Probabilistic Modeling for Decision Support in Integrated Operations

Abstract

Summary We report work carried out within the project on collaborative decision support for integrated operations (CODIO). As one part of this project, we have designed a system to provide assistance in operational decisions based on real-time sensor readings in a typical scenario: While drilling close to the transition to a high-pressure formation, a gas influx is observed. The drilling team needs to decide whether to circulate, increase the mud weight, plug back, or set a casing. After a brief description of the technology we used to model decision problems [Bayesian networks, influence diagrams (IDs)], we describe our case study and discuss the particular challenge of applying decision analysis to the kind of operational decision that is central to the case study. We then proceed to apply the method of decision analysis to our case study. We discuss how the resulting ID was tested using a simulation and discuss challenges in applying the technology, as well as lessons learned in the process. We also give a list of desiderata to establish better decision-making practices in the petroleum industry.