多跨转子系统多频传递力的神经网络自适应PD控制  

An adaptive neural network PD control of multi-frequency transmission force of multi-span rotor system

作　　者：王金健 祝长生[1] WANG Jin-jian;ZHU Chang-sheng(College of Electrical Engineering,Zhejiang University,Hangzhou 310058,China)

机构地区：[1]浙江大学电气工程学院,浙江杭州310058

出　　处：《振动工程学报》2022年第1期148-158,共11页Journal of Vibration Engineering

基　　金：国家自然科学基金资助项目(11632015);国家重点研发计划(2018YFB0905500)。

摘　　要：针对大型旋转机械在运行过程中,由于自身不平衡量以及复杂的外部环境激励,导致转子系统振动,进而对基础和外部结构产生多频传递力的问题,提出了一种基于BP神经网络的自适应PD控制算法。采用一种电磁执行器与固定瓦滑动轴承集成的混合轴承结构,分析了该混合轴承的动力学特性;针对一个多跨转子系统,用有限元法建立了系统的动力学方程,从原理上分析了PD控制方式下传递力的主动控制;针对传统PID控制参数获取困难的问题,提出了基于BP神经网络的自适应PD控制算法;在一个四轴承双跨转子系统仿真模型上,分别对基于BP神经网络的自适应PD控制、BP神经网络控制及LMS控制的效果进行了对比分析。结果表明,基于BP神经网络的自适应PD控制对转子系统多频传递力具有更好的抑制效果。For solving the problem that the large rotating machinery was disturbed by the rotor imbalance and external complex environment excitations during the operation,which caused the rotor system to vibrate,and generated multi-frequency transmission forces to the base and external machinery,an adaptive PD control based on BP neural network algorithm was proposed.A hybrid bearing structure integrating electromagnetic actuator and journal bearing was applied,and its dynamic characteristics were analyzed.The finite element method was used to establish a multi-span rotor system dynamic equation,and the transmission forces control under PD control was analyzed in principle.For the defects of traditional PID control,an adaptive PD control algorithm based on BP neural network was proposed.Numerical simulations were carried out in a four-bearing double-span rotor system and effectiveness of adaptive PD control based on BP neural network,BP neural network control and LMS control were compared respectively.The results showed that the adaptive PD control based on BP neural network has better suppression effect on the multifrequency transmission forces control of the rotor system.

关 键 词：转子动力学 自适应PD控制 多跨转子 混合磁轴承 神经网络 

分 类 号：O347.6[理学—固体力学] TB535[理学—力学]