Saliency-tracking-based sensorless control of AC machines using structured neural networks

The focus of this paper is the use of structured neural networks for sensorless control of ac machines using carrier-signal injection. Structured neural networks allow effective compensation of saturation-induced saliencies as well as other secondary saliencies. In comparison with classical compensation methods, such as lookup tables, this technique has advantages such as a physics-based structure, general scalability, reduced size and complexity, and correspondingly reduced commissioning time. When compared with traditional neural networks, structured neural networks are simpler, physically insightful, less computationally intensive, and easier to train. All make the proposed method an improved implementation for sensorless drives.