 wells/year, excellent, butmuch lower rate wells than Troika

Well#

CumProd.,MM Bbls

Months onProduction

No. ofInterventions

ProdRate,M Bpd

Anglethrupay

MechSkin

GlobalSkin

TA1 10.29 32.5 0 7.5 <20o +216*

+502*

TA2 16.82 29.4 0 12 <20o +116*

+243*

Troika Data

*

TA3 31.95 33.9 0 28 <20o +64 N/A

TA4 18.4 17.9 1 30 <20o +50 N/A

TA5 28.3 26.3 0 29.2 54o +45.8 +92

GOM DW Intervention FrequencyNumbers ????

• Pod Failures (minor) – 0.05 to 0.33

• Flow assurance problems – 0.01 to 0.15

• Subsurface safety valves – 0.027 to 0.1

• Profile control - 0.025 to 0.1• Profile control - 0.025 to 0.1

• Seal problems (minor?) – 0.05 to ?

• Sand control failures (TSO)– 0.085 to 0.12

• Corrosion – 0.01 to 0.04

Intervention Comments

• Intervention Frequency Lowered by:

– Good cleanup (UB perf and flow)

– Effective stimulation (TSO event = capacity)

– No tolerance test program during installation– No tolerance test program during installation

– Low damage/high productivity as only completionbenchmark (forget rig time !!!!)

– Redundancies in operating system

– Flexibility in design

Summary Comments

• Intervention Frequency Raised by:

– No surge cleanup (OB perf and no flow)

– Slurry packing (no TSO event = low capacity)

– Less than perfect installation– Less than perfect installation

– Efforts to cut rig time

– Prototype and first edition equipment

– Non flexible designs

Intelligent Completions???

• Downhole gauges – great!

• Zone control by downhole valves? – expectfailures early (leak paths), but these may payout for control – it is a gamble.out for control – it is a gamble.

• Most reliable completions are big, dumb(simple) completions with flexibility for repairand recompletions.

Risks

• Low well numbers

• Injector/producer communication? In faulted & layeredreservoirs, expectation is an extreme risk.

• GOM DW plans do not include artificial lift. Most experienceddeep water operating company, Petrobras, studying threedeep water operating company, Petrobras, studying threeforms of DW lift.

• Upper completion design configurations should be flexible.This should include annulus pressure and temperaturemonitoring, annular venting capability and redundancies.

Tubing Pressure Testing:Why it was stopped at BP.

• There have been NO tubing leaks of premium threadsdiscovered after over 180 tubing tests performed on 60completions in the North Sea since the early 1990’s.

• Operations to support pressure testing of tubingaccount for 25% of completion time in pressure testing,account for 25% of completion time in pressure testing,running and pulling plugs, and special packer testing.

• At this high proportion of completion time, tubingtesting would have to find at least one tubing leakevery four completions to be cost effective.

• This project reviewed over 587000 feet of mild steeland chrome tubing with over 19500 connections.

Dan Gibson, 1998

Well Interventions – Planning Methodology

Tubing

retrievable

SSSV's

FishRecompletions

Logging

Gas Lift valves

(inc. back check

repair)Sealing

devices

Screens

Tubing

Stimulation

PER WELL FAILURE FREQUENCY ANALYSIS

Production Tree Water injectiontree

Gas DisposalWell

Cause of intervention Classed

'Intervention'

Classed

'Workover'

Classed

'Intervention'

Classed 'Workover' Classed

'Intervention'

Classed

'Workover'

Failure rate Failure rate Failure rate Failure rate Failure rate Failure rate

[ / in-service year] [ / in-service

year]

[ / in-service year] [ / in-service year] [ / in-service year] [ / in-service

year]

Subsea X-mas tree 0.052 0.038 0.035

Well Interventions – Planning Methodology

• Schiehallion Well Intervention Frequency Predictions (D Driesen 1997) basedon SINTEF WellmasterTM Database…

Subsea X-mas tree

Failure

0.052 0.038 0.035

Tubing Failure 0.00215 0.00215

DHSV Failure 0.05 0.05 0.05

Gaslift System Failure 0.006

Elastomeric Seal Failure

Sandcontrol System

Failure

0.0198 0.015

Consequential Failure 0 0.003 0 0.003 0 0.003

Reservoir Surveillance 0.02274 0.017145 0.0195 0.012945 0.015 0.0114

Dynamic Data Aquisition:

Before Year 2000 0.007

After the year 2000 0

Water Shut-off @

receiver:

Before the year 2000 0

After the Year 2000 0.007

Water Shut-off @ source:

Before the year 2000 0

After the Year 2000 0.019

Gas Shut-off 0

Side track/re-completions 0.0132 0.01

Stimulation 0 0

TOTAL = 0.0898 0.07035 0.084 0.05315 0.05 0.038

Well Interventions – Field A: Actual vs. Prediction

Why are there fewerinterventions and why isthe timing different?

