Supervisor:
Zhe Chen

Co-Supervisor:

Assessment Committee:
Sanjay Chaudhary (Chair)
Kai Strunz, Technische Universität Berlin
Tomi Roinila, Tampere University

Moderator:
Sanjay Chaudhary

Abstract:

Environmental concerns and the depletion of fossil fuel resources have spurred the development of various renewable energy sources, such as wind and solar power. As the penetration of renewable energy sources increases, the traditional power grid is gradually evolving from an AC-dominated network to a DC microgrid. In this context, the bidirectional isolated DC-DC (BIDD) converter emerges as a critical component in connecting and integrating renewable energy sources within DC microgrids. Among BIDD converters, the dual active bridge (DAB) converter stands out as a preferred option due to its advantages of galvanic isolation, high efficiency, and power density, which facilitate efficient power transfer and enhance the reliability and stability of the energy system. Consequently, the modeling and performance optimization of BIDD converters are crucial for improving DC microgrid operation and reliability.

Traditional methods for eliminating DC bias in DAB converter inductors often overlook passive parameters, hindering the effective elimination of DC bias under specific conditions. To address this challenge, Chapter 2 proposes a novel DC bias suppression method for DAB converters based on triple-phase shift (TPS) modulation. An accurate equivalent circuit model is established by incorporating the equivalent series resistance (ESR) of the inductor. Theoretical analysis elucidates the relationship between transient variables and passive parameters, including ESR, inductance, and switching frequency, providing insights into effective DC bias suppression.

Furthermore, Chapter 3 focuses on the dynamic performance of three-phase dual active bridge  (DAB3) converters, addressing the issue of DC bias in the inductor current. The complex structure of DAB3 converters and multiple control variables in variable duty cycles (VDC) modulation present significant challenges in analysis and deduction. To overcome these challenges, this chapter proposes an artificial neural network (ANN) method to suppress DC bias current without human intervention. An optimal objective function is established by analyzing the converter's behavior, focusing on the deviation of the peak inductor current between the transient and steady states. Using PLECS software, optimal datasets of transient variables are obtained via particle swarm optimization (PSO). These datasets are used to develop ANNs, providing data-driven models to effectively eliminate DC bias and reduce transient current stress.

Finally, Chapter 4 focuses on the modeling and control of the DAB-based input series and output parallel (ISOP-DAB) converter, which is a preferred choice for DC transformers in DC microgrid. To enhance the performance of ISOP-DAB converters, the TPS modulation strategy is incorporated into the converter. However, the multi-mode and multi-variable characteristics of TPS modulation present significant challenges for modeling and control. This chapter conducts a frequency domain analysis of the ISOP-DAB converter using optimized TPS modulation to manage current stress, considering the variability of passive components. The inherent multi-input multi-output (MIMO) nature of the ISOP-DAB converter is addressed with a decoupling algorithm. Additionally, the chapter develops a tailored control scheme to ensure robust phase margin and bandwidth across the full power range.

Overall, the results of this thesis collectively enhance the dynamic performance, reliability, and stability of DC microgrid systems, driving the evolution of modern power systems towards greater sustainability.