Absolute Position Acquisition for Linear Synchronous Motor With Passive Mover

Tuanshan Zhang,Xuesong Mei,Xin Du

doi:10.1109/access.2021.3096957

Abstract

The winding segmented permanent magnet linear synchronous motor has restrictions on the last absolute position acquisition when the power goes from off to on, and the traditional absolute position acquisition systems depend on complex battery and calibration systems. Therefore, a true absolute magnetic encoder based on the Vernier principle with error rejection and decoding technology is studied in this paper. First, the output signal of the magnetic encoder with noise was subjected to the adaptive ellipse parameter correction algorithm with recursive least squares (RLS), and the source output signal predicted and reconstructed. Furthermore, the RLS was optimized by the speed weighting factor and the forgetting factor to reduce the computational pressure and the iterative process. Second, one improved double second-order generalized integrator(DSOGI) decoding technology was investigated by extracting and separating the symmetric-positive sequence components in the unbalanced signal, and, using the phase-locked loop technique for tracking calculation, accurate speed and angle obtained. The adaptive ellipse parameter fitting algorithm effectively reduced the problem of overshoot and steady-state error of the DSOGI, which provides a practical and theoretical basis for the future application of the magnetic encoder. This trajectory arrangement not only avoids the influence of accumulated error on the accuracy of encoder, but it also solves the contradiction between high resolution and small volume, and the advantages over the special decoding chip IC-Haus were also verified by comprehensive experimental results.

Highlights

Linearelectrical motors are presently gaining increasingly widespread use in transportation [1] and industry
EXPERIMENTAL RESULTS In addition to the simulations, recursive least squares (RLS)-dual second-order generalized integrator (DSOGI) decoding technologies were tested, and the decoding results were compared with the IC-Haus decoding results to further study the decoding accuracy and transient performance of the RLS-DSOGI decoding technology
The RLS-DSOGI and other decoding algorithm were completed in a RT-MATLAB environment constructed by the semi-physical real-time system provided by LinksRT (Beijing LINKS Technology Co., Ltd., Beijing, China), including an A/D board (Ni6259) for analog input and a D/A board (Ni6216) for digital output for encoder signal simulation

Summary

INTRODUCTION

Linearelectrical motors are presently gaining increasingly widespread use in transportation [1] and industry (e.g., machining and actuators). Aiming at how to improve the calculation accuracy and error suppression ability of the decoding algorithm, a double synchronous reference-frame-based phase-locked loop (DSRF-PLL) method was proposed previously [9]. This method has high precision and can suppress the error signals, but its disadvantage is that it will increase the response time when decomposed to the double rotation coordinate system. The accuracy of absolute magnetic gate scales consists of two parts: the position accuracy of the whole length and the subdivision accuracy of the moire fringes in the signal cycle In this application, a high-precision, single-channel, absolute-position-coding, absolute-grating-ruler (Fig. 3) is unrealistic.

ABSOLUTE POSITION PROCESS

RLS-DSOGI SIMULATION ANALYSIS

EXPERIMENTAL RESULTS

CONCLUSION