부산대학교 도서관

The modular multilevel converter is a good alternative for high-power, medium-voltage drive applications due to its modularity and scalability. However, the control is complex and typically involves several highly coupled circulating current feedback loops implemented using single-input single-output design tools. In addition, each circulating current has many different frequency components and electrical sequences to increase the degrees of freedom in the controller. In this work, the use of continuous control set model predictive control is proposed in order to include cross-coupling and interactions between the state variables and to consider the system constraints, such as maximum current and maximum output voltage. The controller is intended for drive applications and is designed to operate with good dynamic performance over the entire speed range of the machine. The control methodology proposed is experimentally validated using an 18-cell MMC prototype driving a cage machine. Additional experimental tests are performed using PLECS-RT HIL platforms.