Well Integrity Technical Section

 View Only
Back to discussions
Expand all | Collapse all

Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

  • 1.  Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-18-2023 04:02 AM

    Good morning community,

    We are about to start acquiring corrosion / dimensional logs (SL, memory) in some old wells with a 9-5/8" 47# production casing (through tubing, memory logs, 4-1/2"x3-1/2" tubing) with a view to assess internal yield in the upper sections of some wells and MinAP in others.

    Once the log data becomes available we hope to have in hand:

    • Caliper data (56 finger) in some sections, hopefully allowing us to assess localized corrosion (capturing pitting corrosion if we see any)
    • Wall thickness (average, magnetic) - this will hopefully give us an average around the circumference, and we can then re-apply some calcs (the way I see it it will mostly be comparing the original WT compared to the one as-logged...)
    What we actually do with that info is a bit more complicated...

    • If we see evidence of pitting corrosion we'll have not other option but to blindly apply a "healthy" derating on the internal yield pressure for the casing (ex. SF of 3 or more...) - or acquire more logs specifically to check those pits and assess the minimum remaining WT... armed with that we should be able to revert to models / experiments to anchor the new internal yield in "some" science / This is the short discussion...
    • Collapse can then also be "estimated" using the average WT derived from logs... This is the long discussion!

    Collapse in various sections of the well:

    The upper sections of the 9-5/8" production casing are uncemented (say, surface to 1,500mTVD) - of course they are also shallower and won't be subjected to the highest stresses in case of casing evacuation... I'm not so worried about those: applying some rules of thumbs for WT loss and perhaps some D/t change corrections we should still be able to get something for collapse, and considering stresses will be lower we should still be able to determine a collapse rating / MinAP (and the casing will likely take it).

    The deeper sections (say 1,500mTVD to 2,500mTVD) were cemented during well construction and logged - so the TOC is confirmed and quality of the primary cementation, although variable along the borehole, is sufficient to guarantee hydraulic isolation in most section, and more critically for this discussion, somewhat mechanically supported along the interval.
    This is the section I'm curious about... once we know the average WT (this may have been reduced due to corrosion), and considering the casing can be reasonably expected to be mechanically supported by the primary cementation (ie. less likely to deform easily in its minor diameter axis...) is there industry data (or a position) on how to assess a collapse rating for this? (of course this is also this deeper area is also the one which will see the highest stresses in case of full evacuation).
    I understand collapse (largely governed by dimensional tolerances, D/t and ovality) is a lot more tricky to evaluate than internal yield, and that the majority of experiments may have been conducted on unsupported sections of pipe... but there may have been experiments also conducted on cemented sections... abacus, rules of thumb... etc.

    Hence the call out to the community.

    @Matteo Loizzo / @Philip Wodka / @Simon Sparke - any recommendations / advice? 

    Thanks. ​Best rgds.​​​​


  • 2.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-18-2023 06:26 AM
    Hi Pierre-Edouard,

    Casing collapse is a complex topic. API 5C3 only applies to casing with a fluid in the annulus.
    There is debate on whether a 200 micrometer microannulus could be regarded as a fluid with uniform pressure: even though collapse may happen through a heart-shaped mode rather than the pinched mode (picture you stepping on a can, sideways), collapse still involves large deformations, and these are hindered by cement.

    So I guess your load case involves either large fluid pocket (mud channels) or creeping formation stresses.
    Mud channels are a bane for high-enthalpy geothermal wells, since the expanding or even boiling fluid can collapse casing. The exact failure mode is still debated, but the solution has always been to fully displace all fluids out of the annulus. Completely.

