Modeling, Analysis and Control of DC Hybrid Power Systems.

Abstract

All electric ships are featured with integrated power systems which combine electric propulsion technology with heterogeneous power generation and distribution technologies to form one single electrical platform. The auxiliary and main power generation system form an isolated hybrid power system to feed the ship service loads and to meet the propulsion power requirement. Although for decades, the methodologies for power converter control have been explored in many publications, the modeling, analysis, and control of hybrid power systems with multiple power converters remains an interesting open problem, leading to its exclusive focus in this dissertation.

Along with the opportunities introduced by hybrid power systems, the inter-connectivity and complexity represent a major system analysis, design and optimization challenge, calling for the development of effective tools. Therefore, a comprehensive testbed is developed. Moreover, component level modeling, analysis and modulation strategy development are performed to ensure system level performance. A new power flow model for the dual active bridge converter is derived. The new model provides a physical interpretation of the observed phenomena and identifies other characteristics that are validated by experiments. To overcome the drawbacks of traditional modulation strategies, a novel modulation strategy is developed for the dual active bridge converter. The experimental results verified that, if the new strategy is used to modulate the dual active bridge converter, this testbed can be used as an effective tool for optimal power management algorithm development for the hybrid power systems.

The development of advanced control algorithms, together with the increased computational power of microprocessors, enables us to deal with the control problem from a new perspective. In this dissertation, the voltage regulation problem for a full bridge DC/DC converter is formulated as both a linear and a nonlinear Model Predictive Control (MPC) problem with a nonlinear constraint that captures the peak current protection requirement. The experimental results reveal that both the MPC algorithms can successfully achieve voltage regulation and peak current protection. The successful implementation of the MPC schemes on the full bridge DC/DC converter paves the way for future system-level advanced control algorithm development for hybrid power systems.