Transactions of China Electrotechnical Society  2023, Vol. 38 Issue (12): 3165-3175    DOI: 10.19595/j.cnki.1000-6753.tces.220642
Current Issue| Next Issue| Archive| Adv Search |
Parameter Identification and Estimated Position Deviation Correction of a Three-Phase Transverse Flux Permanent Magnet Machine Based on Inductance Perturbation Injection
Su Youcheng, Chen Zhihui
College of Automation Engineering Nanjing University of Aeronautics and Astronautics Nanjing 211100 China

Download: PDF (3566 KB)   HTML (1 KB) 
Export: BibTeX | EndNote (RIS)      
Abstract  

Transverse flux permanent magnet machines (TFPMM) have been widely used in electric vehicles and other direct-drive applications due to the high output torque. The sensorless control of TFPMM can omit mechanical sensors, reduce cost and improve system reliability. However, the stator resistance, inductance, and flux linkage change from temperature, magnetic saturation, and load disturbance during the operation of the motor, and the nonlinearity of the inverter produce errors between the command voltage and the actual voltage, thus causing deviations between the estimated position and the actual position. This paper proposes a parameter identification method based on inductance perturbations independent of the accuracy of any parameters to correct the position deviation.
Firstly, the two main factors that cause position deviation are analyzed under steady state, model parameter error, and inverter nonlinearity. It can be intuitively obtained that the position deviation is only affected by the inductance parameter error in the sliding mode observer as soon as $i_{\delta}=0$. Secondly, according to the torque equation of TFPMM, the inductance perturbations of +Δd and-Δd are injected separately, and average values of γ-axis current are extracted after the γ-axis current reaches a steady state. Based on the assumption that the load torque changes slowly, the nonlinear equations of inductance error and the permanent magnet flux linkage are constructed. The Levenberg-Marquardt algorithm, which improves the non-convergence of the solution when the Jacobian matrix is ill-conditioned, is used to solve the nonlinear equations. The identification of inductance and permanent magnet flux linkage is realized. Finally, the identified inductance is fed back to the observer to correct position deviation.
Specific operating conditions are simulated and investigated based on a prototype of TFPMM to verify the effectiveness of the proposed method. The inductance and the permanent magnet flux linkage can be obtained by the Levenberg-Marquardt algorithm with errors of 1.38 % and 3.98 %, respectively. Further experiments are carried out to verify the feasibility and robustness of the method. By analyzing the results under different speeds and loads, it can be concluded that the mechanical angle deviation of the proposed method is controlled within 0.178 ° under different operating conditions. However, the identification error increases significantly under load conditions. Thus, the position deviation equation considering the saturation effect is deduced using the active flux concept. In the case of near-zero saliency, the identified inductance and flux linkage are Lq and ψaf, respectively. However, the permanent magnet flux linkage is not used in the observer. The position deviation equation considering the saturation effect shows that the position deviation can be corrected by modifying the observer parameters after identifying Lq.
The following conclusions can be drawn from the simulation and experimental results: (1) The proposed method does not require accurate model parameters and is unaffected by the inverter nonlinearity. Only a priori range of permanent magnet flux linkage is required to ensure convergence of the sliding mode observer. (2) The matrix dimension is low, and the calculation is simple and easy to implement. (3) The identified parameters can be solved accurately, and the steady-state position error can be suppressed effectively. The proposed method is also suitable for permanent magnet synchronous machines (PMSM) with near-zero saliency. Improvement of position estimation for salient PMSM will be the next focus of our research.
Key words： Transverse flux permanent magnet machine      inductance perturbation      Levenberg-Marquardt algorithm      position deviation      parameter identification     
Received: 23 April 2022     
PACS: TM351  
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Su Youcheng
Chen Zhihui
Cite this article:   
Su Youcheng,Chen Zhihui. Parameter Identification and Estimated Position Deviation Correction of a Three-Phase Transverse Flux Permanent Magnet Machine Based on Inductance Perturbation Injection[J]. Transactions of China Electrotechnical Society, 2023, 38(12): 3165-3175.
URL:  
https://dgjsxb.ces-transaction.com/EN/10.19595/j.cnki.1000-6753.tces.220642     OR     https://dgjsxb.ces-transaction.com/EN/Y2023/V38/I12/3165
Copyright © Transactions of China Electrotechnical Society
Supported by: Beijing Magtech