    As for creeping formations, there have been some studies of collapse. Analytical models are interesting, and reveal the key variables involved in pipe failure. Finite element models are very hard to set up: you need to consider formation creep, large pipe deformations, and slip at the interface. If you want to determine the actual load limit, you also need steel plasticity (although for us, the appearance of elastic instability is enough to call a stop).
    Stress exerted by shales is anisotropic, so you always get casing ovality. But I haven't seen collapse so far.
    Salt stress is much higher, but isotropic. You can mix logs and models to assess the state of stress and your safety margin, though you may need very advanced tools and interpretation since you need full annulus geometry and cement properties.
    The worst I have seen is salt (halite and polyhalite) stress in a free pipe or large mud channel section: in this case mud destabilizes the stress field, introduces very high anisotropy and you get ovality of more than 1 cm. This can give you access problems and even casing failure.

    There is a third category of casing deformation and "collapse" i.e., what we refer to as fracture-driven interactions. The exact mechanics is still disputed, and it may have something to do with shear slippage at the formation. Not much collapse modeling required.

    So, after this long and boring explanation, a recommendation: if you're dealing with collapse but your annulus is not totally free, then be very pragmatic, use logs to monitor deformation, and go for trial-and-error.
    Models could help you, but you need both advanced measurements and modeling to drive action. Unless you're ready to mobilize both, then be pragmatic.

    Hope it helps,

    ------------------------------
    Matteo Loizzo
    Well integrity consultant
    matteo.loizzo@mac.com
    Berlin, Germany
    ------------------------------



  • 3.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-19-2023 11:49 AM
    You might find a read of SPE Paper 199577 useful

    ------------------------------
    Olli Coker
    Diamond C Enterprises, Manager
    Former Altus Well Experts, and former ConocoPhillips
    OlliC3@gmail.com
    ------------------------------

    Original Message


  • 4.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-19-2023 01:03 PM
    Thanks Olli,

    A very interesting paper, describing a process similar to what I've been through.
    Except that I was somehow cheating: I had excellent logs that allowed me to map the annulus to a very fine detail, and advanced mechanical models. Plus a humbling, independent review that taught me a thing or two.

    A statement I fully support is that ovality is proportional to stress, and that it can be used - ever so carefully - to gauge the distance from the safe operating envelope.

    There are however a few assumptions in the paper I don't necessarily agree with:
    • Salt (halite and polyhalite, other minerals like carnallite are weirder) has an isotropic stress field - I guess what you call uniform stress. So the horizontal stress is more or less the same in all directions, to within 1 MPa or so. You get measurable anisotropy close to the top of salt, where there is an effect of the 3D salt body shape, but this is a local effect.
      • If there is no cement, salt can be very anisotropic, locally. The reason seems to be that trapped mud fractures the salt borehole and changes radically the local state of stress. When you use Ph and PH, you probably refer to a free pipe section, not a cemented one where Ph~=PH.
      • Ovality in cemented casing through salt is caused by the eccentered pipe. But the ovality values are low and stable over 50+ years. As long as the casing is cemented, of course: large mud pockets and channels can revert to the previous point and cause localized ovality.
    • The FEA models you describe use constant pressure to mimic creep stress. This is very conservative, yet incorrect. To see the difference from slowly creeping solids, just imagine the casing starting to deform: salt must follow it to keep exerting stress, unlike a fluid that can always apply constant pressure regardless of the pipe shape. If you want to study creep-related collapse you have to go the fully Monty: creeping solid, large deformations, and slip boundary conditions.

    Best regards,

    ------------------------------
    Matteo Loizzo
    Well integrity consultant
    matteo.loizzo@mac.com
    Berlin, Germany
    ------------------------------

    Original Message


  • 5.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-20-2023 05:48 AM

    Thanks Olli,

    Good insights... I noted in your paper Fig. 6 (pages 11 and 12) and a collapse resistance significantly higher (5x) for a concentric, cemented string... However I failed to really understand what exactly had beed tested or modelled... The comparison table is between a standard 7" 32# HCL80 pipe and then... a concentric arrangement called 7x5... what would that be? Ideally a model we could immediately refer to / use would be that of a cemented 7" inside a conventional 8-1/2" OH... either centralized (ideal case) or rather non-centralized (real-life). In this case there we could expect a differential effect of cement "arching" over the casing (zero effect on the low side, full effect on the high side). If you know of any papers or research covering this, I'd be happy to hear about it.

