锅炉温度系统的神经网络预测-PID串联控制

Series Neural Network Predictive PID Control of Boiler Temprature System

针对锅炉温度系统的非线性、滞后、时变等特性,提出了一种串联控制策略。提出基于神经网络的微分型预测控制算法,该方法的突出优点是能够加快调节时间。在此基础上结合常规PID控制器构成了预测-PID串联控制,这种串联控制的方法既有基于神经网络的预测控制在实时系统中抗干扰能力强的优点,又充分利用了PID控制方法响应速度快的特点。通过对锅炉温度系统的实时控制实验,证明了所提方法的有效性,极大地提高了系统的控制品质。

Aiming at the characteristics such as nonlinearing, time-delay and timevarying, of the boiler temperature system, a series control strategy is proposed. A differential predictive control algorithm based on neural network is presented, to accelerate the adjustment time. Based on that the conventional PID controller is combined to form a series prediction - PID control. The method combines the advantages of strong anti - jamming ability of neural network predictive control in real - time system and fast response speed of PID. On the real-time control experiments of the boile temperature system shows the effectiveness of the proposed approach and improve ment of the control quality of the system.