State and parameter identification applied to dual gradient drilling with oil based mud

Dual gradient drilling (DGD) is a modern technology that allows for drilling wells for oil and gas in a more precise and efficient manner. Many of the future oil and gas prospects are highly complex and challenging with narrow geo-pressure margins and little room for human errors. Such projects are simply only feasible if they are founded on robust and efficient automated drilling solutions that allow for accurate pressure control. This study focuses on investigation of a real DGD system in the Gulf of Mexico through modeling and application of field data in order to estimate unknown parameters. First an introduction to the basics of drilling is given, explaining the advantages of the dual gradient drilling technology in terms of efficiency as well as safety. Then a hydraulic model for the complete system is developed. The available field measurements delimit a subsystem of the complete system where all parameters are identifiable. This shifts the focus of this thesis solely to the dynamics of the mud return line (MRL), which is the most vital section of the DGD system to know prior to control design. Following the development of the model for the MRL dynamics, field data are applied to estimate the unknown parameters. Obvious discrepancies between observed field data and the outputs of the identified model expressions encourage slight modifications of the model structures until the observed outcomes are replicated by the model expressions to a satisfactory degree of accuracy. Finally, the dynamics of the identified system are simulated. The produced outputs are compared to field data, and the consistency between the simulation results and the field measurements works as a natural measure of the validity of the identified model