Oil IT Journal - dGB recently announced a Consortium* to demonstrate the ‘future of E&P data management’, using machine learning and OSDU. How did this start?
Paul De Groot – I began as something of a skeptic with regards to OSDU having seen similar earlier initiatives from POSC and Open Spirit fail. Things changed when Schlumberger put its weight behind the project and threw its Delfi data infrastructure into the pot as open source software. OSDU now looks to be a game changer. I believe it will be the standard in the near future. So the next step for dGB Is to make OpendTect adhere to OSDU.
The Consortium is billed as a ‘digital transformation demonstrator’ what does this mean?
OK, nobody knows what the ‘digital transformation’ really means, but there is a consensus that machine learning/artificial intelligence is key. So we are setting out to demonstrate ML capabilities with OSDU, ‘the database of the future’.
Isn’t this a takeover of the industry by Schlumberger?
I don’t know. Maybe Schlumberger sees a business model here and is expecting more work via OSDU. But the Delfi formats are now open and will be the platform of the future.
What do you see as a killer app for AI?
I’ve been working for 30+ years in AI for geophysics, in the last 3-4 years on deep learning. DL can do stuff that was not possible previously with shallow nets and accelerators. Now some models can be re-used on other data sets. You can train a model to produce shear logs from sonic, and apply it across the whole of the North Sea. You can build a library of hydrocarbon direct indicator artefacts and apply it everywhere. There is more to this than hype.
Which cloud provider is leading OSDU?
Amazon is probably in the lead. But there is also stuff running on Azure for some clients. OSDU is also on Google Compute Cloud and IBM/RedHat.
Are you backing Open VDS as the new default seismic format?
Yes, we are!
So does that mean that the first step in moving to the OSDU world is migrating SEGY (and other formats) to OpenVDS?
No, OpendTect supports SEG-Y as a native format, which we will continue to support for file-based access. But if you want to store data in the cloud via OpendTect’s implementation of OSDU, the seismic data will be stored in OpendVDS and the well data in Delfi. We expect most vendors to follow this implementation, which is recommended by OSDU. If you want to store SEGY data in the cloud, you can for instance use Osokey’s implementation of OSDU.
*The OpendTect/OSDU Machine Learning Consortium.
Oil IT Journal - What is the rationale behind Ikon Science’s latest offering Curate?
Dennis Saussus - We spent some 20 years developing RokDoc, our flagship package for real-time geomechanics, pore pressure prediction and more, and all this will continue. At the same time, we have learned a lot about the challenges that clients face, knowledge sharing, working with data trapped in specialized desktop apps, transferring data, finding out what has already been done. All this in the context of project management and decision support. These problems have been compounded in the last five years with downturns, people leaving or being let go.
How does Curate relate to your existing portfolio. Is it a rejuvenated iPoint?
In some ways, yes? iPoint helps people find and share data. It started with core data and expanded to a large range of data types. In Curate, we set out to combine iPoint functionality with RokDoc, adding a modern user interface.
What is Curate’s scope, what applications are you seeking to displace?
We are not seeking to displace other applications. Rather to build apps around RokDoc that automatically reuse data. For instance, a web app that supports and democratizes AVO modeling with sensitivity analysis.
The focus is still reservoir geophysics?
Yes. Our aim is to expand with apps for pore pressure, real-time ..
So this is a platform?
Who do you sell to? End users, geophysicists, IT?
A good question! The answer is ‘all of the above’. The Curate workspace integrates application databases including Kingdom, OSDU, Petrel … so it is key to talk to the IT department and leverage Curate’s open API.
Is the API provided to clients or just for Ikon internal use?
It is very much our intent to offer an open API to clients and third parties including other software vendors.
An OSDU purist might argue that vanilla OSDU does much of this already? Is there overlap here?
OSDU’s focus is more on the data management/access end. Curate’s main focus is at the end-user level.
How open is ‘open’? Will this be on Git Hub as open-source?
No, this will not be open-source – but the APIs will be ‘open’ at some point down the road.
The IEA says (today) that oil exploration must end. What is your backup/alternative plan?
Yes, this is one reason for broadening our offering into the data/workflow space. We expect our users to be working in the geothermal/CCUS/renewables areas. Data/Workflow is key to all of these. We do anticipate working outside of traditional oil and gas. The intensity of carbon-neutral pressure will only grow. But the world will not switch off hydrocarbons overnight. Curate opens the door to new products beyond oil and gas.
© Oil IT Journal - all rights reserved.