Well Interventions – Field B Actual vs. Prediction

Why are there fewerinterventions and why isthe timing different?

Losses – What?

What was the problem?

Subsea Field A – impact of failures

• Up to 2003, field will be gas constrained, hence inthe event of production failure, some oil can berecovered by compensation (25%-50%).

• From 2004 field produces flat out: no compensationpossible.possible.

• Immediate water injection losses are only caused byW11 & W13 failures, 4.5 and 12 Mbbls/dayrespectively in 2000 production rates.

• Field has back-up gas disposal well, allowing partialproduction during failure. Oil loss varies from 43%loss in 2000 to no impact in 2007.

10

Field B – impact of failures

• Up to 2004, field is gas constrained; if prod failure,some oil rec by compensation (25%-50%).

• From 2005 field produces flat out: no comp

• Water injection is constrained by injection pumps,hence any reduction in water injection during WIhence any reduction in water injection during WIwell failure will result in production loss: Oil lossvaries from 8 Mbbls/d/injector to 2 Mbbls/d/injector

• In the event of gas disposal well failure, oilproduction can be maintained at 30Mbbls/d byutilizing gas for fuel; allowing minimum flare.

11

Platform Field – impact of failures

• During first three years, back-up wells on platformcan compensate for outage of one gas producer.

• After 2002 no compensation is available in winter.• After 2002 no compensation is available in winter.

• No losses are anticipated during low offtakeperiod

12

Questions?

1. Hashim, T.R., Khalid, M.Z., Kruit, W., Short, D., Duncan, B, Pauzi, N, Haron, J., Low, F.N.: “Implementation of Twin Well Technology Offshore Sarawak,” SPE 50082, SPE Asia Pacific Oil and Gas Conference and Exhibition, Perth, Australia, 12-14 October, 1998.

2. Mayol, J.R., Salazar, A.: ”Dual Completions in El Furrial Field,” SPE 23684, Second Latin American Pet. Eng. Conf., Caracas, March 8-11, 1992.

3. Gabert, R.F., Ghnelm, G. J.: ”Procedures and Practices of Dual-Completion Design in Abu Dhabi,” SPE Production Engineering, February 1991, p20.

4. El Hanbouly, H.S., Saqqa, M.R., Constantini, N.M.,: ”Problems Associated with Dual Completions in SIP Wells: A Case History,” SPE 17985, SPE Mid East Oil Technical Conference and Exhibition, Manama, Bahrain, 11-14 March, 1989

5. Webster, K.R., O’Brien, T.B.:”Deep Duals Simplified,” SPE 3904, Deep Drilling Symposium, Amarillo, Sept 12, 1972.

6. Moring, J.D.: “How Skelly Handles Deep Duals at Warwick,” Pet. Eng., Dec 1974, P66.

7. Austingard, A., Erichsen, L., Vikra, S.: ”Case History: Gullfaks C-36AT3, A Multipurpose Oil Production/Gas Injection Well in the North Sea,” SPE 49107, New Orleans, 27-30 Sep 1998.

8. Wilson, D.J., Barrilleaux, M.F.: ”Completion Design and Operational Considerations for Multizone Gravel Packs in Deep, High Angle Wells,” OTC 6751, Houston, May 6-9, 1991.

9. Hall, D., Gardes, R.: ”Downhole Splitter in the Gulf of Mexico: Field Introduction and Results,” OTC 7905, Houston, 1-4 May 1995.

10. Durham, K.S.: ”Tubing Movement, Forces, and Stresses in Dual-Flow Assembly Installations,” SPE 9265, SPEJ, December 1982, p866.

11. Lambie, D.A., Walton, B.: ”Gas Lift in Multiple Completed Wells,” Southwestern Petroleum Short Course, Lubbock, Tx., April 18-19, 1968.

12. Weatherford Enterra Gas Lift and Design Seminar, Aug 30-Sept 2, 1999.

13. Plathey, Granger, J.L., Agam, A.R., Arnaud, F.: ”Dual Gas Lift Experience in Handil Field D.”Proceedings of the 17 th Annual Convention, Jakarta, October 27, 1988.

14. Davis, J.B., Brown, K. E.: ”Attacking those Troublesome Dual Gas Lift Installations,” SPE 4067, San Antonio, Oct. 8-11, 1972.

