Development and implementation of multi-output multilevel converters

• Main Author: Ahmed Salhin Taha Hussein
• Other Authors: Amer M. Y. M. Ghias
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Electrical and electronic engineering::Power electronics; Engineering::Electrical and electronic engineering::Industrial electronics
• Online Access: https://hdl.handle.net/10356/167984
• Institution: Nanyang Technological University

Future electricity grids will consist of a large-scale integration of both renewable and other distributed energy sources. This increases demand for higher power density and higher power quality power electronic converters. In this thesis, multilevel multi-output converters are explored as a potential solution. They provide several benefits over traditional converters, like a lower power component count and the ability to direct power flow between different power sources and loads. The state-of-the-art with respect to multilevel multi-output converter topologies are lacking in several domains. First, most available topologies have a limited valid operation range to control output terminals independently. Second, the available modulation and control methods are topology specific and are not adaptable to different topologies. Finally, the number of components used is still high to justify adopting such topologies. This thesis focuses on the development and design of new multilevel multi-output converters topologies and their modulation and control. It addresses the limitations of the state-of-the-art and provides several alternatives to each shortcoming. This includes developing a graphical framework to visualize multi-output converters limits, devising generalized modulation and control algorithms, and introducing new multi-output multilevel converter topologies. To understand the limitations and trade-offs involved in using multi-output converters a graphical analysis tool specific to dual-output converters (DOCs) is developed. It is based on the representation of the output voltages of the single-phase converter switching states. The limitations in question are regarding the valid modulation indices and the maximum phase shift of the reference voltage signals. The proposed approach offers a visual representation of dual-output converters operation limits and provides basis for ensuring their safe and secure operation. The theoretical analysis is validated by simulation and hardware experiments. Generalized modulation and control schemes have been developed to facilitate the adoption of multilevel multi-output converters and provide a basis from performance comparison. The work done in this aspect progressed over three stages. First, adopting existing modulation and control techniques like Model Predictive Control. In this stage, MPC was used to demonstrate its superiority over conventional pulse-width modulation (PWM) techniques. Second, improving upon PWM through cost function formulation that targets capacitor voltage balancing and power losses. Finally, devising a new generalized Space Vector Modulation (SVM) technique for existing and proposed multilevel multi-output converters. The uniqueness of the proposed SVM lies in its ability to incorporate conventional SVM algorithms to be used with multi-output converters. Lastly, this thesis introduces four new multilevel multi-output converter topologies with varying features in configurability, size, operation range, and output quality. The Modified Cascaded Multi-Output Multi-level converter can be configured with one or multiple dc sources while being restricted to single-phase systems and a limited operating range in terms of dc voltage utilization. The T-type DOC can be operated in three-phase and has a reduced number of switching devices but still has a limited operating range. The Neutral-Point Clamped DOC is a three-phase converter with a full operation range but a low output power quality.