MOSCOW — More than 30 Gazprom Neft research geologists and IBM Research Brazil machine-learning research scientists have come together for the first time to share their unique knowledge, experience and skills in a research project, “Cognitive Geologist”*. The cooperation agreement signed by both companies aims at the development of state-of-the-art broad AI technologies, such as, knowledge-enhanced machine learning which combines deep-neural networks as surrogate of existing geological models and advanced probabilistic reasoning techniques for modelling and supporting data-driven geological decision making. In so doing, they aim to optimize analytical processes and characterize the value of information related to the most relevant geological models to advance exploration workflows and automate the model creation for exploration. These technologies will help automate routine operations and significantly augment the geological and geophysical data analyses for use in oil exploration or production activities other than deepwater, Arctic offshore, or shale projects.
The system aims at identifying, modelling, integrating, and predicting critical prospective geological object information in support of decision-making, using data from already discovered analogues and big data technologies for processing extensive data arrays. Thanks to this initiative, the typical geological exploration cycle could be significantly reduced, with processes previously taking up to half a year now completed within a month. In addition, forecasting quality would be further improved with each new project, through the use of cognitive technologies and continuous-learning models.
Mars Khasanov, head of technology directorate, CEO, Gazprom Neft Science and Technology Centre: “As part of Gazprom Neft’s Technology Strategy, we are paying particular attention to developing digital projects in exploration and production. Together with leading international companies we are developing the oil production tools of the future.”
Alexei Vashkevich, director for geological exploration and resource base development, Gazprom Neft: “Data volumes are becoming ever larger, and reserves ever more challenging — which is why we have to improve processes in extracting, sourcing and analysing information, and find new opportunities for using machine learning methodologies. To that end we are developing a unique digital solution that could lead to a breakthrough in working with geological information.”
Ulisses Mello, director, IBM Research Brazil: “At the IBM Research Natural Resources practice, it is our mission to collaborate with leading Oil and Gas partners to ensure that the industry benefits from the application of the most advanced AI technologies by maximizing its ability to deal with the challenges and opportunities presented by the ever growing amounts of data generated in the exploration and production activities.”
Source: World Oil