Reflected wave phenomenon in inverter-fed machines with fast-switching (high dv/dt) wide bandgap power converters
: causes, consequences and solutions

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

Power density, efficiency and reliability are key design drivers and central concerns for adjustable speed drive systems in a plethora of applications including industrial automation and robotics, and transportation systems. The commercially available of Wide Bandgap (WBG) power semiconductors such as Silicon Carbide (SiC) MOSFETs with extremely fast switching speed are pushing the boundaries of power converter performance to meet the aggressive power density and efficiency targets (e.g., 25kW/kg, 99% +) for existing and emerging applications.
However, these steep voltage transients (high dv/dt) of SiC inverters are expected to accelerate the degradation of the connected motor stator winding insulation, which significantly affects the reliability and lifetime of the motor drive system. In a typical SiC-based cable-fed motor drive system, the fast-fronted Pulse-Width Modulation (PWM) voltage pulses generated by the SiC inverter make cables act like transmission lines, with waves travelling along the cables back and forth. Since the characteristic impedance of the motor is much higher than that of the cable in the motor drive system, the inverter PWM voltage pulses experience reflections at motor terminals, leading to excessive overvoltage oscillations that can be twice the inverter voltage. This phenomenon is known as the reflected wave phenomenon (RWP). The resultant overvoltage stress adversely affects the reliability and lifetime of the motor by accelerating the motor stator winding insulation ageing through the inception of partial discharges, progressively yielding to the degradation of motor winding insulation, while raising the Electromagnetic Interference (EMI) problems.
The aim of this PhD thesis is to investigate the RWP in SiC-based cable-fed motor drive systems and explore the overvoltage mitigation techniques without compromising the benefits of SiC switching devices. The RWP is systematically investigated in both the time domain and frequency domain, providing the foundation for developing the active waveform shaping techniques. The impact of parasitic of SiC switching devices on the motor terminal overvoltage due to the RWP in a three-phase motor drive system is investigated and experimentally verified. Moreover, two active waveform shaping techniques, i.e., the Quasi-Three-Level (Q3L) PWM scheme and the voltage slew-rate (dv/dt) profiling, are proposed to address the motor terminal overvoltage oscillations in SiC-based long cable-fed motor drive systems. The proposed active waveform shaping techniques are supported by theoretical analysis and experimental verification. The essence of the overvoltage mitigation mechanism of the proposed active waveform shaping techniques is crystallized in both the time domain and frequency domain. The philosophy used here is addressing the motor terminal overvoltage oscillations at the source by actively mitigating the excitation source for the overvoltage oscillations. The theoretical and experimental results indicate that when the dwell time for the Q3L PWM scheme is set as 2t_p (t_p is the wave propagation time), the motor terminal overvoltage oscillations can be attenuated since there is no excitation source invoking the overvoltage oscillations. In addition, the proposed Q3L PWM using the SiC module-parallel inverter can extend the switching devices’ current capacity. The proposed voltage slew rate (dv/dt) profiling is implemented on the SiC Auxiliary Resonant Commutated Pole Inverter (ARCPI) to verify its effectiveness. The theoretical analysis and experimental results show that the ARCPI can entirely mitigate the motor terminal overvoltage oscillations when the rise and fall times of the PWM voltage pulses are shaped as 4t_p. Also, the voltage slew rate profiling inherits the advantages of the soft-switching inverter including high power efficiency and EMI performance.
Date of Award21 Jun 2022
Original languageEnglish
Awarding Institution
  • The University of Bristol
SupervisorXibo Yuan (Supervisor) & Phil H Mellor (Supervisor)

Cite this

'