Despite its enormous promise, the digital oil field concept is no panacea, a trendy cure-all to be implemented carelessly for instant benefits. Realising the potential of a fully connected offshore operation is no easy feat, and presents challenges all of its own. Finding the right workforce and putting into action effective training in digital technologies is a start, but one of the major issues for the industry to tackle is data overload – the sheer breadth of data that can be harnessed from offshore oil platforms and their associated infrastructure can do as much to paralyse decision-making as liberate it.
Cherry-picking actionable information from the data stream, therefore, has become the key to truly 21st century oil and gas operations, and the industry is opening up to new partners to help make this happen. Madrid-based oil and gas multinational Repsol, for example, recently launched a partnership with IBM Research’s Cognitive Environments Lab (CEL) to develop two new cognitive technology prototypes to help the company make better exploration and production decisions. The prototypes, which leverage IBM’s experience developing the artificially intelligent system it has dubbed Watson, are intended to improve the interface between humans and computer systems to get the right data to the right people.
Repsol is clearly serious about the project, having committed to an investment of $15-20m to develop the prototypes, with early results expected by the end of 2015. Both Repsol and IBM will dedicate up to ten workers on the project, which will be carried out at CEL in New York and Repsol’s Technology Centre in Madrid. We spoke to CEL senior manager, Brian Gaucher, to discuss the collaboration with Repsol, the potential of smart data analytics and the benefits of the ‘cognitive environment’.
Chris Lo: To what extent are cognitive technologies used in the oil and gas industry currently?
Brian Gaucher: The research collaboration with Repsol is the first to apply cognitive technologies for oil and gas applications. Repsol are really being thought leaders here in trying to drive this agenda forward, investing in innovations that can gather input from myriad stakeholders as well as the data from across both the business and operations.
CL: What have been the issues so far that have stopped the industry getting the best use out of big data?
BG: Not only is the explosion of big data a challenge for any industry, the oil and gas industry bears the challenge of an enormous variety of data and how to best access it. The industry has a tremendous amount of data coming in from and being used across multiple stakeholders – from operations, the line of business and general IT. The data is very heterogeneous, it comes in many formats that are incompatible with some systems; it can be very ‘noisy’ and requires careful filtering in order to help and not hurt the overall process, and in the end it has to bring value across a multi-disciplinary group of people. In a way, new tools and systems are needed that manage it and that can scale and magnify each stakeholder’s abilities by bringing together those disparate data sets and points of view into a cohesive whole.
What could Lockheed bring to the cyber security situation in the oil & gas sector?
CL: Could you describe the two prototype cognitive applications that IBM and Repsol are co-developing at the moment?
BG: We are working with Repsol to enhance their capabilities in two major areas, but I’d also like to point out that our view of cognitive systems is not just a software-centric point of view, but also the collaborative aspects of team decision-making in specific cognitive environments.
We feel that cognitive environments can enhance the collective intelligence of the group and influence the direction of strategic decisions for better outcomes. For Repsol we will jointly create this new cognitive environment with a specific focus on these teams and big data to address two main areas: firstly, optimising reservoir production and secondly, enhancing the decision-making process for the acquisition of new oil fields.
Optimising reservoir production in a CE will address the needs of a diverse set of technical experts and provide them with new tools intended to enhance their abilities to mine data from many different sources and perform deep computational analysis to provide evidence-based insights to the experts who are guiding the system. They can bring in existing production models, data and information and adjust the models to more accurately match current production as time goes on.
Enhancing the decision-making process for the acquisition of new oil fields involves strategic decisions across another group of technical experts. The purchase of new land requires an immense investment based on highly uncertain data. Gaining deeper insights at reasonable cost that reduce uncertainty can be extremely valuable.
No expert can read every journal article from even a single source, but a cognitive system can do that and more through the use of natural language processing. An expert can guide the system to quickly produce evidence-based summaries for specific queries that can aid the decision-making process. Leveraging that inside a cognitive environment allows multiple stakeholders to similarly find and use data, run simulations and factor in decision-making biases to reduce uncertainty.
CL: Do you think cognitive tech will unlock the full potential of big data and the Internet of Things by helping users get the most out of all this information?
BG: I believe the Internet of Things will be one of the biggest generators of data in the future. I think cognitive environments and technology will first evolve around specific domains like oil and gas, finance and others who will be the first beneficiaries. I imagine that cognitive environments are places that provide humans with enhanced abilities to find, rank, analyse and filter big data to make it more consumable. In a way, such environments and capabilities will democratise big data making the otherwise endless noisy data manageable and speed access to insights.
CL: How can the interaction between humans and software become more natural and intuitive?
BG: Cognitive environments are places and capabilities that work the way humans like to work and communicate where one can speak, gesture and/or touch information. The environment will be able to recognise us, listen as needed, ask questions and proactively suggest insights much as another person or collaborator might do. So natural language processing, both spoken and written, will be important, as well as knowing who might be using the system to help them in ways they have trained the system to recognise.
The overall use and experience will be much more personalised to our needs and our individual ways of communicating and use of information. In our Cognitive Environments Lab, we are experimenting with quantifying the performance of and determining the right balance between interaction modalities. While speech may be the only acceptable and safest form of interaction in a car, a cognitive board room or meeting room opens up the aperture for combining modalities that include gesture and advanced visualisation and navigation techniques.
CL: In what ways might cognitive technologies improve the task of offshore exploration specifically?
BG: Offshore exploration is extremely expensive. A single well may cost $200-$400m and the data used to determine the best location is very limited. The decision to drill is based on highly uncertain data where only one in four or one in five wells drilled are successful. The cost of gaining more data is also very expensive; for example millions of dollars may be spent on seismic soundings. We believe that cognitive environments and technologies can bring decision makers together, help them seamlessly share insights, bring in heterogeneous data sets more fluidly, and enable target analysis and simulation. These capabilities can reduce the uncertainty, address the value of information in prescriptive terms and in the end reduce risk.
Reliable, high-speed communications technologies are transforming the offshore environment.
CL: How can human users trust that intelligent software is really highlighting the correct data to make the best decision?
BG: We view that cognitive systems will partner with humans to enhance their expertise and capabilities. Any user will be informing and training the system as they go and the system will be a reflection of the individuals using it, where each is learning from the other. So validation of the right data and insights for users will be happening not at the end, but along the whole journey.
CL: How can the cognitive environment bring together oil and gas specialists from different fields to enhance the decision-making process?
When experts leverage a cognitive environment, it will be analogous to being in a place where if one person speaks another language, the environment can be used to translate critical insights into a language or visual insight that others can see and understand. It can also be used to pull in real-time insights to help the experts answer questions or provide data or references to support claims or to run simulations or analyses to help provide clarity to others.
CL: How long do you think it will take to start seeing these innovations in Repsol’s operations? Do you have a timeline in mind for this collaboration?
BG: This is a research collaboration and we have a three-year timeline with Repsol on our joint development. We expect that by the end of year two we will have a cognitive environment and capability that addresses the optimisation of oil fields and by year three the acquisition of new oil fields.