– It’s not precise at all, so please handle with care, is one of the first things Anita Hansen, Geoscientist at ResultatGeo, tells me when we meet on Teams.
We spoke about cuttings data from wells.
Drill cuttings (rock chips) provide an analogue data set that is collected in large portions of a well at selected sampling intervals.
– It is important to remember that the samples do not represent a specific depth, but rather a 10 to 3 m range, depending on where in the well the sample is taken, Hansen continues.
Cuttings are formed from the action of the bit and are transported up to the drill floor with drilling mud, so there is every possibility of contamination and significant uncertainty.
– Cuttings data always have to be integrated with well log data and drilling parameters in order to get a more complete picture of lithological variation with depth, she says.
A conglomerate or basement?
As so many people of the “older” generation in Norway, geologist Anita Hansen started her career at Saga Petroleum in 1989 after two years working as a mudlogger. For Lundin, she was the operations geologist involved in all exploration wells the company drilled in the Utsira High area from 2007 to 2018, which has given her a lot of good memories.
Hansen remembers drilling the first well on what was to become the Edvard Grieg discovery.
– We had a gas response in the well, so we decided to cut core. When the core was inspected on the rig, we found granitic boulders about 10 cm in diameter in the core barrel. Some people argued it was a sign that we hit basement and should therefore give up on the well.
– However, we saw that the boulders were rounded, and argued that it may represent a conglomerate instead. It was decided to continue drilling, which proves to be a wise decision in hindsight.
DIGEX 2023
At the upcoming DIGEX Conference, taking place in Oslo from 28-30 March, Anita Hansen will present about uncertainties in drilling data and how integration of data sets and interdisciplinary knowledge is key to arriving at a complete picture.
Cutting size ≠ grain size
Back to cuttings. Hansen emphasises that cutting size is not synonymous with grain size.
– Single cuttings can be bigger than the individual grains in the rocks, which means that it’s possible to greatly overestimate the permeability and porosity of the formation. The mitigation is, again, the integration of data sets.
This also has implications on how photos from cuttings should be used for machine learning purposes.
– Using machine learning, it is possible to detect and map the individual cuttings, but then another correction will be required to translate cutting size to grain size in order to give a meaningful result, she adds.
Another aspect worth stressing is that it is important to remember that samples are being washed and sieved. Therefore, the sieve size needs to be known in order to calculate the maximum size of fines in the original samples.
– Another way to get a better understanding of the original sample is to record weight before and after washing.
– If from a 50 g sample there is only 25 g left after washing, we know that there is much more fine material in the sample than what we see in a picture of a washed sample. Therefore, it would be wrong to use cuttings size from these images as grain size but in combinations with other data sets it is a “piece in the puzzle” for better understanding, Anita Hansen concludes.
HENK KOMBRINK