Below are brief descriptions of the project topics for fall 2022, supervised by professor Lars Imsland, typically together with others. Contact lars.imsland@ntnu.no for further information. Most topics can be continued in a master thesis project if the candidate and supervisor(s) so wishes. Some of the descriptions are rather advanced/involved for a single project, and should be considered descriptions for a combined project/master project.

Drillbotics (http://www.drillbotics.com/) is an international student competition related to autonomous drilling. There are two categories in the competition:

• Group A: Design a full-scale virtual drilling rig with autonomous drilling control system. Sponsored by DNV.
• Group B: Design a model-scale physical drilling rig with autonomous drilling control system. Sponsored by Equinor.

NTNU has previously been part of Group B (winning in 2018 and 2021), but new this year is that we will also try to put together a team for Group A. For Group B, we aim to have 1-2 cybernetics students, while for Group A, we aim to have 2-3 cybernetics students.

For both groups, we aim to put together a multidisciplinary team of 3-5 students with expertise in Petroleum Engineering (drilling), Cybernetics and, possibly, Mechanical/Electrical Engineering. These will work for a year (specialization project + MSc project) on developing, building, programming and testing/tuning the virtual/model-scale drilling rig.

It is a large-scale project being a replica of an industrial technology development project. By the end of the year (in ca. June 2023), the teams will participate in the competition. According to the students from the previous teams, it was the best learning experience from their whole NTNU studies. You can read more about this year's NTNU Drillbotics team (Group B) here: https://www.ntnu.edu/bru21/drillbotics .

The teams are supervised by a team of professors from Departments of Geoscience and Petroleum and Engineering Cybernetics. The supervisors from ITK are Lars Imsland and Børge Rokseth.

Carbon dioxide (CO2) heat pumps are promising technologies to provide environmentally friendly heating to buildings. The focus of this project is on the control of a heat pump for residential heating purpose. Traditional heat pumps use hydrofluorocarbons and chlorofluorocarbons, which seriously damage the ozone layer. Compared with traditional refrigerants, carbon dioxide (CO2) with the advantage of non-flammability and environmental friendliness is advanced to be used in the heat pump. Well-functioning control methods for CO2 heat pumps are still lacking for improving the energy efficiency of the entire system including the CO2 heat pump and the building heating system. The problem in improving the energy efficiency of the CO2 heat pumps is in improving the control between the heat pump and the belonging heating system. In the laboratory at the Department of Energy and Process Engineering, there is installed a CO2 heat pump with the heating room. The entire system is well instrumented and connected to be controlled via LabVIEW. In LabVIEW, manual, PID, and MPC controllers for different elements are enabled.

The aim of this project is to test and develop relevant controllers for respective components to increase energy efficiency of the CO2 heat pump with the belonging heating system. In this case, it means to develop controllers that may help to decrease the return water temperature from the heating system to consequently increase the coefficient of performance (COP) of the heat pump. Further, performance improvement of other components through control improvement is welcome. The model predictive control should be compared with PID controller for different components. The entire project may be performed by direct use of LabVIEW and control improvement by using the monitoring data from the laboratory or by combining monitoring data and simulation.

Co-supervisor: Natasa Nord, EPT/NTNU.

Equinor has been developing advanced process control solutions for more than 20 years using mainly the inhouse MPC software SEPTIC. More than 100 applications have been developed both for onshore and offshore facilities contributing both to increased production, improved quality and reduced energy consumption. In the coming years Equinor will develop advanced process control solutions also within renewable and low carbon energy systems.

This task will study MPC solutions with integer control variables.

1. Literature study.
2. Implement in Python/Matlab.
3. Case study with offshore process controlled by step chokes and/or on/off routing valves (to be given by Equinor).

Co-supervisors: Researchers at Equinor's research centre.

Equinor has been developing advanced process control solutions for more than 20 years using mainly the inhouse MPC software SEPTIC. More than 100 applications have been developed both for onshore and offshore facilities contributing both to increased production, improved quality and reduced energy consumption. In the coming years Equinor will develop advanced process control solutions also within renewable and low carbon energy systems.

This task will study nonlinear MPC using genetic algorithms for optimization.

1. Literature study.
2. Implement in Python/Matlab.
3. Case study with nonlinear model (to be given by Equinor) – comparing with traditional MPC.

Co-supervisors: Researchers at Equinor's research centre.

Equinor has been developing advanced process control solutions for more than 20 years using mainly the inhouse MPC software SEPTIC. More than 100 applications have been developed both for onshore and offshore facilities contributing both to increased production, improved quality and reduced energy consumption. In the coming years Equinor will develop advanced process control solutions also within renewable and low carbon energy systems.

This task will study MPC solutions with neural networks for modeling and optimization.

1. Literature study.
2. Implement in Python/Matlab.
3. Case study with nonlinear model (to be given by Equinor) – comparing with traditional MPC.

Co-supervisors: Researchers at Equinor's research centre.

This task will study linear MPC implementations, in C/C++, using open-source QP solvers (e.g. OSQP, https://osqp.org/, or/and qpOASES https://github.com/coin-or/qpOASES). Based on input files specifying the MPC problem (model, weights, horizon, simulation scenario, etc.), the code should generate the MPC controller, and run test simulations. Models may be both linear state-space models, and/or input-output (step-response) models typically used in industry.

The longer-term purpose is to create a light-weight tool for tuning and testing MPC, that can be run e.g. in a browser/app.

Co-supervisors: Researchers at Equinor's research centre.

Numerical simulations of sound propagation or scattering often need to be compared to measurements on real objects. A bottleneck for such measurements is that the loudspeaker used for the measurements does not behave as a monopole, but in many types of numerical simulations a sound source needs to be modeled as a monopole. However, a real loudspeaker could be modeled as a set of monopoles, and the amplitudes/weights of those monopoles can be adjusted to fit a measured directivity as well as possible. In this project, the directivity of one or more rotationally symmetrical loudspeaker should be measured, in the anechoic chamber at NTNU, and some numerical method such as Least mean squares (LMS) should be used to find a set of monopoles that give the same directivity. The modeling should be applied to measured impulse response, that is, a multi-channel LMS algorithm should be implemented for this problem.

In many engineering optimization problems, the objective function is only available as the output of a simulation (black-box), that does not provide derivative information. Optimization methods in this case are often called Derivative-Free Optimization (DFO), only working on samples of the objective function.

In this task, the candidate will test various state-of-the-art DFO code on the «optimal gas-lift allocation» problem, arising in optimization of oil- and gas production. The task involves implementation in Python.

1. Literature review on derivative-free optimization
2. Test selected algorithms on a model of gas-lift optimization
3. Compare algorithms, also with a method requiring derivatives
4. Discuss, and write report

Co-supervisor: Knut Vannes, Lundin

Related to a cooperation project we have with two top Brazilian universities (in Florianapolis and Rio de Janeiro), we seek 1-2 adventurous students to do a combined project/master. The specialization project (fall) will be in Norway, while the master thesis work (spring) will be done in Brazil. The topic will be related to machine learning and control for production optimization in offshore oil and gas production, for instance 'physics-informed neural networks' used for data-driven control of gravity separators or electrical submersible pumps. Some financial support for the stay in Brazil is available.