Non-linear model predictive control for an oil production network based on gas-lift

Control and optimization of an oil production network based on gas-lift is a difficult task with challenges including well fluid flow oscillations and pipeline slugging which can lead to instabilities in the system. A possible control solution to this is Model predictive control, which requires a model capable of capturing the main characteristics of the controlled process, such as oscillations and input dependencies. In this thesis models for wells and flowlines are instantiated in the detailed multiphase dynamic fluid flow simulator OLGA. Further, simpler dynamic models are attempted fitted to these models, and the comparison between the resulting models is discussed. In addition state estimation using an Extended Kalman Filter is performed on the flowline model, based on measurements proposed in \cite{Lund}.Based on the fitting of models and implementation of state estimation for the flowline system, methods for NMPC are attempted implemented using the JModelica.org environment, and limitations and advantages associated with this environment are assessed