Well Integrity Technical Section

Let me start by saying that this post is not a question, rather an attempt to start an active discussion. OK, maybe it is full of questions, but I doubt there are obvious answers.

Microannuli are very common, causing leaks in maybe one well in five. They have a disastrous effect on CBL, and any suggestion that a microannulus is not as nasty as it looks tends to be met by skepticism by regulators. But are these hair-wide, mile-long defects bad? And how exactly do they appear?

I have been observing microannuli on logs for years, and I'm trying to understand them through experiments and models (including a recent in-depth study of a data-rich microannulus leak), so I'd love to hear your experience, ideas or opinions on microannuli.

Here are some questions with still too many answers, but feel free to add your own:

* How does a microannulus form? Is it because of local conditions at each depth or because of a single debonding event progressing as a hydraulic fracture, or both?

* What is the initial state of stress in cement, i.e. what is the net stress that determines the defect’s opening?

* And how about the role of normal stresses created by formation creep in rocks such as shales, marls and salt?

* Is the width of a microannulus the result of a tug-of-war between fluid pressure and the elasticity of the well materials?

* Are (relatively) simple mechanical models sufficient to predict flow through microannuli? Or does one have to mobilize 3D finite-element models?

* Are cement-formation microannuli common? Do microannuli switch between interfaces?

 

There is also a project proposal to gather a public library of microannuli on my LinkedIn profile: your suggestions and comments on the document would be really useful. Please let me know if you'd like me to send you a copy.

------------------------------

Matteo Loizzo
Well integrity consultant
Berlin
------------------------------

Matteo,

In my experience a lot of these problems are the result of the inherent drawbacks of using Portland type cements. They all generate gel strengths between the fluid and set states. In pressured environments the resultant reduction in hydrostatic can , and ofter does, create channels.

Leaking wells as a result of poor cementing is a World wide problem.

There are now alternatives to using Portland cements where there is no gel strength generation .But getting the industry to change is very difficult . How often I hear 'cement is cheap' . Not in the long run when faced with expensive remedial operations which often just repeats the process by using cement. 

------------------------------
Brian Tomlinson
President Technology Development
Deep Earth Solutions
Bacou
------------------------------

Hello Brian T and Matteo L, 

A great topic "shedding a light on microannuli"  With it there is a matter of getting through the "red tape" that is blocking seeing the root cause for the failures that occur.  Transparency of operational and logging logs.  

So if I may suggest we first work the issue  " * How does a microannulus form? Is it because of local conditions at each depth or because of a single debonding event progressing as a hydraulic fracture, or both?"

Can I start by suggesting it has to do with both "local conditions at each depth" and "a single debonding event" but seldom to do with "hydraulic fracture" unless mud weights and casing seats are improperly designed which they were on Lusi Mud Volcano, Montara and Macondo where there has been data made public because of massive well integrity failures. 

A common factor to a bonding / well integrity failure being revealed is in most cases related to gas (the Bladder Effect) and to a much lesser degree related to water (killing the well flow and corrosion for example).   Point being the invasion of these two fluids occurs where there is a annular bonding failure.  

On the research I was involved with in the 70's with Esso (now ExxonMobil) and Halliburton, the two issues of failure noted were:

1. if casing isn't rotated where there is spiraling and rugose / washed out hole conditions, there would not be a bond because the spacer and cementing fluids would channel through the more open side of the casing / hole annuli even when in turbulent flow.   Later if a pressure imbalance is occurring between two zones the gas or water will push through the micro annulus and produce unwanted fluids and in some cases migrate to where there is cross flow behind the casing. 

2.  with the cementing setting process (as discussed by Brian T) there is a period in the hydration process where cement solids are supported but there is still a liquid phase and the hydro static of the liquid column to formation gas becomes under balanced and the gas comes into the annulus and then my migration works it way through any un-cemented micro annulus.  Another situation for the gas to migrate is when a high overbalance of drilling fluid exists and the "bottom" of the reservoir is unable to hold the hydro static of whole mud and so when the annulus at the well head is sealed, the hole cannot fill so the balance of hydro static at the bottom leads to the under balance of a gas zone at the top and once a gas bubble enters the hole the "bladder effect" prevails.   

In many cases a casing design, that may include the uses of SET technologies can be reliable with a Class C to G cement but for sure there are times where alternatives such as expandable casing packers and non cement based adhesives.
------------------------------
Wayne Needoba
LIS Thailand Co. Ltd.
KarawaraWA
------------------------------

Original Message:
Sent: Sep 01,2015 05:28 AM
From: Brian Tomlinson
Subject: Shedding a light on microannuli

Matteo,

In my experience a lot of these problems are the result of the inherent drawbacks of using Portland type cements. They all generate gel strengths between the fluid and set states. In pressured environments the resultant reduction in hydrostatic can , and ofter does, create channels.

