BP神经网络PID控制仿真在挤压机中的应用  

The Application Based on PID Control Simulation of BP Neural Network to Extruder

作　　者：王建鑫[1] 申德超[2] 马跃进[1] 

机构地区：[1]河北农业大学机电工程学院,河北保定071000 [2]山东理工大学农业工程与食品科学学院,山东淄博255049

出　　处：《农机化研究》2011年第3期194-197,共4页Journal of Agricultural Mechanization Research

基　　金：山东省自然科学基金项目(Y2008B10)

摘　　要：为了减少挤压机运行前的繁杂工作,针对自动供给系统提出了一种基于神经网络的PID控制方案,利用神经网络的自学习能力对PID控制参数在线整定,使PID控制器具有自适应性。采用动态BP算法,达到了在线实时控制的目的,显示了BP神经网络的PID控制方法很强的鲁棒性,同时也显示了神经网络在解决时变性和严重不确定系统方面的潜能,并改变了运用试凑法调节参数的方法。运用Matlab软件对挤压加料、加水和加酶制剂自动控制系统进行仿真研究。仿真结果表明,神经网络PID控制器优于传统PID控制器,具有较高的精度和较强的适应性,可以获得满意的控制效果。Before order to reduce the extruder movement the numerous and diverse work,in view of the automatic supply system proposed that one kind based on neural network＇s PID control plan,using neural network from the learning capability to the PID controlled variable online installation,enables the PID controller to have auto-adapted.Using the dynamic BP algorithm,has served the online real-time control purpose,had demonstrated BP neural network＇s PID control method very strong robustness,simultaneously also had demonstrated neural network when solution denatured and the serious indefinite system aspect＇s potential,and changes the utilization cut and try method adjustment parameter the method.Then extrusion add powder,add water and add enzyme preparation of automatic control system for simulation with the Matlab software to conduct the simulation research.The simulation result indicated that the neural network PID controller surpasses the traditional PID controller,has the high precision and the strong compatibility,may obtain satisfaction the control effect.

关 键 词：神经网络 BP算法 PID控制 挤压机加料、加水和加酶制剂系统 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置]