15. Telfer, G.: ”Downhole Packers for Use with ESP Completions,” OTC 7066, May 4-7, 1972.

16. Tunstall, K.N.: ”Artifical Lift as Applied to the Multiple Completion Choke Assembly,” Southwestern Petroleum Short Course, April 21-22, 1966.

17. Houlb, L. D.: ”Application of High Pressure Dual Tubingless Completions and Low Cost Concepts in an Environmentally Sensitive Area,” IADC/SPE 35123, New Orleans, 12-15 March 1996.

18. Soetedja, V., Hunter, D.L.: ”Production Optimizating with Coiled Tubing and Other Rigless Techniques,” SPE 36963, Asia Pacific Mtg, Adelaide, Australia, 28-31 Oct. 1996.

19. Sutton, D.L., et.al.: ”Well Cementing Process and Gasified Cements Used Therein,” US Patent 4,340,427, (1982)

20. Cerruti, S.E.: ”Dual-Completion Design for HP/HT Corrosive Oil Well, Villafortuna-Trecate Italy,” SPE 28892, Euro Conf, London, 25-27 Oct 1994

21. Grassick, D.D., King, S.D.J.: ”Blowout Risk Analysis of Gas-Lift Completions,” SPEPE, May 1992.

22. Grassick, D.D., Kallos, P.S., Jardine, I.J.A., Deegan, J.: ”Risk Analysis of Single and Dual-String Gas-Lift Completion,” JPT November, 1980.

23. “King, G.E.: Surface Controlled and Subsurface Controlled Subsurface Safety Valves: Use in OnShore Wells,” an internal Amoco report based on industry literature.

24. Molnes, E.L., Rausand, M., Linquist, B.: ”ScSSV Reliability Tested in the North Sea,” PEI, Nov 1987.

25. Busch, J.M., Policky, B.J., Llewelyn, D.C.G.: ”Subsurface Safety Valves, :Safety Asset or Liability,” SPE 12193.

26. Sintef: Reliability of Surface Controlled Subsurface Safety Valves-Phase IV, STF75 F91038.

27. Deepstar II CTR 1010-1,” Surface Controlled Subsurface Safety Valves Risk Assessment,”

28. Molnes, E., Sundet, I.: ”Reliability of Well Completion Equipment,” SPE 26721.

29. King, G.E., Corgan, J.M., Perry, J.T., Powers, B.S.: ”Initiating Report on Pipe Cut-Off Techniques: Potential for Improving the Technology,” An internal Amoco Report.

30. King, G.E.: ”Evaluation of Downhole Cutting Methods for Severing Drill Pipe in Trinidad Well MA-08,” F99-P-8, February 11, 1999.

31. King, G.E.: ”Development of Technology Risk Values for Weighting Workover Economics, 1997 Update,” An Internal Amoco Report

32. Kirkpatrick, J., Cloud, M.: ”Tubing Conveyed Perforating, A Performance Evaluation,” Amoco Internal Report.

33. King, G.E.:”Evaluation of Tubing Perforating Puncher Charges in Tubing Selective Completions Through Gravel Packed Intervals,” An Internal Amoco Report33. King, G.E.:”Evaluation of Tubing Perforating Puncher Charges in Tubing Selective Completions Through Gravel Packed Intervals,” An Internal Amoco Report

34. Conversation and data from Dan Gibson, North Sea Operations

35. Barrilleaux, M.F., Colbert, J., King, G.E., Murugappan, B.: “Real Time Reservoir Management Using Intelligent Completions to Gather Data, Modify Flow Paths, Accelerate Recovery, and Optimize Flow in Multiple Pay Wells; Notes from an SPE Forum,” F98-P-19, July 16, 1998.

36. Soeprijadi, Hervochon, J., Nugraha, A., Noiray, J.M.: “Light Workovers in the Handil and Bekapai Fields,” Indonesia Pet. Assoc., 19 th Annual Convention, October, 1990. IPA90-174.

37. Aldridge, D.: “Multi Zone Completion Design,” SPE 35229, Permian Oil Conference, Midland, 27-29 March 1996.

38. Barua, S.: “Computation of Heat Transfer in Wellbores with Single and Dual Completions,” SPE 22868, Annual Tech. Conf., Dallas, Oct 6-9, 1991.

39. Davidson, J.A., Victors, C.M.: “An Overview of the Performance of Gas Lift Operations in the Ardjuna Fields, Indonesia Petroleum Proceedings, 15 th Annual Conv., Oct 1986.

