基于遗传算法的无头轧制技术闪光焊机行走系统  

Research of motion system of flash butt welder based on genetic algorithms in endless rolling technology

作　　者：王睿[1] 张军伟[1] 孙立雄[1] 高忠林[2] 

机构地区：[1]河北工业大学控制科学与工程学院,天津300130 [2]天津七所高科技有限公司,天津300409

出　　处：《电焊机》2013年第8期8-11,共4页Electric Welding Machine

基　　金：天津市应用基础及前沿技术研究计划资助项目(11JCYBJC06500)

摘　　要：介绍了无头轧制闪光焊机的行走系统的组成及其工作原理,建立了闪光焊机行走系统的数学模型。提出了一种基于遗传算法和神经网络PID相结合的控制方法,该算法用GA(遗传算法)优化BP网络的初始连接权系数,得到一组较佳的BP网络初始权值的参数组合,再通过BP算法,实现PID参数的在线调整。仿真结果表明,经过GA对网络初始权值优化后,闪光焊机行走系统的输出响应能够快速的跟踪输入的信号,使系统具有精确的同步性能。This paper presents the configuration and principle of motion system of flash butt welder in endless rolling technology.It establish mathematic model of motion system of flash butt welder. PID control method which is based on genetic algorithms and neural networks is discussed in this paper. The method uses the initial connecting weights according to the genetic algorithm optimization BP network and gets a group of combination parameters closed to its initial values. By BP algorithm, it can adjust the PID parameters on line. So the welding machine and the rolling billet are simultaneous. The dynamic tracking performance is improved.

关 键 词：闪光焊机 无头轧制技术 行走系统 遗传算法 

分 类 号：TG431[金属学及工艺—焊接]