神经网络分数阶PID控制器在纸浆浓度控制中的应用

分数阶PID控制器继承了常规PID控制器的优点,并且具有更高的控制精度和更强的鲁棒性。针对常规PID控制器在纸浆浓度控制过程中存在的问题,设计了一种基于神经网络的分数阶PID控制器。用分数阶PID控制器代替常规PID控制器,并通过神经网络调节分数阶PID控制器的5个控制参数,实现一种参数自整定的PID控制器。仿真实验结果表明,神经网络分数阶PID控制器比常规PID控制器的控制精度高,对纸浆浓度的控制更稳定；采用神经网络分数阶PID控制器控制纸浆浓度是切实可行的,具有很好的推广应用前景。

Application of Fractional Order PID Controller Based on Neural Network to Pulp Consistency Control System

Since fractional order PID inherits the advantages of traditional PID and has better control quality and higher robust, a fractional order PID controller based on artificial neural network was proposed and applied in pulp consistency control system. Using fractional order PID instead of the traditional PID, a self-tuning PID controller with five control parameters was realized by using parameter adjustment strategy of neural network. The simulation results showed that neural network fractional order PID controller had higher controlling accuracy and realized more stable control of pulp consistency than traditional PID controller. Control curve proved that the new controller was feasible and had popularizing value.