    Could you pls explain how Fig. 6 was obtained? In any case it would be great to have solid experimental testing showing using a multiplier for the cemented sections is legitimate (intuitively, I think it is)

    @Matteo Loizzo I also found some insight in SPE paper 81820 (page 17, just before conclusions)​​. This is quoting that:

    "Analyses were also performed with the casing annulus cemented. The same configurations for hole geometry were used (e.g., ellipticity, casing sizes, etc), and the essential effect of the cement is to change a potentially nonuniform loading condition into a uniform one. With the cement in place from the start, there is little possibility for unequal deformation to occur. Very small or negligible casing ovalization occurs (<0.3 in. for the 16-in. casing considered)."

    So:

    1. Ovalization and WT (and loss of WT due to corrosion) seem likely to be driving factors for the uncemented sections (logs can help us find out more about both... but ovalization cannot easily be obtained via mag tools... US tools might be better, not sure we can do that TT at the moment)
    2. Arching due to cement and the potential positive effect on collapse seem to become the driving factor on collapse for our deeper, cemented sections... what multiplier we can use is the question (let's see if Olli's experiment can shed some light)... but it seems reasonable that we can expect a positive effect
    Rgds.
    Original Message


  • 6.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-20-2023 02:55 PM
    My heritage company had some major problems with collapsing casing in the Gulf of Suez in the late 1970's.  One of the smartest guys I know, Phil Patillo, did the analysis and developed a solution to the moving salt damaging wells.  

    A few years later, we had a "boom" going in the US overthrust and experienced similar problems with flowing salt and used the same techniques to solve the problems.  

    We assumed we could not/were not getting good cement through the "washed out" sale zone (washed out during the drilling process but would flow through and damage the well when things were static).  The solution that work in Egypt then in Wyoming and Utah was to "double case" the flowing salt zones with "good cement' in the casing annulus.  This is the only mechanical means to overcome the flowing collapse forces of the flowing salt.  It worked, if we got the 9-7/8 and 7" casing annulus cemented!

    Doug White
    Sugar Land
    Original Message


  • 7.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-20-2023 06:27 PM
    Yup.  Learned a lot from Phil.  One of the oilfield's greats.
    Olli

    ------------------------------
    Olli Coker
    Diamond C Enterprises
    Manager
    OlliC3@gmail.com
    ------------------------------

    Original Message


  • 8.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-21-2023 03:40 AM
    Edited by Pierre-edouard Vincent 01-21-2023 08:11 AM

    Thanks guys,

    @Douglas White - the case you're referring to is similar to what Olli explained: the pbm to solve was collapse resistance against creeping formation stress, and a good fix found was to go pipe in pipe & cement... In the paper quoted by Olli the higher collapse of the 5" casing/tubing has to be considered... a simple 5" 18# L80 has a published collapse of 10,490psi... ie. already higher that the initial collapse of the original 7" #32 80ksi casing (8,600psi). While it points at the potential for PIP to be even stronger (in this case, considering 2 casings and cement in-between would definitely affect the D/t ratio, so recalculating a new theoretical collapse would be useful, to then compare this to experimental results...).
    ​
    I have tried to get in touch with Steve Willson at Apache, about the other paper I found (pending).

