Reliability of Modular Multilevel Converters

Abstract

This thesis focuses on advancing modular multilevel converters' (MMCs') reliability and fault tolerance. MMC is a popular converter in high-power applications due to its scalability, efficiency, and modular design. Despite their advantages, MMCs are susceptible to reliability issues, which can compromise continuous operation. To address these challenges, the research introduces a series of strategies and methodologies to enhance the resilience of MMCs through improved reliability assessments, optimized redundancy strategies, and fault-tolerant reconfigurability techniques.

The work begins with a comparative analysis of reliability assessment methods, including the Military Handbook, FIDES, and Mission Profile approaches. By examining the strengths and limitations of each, the thesis provides a foundation for selecting the most suitable method for evaluating MMC reliability in diverse operational contexts. This step is crucial as it addresses the challenge of accurately predicting MMC lifespan under variable conditions, paving the way for more reliable converter designs.

The Monte Carlo simulation framework is developed to evaluate the effectiveness of different redundancy strategies. This framework allows for modeling complex, real-world operational stresses and testing redundancy schemes. Through this approach, the thesis explores various configurations, such as Fixed-Level Active Redundancy and Standby Redundancy, demonstrating how each impacts overall reliability. The insights from these simulations provide a data-driven foundation for optimizing redundancy strategies tailored to specific application requirements.

The thesis proposes a cost-effective design methodology to address the trade-off between modularity, redundancy, and cost. By optimizing switch voltage ratings, this methodology balances capital and operational expenditures with reliability goals, allowing for scalable, modular designs without compromising durability. This approach reduces the initial costs associated with MMCs and ensures that performance standards are maintained under typical operating conditions.

To optimize the reliability of MMCs, a Mixed Redundancy Strategy (MRS) is introduced, combining active and spare redundant submodules (SMs) within the converter structure. This strategy offers an optimal balance between reliability improvement and cost containment. Through sensitivity analyses and simulations, the MRS is validated as a practical approach to enhancing MMC resilience, ensuring continuous operation despite faults without high redundancy costs.

Finally, the thesis develops a reconfigurability method enabling MMCs to operate continuously under fault conditions by dynamically bypassing faulty SMs or reconfiguring. This fault-tolerant reconfiguration technique is experimentally validated, demonstrating its effectiveness in detecting, localizing, and isolating faults in real-time. The proposed method provides a practical solution to enhance fault tolerance, enabling the MMC to maintain functionality with minimal impact on performance.

Overall, this thesis presents a comprehensive framework for advancing the reliability and cost-efficiency of MMCs through targeted improvements in reliability assessment, redundancy management, and fault-tolerant design. These contributions represent a significant advancement for power electronic systems where robustness and continuous operation are paramount, providing valuable insights and methodologies for developing resilient power converters.