论文部分内容阅读
针对板形板厚综合系统具有强耦合、非线性、含纯滞后环节的特点,建立了板形板厚耦合模型,并在对其进行神经网络解耦设计的基础上提出了基于对角递归神经网络(DRNN)整定PID的板形板厚解耦控制方法,然后根据带钢冷轧情况提出神经网络解耦对不同塑性刚度参数的实际适用范围。仿真结果表明,该解耦控制系统具有比传统前馈补偿解耦PID控制效果好、响应速度快、自适应跟随能力强等优点,并且符合实际轧制要求,有效地提高了板形板厚的控制精度。
Aiming at the characteristics of the plate-thickness integrated system with strong coupling, nonlinearity and pure hysteresis, a plate-thickness coupled model was established. Based on the neural network decoupling design, Network (DRNN) is used to set the decoupling control method of plate thickness of PID. Then, the practical application of neural network decoupling to different plastic stiffness parameters is proposed according to the strip cold rolling conditions. The simulation results show that the decoupling control system has the advantages of better effect than the traditional feedforward compensation decoupling PID control, fast response, adaptive ability to follow, and meet the actual rolling requirements, effectively increasing the plate thickness control precision.