40. Geyelin, J.L.: “Field Experience in Down Hole Annulus Safety Valves,” SPE 19278, Offshore Europe, 1989.

41. Tamimi, Y.K.: “Rigless Acid Fracturing Through Dual Completions: An Offshore Experience in the Upper Zakum Field,” ADNOC/SPE 21313,

42. VT between BP-Kuwait and Abu-Dhabi offices, 9/23/98

43. Information from E-mail from Mark Barrilleaux, experience on Exxon wells

44. Information from E-mail from Graham Beadie, experience on Abu Dhabi wells

45. Gyani, Angela, presentation on Asphaltenes

46. J. Faber, G. J. P. Joosten, K. A. Hashmi & M. Gruenenfelder, SPE 39633, presented at the SPE/DOE Improved Oil Recovery Symposium, Tulsa, OK, 22/04/1998. “Water shut-off field experience with a relative permeability modification system in the Marmul feld (Oman)”.

47. F. Legere, J. Cdn. Pet. Tech, October 1978, 51 - 60, ”Reduction of water - oil ratio using polyacrylamides in vugular carbonate reservoirs”

48. P. Hessert & P. D. Fleming, Third tertiary oil recovery conference, Wichita

49. D. Sydansk & T. B. Smith, SPE 17383 presented at the SPE/DOE EOR Symposium, Tulsa, OK, April 17 - 20, 1988, “Field testing of a new conformance - improvement - treatment Chromium (III) gel technology”

50. N. Senol, R. Gulumser & N. Tekayak,

51. D. Sydansk & P. E. Moore, 1991, SPE 21894,

52. D. Sydansk & P. E. Moore, Oil & Gas Journal, Vol. 90, Number 3, 40 - 45, January 20th 1992, “Gel Conformance treatments increase oil production in Wyoming”.

53. D. Moffitt, JPT, April 1993, 356 - 362

54. S. Seright & J. Liang, DOE/BC/14880-5, 95 - 140, December 1993, “Improved techniques for fluid diversion in oil recovery”

55. Gel Technologies Corporation treatment summary report for Marathon Oil Company Big Lake Field, Reagan County, Texas University NO 208, January 1994.

56. P. Southwell & S. M. Posey, SPE 27779, SPE/DOE 9th Symposium on IOR, Tulsa

57. C. Borling, SPE 27825, SPE/DOE 9th Symp. on IOR, Tulsa , April 17 - 20, 1994, “Injection conformance control case histories using gels at the Wertz field CO2 tertiary flood in Amoco Wertz field, Wyoming”.

58. D. Sydansk & G. P. Southwell, SPE 49315, Ann. Tech. Conf. & Exhibn., New Orleans, Sept 27 - 30, 1998. “More than 12 years of experience of successful Conformance - control polymer gel technology”

59. I. Tweidt et al, SPE 38901, SPE Ann. Tech. Conf. & Exhibition, San Antonio October 5 - 8, 1997. “Improving sweep efficiency in the Norman Wells naturally fractured reservoir through the use of polymer gels”. Plus notes.

60. Randy Seright, New Mexico PRRC 1997

61. Personal communication from BP Kuwait.

62. Grassick, D.D., Kallos, P.S Dean, S., King, S.D.J.: “Blowout Risk Analysis of Gas-Lift Completions,” SPE Prod Eng., May 1992, p172.

63. Grassick, D.D., Kallos, P.S., Jardine, I. J. A., Deegan, F.J.: “Risk Analysis of Single and Dual-String Gas-Lift Completions,” JPT, Nov. 1990.

64. 13 Cr Running Time - Cross Assets Comparisons - Forties Case

65. Cameron, Paul: “Total-Fina-Elf - BP Northern BU,” Meeting Slides, 14 March 2002.

66. Gibson, Dan: “Data on 60 well study in North Sea (1999) and 25 well study in Forties (2001)

Intervention Data – Stat. Fail. Freq.

The following average MTTF are used (* indicates initial valuesgiven prior to adjustment to reflect ‘critical failures’ only):

Read the definitions and explanations of MTBF and the limits ofStatistical frequency before inferring rates from this data.

Xmas tree replacement: 250 yrs (* 50 yrs) Xmas tree replacement: 250 yrs (* 50 yrs)

SCSSV wireline insert: 83 yrs (* 17 yrs)

SCSSV replacement: 111 yrs

GLV critical failure: 50 yrs (*10 yrs)

Upper completion repair: 100 yrs

Upper completion replacement: 100 yrs

Lower completion repair: 50 yrs

Lower completion replacement: 33 yrs

15