Experts meet in Houston to assess AI progress in oil and gas industry   

The Future Digital Twin and Generative AI Conference USA, held in Houston on May 23^rd 2024, brought together energy and technology professionals from 120 organisations to discuss how the energy industry can build on the significant strides already being made  to incorporate digital twin tech and artificial intelligence into its operations.

The event, organised by Cavendish Group International, provided a forum for vigorous  debate and an opportunity to hear fascinating perspectives from innovators and thought leaders on the future of a fast-evolving sector.

High on the agenda were hot topics such as how best to smoothly incorporate large language models (LLMs) and generative AI into the industry, the need for interoperability between a growing array of diverse applications and improving data sharing within and between organisations.

The conference heard how all this must be underpinned through establishing trust – trust in the accuracy of the data being fed into AI applications, trust in the results being generated by those applications, and trust that your data is not going to be misused by others in a new era of where greater collaboration and data sharing is a necessity.

Adam Soroka, CEO of Cavendish Group International, said “the Future Digital Twin & GenAi conference highlights the value of digital twins, emphasizing greater collaboration and best practices from other sectors like aviation. AI and digital twins can enhance interconnectivity, though we must address trust, ethics, and data limitations. AI should augment roles and automate processes, improving the user experience and facilitating digital transformation”.

Digital Twins enable organisations to create digital replicas of physical assets, processes or systems, carry out predictive modelling and improve their operations. In the oil and gas industry that provides the opportunity to make operations safer, and more efficient and to decarbonise them faster. The addition of Generative AI is helping to speed up processing of the vast amount of data being produced by the industry to feed digital twins and other applications, and to provide valuable support for decision makers.

Speakers agreed the hydrocarbons sector could learn valuable lessons from other industries that are further along their ‘AI journeys’. However, a recurring theme at the event was surprise at the rapid speed with which some parts of the oil and gas industry have embraced digital twin technology and AI more widely, in light of the sector’s traditionally more conservative approach to operational change and data sharing.

Shane McArdle, CEO of Kongsberg Digital and Chairman of the Conference Advisory Board, told delegates the industry had moved rapidly from proof of concept to the proof-of-scale stage in terms of digital twin and AI adoption.

“Companies that invested in the foundational capabilities of digital twins or data platforms have been able to take advantage of this technology very quickly as it  has scaled up. Those capabilities have allowed us, as an industry, to develop the core capabilities you need to connect something like generative AI to industrial data,” he said.

Here's what some of the other panellists at Future Digital Twin and Generative AI Conference USA had to say:

Amit Jain, an advisor on Digital Twin development and deployment at Chevron, also highlighted the shift in working practices in the industry, noting that the situation prevailing five years ago where vendor applications were working in separate silos and vendors were reluctant to collaborate had been turned on its head.

“The digital transformation has helped us to break those silos. Now, when I'm speaking with partners, they're saying: our solutions are open, we can send and receive data from everywhere. So, it has really transformed how data is treated both within and between organizations,” he said.

Andrea Course, Digital Innovation Program Manager, at Shell urged companies to build on a new spirit of cooperation to speed up the digital transition.

“Collaborate, collaborate, collaborate: pick good partners that will take you along the road that you need to go on. Once you have your vision, find out who can help you,” she told delegates.

Justin Piwetz, Asset Management Lead, Research & Innovation in Virtual Technologies at bp told the conference how interest in generative AI had been fuelled by the need to augment human capabilities to process the petabytes of information being generated by the upstream industry for use by digital twins.

“Digital twins provide structure and gen AI allows you to search in a more abstract way than you normally would – it’s critical that the two grow together. Without that, you're going to get to a point where the human aspect is the limiting factor with both the digital twin and the AI,” he said.

Michael Hotaling, Operations Excellence Digital Manager, at ExxonMobil stressed the importance of interoperability, allowing data to be shared independently of the diverse range of technologies and applications deployed by the industry. 

“We need separation of hardware from software, separation of software and data and separation of the AI Model from the platform,” he said. “Data needs to be in a model that can be interrogated no matter what generative AI, or NLP or LLM you are using. We've got to keep aligned as an industry and get to a modular plug and play approach – the additional use cases we can build on that will absolutely move exponentially.”

Jason Gislason, Global Technology Manager at Chevron Phillips emphasized the balance to be struck between using the industry’s most valuable resource – it’s people – and AI.

“There's a lot of really good applications for LLMs looking at very specific things, but you want to keep people in the loop, and you want to do it in such a way that that you add value to the organisation, but not put it at risk,” he said.