High-Fidelity Memory-Compact Dynamic Model of Interior-Permanent Magnet Synchronous Machines for Motor-Drive Transient Simulations

Interior permanent magnet synchronous machines (IPMSMs) are often considered in electric vehicles, servomechanisms, aerospace, robotics, industrial automation, etc. In IPMSMs, the flux linkage may vary significantly depending on the design and operating conditions. Detailed models of IPMSMs are required to study their operation in motor-drive applications, where computationally efficient and memory-compact models are highly desirable. The conventional
$qd$
models may typically use large lookup tables to accurately represent the flux-current relationship over a wide range of operating conditions. Such lookup tables may be obtained from the finite-element analysis at the motor design stage or experimentally for a considered motor prototype, but in general, they have many data points and require significant memory allocation. This paper presents the qd-model of IPMSM, where the lookup tables are replaced with bivariate polynomials, significantly reducing the data points needed to accurately capture the flux-current (and its inverse, current-flux) relationships. The proposed method is demonstrated on simulations of the IPMSM drive with a vector-PI-based field-oriented control (FOC). Using bivariate polynomials reduces the number of data points from around 10,000 to about 100 while achieving similar accuracy, representing a significant improvement over the traditional methods.