Model Reference Adaptive Control of Asynchronized Synchronous Motor Based on Neural Network On-Line Identification

From system control point of view, asynchronized synchronous motor is a complex nonlinear plant with uncertainties, and is hard to be controlled by traditional control method. An excitation control method based on model reference adaptive control of asynchronized synchronous motor using neural network on-line identification is proposed in this paper. This method utilizes the mapping abilities of neural networks to construct the reference model of the plant, and self-learning is achieved by means of neuron. The simulation results show a better robustness and control accuracy of the proposed method.

I. Introduction

Modern power system has been in the stage of huge capacitance, super high voltage and remote distance transmission. The problem of sustained power frequency over-voltage caused by surplus reactive power of long transmission line risks the isolation of system and increases the loss of transmission line, even leads to generators displacement. Also, the stability of long line transmission system is an important problem. The application of asynchronized synchronous motor is not only the popular method of solution of this problem, but also could improve system stability.