The outstanding features of modular multilevel converters (MMC) have recently gained much attention in the high-voltage direct-current (HVDC) transmission field. Power quality, converter cost and system performance are three crucial aspects of HVDC MMCs which are directly linked to the converter modulation and switching schemes. High power quality and performance require high switching frequency and large cell capacitor whereas low switching frequency and small cell capacitor are needed to reduce the converter cost.

The main objective of this thesis is to propose a practical switching method for HVDC MMCs which balances the aforementioned contradictory requirements. A mathematical analysis of the converter switching pattern, against the power quality and converter cost, has been conducted to formulate an optimization problem for MMCs. Different objective functions are studied for the formulated optimization problem such as converter loss minimization, voltage imbalance minimization and computational burden minimization. This thesis proposes three methods to address different objective functions. Ultimately, a real-time simulator has been built to practically verify and investigate the performance of the proposed methods in a realistic point-to-point HVDC link.

The most significant outcome of this thesis is the tolerance band-based switching scheme which offers a direct control of the cell capacitor voltage, low power losses, and robust dynamic performance. As a result, the converter switching frequency can reach frequencies as low as 70 Hz (with the proposed cell tolerance band (CTB) method). A modified optimized CTB method is proposed to minimize the converter switching losses and it could reduce the converter switching losses by 60% in comparison to the conventional phase shifted carrier modulation method.

It is concluded intelligent utilization of sorting algorithm can enable efficient HVDC station operation by reducing the converter cost.