Accurate electrical and thermal modelling of power semiconductors for transient simulation

Li, Chen 2025. Accurate electrical and thermal modelling of power semiconductors for transient simulation. PhD Thesis, Cardiff University. Item availability restricted.

	PDF (Cardiff University Electronic Publication Form) - Supplemental Material Restricted to Repository staff only Download (190kB)
	PDF - Accepted Post-Print Version Restricted to Repository staff only until 12 August 2027 due to copyright restrictions. Download (5MB)

Abstract

Semiconductor device modelling is a critical step in the design and optimisation of power electronics systems. Its primary objective is to accurately characterise the electrical, thermal, and dynamic behaviours of devices using high-precision modelling techniques. The rapid advancement of power semiconductor technologies has introduced several new challenges to device modelling, including the complexity of package parasitic inductance, chip-level dynamic effects, and thermal behaviour. Consequently, to ensure both high accuracy and wide applicability, this thesis proposes a unified modelling methodology that addresses all three challenges. For each aspect, the device most sensitive to the corresponding phenomenon is selected for targeted experiments and simulations. To address the challenge of accurately modelling the parasitic inductance of packages, this thesis proposed a non-destructive extraction method tailored for high-speed power devices. High-speed SiC MOSFET modules are typically designed with tightly coupled commutation paths to minimise effective loop inductance. Conventional two-port extraction methods that neglect mutual inductive coupling and fail to reproduce the actual switching current paths cannot accurately capture the effective loop inductances of such packages. The proposed method reduces the maximum relative error from 60.2% with the conventional approach to only 0.9%. The inductance model is further validated through 650V double-pulse experiments, demonstrating high modelling fidelity. The dynamic switching behaviour, characterised by voltage–current transitions and tail current effects, influences device performance under varying operating conditions. To enable accurate simulation of these behaviours, this thesis developed a compact behavioural model for dynamic switching. Among typical device types, the IGBT is chosen as the representative platform because it exhibits comprehensive dynamic characteristics. The model is implemented in Verilog-A, allowing efficient integration into circuit simulators with low computational overhead. Comparisons of the measured and simulated waveforms show good agreement in current and voltage slopes, while the turn-on, turn-off, and total switching-energy errors remain below 10% across the tested operating conditions. Accurate thermal characterisation is essential for modelling power semiconductor devices, as it underpins reliable electrothermal predictions and safe operation. In particular, accurate modelling of the early thermal response requires the measurement blind window to be minimised so that the most informative part of the transient is not lost. This thesis, therefore, focuses on GaN HEMTs as a demanding case study, because self-heating is highly localised in the 2DEG active region and a substantial proportion of the useful thermal information is concentrated in the early microsecond range. A temperature-sensitive electrical parameter (TSEP)-based thermal characterisation methodology is proposed to reduce the measurement blind window and capture these early temperature variations. The proposed method is validated through finite element method (FEM) simulations and further corroborated by infrared thermography and power-tester measurements. The results show that the method can reproduce the early thermal transient response. In summary, the proposed methodology ensures accuracy, scalability, and applicability across device types, supporting the design and optimisation of modern power electronic systems.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Schools > Engineering
Uncontrolled Keywords:	1. Power semiconductors 2. Electrothermal modelling 3. Wide-bandgap (WBG) 4. Parasitic inductance 5. Thermal networks
Date of First Compliant Deposit:	12 August 2026
Last Modified:	13 Aug 2026 09:01
URI:	https://orca.cardiff.ac.uk/id/eprint/188933

Actions (repository staff only)

Edit Item