Adaptive Determination of Optimum Switching Frequency in SiC-PWM-Based Motor Drives: A Speed-Dependent Core Loss Correction Approach

This study focuses on identifying the optimal switching frequency for silicon-carbide (SiC)-based motor drives across a wide range of operating conditions using a loss minimization strategy. The results are then compared with those of traditional silicon-insulated-gate bipolar transistor (IGBT) systems. The approach involves conducting a comprehensive real-time finite element method (FEM) analysis of losses induced by pulsewidth modulation (PWM) voltages in an interior permanent magnet synchronous machine, compared to conventional sinusoidal current excitation feeding. The analysis integrates electromagnetic field simulations in Ansys Maxwell with the drive system control algorithm in Ansys Twin Builder, ensuring an accurate representation of their interactions. In addition, a method utilizing speed-adaptive core loss coefficients, which account for variable frequencies, is implemented for a more precise core loss estimation. The results reveal a notable discrepancy of up to 80$\%$ in the core loss calculations when using speed-adaptive coefficients versus fixed coefficients. By employing the real-time coupled simulations, the higher switching capabilities of SiC mosfets could be effectively realized to optimize the PWM frequency over a broader range (10–50 kHz), particularly in the main drive region of electric vehicles, with differences of up to 20 kHz compared to IGBT systems. Furthermore, applying the proposed optimal PWM frequency profile in the worldwide harmonized light vehicle test cycle leads to a reduction of up to 22$\%$ in accumulated energy losses in the SiC motor drive compared to its IGBT counterpart.