摘要
永磁直线同步电动机(PMLSM)提升系统,在结构上、控制机理上均与传统的提升模式不同。电机存在着铁心开断、三相绕组分布不对称以及运行过程中参数变化较大等因素。采用理想的解析模型难以准确地反映该系统的运动特性。本文运用BP神经网络,建立了该系统的动态模型。侧重介绍了网络的学习算法,模型的构建、训练样本的获取、训练系数的选取等方法。仿真和实验结果表明,该模型比用解析法建立的数学模型更能逼真地反映出PMLSM提升系统的基本运动特性。对该系统的运行特性分析及控制策略的研究都具有实际的应用价值。
The hoisting system driven by Permanent Magnet Linear Synchronous Motor (PMLSM) is different from the traditional hoisting model in construction and control mechanism. It抯 difficult to reflect the kinetic characteristics of the system accurately on the base of analytical model because of the influence of many factors, such as the ferrite core抯 disconnection, the asymmetry of three-phase winding and the big fluctuating of parameter in the course of work etc. In this paper, a dynamic model of this system was constructed on the basis of BP network, meanwhile it抯 mainly introduced about the network抯 learning algorithm and the method of modeling, the gain of training data and selection of the network training parameter. Simulation result and experimental verification show that this model reflects the basic kinetic characteristics of the system more realistically than the analytical model. It has practical value in analyzing characteristics and control tactics of the system.
出处
《系统仿真学报》
CAS
CSCD
2002年第9期1249-1251,1254,共4页
Journal of System Simulation
基金
国家自然科学基金(编号:69674021)
河南省自然科学基金(编号:0211060500)
河南省自然科学基金(编号:004040500)
河南省重大攻关(编号:991120429)
河南省高校杰出科研人才创新工程(编号:2000KYCX009).
