Recursive Least Squares based Dynamic Model Predictive Control for Grid-Tied Active Front End Converters

The classical current control techniques for grid connected LCL type active front-end converters (AFEs) are usually interacts with the phase locked loop (PLL) nonlinearities and grid impedance, thus deteriorates the stability of the converter which leads to system instability. In order to avoid the stability issues, this manuscript provides a simpler and much better recursive least square based dynamic model predictive current control technique to approximate the attributes of physical model to their true value. In this control structure the predicted variables are evaluated using the cost function at each sample interval and the most efficient switching state with the lowest cost function value is chosen. According to the estimated switching states, the gate pulses are generated for the AFEs switches. The performance of the proposed control structure is tested using simulation platform under a widespread range of variations in grid impendence and system's nonlinearities. Under these conditions, the suggested method which considers the system's nonlinearities provides outstanding resilience in high and medium frequency grid dynamics. Furthermore, the simulations results have been varied through hardware platform using Typhoon HIL to show the advantages of proposed control over classical controls.