International Conference
| Learning-based Position Sensorless Control in Low-speed Region for SMPMSM | |
|---|---|
| Year | 2022 |
| Month | October |
| Journal | 2022 IEEE Energy Conversion Congress and Exposition (ECCE) |
| Author | Jaehoon Shim; Byung Ryang Park; Sunghyuk Choi; Jung-Ik Ha |
| Link |
http://ieeexplore.ieee.org/document/9947561
100회 연결
|
Many studies have been conducted on position-sensorless control of Permanent Magnet Synchronous Motor(PMSM). However, since there is no salient polarity in the case of a Surface-Mounted Permanent Magnet Synchronous Motor(SMPMSM), low-speed sensorless control through signal injection is inevitable. However, it is hard to implement. Further, it has a noise problem. This paper proposes a low-speed position-sensorless method using machine learning to overcome these problems. Among many machine learning models, denoising autoencoder[1] is used. It could restore the noisy input data well. And middle layer could be considered to imply the compressed characteristics of the original data. A neural network module that infers rotor angle is added for using this characteristic. The effectiveness of the proposed model is verified by experimentally showing that it is possible to drive over a wider low-torque and low-speed conditions better than the conventional method using a back Electro-Motive Force(EMF). |
|
Seoul National University, 1, Gwanak-ro, Gwanak-gu,
Seoul, Republic of Korea, 08826 Bldg. 301, Rm. 606