Petroleum Systems Through Economic Assessment of the Permian Basin, Texas Aided By a Cognitive System for Geosciences
Current pressures within the Oil & Gas Industry demand a step change in productivity. On one hand, leaner teams have to deliver increasingly reliable evaluations at a faster pace. On the other hand, data of varying qualities are more abundant and available than ever. Workflows, from data retrieval to final interpretation, remain lengthy, limiting iterations within and between disciplines. This in turn limits completeness of the resulting volume, risk and value/economic scenarios. Recent progress in Artificial Intelligence (AI) offers an unprecedented opportunity for improvement by accelerating workflows so that more time can be spent on the final technical & economic integration. We present a novel framework to take advantage of both principled, physics based bottom-up and data-based top-down approaches, powered by a cognitive system blending symbolic AI with deep learning methods. We illustrate the method using an analysis of the Permian basin, Texas. At the core of the system resides a knowledge base of the basin linking millions of data, facts, interpretations and concepts, gathered from sources as diverse as raw data, technical documents or modeling results. Incomplete, imprecise and sometimes contradictory information reported by thousands of operating companies is consolidated automatically into a consistent geological framework using machine learning and fuzzy logic algorithms. Evaluation of regional thermal stress, pore pressure and petroleum generation is accelerated thanks to a novel physics based AI, trained to learn the impact of basement type, erosional events, lithology, burial rate, source rock organo-facies and expulsion potential. Initial production data, GOR, water cut and API gravity from thousands of wells can be used to constrain volume and composition scenarios. This leads to high resolution prediction of GOR, saturation and bubble point, by play, at the basin scale. Predictive analytics link these parameters to production through time. The cognitive system manages automatic updates as soon as new data or hypothesis is added. Leveraging the AI promise of decreasing the cost of prediction, this technology is a solution for the fast evaluation of undrilled opportunities or development plans. It has also deeply transformative capabilities for organizations as it enables a true collaboration all along the value chain; leaving human beings to understand and weigh complex decisions rather than themselves performing the role of ‘computer’.
AAPG Datapages/Search and Discovery Article #90332 © 2018 AAPG International Conference and Exhibition, Cape Town, South Africa, November 4-11, 2018