ENERGY

Why log cement?

 

Published 15 January 2020

We've all had a blood test: a sheet with a bunch of abbreviations and numbers, and helpful ranges that tell us whether we should worry or not. And what if a value sticks out, say the bad cholesterol?

Contributor

Matteo_Loizzo_400x400

Matteo Loizzo
Well Integrity Consultant, Germany

Then I, with the advanced medical knowledge of a 5th grader, would go see a specialist who can help answer a few key questions:

  • Is it really bad?
  • Will it go away on its own (read: do I really need to forgo deep-fried Mars bars?)
  • Anything I must do to fix it now, e.g. do more tests, take drugs, get operated?
  • Any changes to my lifestyle that can prevent it happening again, say, a change of diet or even exercise?

 

But most of all I would look at the wise person and ask: "Hey doc, what exactly is it that I have?". I want to know the diagnosis, the reason why something's wrong, which is the necessary step to figure out what to do next.

That's what logs are for

With cement logs we're not that far off: our goal is to understand what is happening outside of the casing so that we can take the right decisions. As Didier Rouillac put it in his Cement Evaluation Logging Handbook "...the primary use of [cement logs] is not to decide when and where [to squeeze,] but to try and improve the primary cement job on the next well of the same kind".

 

In fact, logs serve a dual purpose:

  • They help us reveal the presence and geometry of defects in the cement sheath. These can be rigid, pipe-like voids left over from the construction process (mud channels and chimneys created by fluid migration during cement setting); or can be hair-width elastic gaps that can appear at any time during the life of the well (microannuli, and possibly radial cracks). Or we could have the trivial defect, when cement doesn't cover the zone of interest - after all verifying the Top of Cement is still a key task in evaluation.
  • They also provide measures to quantify the properties of cement and the shape of the borehole. On top of it, ultrasonic imaging tools are an elegant an inexpensive way of obtaining formation creep stress and thus characterise geological barriers (for more info on this passionating topic, you can take a look at a LinkedIn article, an SPE webinar or a conference paper). Technology has flooded us with data yearning to be distilled into precious information on steel, cement and rock - it's not unusual to get 1 GB (gigabyte) per kilometer. That is real Big Data, and a few squiggly lines are not enough to make sense of it: you need sophisticated statistics (yes, including machine learning) and models to get value out of your logs.
BATD421_article_pic
BATD421_article_grafic

In fact, logs serve a dual purpose:

  • They help us reveal the presence and geometry of defects in the cement sheath. These can be rigid, pipe-like voids left over from the construction process (mud channels and chimneys created by fluid migration during cement setting); or can be hair-width elastic gaps that can appear at any time during the life of the well (microannuli, and possibly radial cracks). Or we could have the trivial defect, when cement doesn't cover the zone of interest - after all verifying the Top of Cement is still a key task in evaluation.
  • They also provide measures to quantify the properties of cement and the shape of the borehole. On top of it, ultrasonic imaging tools are an elegant an inexpensive way of obtaining formation creep stress and thus characterise geological barriers (for more info on this passionating topic, you can take a look at a LinkedIn article, an SPE webinar or a conference paper). Technology has flooded us with data yearning to be distilled into precious information on steel, cement and rock - it's not unusual to get 1 GB (gigabyte) per kilometer. That is real Big Data, and a few squiggly lines are not enough to make sense of it: you need sophisticated statistics (yes, including machine learning) and models to get value out of your logs.

Knowing the mechanical properties and geometry of the cement sheath is essential if we want to explain the defects we see, and even more important if we want to predict the future evolution of our well and its integrity. Have you ever been staring at a log a few weeks to many years old and wonder if everything is still the same out there? (Hint: no, it isn't; elastic defects are driven by pressure, and they come and go as operating conditions change).

Which brings us to the sad truth of cement evaluation logs: alone, they are close to worthless. You need to put together a complete picture of the well, and especially of the geology on the outer side of the annulus: rocks are driving most leaks and (again, back to geological barriers) help you control many of them. Rehashing the medical analogy, have you ever met a doctor that tells you what's wrong with your blood test without a lengthy and puzzling form about your ancestors and ghoulish diseases? And without taking a long probing look at you?

