论文部分内容阅读
常规PID控制器是在冷轧机厚度控制中应用广泛的一种方法,但当AGC(厚度自动控制)系统特性或运行条件发生变化时,需要重新整定控制器参数,才能保证系统正常运行,使系统处于最佳工作状态。针对冷轧机AGC系统中的滞后环节,运用神经网络的自学习能力在线调整积分控制器参数值,该神经网络的权值与积分参数值相对应,可根据被控系统的动态特性调整积分参数,提高了调节器的自适应能力。
Conventional PID controllers are a widely used method in thickness control of cold rolling mills. However, when AGC (Thickness Automatic Control) system characteristics or operating conditions change, controller parameters need to be readjusted to ensure the normal operation of the system. The system is in the best working condition. Aiming at the hysteresis in the cold rolling mill AGC system, the self-learning ability of neural network is used to adjust the integral controller parameter value online. The weight of the neural network corresponds to the integral parameter value, and the integral parameter can be adjusted according to the dynamic characteristics of the controlled system , Improve the adaptive ability of the regulator.