My research interests are at the intersection of power electronics, constrained optimal control and optimization. Specifically, I am interested in electrical drives, grid-connected converters, model predictive control, hybrid systems and mathematical programming.
On the power electronics side, I focus on new control and modulation schemes for high power converters, including medium-voltage drives, FACTS and modular multilevel converters. The goal is to devise control and modulation schemes that during steady-state operation provide very low distortions per switching losses thus effectively resembling offline optimized pulse patterns, while providing a very fast controller response time during transients.
Three such approaches have been developed over the past years, namely Model Predictive Direct Torque Control (MPDTC) and its derivaties, Model Predictive Pulse Pattern Control (MP3C), and long-horizon finite control set MPC. In the past, I had also worked on switch-mode dc/dc converters and a few other application domains including flow control for cascaded river power plants, supermarket refrigeration systems and wide-area control schemes for power systems.
On the control side, I am interested in model predictive control (MPC), hybrid systems (particularly piecewise affine systems), and the underlying mathematical optimization problems. I have investigated techniques to reduce the online and offline computational burden of MPC through optimal complexity reduction and Lagrangian decomposition, and worked on modelling tools for piecewise affine systems. The Matlab code is included in the Multi-Parametric Toolbox (MPT), which is available free of charge under the GNU license.