神经网络控制器用于永磁电机的伺服驱动

Servo Drive of Permanent Magnet Motor Using Neural Network Adaptive Controller

提出用一种遗传算法（ＧｅｎｅｔｉｃＡｌｇｏｒｉｔｈｍ，简称ＧＡ）训练的多层神经网络结构（ＮＮ），并把它用于交流直线永磁同步电动机（简称ＬＰＭＳＭ）的辩识与控制，使其运行在伺服驱动环境．ＮＮ可随时辩识被控系统的动态待性，而不需要ＬＰＭＳＭ的精确的数学模型．

In this paper, a multi loyer neural network(NN) architecture based on GA is proposed for the identification and control of LPMSM operating in a servo drives environment.The NN in the proposed structure can identify the controlled system dynamics at all times.Hence, detailed and elaborate models for the LPMSM are not necded.The simulation results are satisfactory.