Design and Performance Optimization of Power Converters for Energy Storage Systems

Abstract

Energy shortages and power outages are key concerns today, exacerbated by rising energy demands and the global imperative for clean energy and decarbonization. Energy storage systems (ESS) traditionally address both long-term and short-duration outages, with renewable energy sources and suitable power converter interfaces (PCI) tackling the shortages. While battery-based ESS (BESS) are commonly used for short-term blackouts, ultracapacitor (UC)-based ESS are increasingly preferred due to their higher power density and superior lifecycle characteristics. Although large-scale BESS are still being deployed, hydrogen-based fuel cells and electrolysers are being explored for long-duration needs owing to their higher energy density. A hybrid strategy that integrates hydrogen-based systems with UC storage, each optimized for their respective dynamic ranges along with appropriate PCIs, offers a promising solution to span a broader frequency spectrum while improving overall system efficiency. This thesis aims to achieve twin broad objectives: to optimize the design of an ultracapacitor stack along with its associated bidirectional half-bridge power converter for contingency requirements; and to develop efficient power converter interfaces, specifically a phase-shifted full-bridge (PSFB) converter, for high-power, long-duration applications such as electrolysers and hydrogen-based fuel cell systems. In the UC-based ESS, the inherent non-linear behaviour of ultracapacitors is analysed, leading to the development of a framework for accurately characterizing the effective stack capacitance and voltage profiles. This framework is used to propose a systematic design procedure that optimizes the discharge ratio and iteratively selects stack parameters, thereby minimizing overall system cost while meeting performance specifications. An additional stability check is included, considering the maximum power transfer constraint imposed by the equivalent series resistance. Further, the research investigates the PSFB converter for both low and high-power applications. A detailed analysis of the topology is presented, examining the influence of key circuit parameters such as parasitic inductances and capacitances on converter behaviour and design trade-offs. This culminated in the development of a two-level, loss-optimal, iterative design algorithm that yields a unique set of design parameters across a wide specification range from the design space. For high-power applications, a modular system of PSFB converters configured in an input parallel output-parallel (IPOP) topology is explored. The limitations of conventional equal power-sharing schemes are identified, and the sensitivity of load sharing to design parameter variations is studied. Based on this, the work proposes an asymmetrical module design with two design-sets, coupled with a Lagrangian-based loss-optimal load-sharing control strategy to ensure high-efficiency and achieve soft-switching across the entire load range. The proposed models, analyses, and design algorithms for the ultracapacitor stack and PSFB converter, including their modular configurations, are validated through experimental results on 1-3 kW hardware prototypes