Model predictive control of a current source inverter together with its current source

Over the past years a great amount of research has been done to interconnect different energy sources with the power grid or different types of loads. Current source inverters have proved to be a good option due to its high reliability, fault tolerant capabilities, quasi soft switching and the use of lower capacitor values among others. Most of the current source inverters topologies have a current source as input, which could be implemented with a buck converter, and they are controlled with cascaded linear control (usually PI controller), rotating frame coordinate transformation and a modulation stage. In this paper a predictive control strategy of a current source inverter together with its associated current source is presented. This strategy allows to track not only the output voltages at the load but also the current source within a single controller. The control algorithm makes use of a discrete time model of the whole system so as to predict its future behaviour for each one of the available switching combinations. Each one of the predicted values are used to minimize a set of predefined control goals within a multiple term cost function that includes cost associated with the commutation frequency and the reference tracking. Simulation results show a good behaviour and fast dynamics with a low switching frequency of all the switches involved. These characteristics make the proposed controller a suitable option to use with high power inverters.