Neural network based speed observer for interior permanent magnet synchronous motor drives

Abstract

This paper introduces an artificial neural network (ANN) based observer for accurate speed control of an interior permanent magnet synchronous machine (IPMSM). A multilayer perception is trained online using back-propagation learning algorithm to estimate the rotor speed. Unlike other observers, no a priori offline training or weights initialization is required. Experiments for different situations highlight the performance of the proposed observer in transient, steady-state, and standstill conditions. The proposed observer is reliable and effective for IPMSM speed control. Furthermore, the neural networks inherent parallelism makes them a good candidate for implementation in real-time PMSM drive systems.