气动比例阀控系统死区特性及仿真研究

The Dead Zone Specialty and Simulation Study of the Pneumatic Proportional Valve System

针对气动比例系统中存在死区的情况,采用神经网络自学习的方法,解决由于死区的存在引起的系统定位精度问题。选定一死区补偿初值,以定位精度为目标,根据系统动态响应过程中的误差及误差变化,通过在线修正神经网络权值来调整死区补偿值。

According to the dead zone existing in the pneumatic proportional system, a neural network self- study method is introduced to solve the problem of the system location precision caused by the dead - zone existence. The dead- zone compensation value is adjusted through sel^ting an initial dead - zone compensation value, making location precision as the objective, the errors and error varieties in the process of the system dynamic response and modifying the neural network authority value online.