料筒温度RBF神经网络PID控制器设计及仿真

Design and Simulation of Barrel Temperature RBF Neural Network PID Controller

针对PID控制器具有参数整定不良、性能欠佳、温度控制精度较低,无法满足当今高精密挤出成型加工需要的问题,设计了一种基于RBF神经网络的PID控制器,该控制器将神经网络能无限地逼近非线性系统、运算量小、收敛快的优点和PID控制技术有机地结合起来,获得较高的温度控制精度。仿真结果表明,神经网络PID控制器能有效地缩短过渡过程时间,具有很好的稳定性和快速响应性,比普通PID控制具有更好的控制效果,可改善料筒温控系统的动、静态性能。

To the problem that the PID parameters are difficult to tune, the performance is unsatisfied the control accuracy is lower, and the PID controller can not satisfy the high precise extrusion processing, a PID controller based on radial basis function neural network （NN）is designed. The merits of NN such as infinity approaching nonlinear system, httle operation quantity, speedy constringency, and high control accuracy are combined. The simulate results show that the NN PID controller makes the transient response short obviously. The system holds good stability and has better control effect than general PID controller. It could improve the performance of the dynamic and static state of the barrel temperature control system and gain a high temperature control precision.