摘要
根据当代物流的高速高效特点,对翻车机的转速进行基于径向积函数(RBF)神经网络辨识的自适应PID的控制,该方法是利用神经网络的自学习能力来弥补PID控制的在线运算能力,即利用RBF神经网络对PID控制中的3个重要参数实时调整,使得该种控制方法集合了神经网络的实时自适应性和PID控制快速方便的特点,比传统的单纯PID控制更适合当代大型工业系统控制,其能够针对受控对象的某一时刻的特征迅速进行在线运算,并且提出利用递推极大似然估计法对RBFNN网络权值进行训练,可以迅速达到稳定状态,超调量很小,渐进性好,克服了实际系统控制中的稳定性与响应速度的矛盾。将其应用于翻车机角速度的优化控制中,并进行仿真,得出该控制方法的准确性与鲁棒性可以达到满意效果。
According to the characteristics of current labour exchange which are high speed and high effective, in this paper we de- scribe a adaptive PID control method of the car dumper＇ s rotational velocity based on RBFNN, this method can improve the online op- erational capability of the conventional PID controller by using the neural network which has the learning capability. In other words, the parameters of the PID controller are adjusted by the RBFNN. Thus, this control method has gathered the characteristics of the neural network and the PID control. Compared to the traditional pure PID control, this method is more suitable for the major industry systems control of the present age. It can be able to aim at the controlled object＇ s characteristic of certain time and carry on the online opera- tion rapidly, moreover, proposed using the recursive maximum likelihood estimation to train the network weights of RBFNN, and then may arrive the steady state very quickly. At the same time, the overflow is very little and well Progressive. Therefore, it has gathered the control stability and the speed in the reality systems. Applies it in the tripping device angular speed optimized control, and carries on the simulation, obtains this control method the accuracy and robustness may achieve.
出处
《控制工程》
CSCD
北大核心
2012年第5期850-854,859,共6页
Control Engineering of China
基金
高校博士学科专项科研基金资助项目(20060216008)
关键词
翻车机
转速
神经网络
智能优化
径向积函数
PID控制
car dumper
rotary velocity
neural network
intelligent optimization
（ radial basis function） RBF
（ proportional integral differential ） PID
