Using a cement bond log, a variable density log, and ultrasonic tools provides an indispensable (but different) set of evaluative data. The objective of cement bond logs like these is to provide information to understand if we have achieved zonal isolation. If not, they can help to asses if a cement squeeze will be possible to restore it.
The challenge of variable density and cement bond log interpretation
The first paragraph covers how cement evaluation logs are mostly accepted, however…
There is a big, big BUT here…
Cement logs, i.e., cement bond logs (CBL), variable density logs (VDL), and ultrasonic tools do not show or measure zonal isolation…
With them, we can only see cement presence in contact or close to the casing (CBL and Ultrasonic logs), and in contact or near the formation (only VDL). If we see that on average or, better, circumferentially then we infer that isolation in likely there. But, if we do not see sufficient cement, then isolation cannot be verified by logs. But by interpretation, if we are lucky enough.
Everything I will mention from here assumes that the cement evaluation logs have been QA/QC’ed diligently, in other words, there are no tool setting/calibration/configuration issues affecting (negatively or positively) the look of the logs.
Most people would agree that there are only two situations when the log totally speaks by itself and does not require a ‘log whisperer’, that is: free pipe and fully bonded cement. For practical purposes, I would agree. However, certain situations can blur that assumption:
1) On the one hand, poor mud properties (flocculation, leaking & thick mud cake, etc.) can lead to a similar appearance to free pipe. (When gaps beyond the depth of investigation (usually > 0.25 mm liquid filled gaps) invisible to ultrasonic tools are created) ; and
2) On the other hand, formations debris (from the drilling process or the interaction with drilling mud), fast formations and even creeping formations can give a too optimistic image on the log.
My point of view
As a cementer, I rely on pumping pressure. That is the only direct measurement we have that proves we put cement in the annulus. However, there are two circumstances here:
1) Well in control (no losses no gains while cementing) and final circulating pressure matches the expected value. In this case, doesn’t matter what the log says. The cement is there, so sooner or later, isolation will be in place.
Situations where you should be worried when the log is not answering important questions:
- Gas wells.
- A significant pressure differential between two adjacent formations, with a natural barrier in between. Like an upper lost zone and lower high-pressure zone.
- Interval-specific stimulation candidates.
Why you should worry about these situations?
Because in addition to lacking isolation, which means there is a path, for fluids to move they must overcome the friction pressure of the path; (if there is a gap between cement and casing or annulus, like a micro-annulus); or channel (if there is a gap parallel to the casing). Gas and significant differential pressure can undoubtedly create that trick.
2) Well is not in control (loses or influx during cementing), forget about pumping pressure because now its true meaning has been concealed … and you are entirely in the hands of the ‘log shaman’.
In both cases, post-job analysis done by a qualified cementing technical person can provide significant help to explain the look of the log. But, only if sufficient pre-job (caliper, rheology of fluids, hole condition, etc.) and job data (returns, mud logging, time log, etc.) is available. Otherwise, the analysis would become more of an educated guess than an objective fact-based report.
Now, what about what we see in the log: ‘cement coverage defects.’
1) Geometry-induced channel. It is the result of hole geometry and pipe centralization
2) Rheology-induced channel. It is the result of the improper design of cementing fluids. Or the desperate attempts by some cementing companies to add expensive and sometimes useless pre-flushes and/or spacers
If you only have a CBL/VDL, 1) and 2) would look exactly the same, but comparatively a rheology-induced channel would tend to have a lower amplitude reading. In both cases, the VDL would show formation arrivals.
3) Contaminated cement. Probably a combined effect of 1) and 2).
Contaminated cement is a form of unset cement caused by inter-mixing with mud or spacer. It would affect the CBL log with amplitude values ranging 10 to 30 mV, but lower than free-pipe value. In ultrasonic logs, it would appear as areas/spots with lower acoustic impedance values.
Two important things here:
- A cement log is only a snapshot. A cement log days after can be entirely different
- Cement will always set. If contamination does not allow setting, segregation takes place. Separation means mass transport (gravity/settling for particles and buoyancy/floating for lighter fluids).
Contaminated cement is common cause when no bottom plug is present, e.g., liner jobs, two-stage jobs, etc.
4) Unset cement. Long cement column, lead cement, abnormal lab test (wrong temperature or heating/pressure schedule) or insufficient WOC.
Unset cement looks similar to contaminated cement, but the origin is different. Unset cement is a design failure when the cement is placed in a lower than expected/assumed temperature or insufficient WOC. In ultrasonic logs, it would appear as near-mud acoustic impedance values and would primarily affect the definition of the TOC (pressure-based estimated TOC and cement log needs to be used jointly in these situations).
