[1]句建国,邢进生,王冬冬.基于BP-NN的热连轧产品性能自适应逆控制模型[J].计算机技术与发展,2018,28(12):185-189.[doi:10.3969/j.issn.1673-629X.2018.12.039]
　JU Jian-guo,XING Jin-sheng,WANG Dong-dong.An Adaptive Inverse Control Model of High-dimensional Hot-romng Performance Based on BP-NN[J].,2018,28(12):185-189.[doi:10.3969/j.issn.1673-629X.2018.12.039]
点击复制

基于BP-NN的热连轧产品性能自适应逆控制模型()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
28
期数:
2018年12期
页码:
185-189
栏目:
应用开发研究
出版日期:
2018-12-10

文章信息/Info

Title:
An Adaptive Inverse Control Model of High-dimensional Hot-romng Performance Based on BP-NN
文章编号:
1673－629X(2018)12－0185－05
作者:
句建国;  邢进生;  王冬冬
山西师范大学数学与计算机科学学院,山西临汾,041000
Author(s):
JU Jian-guo; XING Jin-sheng; WANG Dong-dong
School of Mathematics and Computer Science，Shanxi Normal University，Linfen 041000，China
关键词:
热连轧产品;  内模控制;  BP神经网络;  轧制工艺参数;  自适应逆控制
Keywords:
hot rolling products; internal model control; BP neural network; rolling process parameters; adaptive inverse control
分类号:
P39
DOI:
10.3969/j.issn.1673-629X.2018.12.039
摘要:
为了满足热连轧产品用户对性能的不同需求,钢铁企业需要逆控制模型调整生产工艺参数.以某钢铁企业热连轧产品质量为研究对象,运用BP神经网络构建钢铁热连轧产品性能指标和钢铁化学成份与轧制工艺参数的逆模型,实现了根据给定钢铁性能指标求轧制工艺参数的目的.结合BP神经网络、自适应逆控制与内模控制理论,建立了基于内模控制的多输入单输出(MISO)的BP神经网络逆模型,实现了BP神经网络输出、输入变量的逆映射,根据模型的输出变量可以求解出输入变量,并且给出逆模型求解的具体算法步骤.将所建模型应用到钢铁热连轧产品质量控制系统中,设置热连轧产品性能指标,求解轧制工艺参数-轧制卷曲温度,实现轧制工艺参数的可控性.使用热连轧产品质量控制正系统验证,误差在0.05范围之内,符合企业生产要求.
Abstract:
In order to meet the different needs of users of hot tandem rolling products，steel enterprises need to reverse the control model to adjust the production process parameters．Taking the quality of hot rolling products of a steel enterprise as the research object，the BP neural network is used to construct the inverse model of the product properties of the hot strip mill and the chemical composition and rolling process parameters of the steel．The rolling process is achieved according to the given steel performance index for the parameters．Based on BP neural network，adaptive inverse control and internal model control theory，we establish a multi－input single－output (MISO) BP neural network inverse model based on internal model control to realize the inverse mapping of BP neural network output and input variables．According to the output variables of the model，the input variables can be solved，and the specific algorithm steps of solving the inverse model are given．The model is applied to the product quality control system of the hot strip mill．The performance indexes of the hot strip mill are set up to solve the rolling process parameters－the rolling curl temperature and the controllability of the rolling process parameters．The use of hot tandem mill product quality control system is verified，the error within the 0． 05 range，in line with production requirements．
更新日期/Last Update: 2018-12-10