电液伺服协调加载系统的神经网络自学习 PSD 控制

Neural network self-learning PSD control for electrohydraulic servo synchronous loading system

从电液伺服协调加载系统的特点出发，提出一种基于神经网络的在线自学习ＰＳＤ控制方法。该方法简单、实用，便于在线实现。控制器无须事先训练，参数选取极具一般性，适用范围广，控制精度高且鲁棒性强。采用这一方法实现了对某大型电液伺服结构试验装置的协调加载控制，控制品质优良。

This paper presents a neural network based self_learning PSD control strategy with respect to the control requirements of electro_hydraulic servo synchronous loading system.The control strategy is simple ,utility,and easy to be realized on_line.Compared with traditional methods,the controller is designed without specific pre_training stage.It is more suitable and robust with higher control precision.By adopting this strategy,an on_line selflearning control system is developed for the synchronous loading control of a large_scale electro_hydraulic servo structural testing machine.The tracking results show the controller has very good control performances.