The well-tempered log

If a fundamental role of logs is giving us precise and accurate measurements about the casing and the world beyond, in the same way that open hole logs are used by geophysicists, then it is clear that we need valid measures (i.e. they have to pass Quality Control) and that we must calibrate them exactly. Even if you're not the quantitative kind of guy, and care mostly about shades of black on a map or low-high amplitudes on a CBL, comparing logs across wells or between repeat runs demands repeatable calibration. How many debates have raged over whether a microannulus is gas-filled (0 MRayl) or water-filled (1.5 MRayl)?

Some tools are calibrated in the shop, such as the Isolation Scanner, while some have a bootstrapping method that should in theory deliver reliable parameters in the well. Pulse-echo tools are famously tricky, since they depend on a seemingly innocent physical property of the logging fluid, its acoustic impedance (ZMUD). Trouble is, ZMUD is not only a measurable quantity, but also a subtle way of correcting for each transducer's parasitic resonances. Free pipe amplitudes for CBL tools are also delicate to determine, even more so when multiple parameters are used by the processing algorithm, unbeknownst to most wireline engineers (the best available CBL tool comes to mind).

The most sensible approach is to follow a reliable calibration protocol, something I call "absolute calibration" - without any reference to Kelvin or vodka.

Through the looking glass

Quantitative analysis of a well-calibrated log is how life should be. The reality is often closer to the nightmare scenario I've experienced a few times.

Picture it: you're a busy (make it very busy) drilling or production engineer, and a log doesn't look right. Then you call the wireline company and ask them for help, and that's when it starts going downhill.

After all, they did run the logs and they look real smart. Like the guys taking an NMR scan, or those testing your blood: why not asking them to interpret the results for you? (Except they don't and, with a varying degree of politeness, direct you to a specialist.) But the wireline engineer tries nonetheless and gives you a first set of answers, which tend to be replete with platitudes ("that's a microannulus"), fuzzy jargon ("oh, that? It's gas-cut cement") or annoying acronyms.

Still not an answer that can tell you whether there's a problem with your well, or what you should do about it.

So you grumble and the wireline engineer escalates and produces a being from a higher plane of existence - either an expert or a manager (a couple of times I've been shown a marketing guy, but that's mercifully rare). They normally have a collection of plausible-sounding theories that range from the crocodile-in-the-sewer meme, to the stuck pipe gremlins of my North Sea youth. One of my favorites is this guy explaining to me that this is exactly the noise of gas flowing through the cement matrix.

Still not happy. Then you make a second mistake, and bring the cementing company to the table. It is a tempting thought: they pumped the grey soup, so they should know what's wrong with it. Actually, their job is to get the slurry in the annulus without cementing the casing the wrong way around - and they are pretty good at it. They know precious little about what goes on once they shut down their pump.

So you witness a shouting match, with service companies hurling "it's your fault" at each other. This is interesting if you're into gladiator fights, or if you're a psychologist studying how people cope with uncertainty and the fear of losing face. If, like me, you're an engineer focused on solving problems, it quickly smacks of a waste of time. And it is: even if you don't count the inordinate amount of time spent by service companies, the time you and your colleagues pile up in meetings, reviews, discussions and other not terribly useful endeavors can total many thousands of dollar. And still no answer...

 

The thing is, economy tells us why this nightmare keeps happening: log interpretation is not the core business of wireline companies (or cementing companies for that matter).

 

It is not that they couldn't possibly do it if they put their heart to it, but since you're not paying for interpretation, you won't get it. Even if you seriously think that interpreting logs is part of their condition of service, their lawyers and your supply chain will point out that it isn't. Neither you, nor the service company, want to waste money in creating a complex ecosystem of geeks that doesn't generate revenue. And even if by accident there was one before the crisis, they have all been fired by now.

On top of it, I have serious doubts (backed by a fair amount of experience) that wireline or cementing companies could deliver half-decent interpretations with their current business model and structure. But this is a moot point if we keep insisting on having value for free.

Do you need to do a quantitative in-depth analysis every time? As with blood tests, if everything looks normal then you can get on with your job. Even if I don't get to see many good logs, I'm sure there are a few of them around.

However, if you ever come across something weird or funny, or you want to learn something from your logs, you should be seeking professional advise from a specialist, somebody who can dig through data and give you answers. You may have someone in-house, but chances are that she's overwhelmed with the low-hanging fruit and lacks the computing horsepower and the models (not to mention time) to extricate answers from the data jungle.

Now I will not pull an Oleg on you and ask "do you agree?". But I'll appreciate any comment, question or opportunity for an enriching discussion. And if there's any topic you'd like more info about, please don't hesitate to let me know.