Design, Analysis and Optimisation of DC-DC Converters for Fuel Cell Hybrid Vehicles

  • Noass Kunstbergs

Student thesis: Doctoral Thesis

Abstract

The impact of climate change on our environment is rising, which is why several countries aim for ambitious climate targets to reduce greenhouse gas emissions. Because of this trend, automobile manufacturers focus on CO2-neutral vehicles for urban traffic, such as the fuel cell hybrid vehicle that combines battery and fuel cell in the powertrain. For the DC bus of the powertrain, it is advantageous to connect this battery with a bidirectional DC-DC converter for electrical performance. The design of this DC-DC converter is the focus of this study. Generally, there are several steps to take for converter design. Among others, the topology is selected, suitable power semiconductor material is identified, and methods are investigated to enhance converter efficiency and power density. Unfortunately, published literature often presents complex topologies without considering the practicability of implementation. Furthermore, it is frequently reported that wide bandgap power semiconductors are superior for every application. However, standard design procedures use unreasonable switching frequencies for wide bandgap applications. Authors usually select switching frequencies without trade-off consideration for the impact on converter losses and volume. Certainly, wide bandgap materials reduce the volume of filter components using high switching frequencies. However, these materials currently have several challenges, such as low availability and high price. Further, the design of a converter with these materials can be quite challenging due to the fast switching transients. These fast switching transients worsen EMI and DC bus ringing. Regarding converter efficiency, complex approaches require parameters not available in datasheets or additional tests for loss characterisation; and are often too high in their effort to be feasible. To conclude, several questions remain to be answered. For example, whether a simple converter topology could achieve high efficiencies over the entire power range with a reasonable power density. This Thesis aims to find the reasonable power range of power semiconductor materials by optimum design procedures. In addition, the Thesis investigates if the enhancement of converter efficiency is possible in a simple way without relying on too many parameters. This way, this Thesis contributes to engineers facing issues in the early design stage that do not allow complex approaches, which is the limited amount of parameters in datasheets for loss estimations. The Thesis presents iterative approaches for converter design. The reason for this methodology is the trade-off consideration. As magnetic cores contribute The impact of climate change on our environment is rising, which is why several countries aim for ambitious climate targets to reduce greenhouse gas emissions. Because of this trend, automobile manufacturers focus on CO2-neutral vehicles for urban traffic, such as the fuel cell hybrid vehicle that combines battery and fuel cell in the powertrain. For the DC bus of the powertrain, it is advantageous to connect this battery with a bidirectional DC-DC converter for electrical performance. The design of this DC-DC converter is the focus of this study. Generally, there are several steps to take for converter design. Among others, the topology is selected, suitable power semiconductor material is identified, and methods are investigated to enhance converter efficiency and power density. Unfortunately, published literature often presents complex topologies without considering the practicability of implementation. Furthermore, it is frequently reported that wide bandgap power semiconductors are superior for every application. However, standard design procedures use unreasonable switching frequencies for wide bandgap applications. Authors usually select switching frequencies without trade-off consideration for the impact on converter losses and volume. Certainly, wide bandgap materials reduce the volume of filter components using high switching frequencies. However, these materials currently have several challenges, such as low availability and high price. Further, the design of a converter with these materials can be quite challenging due to the fast switching transients. These fast switching transients worsen EMI and DC bus ringing. Regarding converter efficiency, complex approaches require parameters not available in datasheets or additional tests for loss characterisation; and are often too high in their effort to be feasible. To conclude, several questions remain to be answered. For example, whether a simple converter topology could achieve high efficiencies over the entire power range with a reasonable power density. This Thesis aims to find the reasonable power range of power semiconductor materials by optimum design procedures. In addition, the Thesis investigates if the enhancement of converter efficiency is possible in a simple way without relying on too many parameters. This way, this Thesis contributes to engineers facing issues in the early design stage that do not allow complex approaches, which is the limited amount of parameters in datasheets for loss estimations. The Thesis presents iterative approaches for converter design. The reason for this methodology is the trade-off consideration. As magnetic cores contribute significantly to overall converter volume, the study introduces a way to estimate core volume for a switching frequency range. This range is then used for the iterative design procedures to identify the optimum switching frequency for the DC-DC converter. An analysis of silicon and silicon carbide power semiconductors is carried out by this optimum design, while availability is also considered. By doing so, this Thesis identifies the power semiconductor materials’ reasonable power range. In order to address the issue of limited data and complexity for improvement in converter efficiency, this study focuses on the power loss distribution of all components (power semiconductors, inductors, and capacitors) of the selected DC-DC converter topology. By straightforward loss analysis, the approach balances the power loss distribution using just datasheet parameters without any additional hardware and moderate computational effort. This balancing is achieved with a switching frequency modulation method with powder cores and their load-dependent inductance. To validate the proposed procedures and methods, this Thesis introduces two 27 kW bidirectional DC-DC converters with silicon and silicon carbide power semiconductors. The requirements for these converters are based on a literature review and are appropriate for fuel cell hybrid vehicle applications. Experimental tests demonstrate that the optimum design with the proposed procedures enables similar efficiencies for silicon based and silicon carbide based DC-DC converters. Accordingly, the following proves that the optimum switching frequency for the silicon based converter is 20 kHz and that of the silicon carbide based one 60 kHz. Furthermore, the proposed modulation method adjusts the switching frequency to avoid discontinuous conduction mode while considering power loss balancing between power semiconductors and the inductor. By doing so, this method significantly reduces overall converter losses. As a result, the silicon based converter reaches a peak efficiency of up to 98.99% and the silicon carbide based converter 98.78%, reducing power losses by up to 30% compared to a fixed switching frequency implementation.
Date of Award2 Feb 2024
Original languageEnglish
SupervisorNigel Schofield (Main Supervisor)

Cite this

'