基于BP神经网络的纸机加压网前箱的建模研究

Modeling of Headbox of Paper Machine via BP Neural Network

针对采用传统的建模方法建立造纸工业过程中的纸机加压网前箱的数学模型很难得到满意结果的问题,尝试用BP神经网络对其辨识建模。通过对物理模型的多次试验,获得了大量的输入输出数据;在此试验数据的基础上对该结构进行BP神经网络模型辨识。为了验证辨识得到的模型的准确性,对该模型采用PID闭环控制,并在Matlab上对结果进行了仿真。实验结果表明,BP神经网络在纸机加压网前箱系统建模方面有着较高的辨识精度。

To the problem that the mathematical model of the pressured headbox of the paper machine in the process of papermaking is very difficult to construct by using the conventional method,BP network is used in model identification of pressured headbox.The input and output data are got from the lots of experiment of the physical model of the headbox.Based on the data,the structure is modeled and identified using BP networks.In order to validate the veracity of the identification model,the PID closed-loop control method is adopted.The simulation results show that the method of BP networks has high identification accuracy in the field of the pressured headbox of the paper machine.