Reduced-Order Extended Observer and Correction Discretization Method for High-Speed Permanent Magnet Synchronous Motor With LCL Filter

The forward Euler method is well known for its simplicity and ease of implementation in permanent magnet synchronous motor (PMSM) controls. However, it introduces significant errors, which become more severe when establishing the discrete-time model of a high-order system for a high-speed PMSM with an LCL filter. Traditional discrete methods, which have smaller errors, are too complex and difficult to apply in high-speed systems. To address these issues, this article proposes a sensorless control method for surface-mounted high-speed PMSM with an LCL filter based on a predictive correction Runge–Kutta (PCRK) discrete method. High-precision extended Luenberger observer and strongly robust sliding mode observer (SMO) are established to estimate motor side currents and rotor position. PCRK is proposed to reduce discretization errors and computational load. Implicit equation issues in numerical solutions are eliminated through predictive-corrective steps to implement the algorithm in digital circuits. Several simulations and experiments are conducted in three methods: the traditional Euler method, the bilinear transform method, and the proposed method. Comparisons demonstrate that the proposed method has good rotor position estimation performance and a wide pole configuration range in sensorless control of HSPMSM with an LCL filter.

I. Introduction

High-speed permanent magnet synchronous motors (HSPMSMs), known for high power density advantages, often exhibit a design choice with a smaller stator inductance, leading to significant harmonic issues [1], [2]. To enhance control performance, LCL filters, equipped with the motor, are frequently used in HSPMSMs for efficacy in suppressing high-frequency harmonic [3], [4]. However, the use of LCL filters causes phase delays, particularly between voltage and current on the inverter and motor sides. In traditional control systems, current sensors are installed on the inverter side. To achieve precise current control of HSPMSM with an LCL filter, it is necessary to add motor-side current sensors and filtering branch voltage sensors. This approach increases system costs and reduces reliability. Achieving direct control of motor-side currents in high-order systems without altering the hardware has become a research focus in the field of HSPMSM. Moreover, position sensors in HSPMSM are plagued by numerous issues such as high cost, installation difficulties, and low reliability. Consequently, there is a need for sensorless techniques to estimate rotor position and speed. Researchers have conducted extensive studies on sensorless technologies for HSPMSMs. Among the common methods employed are those based on back electromotive force (back EMF) or flux linkage. This approach is essential in applications where physical sensors are impractical due to cost, space, or environmental constraints or where enhanced reliability and performance are desired.