Artificial Intelligence for Heavy Oil Assets: The Evolution of Solutions and Organization Capability

摘要

Artificial Intelligence (AI) techniques have been successfully applied in the petroleum industry since the early 1990s initially attempting to solve simple tasks and more recently evolving into hybrid systems taking on complex optimization and modeling problems. Artificial Intelligence has become an integral part of our business in the last 10 years with applications ranging from reservoir characterization, production optimization to surrogate models used in reservoir simulation. Chevron's San Joaquin Valley Business Unit has been implementing AI technologies through Chevron's i-fieldTM program since 2003. With the majority of the production coming from heavy oil assets, most of the intelligent applications developed using AI target the optimization of thermal operations. The paper describes the operator's experience and continuing efforts towards integration of Artificial Intelligence technologies in the Business Unit. The manuscript covers two topics: a review of several successful Artificial Intelligence applications implemented in the heavy oil assets while focusing on solutions and value creation, and an internal training program aimed at developing the company's organizational capability in this domain. The first part discusses the successful implementation of AI techniques, neural networks, genetic algorithms, fuzzy logic, case- based reasoning and hybrid systems to solve complex projects. Case studies covering well candidate selection, optimization of cyclic steam scheduling, increased production opportunities identification, well failure diagnostic and job planning are presented. The review concludes with lessons learned, challenges and business value creation (increased production, NP PV, time savings, etc). The second part highlights the efforts of initiating and implementing a training program aimed at building awareness and organizational capability. In 2009, a few AI enthusiasts and practitioners formed an Interest Group with the mission of enhancing reservoir management workflows by developing innovative solutions using Data Mining and AI technologies. The section describes the target audience, training material, workshops structure, brainstorming exercises and lessons learned from the sessions completed so far. Ultimately, the review is intended to demonstrate the value of applying AI, and how to grow AI organizational capability. A brief overview of future plans concludes the paper.