
Neural network compensation method for improving detection accuracy of rotor position from magnetic encoder
Author(s) -
Yin Zhipeng,
Xu Jiaqun
Publication year - 2021
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
eISSN - 1350-911X
pISSN - 0013-5194
DOI - 10.1049/ell2.12282
Subject(s) - rotor (electric) , control theory (sociology) , encoder , computer science , position (finance) , compensation (psychology) , artificial neural network , resolver , rotary encoder , switched reluctance motor , torque ripple , backpropagation , engineering , artificial intelligence , induction motor , voltage , direct torque control , electrical engineering , psychology , control (management) , finance , psychoanalysis , economics , operating system
With the advantages of low cost, small size and high resolution, the magnetic encoder is suitable for permanent magnet synchronous motor. However, due to the installation accuracy and the motor vibration, the rotor position error from magnetic encoder can seriously impact the performance of the motor. Therefore, it is necessary to reduce the detection error of the rotor position from the encoder. In this letter, a new rotor position detection method is proposed based on neural network. The characteristic of the position error from magnetic encoder is analysed, and the backpropagation neural network is presented to obtain the error compensation function. On the basis of the compensation function, the rotor position error can be corrected and thus the precise rotor position can be detected. The experimental results show that not only the rotor position error but also the related total harmonic distortion of the phase current and the torque ripple can be reduced effectively.