5) Formation-induced defect
Predominantly the effect of high permeable or abnormally pressurized formations. Basically:
a) The cement liquid phase filtrates into the permeable formation or;
b) Formation fluids invade the cement during the transition (liquid to solid) stage.
In the cement logs they will look different:
For a) considering liquid volume is lost, the resulting set cement is unable to fill the annular gap resulting in radial cracks and de-bonding. This is sometimes confused with micro-annulus, because of a similar signature in the CBL/VDL. In the ultrasonic log, because all gaps and cracks are filled with fluid, the image is erratic or patchy with intermittent highs and lows acoustic impedance; the average, however, will be closer to good cement value.
In b) the setting process is delayed showing a similar pattern to 3) and 4) in the CBL/VDL and ultrasonic logs.
A characteristic sign of formation-induced defect is a sudden change or step change in front of the permeable or high pore pressure formation. The cement log would quickly change from good to bad; it would stay bad in front of the problem zone and then quickly changing back to good in the next formation.
6) Pressure-induced defect
This would be the result of the excessive pressure differential a) while cementing (logging fluid lower density than displacing fluid) or b) after cementing (casing pressure test, confirmation of micro-annulus – this is a thoughtless, destructive test – formation integrity test). a) is the classical definition of micro-annulus. Theoretically, this micro-annulus would be either dry or wet.
Depending on the size of it, a dry gap could have a lesser effect on CBL (VDL would show formation arrivals in either case) than a wet gap (CBL can appear almost as free pipe). For ultrasonic logs, its the opposite, a dry gap would have a greater effect than a wet gap (most ultrasonic logs can see thru a wet micro-annulus as long as it remains less than 250 microns).
In b) the cement fails as a result of the stress condition derived from pressurizing the casing. The extension of the damage would be mainly influenced by the magnitude of the pressure, the density of the cement (compressive strength) and the compressive strength of the supporting formation. This is similar to 5.a) and as well mostly confused with micro-annulus. However, in general, the CBL will be less affected as it will take some time for the gap and cracks to be filled with liquid. The ultrasonic log, the image will be erratic or patchy as well but the average, in this case, would be closer to bad cement value.
7) Mud-induced defect
This is the case when the mud is unattended after the drilling is over. Typically flocculated mud and/or low-quality mud cake (permeable) for example can make any cementing fluid placement design obsolete. The problem here is that the excessive mud filtrate dynamically leaking causes a viscosity increase or gelling in the mud; this forces the cement to flow or channel circumferentially creating, once set, a big gap between cement and formation that quickly translate to immediate de-bonding of the cement attached to the casing.
This is a typical case when the well was in control and final circulating lift pressure is OK, but there is no cement in logs… However, cement is there, it is just that the gap is beyond what the logs can see. In a way, the cement is floating in between the casing and formation.
8) Lightweight cement
The only thing I can say about low-density cement (cenospheres, foamed cement, extended systems) is that it worsens the effect of all the previous defects (from 1 to 7); this makes the assessment of the logs particularly tricky and gives more importance to the post-job analysis and joint work with a log interpreter.
NOTE: I have worked with countless cement log experts in my career and in all cases, the interaction has been positive and a learning experience for the parties involved. The most important thing to understand is that a good cement evaluation log is not achieved by arguing, underestimating or ignoring the assessment of the log interpreter. A ‘good log’ is the result of a dynamic and strong drilling team.
Please let me know your comments and suggestions in the space below. Hope that this has helped you understand cement bond logs and variable density logs more!
Cheers
L. Diaz
one of factors that could misled your cement bond is If we run CBL too early before cement is set, we may see poor bonding. It may lead to unnecessary squeeze operation. If possible, estimate the wait on cement time before you log the CBL. Often, if you rerun a CBL log for the second time, you will see improvements in your cements.
Dear Mr.diaz
I decide to interpret cement quality charts in a gas field. I may use your guidance in this field. Please let me know if possible.
amir_naqibi87@yahoo.com
Yes, Amir. Any way I can help. Please let me know
I fully agree with you Mohamed, Thanks
CBL is always make some cementer can not sleep hehe
Impressive post Lenin Omar Diaz Torres! I’m kind of a crossbred between a cementer and a log whisperer so I always look at cement from the perspective of what you see and how you see it. There’s a few points where I’d love to start a discussion and exchange experience:
* You propose that the “microdebonded” cement seen on ultrasonic log is caused by radial cracks (5.a and 6. in your list). This is a new approach and an interesting one, but logs are in general fairly devoid of cracks, even when you look for them with the right logs and interpretation tools. Plus the exchange of fluids with the formation, and the effect on hydration and setting, is not simple. A subject for a new topic/challenge?
* 1. and 2. are really the same thing: pipe eccentering coupled with rheology causes channeling. This can be dynamic (such as the long CBL slope at ToC), or static.
