Position Sensorless Control for Doubly Salient Electromagnetic Machine with Improved Startup Performance

Affiliation of Author(s):自动化学院

Journal:IEEE Trans Ind Electron

Abstract:It is an effective approach to estimate rotor position of doubly salient electromagnetic machine (DSEM) from self-inductances, which changes with rotor position. To overcome the restrictions of low estimation accuracy as well as the inevitable commutation lag in the conventional sensorless startup method, this paper proposes a novel rotor angular position estimating technique for the DSEM sensorless startup. It starts with injecting three test pulses A + C- (phase A is connected to the positive rail of dc bus, and phase C is connected to the ground rail), B + A- and C + B- on armature windings, so two-phase series self-inductances can be identified according to the resultant currents. Then, the estimating expression of the DSEM rotor angular position is further obtained based on the geometrical relation in the rectangle composed of self-inductances, and the inevitable commutation lag in startup process is likely to be eliminated with the estimated rotor angular position. Moreover, the error of rotor angular position introduced by the nonideal linearity of self-inductances are analyzed in detail, and the boundaries of neighboring sectors are distinguished from normal sectors to compensate the estimation error of rotor angular position. Finally, to enhance the output torque performance during the DSEM startup process, a novel selection method for test pulses concerning rotor sector is developed. Compared to the conventional method, a quicker and superior startup performance is achieved since no negative torque is generated by the test pulses. The experiments on a 12/8-pole DSEM validate the correctness and feasibility of the proposed sensorless control methods. © 1982-2012 IEEE.

ISSN No.:0278-0046

Translation or Not:no

Date of Publication:2020-03-01

Co-author:Zhou, Xingwei,Wang, Kaimiao

Correspondence Author:Bo Zhou

Date of Publication:2020-03-01