基于Smith预估补偿与RBF神经网络的改进PID控制

Improved PID Control Based on Smith Predictive Compensation and RBF Neural Network

由于工业界普遍存在且难以很好地解决恒温控制的大滞后和非线性问题,特提出了将Smith预估补偿和RBF神经网络与PID控制相结合的改进PID控制算法。该算法利用Smith预估补偿对温度滞后问题进行处理,利用RBF网络在线学习能力进行PID参数的动态调整处理非线性问题,进而保证恒温控制使系统处于最佳状态。

As the problems of big delayed time and nonlinear are normally exist and difficult to solve in the processing industry which used the constant control of temperature,an improved PID algorithm based on Smith predictive compensation and RBF neural network is proposed.The algorithm used the Smith predictive compensation to deal with the big delayed time,and adopted the online learning of RBF neural network to dynamically adjust the PID parameters to deal with the nonlinear,thereby the constant control was ensured to make the system in a good state.