* 3. (contamination) actually happens mostly on the way down the pipe, between fluids not separated by plugs. (next post)
(…)
* For CBL, wet or dry microannuli are almost the same: amplitude shoots up to ~80% of free pipe. Its ultrasonic tools that read very different: any thickness of “gas” is seen as pure gas (no info from cement & formation), whereas liquid-filled gap give an impedance that oscillates between solid and gas, and you get reflection from the formation as well. So I’d rather say that the CBL is less affected than the catastrophic failure of ultrasonic tools”, but still sees the effect of cement if you know what to look for.
* 7. (mud effects) is also interesting: first, I don’t really look for effects of mud cake since they appear across permeable formations, where cement is not needed. Second I have been looking for evidence of mud-related casing-cement microannulus, but found none (unlike the effect of casing coating and possibly rust). Bernard Piot has always said that PHPA mud debonds cement, unless properly removed, and I tend to believe him.
Matteo Loizzo it is excellen to have you onboard and ready to further discussions on these matters; this group in particular is always ready for great discussion and exchange.
In the last 9 years I had the opportunity to work in projects where there has been great debat on cement log results. In these projects achievin zonal isolation in intermediate sections was particularly important. It is amazing what you can learn from something like this and how difficult it can be to get great cement coverage sometimes. Now you can understand why cement in front of permeable formations was needed in these projects. On one occasion we where in case of “cement disappearance”. Before logging hopes were high due to a perfect cementing program deployment; pressure, returns all was great. Then a series of logs and processing were unable to find any cement. …. Ok, a portion was foamed cement (20%FQ) but what about the rest of the unfoamed cement … while everybody was looking at the N2-effect,
I was taking a close look at the mud … and I found that due to cement contamination the mud for the 2nd stage (focus of the debate) had its properties gone all over the place with the obvious impact on rheology and mud cake … the 1st thing, there was no cement disapperance, with logs we can only see a tiny portion (up to 0.25 mm) of a 12 1/4 x 9 5/8″ annulus … cement was in between 2 micro “macro” annulus .. My theory is that a leaking filter cake and subsequent dehydration of the mud (worse closer to the formation) caused a gap between formation and set cement that later debonded whatever set cement was in contact with the casing.
Lenin if you are saying you were looking at logs between two strings of casing that open up a whole new ballgame and one that will likely be debated forever with the limited technologies we now have for cement evaluation. However, two things I believe that must be taking into consideration (perhaps 3 of you want to consider eccentering of ultrasonic tools) but one that is always in play is that any reading from any log will be reacting to where its seeing very near the casing. That may be what you are referring to with your 0.25 mm. But there are effects from two strings of pipe that can include both constructive and destructive interference. Hence why there are the alternating bands on ultrasonic logs when cement thickness changes (“cat eye” effect). Rarely would the one string be well centralized inside the second string. Then there are the possibility of reflections returning from the second string of casing which can cause interference with the measurement depending on timing
Apparently I am too wordy haha. Depending on the timing in which the signals are evaluated. Back in the simple days of only 8 well centralized receivers on ultrasonic logs these were identified as secondary reflections. The long and the short is its really difficult to evaluate anything between two strings of casing with our technology. Finally, it has been realized that there is a significant micro annulus effect on ultrasonic cement logs and quite often that’s mis interpreted as “dry micro annulus” or something that indicates gas even when there is NO possibility of gas anywhere from the cement or the formation and you run them under enough pressure that so called “gas” will magically disappear.
That’s interesting, Lenin! Let’s see if I understand it correctly: you were cementing a second stage 9⅝”-12¼” across a permeable zone into the previous casing. Had 20% FQ lead and a neat tail that was supposed to be across the PZ. I would agree with Gary that logging foam in double string is not obvious, but let’s focus on the OH section:
* What was the tiny portion of cement you saw? 250 microns across the thickness vs. 33 mm? Or azimuthal coverage? How did you take a thickness reading? I’m fairly versed in measuring cement and microannulus thickness with ultrasonic tools, but it’s not trivial.
* I can understand filter cake at the formation side, but why should you have the same at the casing? Mud too thick to fully remove (mind that lab experiments say that this is possible)?
* Can you explain further why – indeed how – you want to seal across a permeable formation? Maybe distinct permeable zones in a thin-bed formation? Or the old folly of sealing across cement thickness?