Leaking wells as a result of poor cementing is a World wide problem.

There are now alternatives to using Portland cements where there is no gel strength generation .But getting the industry to change is very difficult . How often I hear 'cement is cheap' . Not in the long run when faced with expensive remedial operations which often just repeats the process by using cement. 

------------------------------
Brian Tomlinson
President Technology Development
Deep Earth Solutions
Bacou
------------------------------

Thanks Wayne and Brian,

You're spot on on how a type of defect appears, but that's rather a chimney: the path left behind by a migrating fluid through setting cement (the result is a thin "tube", hence the name). A chimney can't be spotted on a CBL, but it's obvious on a high resolution ultrasonic log. With an SBT now you see it, now you don't.

From experience and theory there are three main differences between microannuli and chimneys:

• Microannuli appear after cement set, and chimneys during setting.
• Chimneys let through far more fluid than microannuli do.
• Microannulus opening depends on elasticity (i.e. they are non-linear), whereas chimneys don't. That's due to their shape: a microannulus is a thin crack enveloping the casing and a chimney is a roundish tube.

I think I understand Wayne's point about mud channels and imperfect cement placement and the creation of microannuli, but I cannot see how the two can be related: after all if you have mud, you don't have bond. And if there is overpressure, then the mud channel will "fail" before the cement bond does. So I would tend to put the first mode of failure under the "mud channel" heading, and the second one under the "chimney" heading.

By the way: if you're into history, here are three papers that pioneered the idea of microannulus as fracture:

• Carter & Evans (Halliburton), SPE 764 PA who in 1964 clocked the debonding vertical progression at 1.25 ft/min with water and 20 ft/min with gas
• Dusseault et al., SPE 64733
• Lecampion et al., "Interface debonding driven by fluid injection in a cased and cemented wellbore: Modeling and experiments", IJGGC 2013.

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------

Original Message:
Sent: Sep 01,2015 08:16 AM
From: Wayne Needoba
Subject: Shedding a light on microannuli

Hello Brian T and Matteo L, 

A great topic "shedding a light on microannuli"  With it there is a matter of getting through the "red tape" that is blocking seeing the root cause for the failures that occur.  Transparency of operational and logging logs.  

So if I may suggest we first work the issue  " * How does a microannulus form? Is it because of local conditions at each depth or because of a single debonding event progressing as a hydraulic fracture, or both?"

Can I start by suggesting it has to do with both "local conditions at each depth" and "a single debonding event" but seldom to do with "hydraulic fracture" unless mud weights and casing seats are improperly designed which they were on Lusi Mud Volcano, Montara and Macondo where there has been data made public because of massive well integrity failures. 

A common factor to a bonding / well integrity failure being revealed is in most cases related to gas (the Bladder Effect) and to a much lesser degree related to water (killing the well flow and corrosion for example).   Point being the invasion of these two fluids occurs where there is a annular bonding failure.  

On the research I was involved with in the 70's with Esso (now ExxonMobil) and Halliburton, the two issues of failure noted were:

1. if casing isn't rotated where there is spiraling and rugose / washed out hole conditions, there would not be a bond because the spacer and cementing fluids would channel through the more open side of the casing / hole annuli even when in turbulent flow.   Later if a pressure imbalance is occurring between two zones the gas or water will push through the micro annulus and produce unwanted fluids and in some cases migrate to where there is cross flow behind the casing. 

2.  with the cementing setting process (as discussed by Brian T) there is a period in the hydration process where cement solids are supported but there is still a liquid phase and the hydro static of the liquid column to formation gas becomes under balanced and the gas comes into the annulus and then my migration works it way through any un-cemented micro annulus.  Another situation for the gas to migrate is when a high overbalance of drilling fluid exists and the "bottom" of the reservoir is unable to hold the hydro static of whole mud and so when the annulus at the well head is sealed, the hole cannot fill so the balance of hydro static at the bottom leads to the under balance of a gas zone at the top and once a gas bubble enters the hole the "bladder effect" prevails.   

