神经网络PID控制器在山梨醇生产中的应用研究

ON APPLYING NEURAL NETWORKS BASED PID CONTROLLER TO SORBITOL PRODUCTION

在山梨醇化工生产过程中,针对结晶过程温控系统存在大滞后和模型不确定的特性,利用BP神经网络所具有的自学习和任意非线性表达能力,提出了用BP神经网络自整定PID参数的控制策略,对结晶过程温控系统进行控制,并用MATLAB软件进行仿真研究。仿真结果表明:基于BP神经网络的PID控制器具有很强的适应性,可获得满意的控制效果。

As the properties of large time delay in temperature controlling system and uncertainty model both exist in crystallizing process of the sorbitol production, according to this situation and combing with the self-learning ability and arbitrary non-linear representing ability of BP neural network, a controlling policy of self-tuning PID controller＇ s parameters with BP neural network is proposed. It is used to control the temperature controlling system in crystallizing process,and the study of its simulation has been carried out with MATLAB software. Results of experiment show that the PID controller based on BP neural network has strong adaptability so as to achieve satisfied controlling effect.