Control Methods for Multilevel Converters

This thesis aims to develop new control methodologies for the Packed U-Cell (PUC) converter. The main problem lies in the structure of the PUC converter, where it is a hybrid type of multilevel converter (MLC) similar to cascaded H-bridge converters. From the control prospective, the topology has discrete digital control inputs, thus conventional modulation techniques cannot be directly applied to the converter. Proper control methods have to be introduced in order to make use of the great advantages of the PUC converter. These control methods have to be simple in implementation for industrial environment with low switching frequency. Additionally, the newly proposed control methods should have advantages over the existing ones. In the first attempt of this thesis, a modified version of model predictive control (MPC) has been integrated to control grid connected PUC inverters. Mainly, the cost function is formulated to guarantee the stability of the controlled system. Then, the gains associated with the controlled variables are eliminated in order to simplify the control. The proposed controller has shown great features in terms of stability and low average switching frequency. In the second attempt of this thesis, the interesting features of sliding mode theory are utilized to serve the control problem in a better way. The proposed controller is very simple to implement with low computation time, which reduces the computational burden on the controller. The controller aims to allocate the control input which stabilizes both the grid current and the auxiliary capacitor voltage in the grid connected PUC inverter. Lastly, the control problem of the PUC rectifier with a dual output is targeted using Lyapunov-based MPC. This attempt is done in order to prove the workability of the controller for three control variables. In this method Lyapunov-based MPC doesn’t require gain tuning as in the PUC inverter problem. Load current measurements are eliminated by making use of the mathematical model of the PUC converter. This significantly reduces the cost and simplifies the control algorithm, where no current sensors are used for the load measurements. For all proposed methods, simulations and experimental tests were done in order to justify the correctness of the proposed methods. Dynamic tests and parameter mismatch tests are carried out to measure the controller response and to show the robustness feature of the proposed controllers. Keywords: Finite control set, Lyapunov control, model predictive control, multilevel converter, packed U-Cell converter, sliding mode control.