In many cases a casing design, that may include the uses of SET technologies can be reliable with a Class C to G cement but for sure there are times where alternatives such as expandable casing packers and non cement based adhesives.
------------------------------
Wayne Needoba
LIS Thailand Co. Ltd.
KarawaraWA
------------------------------

Will follow this with more detail but my thinking in relation to micro annulus, as you define Matteo,  is being in relation to casing compression and expansion.  Like pressure testing the casing where the cement isn't fully set, then the expansion of the casing from pressure leads to a micro annulus when you bleed the pressure.  Some temperature effects would also apply.  

------------------------------
Wayne Needoba
LIS Thailand Co. Ltd.
KarawaraWA
------------------------------

Original Message:
Sent: Sep 01,2015 09:32 AM
From: Matteo Loizzo
Subject: Shedding a light on microannuli

Thanks Wayne and Brian,

You're spot on on how a type of defect appears, but that's rather a chimney: the path left behind by a migrating fluid through setting cement (the result is a thin "tube", hence the name). A chimney can't be spotted on a CBL, but it's obvious on a high resolution ultrasonic log. With an SBT now you see it, now you don't.

From experience and theory there are three main differences between microannuli and chimneys:

• Microannuli appear after cement set, and chimneys during setting.

• Chimneys let through far more fluid than microannuli do.

• Microannulus opening depends on elasticity (i.e. they are non-linear), whereas chimneys don't. That's due to their shape: a microannulus is a thin crack enveloping the casing and a chimney is a roundish tube.

I think I understand Wayne's point about mud channels and imperfect cement placement and the creation of microannuli, but I cannot see how the two can be related: after all if you have mud, you don't have bond. And if there is overpressure, then the mud channel will "fail" before the cement bond does. So I would tend to put the first mode of failure under the "mud channel" heading, and the second one under the "chimney" heading.

By the way: if you're into history, here are three papers that pioneered the idea of microannulus as fracture:

• Carter & Evans (Halliburton), SPE 764 PA who in 1964 clocked the debonding vertical progression at 1.25 ft/min with water and 20 ft/min with gas

• Dusseault et al., SPE 64733

• Lecampion et al., "Interface debonding driven by fluid injection in a cased and cemented wellbore: Modeling and experiments", IJGGC 2013.

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------



Original Message:
Sent: Sep 01,2015 08:16 AM
From: Wayne Needoba
Subject: Shedding a light on microannuli

Hello Brian T and Matteo L, 

A great topic "shedding a light on microannuli"  With it there is a matter of getting through the "red tape" that is blocking seeing the root cause for the failures that occur.  Transparency of operational and logging logs.  

So if I may suggest we first work the issue  " * How does a microannulus form? Is it because of local conditions at each depth or because of a single debonding event progressing as a hydraulic fracture, or both?"

Can I start by suggesting it has to do with both "local conditions at each depth" and "a single debonding event" but seldom to do with "hydraulic fracture" unless mud weights and casing seats are improperly designed which they were on Lusi Mud Volcano, Montara and Macondo where there has been data made public because of massive well integrity failures. 

A common factor to a bonding / well integrity failure being revealed is in most cases related to gas (the Bladder Effect) and to a much lesser degree related to water (killing the well flow and corrosion for example).   Point being the invasion of these two fluids occurs where there is a annular bonding failure.  

On the research I was involved with in the 70's with Esso (now ExxonMobil) and Halliburton, the two issues of failure noted were:

1. if casing isn't rotated where there is spiraling and rugose / washed out hole conditions, there would not be a bond because the spacer and cementing fluids would channel through the more open side of the casing / hole annuli even when in turbulent flow.   Later if a pressure imbalance is occurring between two zones the gas or water will push through the micro annulus and produce unwanted fluids and in some cases migrate to where there is cross flow behind the casing. 

2.  with the cementing setting process (as discussed by Brian T) there is a period in the hydration process where cement solids are supported but there is still a liquid phase and the hydro static of the liquid column to formation gas becomes under balanced and the gas comes into the annulus and then my migration works it way through any un-cemented micro annulus.  Another situation for the gas to migrate is when a high overbalance of drilling fluid exists and the "bottom" of the reservoir is unable to hold the hydro static of whole mud and so when the annulus at the well head is sealed, the hole cannot fill so the balance of hydro static at the bottom leads to the under balance of a gas zone at the top and once a gas bubble enters the hole the "bladder effect" prevails.   