    Hopefully there is some litterature purely looking at the effect of a cement sheath on the collapse of a cement pipe - ideally ex-centered (lying on a low-side). If we had to resort to FEA to model this, we can also do that... (8-1/2" OH, 7" casing on the low side, cemented)

    @Matteo Loizzo / @Oliver Coker - any suggestions on an engineering firm which would be familiar with this kind of work? (not to start from scratch)
    ​

    ​​​Cheers.
    Original Message


  • 9.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-21-2023 08:20 AM
    Hi Pierre-Edouard,

    Creeping formations are not a uniform bunch: shale and marl will not collapse your casing, and neither will carnallite.
    Traditionally, there have been three approaches to cementing halite:
    • What Olli talked about, the pipe-in-pipe solution, has been used by GDR engineers when developing the Altmark field. A reason was, hoping not to irk anybody, poor grasp of well construction processes and the need for a robust solution (i.e., one that works without you necessarily understanding why).
    • An alternative has been using mast casing. One-inch thick steel monsters have been deployed in the Mittelplate oilfield, west of the border and further north along the Zechstein thick salt province.
    • The third approach is to use reasonably normal casing (well, P110 grade or so) and make sure your cementing is spotless. We've recently had a couple of dozens logs of beautiful liners through kilometers of creeping salts (halite and polyhalite) with barely a nick. Same for bigger production casing above. And several intermediate casings though shallower salt with severe deformation (on some, but not all of them) wherever cement was missing. I mentioned the theory of trapped fluids within salt causing stress anisotropy and pipe ovality. Whereas in theory you could simulate this, it takes somebody with brains, experience, and chutzpah. Probably academia rather than engineering firms that will give you the solution of the day. As wrong as yesterday's.
    So here goes my suggestion.
    No prize for guessing the approach I recommend (number 3), but it only works if you can cement your casing properly. No fancy slurry systems, no thoughts & prayers, just superior design and faultless placement.
    So what if you have threats, like lost circulation zones below the salt, or well deviation above 60 deg? Then pipe-in-pipe (but you still have to cement the annulus properly).
    I would not waste money on bad models unless you have a pressing need to show a particular pipe weight works (otherwise you'd have to trash your development concept).
    Mind that there are provinces that have learned to deal with creeping salts: offshore England & Netherlands in the 1980-90, a time of revolutionary progress in cementing, and Brazil in the 2010. If they haven't pitched in in this debate, it is probably because they don't think it's a real problem.

    Best regards,

    ------------------------------
    Matteo Loizzo
    Well integrity consultant
    matteo.loizzo@mac.com
    Berlin, Germany
    ------------------------------

    Original Message


  • 10.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-21-2023 11:20 AM
    Hi @Matteo Loizzo,
    What we are dealing with here (hence the sense of my initial query) are old wells, some of them way over 20yrs so casing design has to be lived with!
    We're converting them to ESP's (from GL) but this will create some new issues and the potential for longer - sustained - periods where the A-annulus pressure will be low... in GL wells the issue was always there (full evacuation) but with vented ESP's the annulus pressure will be low - and always low.
    The question arose re: what we'd do with the info once with acquired logs... particularly once we get an average WT. If this hasn't changed much in >20yrs of service life, easy. If it has been affected by overall corrosion (likely) and we get uncomfortably close to a re-calculated collapse, factoring in the effect of a cement sheath would be nice. I can see how that could alter/increase collapse resistance and counteract a generalized WT loss. That's what I meant with the initial query... salt creep wouldn't necessarily be immediately relevant unless FEA or testing was done assuming a cemented casing (?). In our case lying on the low side.
    Re: FEA I could find engineering companies to do this... but then it would likely be so much easier if they had done previous similar work we could build on.
    I'll touch base with SPE Brazil for networking... It's likely that some of that modelling work will have been done there.
    Cheers.

