热轧板带横向厚度分布神经网络预测分析  

Predictive Analysis of Thickness Transverse Distribution for Hot Rolling Strip based on Neural Network

作　　者：韩世平[1] 

机构地区：[1]太原理工大学机械工程学院,山西太原030024

出　　处：《山西冶金》2016年第1期5-6,共2页Shanxi Metallurgy

基　　金：山西省自然科学基金项目(20140321014-02)

摘　　要：将BP神经网络相结合引入到热轧板带横向厚度分布的预测与控制中,并依据板带轧制的实际数据对网络进行训练,通过比较二辊可送轧机板带横向厚度分布的预测值与实测数据,验证该热轧板带横向厚度分布的神经网络预测模型的准确性。验证表明,该模型的应用不仅能够提高板带轧制质量,而且能够降低设备成本。The BP neural network is used for the forecast and control of hot rolled strip thickness transverse distribution. Based on the rolling data in history, the network is trained on offline and the networks are trained. By comparing measured data of two high reversible hot rolling mill, the accuracy of the neural network prediction mode in transverse distribution thickness of hot rolled plate is verified. Verification shows that the model application can improve the quality of strip rolling with little investment of equipment cost.

关 键 词：神经网络 板厚 预测 热轧 

分 类 号：TG333.7[金属学及工艺—金属压力加工]