神经网络在弹性连杆机构振动主动控制中的应用  

Applications of Neural Networks to the Active Vibration Control of Flexible Linkage Mechanisms

作　　者：宋轶民[1,2] 张策 

机构地区：[1]北京工业大学机电工程学院 [2]天津纺织工学院自动化系,天津300160 [3]天津大学机械工程学院,天津300072

出　　处：《振动工程学报》2000年第1期100-106,共7页Journal of Vibration Engineering

基　　金：国家自然科学基金资助项目! (编号 :5 96 75 0 0 4)

摘　　要：首次将神经网络理论应用于弹性连杆机构的振动主动控制 ,设计、建造了具有压电陶瓷作动器与电阻应变计传感器的弹性连杆机构实验装置及其振动控制系统。根据实验数据离线设计了动态递归神经网络控制器 ,并采用基于神经网络的直接自校正控制策略对弹性连杆机构实施了在线控制。控制后 ,弹性构件输出点的应变峰值降低了 5 0 %左右 。This paper presents the applications of neural networks (NN) to the active vibration control of flexible linkage mechanisms whose flexible links are incorporated with piezoceramic actuators and strain gauge sensors. The experimental mechanism and its control system were designed and constructed. Using the experimental data, the dynamic recurrent neural networks (DRNN) controllers were designed off line. The experimental mechanism was controlled on line by means of the NN based direct self tuning control strategy. As a result, the crest strain of the flexible coupler was suppressed by 50% or so and the dynamic characteristics of the mechanism was improved significantly.

关 键 词：弹性连杆机构 振动主动控制 神经网络 

分 类 号：TH112.5[机械工程—机械设计及理论] TB535[理学—物理]