    ​​
    Original Message


  • 11.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-22-2023 04:35 AM
    Good discussion.
    re: FEA / modelling:
    Blade Energy in Houston modelled concentric cemented liners for an HPHT project I worked on in the North Sea 10 years ago where we were worried about reservoir compaction parting the tubulars on our very expensive wells.
    They built a reservoir model for predicting compaction, and then followed up with the best solution for mitigating / living with it.
    I appreciate it's not completely analogous to your scenario, but the fundamentals are somewhat relevant and they may have helped others in this space.
    Perhaps it's worth a shot.
    Good luck.
    Callum


    Original Message


  • 12.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-22-2023 04:48 AM
    Hi Pierre-Edouard,

    Let's see if we can have some practical recommendations:
    First, collapse is not the only possible mode of failure: shear failure of the casing when von Mises stresses exceed yield stress may be very close to the actual collapse value - if not lower.
    As pointed out in @Oliver Coker's paper, ovality is the main indicator: increases in ovality from the initial state (that can be extracted from the full log) is proportional to the maximum von Mises stress. Slope and intercept of the line can be extracted from simulations that are easier than full-blown collapse one. That would give you a maximum allowable ovality that you can tolerate.
    If you match the current ovality with models, then you can simulate a drop in casing pressure to whatever your ESP will bring about and check if you'll be safe.
    Key is measure ovality, match ovality, predict ovality.

    Second is the power of experiments: if may be cheaper to measure a baseline ovality, lower pressure in the casing, wait a week or so and then relog it. We can come up with a limit to ovality, between Olli's work, modeling we've done, and a few other sources/ Extrapolate linearly from the measurements to when you'll install the ESP and you'll know if it's a no-go.

    Third is corrosion: the majority of wells have exactly zero metal loss. Conditions are actually fairly benign to steel unless you do something untoward. GL is not the worst environment, provided you didn't inject wet, CO2-rich natural gas. Of course, if your tubing outer face is attacked, then so will your casing wall. There's no alternative to logging, but I wouldn't be too pessimistic.

    Fourth, maybe you don't need to model if it doesn't help your decision. It happens often that after 6 months' work and endless debate, your trust in modeling results will not be enough to save you from risk prevention measures anyway. Relining is a possibility, but you could also install an intermediate tubing isolated with a packer. Or even run your ESP with a packer: that will allow you to keep casing pressure to values you know are safe.

    If in the end you want to go and simulate, I could still help out by telling you whether the scope of work makes sense. Free of charge, of course.

    Hope it helps,

    ------------------------------
    Matteo Loizzo
    Well integrity consultant
    matteo.loizzo@mac.com
    Berlin, Germany
    ------------------------------

    Original Message


  • 13.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-21-2023 08:39 AM
    It has been a few years and I did leave out one other example from my Overthurst days:

    The shallow salt (Whitney Canyon field for example) was around 6000 feet (actually, still above MSL) through a shallow salt zone and we would set 13-5/8" casing.  Drill to the ICP and set 9-5/8 casing.  We needed good cement across the 13-5/8 shoe to protect that casing from salt collapse.  

    The example I gave earlier was like the salt we drilled in the Anschutz Ranch Field.  The salt was located at maybe somewhere in the 11-12,000 foot depth and set 9-7/8 inch casing (8.5 inch drift) and then drill to TD and the 7" liner would lap the salt zone above the ICP shoe and cement that area.  

    After some "learning experiences" on how to keep the hole open and in cementing practices, it became pretty routine.  

    Doug White
    Sugar Land
    Original Message


  • 14.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-25-2023 08:51 AM
    Edited by Gang Tao 01-25-2023 08:57 AM

    Hi everyone,

     

    This is a very interesting topic and I can see a lot of valuable discussion here.  I have attempted to offer some relevant comments and additional information on the various questions and challenges as listed below.

     

    Casing collapse with cement support:

    Shell has conducted both FEA and full-scale collapse testing of cemented casing as published in SPE 173069. Their case study covers the pipe-in-pipe case.

