Precise knowledge of magnetic flux-linkage is a fundamental requirement for high-preformance control of electrical machines. Conventional estimation methods − whether based on current look-up tables or voltage integration − are prone to significant errors arising from steep PWM voltage slopes, high-frequency components, and inverter-related switching delays, particularly in SiC and GaN systems. To address these limitations, this dissertation develops and evaluates novel volt-second sensor concepts that directly measure flux-linkage by combining continuous analog voltage sensing with FPGA-based digital processing. The work introduces six new measurement architectures, including feed-forward voltage-distortion modeling for IGBT-based inverters as well as asynchronous and synchronous voltage-to-frequency converter designs. These converters translate continuous phase-voltage signals into discrete volt-second values that can be processed in real time, overcoming the accuracy and implementation constraints inherent to purely analog or purely digital approaches. A comprehensive comparative study evaluates fourteen measurement methods under varying operating points on an industrial half-bridge test bench. The analysis covers measurement accuracy, instantaneousness, and cost, and includes application-oriented validation on a traction-drive test bench. The results demonstrate that the proposed sensor concepts enable precise and instantaneous determination of flux-linkage and corresponding phase voltage, while the underlying hardware can be flexibly adapted to different control algorithms through modifications in the FPGA-based evaluation logic. Overall, the dissertation establishes a unified measurement framework that leverages complementary analog and digital technologies to overcome long-standing challenges in PWM voltage sensing. The introduced volt-second sensors provide a robust basis for enhancing control accuracy in modern electric-drive inverters and open new avenues for future research and technological advancement.
Anne Gisela von Hoegen (Wed,) studied this question.