Modeling and control of managed pressure drilling operations

The upstream oil and gas industry has witnessed a marked increase in the number of wells drilled in areas with elevated subsurface formation pressures and narrow drilling margins. Managed Pressure Drilling (MPD) techniques have been developed to deal with the challenge of narrow margin wells, offering great promise for improved rig safety and reduced non-productive time. Automation of MPD operations can ensure improved control over wellbore pressure profiles, and there are several commercial solutions currently available. However, these automation efforts seldom take into account the uncertainty and complex dynamics inherent in subsurface environments, and usually assume ideally functioning sensors and actuators, which is rarely the case in real-world drilling operations. This dissertation describes a set of tools and methods that can form the basis for an automation framework for MPD systems, with specific focus on the surface back-pressure technique of MPD. Model-based control algorithms with robust reference tracking, as well as methods for detecting system faults and handling modeling uncertainty, are integrated with a novel multi-phase hydraulics model. The control system and event detection modules are designed using physics-based representations of the drilling processes, as well as models relating uncertain variables in a probabilistic fashion. Validation on high-fidelity simulation models is conducted in order to ascertain the effectiveness of the developed methods.Mechanical Engineerin