     

    C-FER has also conducted advanced FEA and full-scale collapse testing of cemented casing (single casing). The test results and FEA had fairly good agreement in terms of the enhancement of collapse strength. However, since the project was proprietary, we unfortunately cannot publish this work.  You can check out my LinkedIn profile (https://www.linkedin.com/in/gang-tao-a0b2c6/), where the picture at the top shows the collapsed casing sample within the cement annulus (on the left side). In addition, we have completed many advanced FEA projects in the past to assess casing integrity under severe load conditions including casing interaction with salt (using a comprehensive salt creep model), substantial compaction (Ekofisk, Valhall, GOM) and/or formation shear (MFHW, CSS/SAGD wells), ultra high temperatures (thermal, geothermal wells), high well curvature, etc.  For example, see SPE Paper 199570-PA (SPE Drill & Compl 36 (02)) which presents analyses which show further details on the effects of high temperatures and loading rates on casing collapse capacity as well as on ductile rupture strength and premium connection performance.  A strain-based design/evaluation criterion is typically employed in such large displacement and non-linear material response scenarios.

     

    For collapse assessments involving casing corrosion, a key consideration is whether the corrosion damage has occurred on the ID or OD of the casing (or both), as this has a significant impact on the collapse strength reduction.  General corrosion on casing OD essentially increases the cement-casing micro-annulus which weakens the radial support.

     

    Uncertainty of downhole logging tools and implementation of Wall Thickness (WT) log in engineering assessment

    C-FER has conducted a few extensive full-scale lab testing programs to evaluate various downhole corrosion logging tools, including MFC, MFL, UT, EM.  Some recent results have been published on the US DOT PHSMA's website and the report is open for public download (https://primis.phmsa.dot.gov/matrix/PrjHome.rdm?prj=747). You may find some relevant information in the report to help plan for any upcoming well logging and casing assessment activities.

     

    Through-tubing casing logging tools (all based on EM technology) tend to be of low resolution and there is a very limited understanding of the actual performance (accuracy) of these tools.  We are currently sponsored by PHSMA and PRCI to conduct a further full-scale lab test evaluation of several through-tubing tools (check this website https://primis.phmsa.dot.gov/matrix/PrjHome.rdm?prj=943). Since these tools can only provide an average wall loss, there is still a question of how to use this information for completing casing integrity assessments for existing wells.  The impact that wall-loss occurrence over a small portion of the pipe circumference vs. over the full circumference has on the casing integrity is significantly different.  The lack of detail on the degree and nature of the corrosion losses will typically lead to large uncertainties when attempting to estimate failure potential under different loading conditions.  Further work/technology development by the industry is needed to address this issue. 

     

    If high resolution logs (MFL or UT) are available for the wells which led to this discussion, the remaining burst strength can be estimated using existing analytical models developed by the pipeline industry (check our PHMSA report mentioned above for some examples). The collapse strength can be assessed by FEA by considering some key factors (wall loss profile, casing elastic-plastic material response, cement support, casing-cement contact interaction, micro-annulus size, axial load, etc.). If only data from an EM log is available which shows uniform wall loss over discrete segments, sensitivity analyses can be performed based on various assumptions, and the application of a risk assessment methodology is probably warranted. Caliper logs (MFC) can only provide the casing ID profile, and therefore such logs may not be reliable for estimating total wall loss (in particular, interpretations of such log data which assume perfectly round pipe and nominal dimensions as the baseline can lead to significant errors in wall loss estimation). The number of fingers in caliper tool as well as the finger geometry represent key factors which can limit the resolution of these tools, especially in cases where the corrosion creates isolated and/or small sharp pits.  C-FER previously developed a software tool which assists in the interpretation and detailed evaluation of caliper log data, especially for situations where substantial casing deformations have occurred. 

     

    Please feel free to contact me to further discuss any of the information provided above.



    ------------------------------
    Gang Tao, PhD, PEng
    Senior Engineering Advisor
    C-FER Technologies (1999) Inc.
    g.tao@cfertech.com




  • 15.  RE: Applying knowledge and log data to assess casing internal yield and collapse in 1- free casing / 2- cemented casing

    Posted 01-26-2023 06:21 AM
    @Gang Tao - many thanks for your contribution and excellent insight/refs. I​'ll get in touch shortly.
    Kind rgds
    Original Message