Abstract:
A two-level voltage source inverter (TLVSI) is commonly used to feed AC load in the automotive, aerospace, microgrid, etc. A conventional TLVSI is designed using Si devices due to its
wide accessibility and low cost. In comparison, the modern TLVSI is designed based on SiC devices due to: (a) the ability to switch at high frequency, (b) availability of the device for higher
voltage/current rating, and (c) high thermal conductivity.
The PWM operation in TLVSI generates the CM and DM noise. The high-frequency CM
noise excites the line to the ground parasitic capacitance path, leading to the propagation of CM
current or ground leakage current. In contrast, the high-frequency DM voltage excites the stray
capacitance between the two or more lines and causes the leakage DM current. The propagation of
these currents can: (a) degrade the motor and cable insulation and (b) cause conducted & radiated
electromagnetic interference (EMI) issues in the system. An electronic appliance is more sensitive
to CM noise than DM noise. The EMI noise can degrade the performance of the electronic system
(like navigation, communication, etc.) in electric vehicles, aircraft, etc. Therefore, the attenuation
of these EMI noises are significantly essential.
The complete elimination of the CM noise is unattainable in a TLVSI. A sinusoidal pulse
width modulation (SPWM) and space vector PWM (SVPWM) are used to drive a TLVSI fed
induction motor drive. An active zero vector PWM (AZPWM) is used to minimize the CM voltage
levels by avoiding the zero vectors. However, the implementation does not guarantee the complete
elimination of the CM noise. Therefore, the design of passive / active CM attenuation methods
has to be explored with reduced CMV PWM methods.
The passive CM noise attenuation method uses a combination of passive elements, viz., inductor, capacitor, and resistor, to attenuate the noise. The passive component are economical, reliable
and easy to adopt/design. Therefore, the passive CM noise attenuation methods are widely used
for noise mitigation in power converters and drives. The passive attenuation methods are often
bulky, and its CM inductor can account for up to 25 % of the total drive volume. The impact of
the AZPWM on minimizing the size of the passive CM noise attenuation method is investigated in
this thesis. To have a similar attenuation in the considered PWM cases, AZPWM requires a lesser
value/size of the passive components compared to SVPWM. Thereafter, a CM impedance reshaping choke is designed to increase the CM noise in the switching frequency region. Alternately, an
active CM noise attenuation method is designed to minimize the power loss in the active circuitry
and lessen the size of the CM transformer employed in active CM noise voltage canceler.A conventional single-stage EMI filter configuration is widely used to mitigate the CM and
DM noise with a sinusoidal output waveform. This conventional EMI filter has a drawback in
terms of the increased volume, weight, and cost of the CM inductor. The aforementioned drawback
is addressed with the modified single-stage and multistage EMI filter topologies and design.
The total DM noise in a system is categorized as intrinsic DM (IDM) and non-intrinsic DM
(NDM) noise. The analysis and measurement method is proposed to separate the IDM and NDM
from the total DM noise. The passive sinusoidal output EMI filter (PSOEF) is designed by eliminating the need for a CM inductor to attenuate the NDM and CM noise. The attenuation performance, circulating current, and filter loss of the designed PSOEF are analyzed with SVPWM and
AZPWM. It is identified that the PSOEF with AZPWM offers low circulating current and filter
loss in comparison to PSOEF with SVPWM.
This thesis investigates the CM and DM noise attenuation methods to reduce the volume,
weight, and cost of the CM inductor in the overall motor drive system. The attenuation methods
are experimentally verified on Si/SiC-based TLVSI.