Nonlinear Model Predictive Control for Modular Multilevel Converter- based High-Voltage Direct-Current Systems

• Applicant: Professor Dr.-Ing. Christoph Hackl
• Subject Area: Electrical Energy Systems, Power Management, Power Electronics, Electrical Machines and Drives ; Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 510267132

This project aspires to develop a high- and a low-level MPC algorithm for MMC-based HVDC systems. Specifically, the high-level MPC - which will replace the conventional PI controllers to impose upper and lower bounds on the stored energy - will control the MMCs in a coordinated manner to jointly manage the energy storage of the module capacitors. The low-level MPC will be localised at each converter aiming for favourable steady-state and transient operation. The combination of both controllers will allow for optimal performance of the controlled system over the whole operating regime (including non-ideal conditions; such as arbitrary grid faults), while it is being operated at its physical limits. This will allow for reduction of the energy storage elements, i.e., module capacitors, resulting in a more cost-efficient system. To summarise, the combined high- and low-level MPC algorithms aim to achieve the following objectives: (O1) achieve favourable steady-state performance: The aim is to produce grid currents and voltages that will comply with the relevant grid codes, without requiring line filters and while operating at a switching frequency as low as possible. Hence, the aim is to improve the well- known trade-off between harmonic distortion and switching frequency (or switching losses). (O2) increase the controller bandwidth to quickly deal with grid disturbances and faults: The goal is to ensure a continuous operation of the system by guaranteeing immunity to a wide range of grid disturbances, such as voltage dips and frequency variations, voltage imbalances and harmonic distortions as well as grid faults. That way, conditions such as the above-mentioned can be compensated for very quickly, and nuisance trips, halts and/or damage due to over-currents are avoided. (O3) reduce the cost of the overall converter system: The coordinated —and thus intrinsically more effective and efficient— constrained control of the converters, along with the constraints on the converter level imposed by the high- and low-level controllers, will allow for minimisation of the power demand. As a result, the storage elements can be reduced in size. Moreover, the absence of line filters will not add extra cost to the system. (O4) reduce the computational burden in order to allow real- time implementation: By appropriately formulating the high- and low- level control optimisation problems of the underlying MPC, the problem can be cast as a convex quadratic programme and integer polynomial programme, respectively, which will allow for adoption of computationally efficient real-time solvers to guarantee an implementation and execution in real-time.

• DFG Programme: Research Grants
• International Connection: Finland
• Major Instrumentation: Dreiphasige Netzsimulationseinrichtung
• Instrumentation Group: 6430 Digital anzeigende Meßgeräte für Spannung, Strom, Widerstand
• Cooperation Partner: Professor Dr.-Ing. Petros Karamanakos