Predictive Control of Power Converters and Electrical Drives

The electronics group has been pioneering this research line since 2004 in the power electronics community. It has been recognized as a promising and powerful tool to control power converters and drives. The reason for this lies in the nonlinear and discrete nature of power converters, which have limited control actions (switching states). This enables to evaluate predictive models of the system over a one or more future time intervals to select the most appropriate control action according to a predefined cost function. The cost function can hold any mathematical term that can model variables, constraints and non-linarites of the system, even in the frequency domain. This flexibility brings many powerful control tools to power electronics, which were before limited to a few degrees of freedom provided by classic linear controllers with modulation stage. Our group has embraced this potential and applied predictive control to a wide variety of power converters, motor drives and applications. A major breakthrough has been the inclusion of different dynamics (electrical and mechanical) into a single predictive control algorithm, eliminating the use of any linear controller. The research carried out by our group in the field of MPC of power converters and drives covers the following topics: current control of three-phase, multi-phase and multi-level inverters; predictive power control, predictive torque and speed control of AC machines, and control of active filters, among others.