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

Data driven MPC for oil production optimization Production of oil and gas as a function of 'well controls' is hard to model from first principles. Therefore, in this task we will investigate MPC approaches that don't need a predefined model. This can be called data driven MPC. The task

1. Literature survey and background regarding data-driven MPC
2. Implementation of a chosen solution for data-driven model construction, based on excitation of the process
3. Test solution using a case with four wells, producing to a separator with a limited capacity for gas handling.

Co-supervisors: John-Morten Godhavn, Equinor.