A Model-Free Predictive Current Controller for Voltage Source Inverters

Abstract

This work proposes a model-free deadbeat predictive current controller using the input and output data of the controlled system. The proposed controller can be used for electric drives or grid-forming inverters. The proposed control scheme does not require knowledge of the system parameters for its correct operation; thus, all load parameter mismatch issues in conventional model-based controllers are avoided. This is performed using an auto-regressive with exogenous input (ARX) model to represent the physical system with a recursive least square (RLS) estimation algorithm for the ARX model parameters identification. Using the identified system parameters, an observable canonical state-space model is constructed. Based on this system, the inverter voltage references using the space vector modulation (SVM) are calculated using the deadbeat control principle. The inherent sensibility of the deadbeat controller to parameters mismatch is avoided. Experimental results confirm the validity of the proposed Model-Free controller in terms of tracking accuracy and waveform quality.