Modeling and identification of iron-less PMLSM end effects for reducing ultra-low-velocity fluctuations of ultra-precision air bearing linear motion stage

• Published in: Precision engineering 2017-07, Vol.49, p.92-103
• Main Authors: Xi, Jingsi, Dong, Zeguang, Liu, Pinkuan, Ding, Han
• Format: Article
• Language: English
• Subjects: Compensation algorithm; End effect; Identification; Thrust force model; Velocity fluctuations
• ISSN: 0141-6359; 1873-2372
• DOI: 10.1016/j.precisioneng.2017.01.016

•A new algorithm to calculate the ironless PMLSM end effects is presented.•A simplified thrust force model including end effects is established.•Both static and dynamic methods are presented to identify the thrust force model.•A compensation control algorithm is presented to reduce speed fluctuations. In this paper, a simplified thrust force model of iron-less permanent magnet linear synchronous motors (IPMLSMs) including end effects is presented. To establish the thrust force model, a magnetic flux linkage calculation algorithm is applied. The thrust force model simplifies all IPMLSM parameters into identifiable constants and demonstrates that the thrust force of the IPMLSM fluctuates periodically with the winding phase position. In addition, both static and dynamic identification experiments are performed to effectively identify the model. It is shown through ultra-low-velocity simulations of IPMLSMs using the identified thrust force model that the end effect causes velocity fluctuations, which corresponds to experimental results on an ultra-precision air bearing linear motion stage. Therefore, a feedback compensation controller based on the identified thrust force model is designed and implemented for reducing ultra-low-velocity fluctuations. Finally, comparative experimental results indicate that the compensation controller is effective and achieves a better performance in terms of ultra-low-velocity errors compared with traditional disturbance observer controllers. Thus, the thrust force model including end effects and identification algorithms are proved to be valid and practical.