In many cases a casing design, that may include the uses of SET technologies can be reliable with a Class C to G cement but for sure there are times where alternatives such as expandable casing packers and non cement based adhesives.
------------------------------
Wayne Needoba
LIS Thailand Co. Ltd.
KarawaraWA
------------------------------

You're right Wayne: reducing pressure or temperature will cause cement to peel off. Nonetheless, cement has to be somewhat set to debond (otherwise you'd get a chimney). A microannulus can form at all stages in the life of a well: I have seen some appear like clockwork as soon as a pump was switched on and pressure in the casing dropped. The late occurrence of microannuli may be why a study found that the failure rate of GoM wells increases slightly with age (reference on request).

About "weird-things-ahv-seen", I once came across a microannulus that was caused by the formation heating up after setting because of a "heat bubble" from drilling that came back towards the well.

Now the real question is: does temperature and pressure cause debonding at the well scale only when there is a source of pressurized fluid (brine, gas, whatever) that pushes upwards at the interface, or microannuli can also appear without fluid migration?

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------

Original Message:
Sent: Sep 02,2015 06:15 AM
From: Wayne Needoba
Subject: Shedding a light on microannuli

Will follow this with more detail but my thinking in relation to micro annulus, as you define Matteo,  is being in relation to casing compression and expansion.  Like pressure testing the casing where the cement isn't fully set, then the expansion of the casing from pressure leads to a micro annulus when you bleed the pressure.  Some temperature effects would also apply.  

------------------------------
Wayne Needoba
LIS Thailand Co. Ltd.
KarawaraWA
------------------------------



Original Message:
Sent: Sep 01,2015 09:32 AM
From: Matteo Loizzo
Subject: Shedding a light on microannuli

Thanks Wayne and Brian,

You're spot on on how a type of defect appears, but that's rather a chimney: the path left behind by a migrating fluid through setting cement (the result is a thin "tube", hence the name). A chimney can't be spotted on a CBL, but it's obvious on a high resolution ultrasonic log. With an SBT now you see it, now you don't.

From experience and theory there are three main differences between microannuli and chimneys:

• Microannuli appear after cement set, and chimneys during setting.

• Chimneys let through far more fluid than microannuli do.

• Microannulus opening depends on elasticity (i.e. they are non-linear), whereas chimneys don't. That's due to their shape: a microannulus is a thin crack enveloping the casing and a chimney is a roundish tube.

I think I understand Wayne's point about mud channels and imperfect cement placement and the creation of microannuli, but I cannot see how the two can be related: after all if you have mud, you don't have bond. And if there is overpressure, then the mud channel will "fail" before the cement bond does. So I would tend to put the first mode of failure under the "mud channel" heading, and the second one under the "chimney" heading.

By the way: if you're into history, here are three papers that pioneered the idea of microannulus as fracture:

• Carter & Evans (Halliburton), SPE 764 PA who in 1964 clocked the debonding vertical progression at 1.25 ft/min with water and 20 ft/min with gas

• Dusseault et al., SPE 64733

• Lecampion et al., "Interface debonding driven by fluid injection in a cased and cemented wellbore: Modeling and experiments", IJGGC 2013.

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------

Gentlemen,
As I said the problem , in my opinion, lies with the cement . We need a different material. The only differece between Portland type cements from aeons ago and today is the additives they put in it.
rgds

------Original Message------

You're right Wayne: reducing pressure or temperature will cause cement to peel off. Nonetheless, cement has to be somewhat set to debond (otherwise you'd get a chimney). A microannulus can form at all stages in the life of a well: I have seen some appear like clockwork as soon as a pump was switched on and pressure in the casing dropped. The late occurrence of microannuli may be why a study found that the failure rate of GoM wells increases slightly with age (reference on request).

About "weird-things-ahv-seen", I once came across a microannulus that was caused by the formation heating up after setting because of a "heat bubble" from drilling that came back towards the well.

Now the real question is: does temperature and pressure cause debonding at the well scale only when there is a source of pressurized fluid (brine, gas, whatever) that pushes upwards at the interface, or microannuli can also appear without fluid migration?

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------

Original Message:
Sent: Sep 02,2015 06:15 AM
From: Wayne Needoba
Subject: Shedding a light on microannuli

Will follow this with more detail but my thinking in relation to micro annulus, as you define Matteo,  is being in relation to casing compression and expansion.  Like pressure testing the casing where the cement isn't fully set, then the expansion of the casing from pressure leads to a micro annulus when you bleed the pressure.  Some temperature effects would also apply.  

------------------------------
Wayne Needoba
LIS Thailand Co. Ltd.
KarawaraWA
------------------------------



Original Message:
Sent: Sep 01,2015 09:32 AM
From: Matteo Loizzo
Subject: Shedding a light on microannuli

Thanks Wayne and Brian,

You're spot on on how a type of defect appears, but that's rather a chimney: the path left behind by a migrating fluid through setting cement (the result is a thin "tube", hence the name). A chimney can't be spotted on a CBL, but it's obvious on a high resolution ultrasonic log. With an SBT now you see it, now you don't.

