Digital Energy Journal- Machine learning may change oil and gas

Join | Existing Member Login

You are Home » News » View Article

Machine learning may change oil and gas - but not the way we think

Friday, November 22, 2019

Accenture published a survey of C-suite and director level IT and business executives, saying that 76 per cent of them believe that 'DARQ' - that's distributed ledger, AI, extended reality and quantum computing - will have a 'transformational or extensive impact on their organisations over the next three years.'

In researching Digital Energy Journal's July issue, we talked to people at the cutting edge of machine learning, in both subsurface and operations domains - and got a somewhat different picture,

There does seem to be an argument for slow and steady improvement in machine learning, and value the industry gets from it. 'Transformational' may be going too far, but you can draw your own conclusions. Machine learning is a subset of AI.

I'm not sure about the other elements of 'DARQ'! Although like with machine learning, the definitions are not completely clear.

'Quantum Computing' seems to mean superfast computers using advances in materials technology at the molecular level. So relevant if you have a need for fast computers. Do we?

'Extended reality' seems to mean computer generated material seen on top of your normal view, for example if you are wearing special glasses. That's useful for people working on top of telegraph poles and other manual work where having hands free is useful. It can be a distraction if you are operating vehicles, and for many tasks the desktop computer, tablet or smart phone is still ideal.

'Distributed ledger' seems to mean a record of something in a way no single party can tamper with. This is something humans have been able to do for millennia, recording transactions on stone, or having central property registers. Bitcoin took the technology to another level but that was to meet some special use cases of a currency beyond the reach of government. My guess is the main obstacles to sharing information are commercial, not technical, in which case a better technology won't change anything.

AI is perhaps the most interesting. I interviewed the founder of perhaps the most advanced asset integrity machine learning company, who told me that he did not personally understand what 'AI' meant, but 'machine learning' was something he did understand. He defined it as using computers to understand the relationships between different variables based on observed data.

The oil and gas industry does have many different data variables to work with across subsurface, production and operations - and there are many efforts to better use machine learning to work with them - and we have many illustrations in this issue across seismic interpretation, well log interpretation and asset integrity management.

Our own theory of the most interesting technology is something you might not think of as a technology - better mapmaking for how all of this stuff should fit together, so it can be carefully designed to provide the right information to the right people which they need to achieve their goals. So we don't need to spend so much time updating each other with e-mails and meetings. We can also carefully design our system so it makes the most of both computer power and brain power, while simultaneously being secure, flexible and maintainable.

Karl Jeffery, editor
Digital Energy Journal, London

Associated Companies
» Accenture

CREATE A FREE MEMBERSHIP

To attend our free events, receive our newsletter, and receive the free colour Digital Energy Journal.