PROJECT TITLE: Model Predictive Control of Multilevel Converters in Multi-Terminal DC Networks
PhD period: 2017.03.01 - 2020.02.29.
Section: Power Electronic Systems
Research Programme: Efficient and Reliable Power Electronics
Supervisor: Tomislav Dragičević
Co-Supervisor: Frede Blaabjerg
Collaborator: Office of Naval Research.
Funding: Aalborg University.
The main objective of the project is the design and implementation of the MPC algorithm for the 3-level NPC converter whose purpose would be to connect the AC grid layer to multi-terminal DC (MTDC) network, a part of the future power systems. MTDC network consists of individual DC links and transmission lines that are being merged with the idea to improve the system performance and reliability. The main objective of the converter is the regulation of the DC voltage and harmonic power exchanges with the AC power system, but also to react rapidly to adverse conditions in the power system in order to protect itself. Secondary objective includes loss minimization, active thermal management and coordination with the other converters in the network. To fulfil all the mentioned converter objectives, the converter control algorithm should be based on the MPC algorithm. Semiconductor switches are operating in two states: cutoff and saturation which means that the number of possible switching combinations for the NPC converter is a finite number. The references of currents, voltages, power, commutation losses, switching frequencies, thermal management, loss minimization etc. can all be included in the cost function of the MPC which will evaluate the difference between the actual values and the references. The possibility to include so many additional terms to improve the behavior of the converter gives the MPC a high flexibility which is difficult to achieve using linear controllers.
Project activities are expected to result with methods for local control of VSC (with inclusion of both primary and secondary functions) and system level control where each VSC will be treated as an agent that belongs to a multi agent system. This research will add great value to the understanding of design and implementation of MPC on parallel connected multilevel power converters in MTDC network.
Publications in journals and conference papers can be found at VBN.