From experience and theory there are three main differences between microannuli and chimneys:

• Microannuli appear after cement set, and chimneys during setting.

• Chimneys let through far more fluid than microannuli do.

• Microannulus opening depends on elasticity (i.e. they are non-linear), whereas chimneys don't. That's due to their shape: a microannulus is a thin crack enveloping the casing and a chimney is a roundish tube.

I think I understand Wayne's point about mud channels and imperfect cement placement and the creation of microannuli, but I cannot see how the two can be related: after all if you have mud, you don't have bond. And if there is overpressure, then the mud channel will "fail" before the cement bond does. So I would tend to put the first mode of failure under the "mud channel" heading, and the second one under the "chimney" heading.

By the way: if you're into history, here are three papers that pioneered the idea of microannulus as fracture:

• Carter & Evans (Halliburton), SPE 764 PA who in 1964 clocked the debonding vertical progression at 1.25 ft/min with water and 20 ft/min with gas

• Dusseault et al., SPE 64733

• Lecampion et al., "Interface debonding driven by fluid injection in a cased and cemented wellbore: Modeling and experiments", IJGGC 2013.

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------

Brian,

I agree with you that SmartSet would greatly reduce the risk for fluid invasion of the annulus while the filler is setting. So no chimneys.

Microannuli, however, are a purely mechanical defect and they don't only affect cement. I have seen shale debonding from the casing (and so have Clarke et al., 2003, "Gulf Coast borehole closure test well near Orange, Texas"), so I doubt that any material is immune to it.

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------

Original Message:
Sent: Sep 05,2015 05:28 AM
From: Brian Tomlinson
Subject: Shedding a light on microannuli

Gentlemen,
As I said the problem , in my opinion, lies with the cement . We need a different material. The only differece between Portland type cements from aeons ago and today is the additives they put in it.
rgds

Matteo, I think you can focus to see solutions around your comment "Now the real question is: does temperature and pressure cause debonding at the well scale only when there is a source of pressurized fluid (brine, gas, whatever) that pushes upwards at the interface, or microannuli can also appear without fluid migration?"

Definitely the microannuli can form without fluid migration.  However, it will only be a process hazard when there is a source of  pressurized fluid (or gravity in the case of gas) to utilize the microannulus.  So step one is to recognize the geological hazards that would be presenting a risk to well integrity, that includes the risk of a microannuli leading to a failure in the bond of the annulus.  Might I suggest that the understanding of what is occurring for every well system as it is cemented should be simulated for understanding and then a planned Vs actual behavior during the D&C process can be used to understand all well integrity risk with certainty.   There is the use of SET technology, expandable casing packers, Smartset, expandable cements that include pre hydrated bentonite and brine mixwaters, and a range of polymer products.  From my experience and the lab work I've witnessed, if you don't rotate the casing during the cementation process there will be channels that will enhance the failure of a microannuli.  So in the end if there is a "pressurized fluid" / "gravity" migration force, there should be redundancy in well integrity design and the tubular rotated, even when setting cement plugs in abandonment.  Channeling is a hazard due to wash outs, spiraling and or rugose hole, and then there is the understanding of what hydro-static changes are occurring during the cementation process.

------------------------------
Wayne Needoba
LIS Thailand Co. Ltd.
KarawaraWA
------------------------------

-------------------------------------------
OriginalMessage:
Sent: Sep 06,2015 02:37 AM
From: Matteo Loizzo
Subject: Shedding a light on microannuli

Brian,

I agree with you that SmartSet would greatly reduce the risk for fluid invasion of the annulus while the filler is setting. So no chimneys.

Microannuli, however, are a purely mechanical defect and they don't only affect cement. I have seen shale debonding from the casing (and so have Clarke et al., 2003, "Gulf Coast borehole closure test well near Orange, Texas"), so I doubt that any material is immune to it

------------------------------
Matteo Loizzo
Well integrity consultant
Berlin
------------------------------
-------------------------------------------
OriginalMessage:
Sent: Sep 05,2015 05:28 AM
From: Brian Tomlinson
Subject: Shedding a light on microannuli

Gentlemen,
As I said the problem , in my opinion, lies with the cement . We need a different material. The only differece between Portland type cements from aeons ago and today is the additives they put in it.
rgds