混合网络在慢时变系统中的应用

Application of mixed neural network in slow time-varying system control

针对工业慢时变系统非线性、调节时间长、参数难以用常规PID控制方法整定等的特点,提出了一种混合神经网络参数自整定复合控制方案。将BP网络和PID结合起来充当系统的控制器,实现PID参数的自整定,RBF网络充当辨识器,实现对受控对象模型的辨识。仿真结果表明,该系统具有较好的性能。

To overcome the defects of the slow time-varying system, a mixed neural network of parameters selfsetting control strategy is proposed. The BP network is combined with PID as the controller. The self-setting of parameters is realized. The RBF network is used as the identifier to identify the controlled object model. The simulation results prove that this control system is effective.