Well integrity for grown-ups


Published 15 January 2020

"The gas storage industry is far more concerned and thorough than oil & gas when dealing with annular leaks", so remarked somebody familiar with both worlds during a chat we were having a couple of weeks ago.



Matteo Loizzo
Well Integrity Consultant, Germany

Indeed, the storage industry tends to ask the right questions, such as: does the annular leak we see account for all of the gas leaving the storage? Why did the cement sheath fail? Is there anything we could - nay, should - do to fix it?

On the other hand, we (alas, I've been in O&G for longer than I care to admit) are rather more cavalier: more than once I faced an abandonment plan where a number of annuli had been "live" since the well was drilled. Nasty minds could argue that API Spec 17D insistence that subsea wellheads not be pierced to probe annular pressure is a case of "don't ask, don't tell", to spare an important (and expensive) class of wells the time-wasting and regulator-alerting scrutiny of sustained casing pressure.

In general, decommissioning is the worst time to worry about annular leaks: production (and revenue) is over, and you have to run a tight ship to deliver on schedule and on a budget a plugged well. You cannot spend days trying to measure and better understand the situation, and the remediation techniques that worked so well in the vendors' case studies tend to have mediocre success rates when the astonishing variability of reality is thrown at them. And yet, somehow, leaks have to disappear before you can relinquish a site...

Granted, truth is seldom concentrated: the storage industry may worry more because they are more likely to mess it up; or because they lack the cool head of the O&G crowd, and they haven't learned yet to focus on things that are truly important.

Is gas storage riskier than production?

On the face of it gas storage could be more leak-prone: there is no depletion, and the frequent, extreme pressure cycling could take a toll on integrity. In reality, it seems the industry is simply better at investigating leaks and gathering statistics than O&G: for instance, the 2009 study by D.J. Evans is unparalleled in its breath and depth, even more so if we consider that the 600,000 well-year of operations or so clocked by gas storage are dwarfed by the experience accumulated in hydrocarbon extraction (about two orders of magnitude more).


One accident that was extensively investigated was the leak of natural gas from the Yaggy salt cavern storage site in January 2001: this caused explosions, fire and two fatalities in Hutchinson (Kansas, USA).


The escaped gas, which was rather precisely estimated at 2,716±437 tons, had been collecting in a permeable zone before a sudden release with a large instantaneous leak rate. If we consider a typical well lifespan of 25 years, the "Yaggy limit" would be around 110 tons per year (t/y): if an undetected leak of that amount were to accumulate in a shallow aquifer before bursting out, the consequences could be catastrophic.

Fire in Hutchinson after the Yaggy leak. Source: presentation from Kansas Geological Survey-University of Kansas

You could argue that this couldn't happen to us, but in fact it does, and with a certain regularity.


Gas leaks from abandoned wells in Colorado have caused explosions in or near homes in 2005, 2007 and again this past April.


Unfortunately the latest event killed two men in Firestone. Yes, the wells involved were old and badly plugged, and there are reasons to suspect Colorado may be more prone to leaks than other geological provinces. Then again, I agree with a couple of researchers when they say that storing gas in salt cavern is inherently safer than poking around clastic reservoirs.

Or is the gas storage industry hopelessly paranoid?

Given that storers are unlikely to be more error-prone than us producers, there's always the possibility that they worry about trivialities. After all, as Nietzsche said "that which does not kill us makes us stronger": we can live with a little sustained casing pressure, there's no need to throw money at a dripping tap.

But let's take a step back: the action level for an annular leak, i.e. the rate above which you have to do something now, varies according to your jurisdiction. There is nonetheless a tendency to use the value proposed by API RP 14B: for gas, this is 15 ft3/min, i.e. around 160 t/y. The fact that the leak rate originally applied to valves hasn't prevented its adoption by a number of standards, among which NORSOK D-010.

If our annular leak rate is below 160 t/y then we tend to use one of two strategies:

  • Either we shut the annulus in and let the pressure build-up to its maximum value.
  • Or, if there are concerns about the maximum allowable pressure, we bleed off periodically (or even continuously).

In principle those are perfectly good approaches, but each has its potential problems:

  • If we shut in, the increasing pressure in the leakage pathway may (will) eventually cause a secondary pathway to open up. Chances are this will be a microannulus at the outer interface of the well envelope, which will cause gas to migrate to a shallow permeable horizon or to the surface.
  • If we flow formation fluids to the annulus top and to surface, we could potentially cause corrosion and an unwelcome failure of outer, weaker, casing strings.

And we should also bear in mind that if none of the two options is safe, we still have the unwelcome alternative of going back and fixing the leak.

All risk management is local

A deeper issue is that the API threshold is not universal: it is true that the logic behind it (a fire that can be put out by a guy with a fire extinguisher) is solid, and that 100 t/y is close to the detection limit, offshore or on land; but it is uncomfortably close to the "Yaggy limit", and could lead to potentially dangerous accumulations. And the offshore crowd shouldn't brush leaks off nonchalantly ("gas accumulating and blowing up a camping ground? Not likely..."): I have come across shallow gas that just appeared after the first few wells were drilled, and looked suspiciously like deeper stuff that accidentally found its way upwards.

So what should we do? (malicious readers may expect an answer like "ask a consultant with a sufficiently deep bag of tricks", but that would be shameless plugging)

First, know your well: fluid sources, completion strategies, geological barriers and sink formations vary significantly. There is no such thing as universal barriers: even the notion in NORSOK D-010 that 30 meter-long sections of "good" cement in the same string are independent barrier elements (i.e. if one fails it doesn't imply the other will as well) is wishful thinking. Indeed, a debonding event that affects the lower element will also cause the upper element to fail.


Second, do proper risk management: don't take action levels and easy solutions for granted, but check that nothing is happening out of the corner of your eye. Is part of the leak flowing outside of the well envelope? If any fluid migrates, where would it go? What is the maximum allowable pressure for geological barriers? Am I letting anything corrosive into the annulus?


Third, use all information at your disposal to build a picture of your well: every piece of data is a symptom that will support a successful diagnosis. Only when you know what is ailing you, a proper cure can be found - or at least that's my point of view: there are plenty of approaches other than rational analysis. So use all logs, but also buildups and venting tests; how leaks react to changes in reservoir or well pressure always offers precious clues. And of course if you want to understand fluid flow you should characterize all liquids and gases involved.


Finally, and this is the reason why I'm advocating an adult approach to well integrity, don't sweep leaks under the carpet. All leaks should be understood, without necessarily wasting money doing so. You can actually tolerate a lot of them, and some will naturally disappear because of depletion and geological barriers gripping your well. However, the best way to manage leaks, including what data to collect and when (if) to intervene depends on a clear-headed evaluation of risks and benefits, and on having a whole production - or injection - life ahead of you to get some return on the investment. Most leak are tame kitten but some are hungry wolverines: you'd better know how to tell them apart before you bring one home with you.