The text below is a brief description of the topics given by professor Lars Imsland. Contact lars.imsland@itk.ntnu.no for further information.

1. Modelling, simulation and control of a future compact separation process
   Co-supervisor: Gisle Otto Eikrem, Statoil Forskningssenter, Rotvoll.
   • Background The process to be studied is a future subsea compact separation system, to be used at deep waters. The system will separate gas and liquid from the wells at the seabed and boost the gas and liquid up to the surface by use of a compressor and a pump. Since this system is much more compact than traditional separation processes, an advanced process control system is required. This will be a challenging and informative project work where the candidate will be given the opportunity to work in an ongoing technology development project with process modelling, programming, dynamic simulation and advanced process control.
   • Objective The project work will cover model upgrade of an existing compact separation process simulator, which is controlled by a model predictive controller (MPC). The existing control system has proved its ability to control the process when small process disturbances are introduced. This project will investigate how such a system can be started up and shut down. Startup and shut down can be viewed as large disturbances for the process. The programming languages to be used are C++ and Matlab. Statoil’s internal MPC tool Septic will be used for process control. Outline of work:
     • Upgrade the dynamic model of the compact separation process.
     • Propose a startup sequence and shut down sequence for the process.
     • Implement the sequences in the dynamic simulator and perform dynamic simulations.
     • Optimize the control system to improve the startup and shut down of the process.
     • Write a report.
2. Observer design for a compact separation process
   Co-supervisor: Gisle Otto Eikrem, Statoil Forskningssenter, Rotvoll.
   • Background A future subsea compact separation system, to be used at deep waters, is under development. The system will separate gas and liquid from the wells at the seabed and boost the gas and liquid up to the surface by use of a compressor and a pump. Since this system is much more compact than traditional separation processes, an advanced process control system is required. A laboratory test unit of this system has been developed and will be used as process in this project work. A dynamic simulator of this process is available with a nonlinear model predictive controller implemented (NMPC). To be able to use NMPC it is required to update the process model by use of measurements from the real process, this requires the implementation of an observer. This will be a challenging and informative project work where the candidate will be given the opportunity to work in an ongoing technology development project with programming, dynamic simulation and advanced process control.
   • Objective The candidate is to develop an observer (state estimator) for the laboratory test facility. Different types of observers should be evaluated to estimate the states. The extended kalman filter (EKF) and the moving- horizon estimation (MHE) are two observers to be evaluated. The compact process system requires a high speed MPC solution which makes computation time a significant aspect. The programming languages to be used are C++ and Matlab. Statoil’s internal MPC tool Septic will be used for process control. Outline of work:
     • Literature study on observers for nonlinear processes/models.
     • Implements the preferred observers.
     • Evaluate the performance of the control system with the different observers applied.
     • Write a report.
3. Methodology for automatic deduction testing of subsea wells
   Co-supervisor: Petter Tøndel, Statoil Forskningssenter, Rotvoll.
   Testing of subsea wells gives a loss of production for the duration of the test and requires a lot of attention from control room operators. Efficient testing requires tight control of the manifold pressure, to ensure that gas- and liquid flows from non-tested wells on a manifold remains constant. By automatically controlling this pressure one can reduce the duration of the well test while improving the test results and reducing operator work-load.
4. MPC for load balancing
   There will probably be external co-supervisors to this task.
   The frequency of the ac voltage is a global measure on planned balance in an interconnected power grid. Frequency deviations outside nominal limits are seen to occur more and more often in the Nordic power system. The objective of this thesis is to investigate the use of Model Predictive Control (MPC) as a method to be applied by transmission system operators to optimize balancing control and use of reserves in a system with increasing forecast uncertainties in both generation and load. Possible elements:
   • Describe today's load balancing procedures
   • Give a description of how an MPC solution can be used in load balancing procedures
   • Implement a prototype solution based on simplified models of net and market
5. Techniques for efficient covariance propagation in the Extended Kalman Filter based on model reduction
   Even though the (recursive) EKF is considered an efficient algorithm for state estimation, it can still be computationally expensive for large/complex systems. The bulk of the complexity stems from the propagation of the state covariance matrix. In this task, the student will consider using techniques from model reduction to make the covariance propagation more efficient, while the full model still is used for state propagation.
6. Parallelization in nonlinear MPC
   Implementation of nonlinear MPC can benefit from parallelization. The candidate should explore strategies for parallelization, and make prototype implementations for assessment and comparison.