摘要
热连轧板坯温度的连续、实时预报是提高带钢产品性能的关键,针对热连轧过程数据维度过高不利于快速、准确预测板坯温度的问题,提出一种基于随机森林-长短期记忆神经网络(RF-LSTM)的板坯粗轧出口温度预测模型。首先,采用改进随机森林算法对特征变量进行选择,通过分析板坯粗轧出口温度的预测结果变化衡量特征的贡献度,进而构造反映过程数据特征与板坯温度的特征选择评价函数;其次,针对热连轧过程数据具有时间序列特性的特点,采用LSTM预测板坯粗轧出口温度;通过钢厂实际的热连轧过程数据特征选择实验验证和对比分析,结果表明:特征选择前后钢坯的温度预测平均绝对误差、均方根误差分别下降了0.21、0.25℃,预测相对误差在±3.0%以内的精度达到了99.07%。
Continuous and real-time prediction of strip temperature is the key to improve the performance of hot strips.The high-dimensional data of hot strip rolling process are not conducive to the rapid and accurate prediction of slab temperature.To address this problem,an RF-LSTM model was proposed to predict the strip rough rolling exit temperature.Firstly,the improved random forest algorithm was used to select the input variables,and the contribution degree of the characteristics was measured by analyzing the change in the predicted results of the strip rough rolling exit temperature,and then the feature selection evaluation function reflecting the process data characteristics and strip temperature was constructed.Then,in view of the time series characteristics of the hot continuous rolling process data,the Long-Short Term Memory(LSTM)neural network was used to predict the strip rough rolling exit temperature.Through the experimental verification and comparative analysis of the data feature selection of hot continuous rolling process in actual steel mills,the results show that the average absolute error and root mean square error of strip temperature prediction before and after feature selection are reduced by 0.21 and 0.25℃,respectively,and the accuracy of prediction relative error within±3.0%reaches 99.07%.
作者
易成新
李维刚
吕立华
赵云涛
YI Chengxin;LI Weigang;LVLihua;ZHAO Yuntao(Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China;School of Information Science and Engineering,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China;Central Research Institute,Baoshan Iron and Steel Co.,Ltd.,Shanghai 201999,China)
出处
《钢铁研究学报》
CAS
CSCD
北大核心
2021年第9期952-959,共8页
Journal of Iron and Steel Research
基金
国家自然科学基金资助项目(51774219)。
关键词
特征选择
LSTM
随机森林
热轧温度预测
feature selection
LSTM
random forest
hot rolling temperature prediction
