【作者】 肖俏； 汪首坤； 王军政；

【Author】 XIAO Qiao,WANG Shoukun,WANG Junzheng School of Automation,Beijing Institute of Technology Haidian District,Beijing 100081,P.R.China

【机构】 北京理工大学自动化学院；

【摘要】 针对电液比例阀中由于电压、流量非线性耦合引起的开环控制精度低的问题,采用基于数据分类的BP神经网络方法建立了比例阀电压-流量-压力静态模型。方法中提出了基于死区、滞环的数据分类原则,有效地降低了比例阀的死区、滞环等非线性因素对建模精度的影响,同时,改进了网络误差函数,提高了系统建模速度,并对模型进行了试验验证。试验结果表明:改进的神经网络建立的模型能准确反映比例阀的死区、滞环的特点,模型泛化性好,系统开环控制精度高。另外利用定期模型自校准的方法整定模型参数,实时修正模型,提高模型准确度和控制精度。更多还原

【Abstract】 This paper aims at controlling the nonlinear couple variables voltage and flow precisely in the electro-hydraulic proportional.A static model of voltage,pressure and flow,which is established from BP neural network based on data classification,ispresented.The data classification principleis given based on dead zone andhysteresis which may cause the low accuracy of model.The experimental results and applications show that the modeling can reflect the characteristics of electro-hydraulic proportional and achieve high precision in pressure controlling.What is more,Depending on real time online parameters tuning,it can improve the model accuracy and control precision.更多还原