Energy Cycle Optimization for Power Electronic Inverters and Motor Drives

Autor: Haque, Md Ehsanul
Jazyk: angličtina
Rok vydání: 2022
Předmět:
Druh dokumentu: Text
Popis: This research proposes low-frequency energy cycle optimization-based control algorithms for reducing the DC-link capacitance requirement and a high-frequency energy cycling-based regenerative output dV/dt filter for motor drives. The proposed methodologies improve the volumetric power density, efficiency, and reliability of the motor drives. Three low-frequency energy cycle optimization methods are proposed for the switched reluctance machine (SRM) drive to reduce the DC-link current ripple and DC-link capacitor requirements. The first method is based on an additional bi-directional converter-based active power decoupling stage, which processes the low-frequency content of the inverter’s input power ripple. A wide voltage swing of the decoupling capacitor reduces the decoupling capacitor requirement. With this topology, the DC-link capacitor is used only to filter out the switching frequency ripple. Therefore, the capacitor requirement in the drive is significantly reduced. The second method reduces the low-frequency content of the DC-link current ripple in the SRM drive by injecting low-frequency harmonics on the phase currents. This control algorithm uses the ripple components of the DC-link current and generates an additional current reference on top of the original current reference. The controller ensures low-frequency ripple energy circulation between the phases of the electric machine and reduces the pulsating energy requirement from the DC source while maintaining the average torque, torque per ampere, and torque ripple performance. Thirdly, a phase collaborative interleaving fixed switching frequency predictive current control method is proposed where energy cycles for both incoming and outgoing phases are collaboratively together in each switching cycle. This technique moves the dominant low-frequency energy concentration towards the high-frequency region. The new energy spectrum is concentrated around twice the switching frequency (2fsw) and reduces the DC-link capacitor requirement significantly. The phase collaborative interleaving method would enhance the performance phase current shaping, or it can work independently. A finite element analysis (FEA) is performed to characterize the SRMs having 18 slots/12 poles for building a dynamic simulation model in MATLAB/Simulink. Detailed simulation results are provided to validate all the DC-link current ripple reduction methods. Moreover, the second and third methods are experimentally verified with Silicon Carbide (SiC) -based 58 kW and 32kW SRM drive setups coupled to a 150kW dynamometer for testing under various operating conditions. All the DC-link control algorithms require a higher current control bandwidth and are therefore implemented with WBG devices. However, WBG devices show high dV/dt, which degrades the performance of the motor insulation and bearing. A regenerative inverter output filter topology is proposed for WBG device-based motor drive system. This work eliminates the requirement of a passive resistor in an LCR filter topology with a GaN-based full-bridge converter. The proposed unit mimics the operation of the passive resistor by cycling the energy of the filter path. The proposed active filter only processes the high-frequency power, so the GaN device's current ratings are significantly lower compared to the main converter's switches. This lossless active resistor allows the use of a lower filter inductor in the series path, reducing the overall filter size. As a result, the proposed regenerative lossless dV/dt filter minimizes the stress on the motor insulation and improves the motor drive system's efficiency and power density. Detailed MATLAB/Simulink-based simulation results are provided to validate the performance of the proposed filter topology. An experimental double pulse test (DPT) results have been performed to validate the operation of the proposed concept.
Databáze: Networked Digital Library of Theses & Dissertations
Externí odkaz: