On-line adaptive neural training algorithm for an induction motor flux observer

This paper presents a new algorithm for induction motor stator flux observation. The novel procedure is based on a neural network with on-line adaptive training. The network topology is a standard multilayer perceptron (MLP) network and the training algorithm is based on sliding mode control (SMC) theory. The main characteristic of this novel observer is the adaptability of the gain (learning rate), which is obtained from sliding surface so that system stability is guaranteed. The neural network stator flux observer employed here does not require previous training or speed measurement. The on-line adaptive training algorithm for the neural network is described, as well as its application to a stator flux observer of an induction motor drive. Neural observer performance is demonstrated by simulations results