Model Predictive Control for Shunt Active Filters With Fixed Switching Frequency

This paper presents a modification to the classical Model Predictive Control (MPC) algorithm and its application to active power filters. The proposed control is able to retain all the advantages of a finite control set MPC while improving the generated waveforms harmonic spectrum. In fact, a modulation algorithm, based on the cost function ratio for different output vectors, is inherently included in the MPC. The cost function-based modulator is introduced and its effectiveness on reducing the current ripple is demonstrated. The presented solution provides an effective and straightforward single loop controller, maintaining an excellent dynamic performance despite the modulated output and it is self-synchronizing with the grid. This promising method is applied to the control of a shunt active filter for harmonic content reduction through a reactive power compensation methodology. Significant results obtained by experimental testing are reported and commented, showing that MPC is a viable control solution for active filtering systems.