[Aug.2022] Ph.D. Dissertation
유지원, “맥동 전압 주입을 이용한 위치 센서 없는 매입형 영구자석 동기전동기의 MTPA 추종제어,” 박사학위논문, 서울대학교, 2022.
J. Yoo, “MTPA Tracking Control of Sensorless IPMSM Based on Pulsating Voltage Signal Injection,” Ph.D. Dissertation, 2022.
(This dissertation won the distinguished dissertation award in Seoul National University.)
Link
As the torque density of the interior permanent-magnet synchronous motors(IPMSMs) gets higher, the more magnetically saturated IPMSM is being developed. Below the base speed, maximum torque per Ampere(MTPA) operation is required for the high efficiency and the maximum torque capability of IPMSM. Therefore, an online MTPA tracking operation considering the magnetic saturation is required.
This dissertation presents an online MTPA tracking algorithm based on the flux model of IPMSM, which can be applied to the sensorless condition. The proposed algorithm consists of two sub-algorithms; one is for the mid-and high-speed region where the stator flux can be observed from the stator voltage, and the other is for the low-speed region.
Prior to the algorithm for the mid-and high-speeds, an MTPA criterion is formulated at an arbitrary reference frame. Furthermore, the estimated MTPA reference frame(EMRF) is proposed, and the MTPA criterion on that frame is derived. On the EMRF, the MTPA condition can be assessed without the rotor position information. Moreover, the proposed MTPA criterion at the EMRF requires only one inductance information, while the conventional criterion at the rotor reference frame is formulated with four inductance information.
Based on the simplified MTPA criterion, an MTPA tracking algorithm for mid-and high-speeds is proposed. The proposed method comprises a flux estimator based on the stator voltage observation, a dynamic inductance estimator based on the pulsating voltage injection, and an MTPA point tracking controller. This dissertation adopts the frequency-adaptive flux observer(FAO) to estimate the stator flux information. The discretization error of the conventional implementation of FAO is analyzed. Based on the analysis, an FAO’s implementation at the synchronous reference frame, robust to the discretization error, is proposed. The dynamic inductance estimator calculates the d-axis inductance at the EMRF based on the pulsating voltage signal injection. To enhance the estimation accuracy, the injection voltage vector is dynamically changed by an injection voltage controller. The proposed mid- and highspeed algorithm does not need the preset inductance/flux look-up table and is applicable to the sensorless IPMSMs.
This dissertation proposes an MTPA tracking algorithm for the low-speeds based on the tilted pulsating voltage signal injection. Instead of the real-time flux estimator, the proposed low-speed algorithm utilizes the inductance and flux information obtained in the mid-and high-speed operation, which determines the direction of the injected voltage vector. This dissertation presents a sufficient condition of IPMSM for the stable operation of the proposed low-speed algorithm. The proposed MTPA tracking algorithm can accurately calculate the MTPA operating point even under the parameter variation induced by the magnetic saturation. The performance of the proposed MTPA tracking algorithm is verified with a 150 kW-rated IPMSM for EV application.
