Interior Permanent-Magnet Synchronous Motors Speed Identification by Using Artificial Neural Networks Left-Inversion Method

摘要

A new speed identification method is proposed for sensorless operation of interior permanent-magnet synchronous motors (IPMSMs). The theoretic invertibility of mathematic model of IPMSMs is derived, and then a speed estimation strategy based on artificial neural networks left-inversion (ANNLI) is proposed. The structure of multi-layer feed-forward neural network is trained by advanced back propagation arithmetic. The effectiveness of the proposed method is verified by computer simulation. The results show that the developed control system can track the rotation speed quickly and accurately.