Matteo Loizzo, Gary Batcheller the previous casing is quite shallow, only 500 mts vs 1500 mts of open hole. cap cement was visble in that area. the foamed lead was quite short in comparison with the tail of the 2nd stage. only cement from 1st stage was visble (same tail blend 1st and 2nd). the only difference between 1st and 2nd was mud conditions as mentioned. in this case, zonal isolation is based on good cement in between and/or across permeable zones. these jobs in 2stages are often not a problem. Job pressures, etc could not be better, only the cement logs (several done) showed no cement, no foamed lead no tail. I believe that when there is no bonding a gap between cement and formation, the cement will just debond from the casing (if any was there at all) almost as soon as the cement sets and on any work/pressure on the casing and remember there is stage tool and cement to remove before logging. The 0.25 mm, i belive is the max the ultrasonic would see. all logs cbl/vdl, ultrasonic always confirm nothing to see. i am a bit far from my initial post but it is indeed quite interesting and worthy to talk about, if you guys dont mind. thanks
That’s getting interesting, Lenin. The depth of investigation of both tools is far more that ¼ of a millimetre, and the ultrasonic imager see routinely out to the formation interface (galaxy patterns). In real life cement doesn’t “bond” to the casing but either contracts around it or is pushed by the formation, so you have a microannulus as soon as you push them apart. It’s fairly easy to spot a microannulus and get its thickness (if you have ultrasonic tools and a fluid-filled defect). From what you say I would suspect contamination on the way down: (possibly) no bottom plug, large casing, unconditioned mud. You should be able to see contamination on the job data. If cement is contaminated it can set VERY late (I’ve seen a month), but you should still see the presence of the slurry on the ultrasonic log with quantitative analysis.
Matteo has some very valid points and from what I am hearing about the operation I am rather suspicious of a micro annulus that you created and did NOT fully remove before logging. One thing for sure once cement is placed it doesn’t go away or disappear so when a log shows apparently “no cement” three possibilities have occurred. 1. A small to significant micro annulus 2. Green cement 3. The cement fell back (then typically you see some evidence of cement on VDL at least. As far a mud contaminated cement (sometimes considered green cement) as Matteo said it can take a while before curing enough (building enough compressive/tensile strength to show up on a log also. I have seen it take several months and the two logs looked like the difference between daylight and dark!
This is certainly a great discussion and I like to thank you all here and those who cantacted me privately. I guess what I have to say is that all your points are obviusly valid and all where already addressed vigorously during the investigation. Regarding the logs even additional processing of the raw data and playback creating scenarios indicated conclusively presence of any of both slurries, lead and tail, there were no losses and plenty of returns were seen, cap cement was pumped with no indication of falling back. it is clear the mud condition had a major role, so apart from what has been already said. How can this be explained? . however, I stand with my theory, which I can not conclusively prove ? but on paper looks still ok to me. This never happen again as controls are in place to ensure proper mud for 2nd stage. So, by default the role of the mud has been confirmed more or less. Any more thoughts?.
Just a quick remark, now that I am reading again my previous comment …. when I said “plenty of returns” was just to highlight that we had no losses, but I actually should have said “as expected” .. to avoid implying channeling … All data, as I mentioned, pointed to a placement as expected with good pressure matching and returns.
Lenin – you mentioned the mud properties were shot. Two clarification questions: 1) What kind of mud was in the hole, prior to the cement job. Water Based? Density? 2) How much Mud / cement spacer was pumped ahead of the cement?
It’s a pity the urgency and opportunity are gone for such an interesting case, Lenin: with logs and job data you can do a pretty good quantitative analysis and figure out exactly what happened. Just give me a shout whenever you come across a similar case and we’ll get some answers.
As Gary and Matteo both stated you can have very long times for cement to develop compressive strength, particularly in areas of high temperature gradients where the use of “retarder intensifiers” are used. If the cement is tested at a BHCT that is higher than the BHST where you are logging expect a very long time for any significant compressive strength development. Days, weeks and as Gary and Matteo said even a month.
Maybe I can add some help in this subject. Firstly, we need to understand that the cement logs (all of these) only measures acoustic impedance, which is the resistance that a system/solid present to the acoustic wave flow, resulting of an acoustic pressure applied to the system. Depend of the resistance of the system, the impedance will be different. Usually, the logger guys and most of the interpreter use a table with different range of impedance and what type of system correspond to such impedance, instead of to evaluate previous and operation data. Here is the first reason why Light Weight Cement looks like in the logs as mud or Foam cement like free pipe.
To a correct interpretation required some previous steps:
1-Verify the centralization of the tools across the well section
2-Verify the invariability of the transit time cross the well section
3-Comparing each section of log with Caliper log and identify the washout sections, as well as, high permeable zones.
Once these three steps are done, we can proceed with the interpretation of the log, which as Lenin say, there are so many alternatives and cases that usually require for a serious analysis, the correct previous (well conditioning process) and operation data, including what type of operation the rig has carried out after the cement job and just before to run the log. The last point has a significant important.
For all the cases with doubt or when complex cement is used, I suggest the consideration of SPE 97186 and its application.