正在加载图片...
工程科学学报,第40卷,第7期:815-821,2018年7月 Chinese Journal of Engineering,Vol.40,No.7:815-821,July 2018 DOI:10.13374/j.issn2095-9389.2018.07.007;http://journals.ustb.edu.cn 基于极限学习机(ELM)的连铸坯质量预测 陈恒志”,杨建平),卢新春12),余相灼),刘青) 1)北京科技大学钢铁治金新技术国家重点实验室,北京1000832)方大特钢科技股份有限公司,南昌330000 区通信作者,E-mail:qliu@usth.edu.cm 摘要针对传统基于BP神经网络建立的连铸坯质量预测模型训练速度慢、适应能力弱、预测精度低等问题,本文提出一种 基于极限学习机的连铸坯质量预测方法,对方大特钢60Si2M连铸坯中心疏松和中心偏析缺陷进行预测,并与BP和遗传算 法优化BP神经网络预测模型的预测结果进行分析对比.结果表明:BP及GA-BP神经网络预测模型对连铸坯中心疏松和中 心偏析缺陷的预测准确率分别为50%、57.5%、70%和72.5%;而基于极限学习机的连铸坯预测模型预测准确率更高,对连铸 坯中心疏松和中心偏析缺陷的预测准确率分别为85%和82.5%,且该模型具有极快的运算时间,仅需0.1s.该模型可对连铸 坯质量进行迅速准确地分析,为连铸坯质量预测的在线应用提供了一种新的方法 关键词连铸坯;BP神经网络:遗传算法:极限学习机:质量预测 分类号TF777.2 Quality prediction of the continuous casting bloom based on the extreme learning machine CHEN Heng-zhi),YANG Jian-ping,LU Xin-chun'2),YU Xiang-zhuo,LIU Qing 1)State Key Laboratory of Advanced Metallurgy,University of Science and Technology Beijing,Beijing 100083,China 2)Fangda Special Steel Technology Co.,Ltd,Nanchang 330000,China Corresponding author,E-mail:qliu@ustb.edu.cn ABSTRACT To solve the problems of slow training,weak generalization ability,and low prediction accuracy in the traditional pre- diction model established in terms of the BP neural network,a method of the quality prediction of the continuous casting bloom based on the extreme learning machine (ELM)was proposed to predict the degree of the center porosity and the central segregation of 60Si2Mn continuous casting bloom produced by Fangda Special Steel.Comparing the prediction models of the BP neural network and the GA-BP neural network,the results show that the prediction accuracy of the model based on ELM is improved to 85%and 82.5% in the center loose and central segregation,respectively,and the operation time is reduced to 0.Is.The model can rapidly and accu- rately analyze the quality of a continuous casting billet,thus providing a new method for the online application of continuous casting bil- let quality prediction. KEY WORDS continuous casting bloom;BP neural network;genetic algorithm;extreme learning machine;quality prediction 连铸坯生产过程中,钢水洁净度较低或工艺操 际生产的连铸坯质量缺陷检测模型,由于检验时间 作不稳定,往往造成连铸坯质量缺陷,如中心偏析、 的滞后性,难以对铸坯质量进行实时精准预测,不利 横纵裂纹等,致使产品质量下降,废品率增加,难以 于钢铁企业高品质钢的生产.如何及时准确地预报 及时满足客户的订单需求.目前,大多数应用于实 和检测铸坯质量是钢铁企业可持续发展过程中亟待 收稿日期:2017-06-12 基金项目:国家自然科学基金资助项目(50874014)工程科学学报,第 40 卷,第 7 期:815鄄鄄821,2018 年 7 月 Chinese Journal of Engineering, Vol. 40, No. 7: 815鄄鄄821, July 2018 DOI: 10. 13374 / j. issn2095鄄鄄9389. 2018. 07. 007; http: / / journals. ustb. edu. cn 基于极限学习机(ELM)的连铸坯质量预测 陈恒志1) , 杨建平1) , 卢新春1,2) , 余相灼1) , 刘 青1) 苣 1)北京科技大学钢铁冶金新技术国家重点实验室, 北京 100083 2)方大特钢科技股份有限公司, 南昌 330000 苣通信作者,E鄄mail: qliu@ ustb. edu. cn 摘 要 针对传统基于 BP 神经网络建立的连铸坯质量预测模型训练速度慢、适应能力弱、预测精度低等问题,本文提出一种 基于极限学习机的连铸坯质量预测方法,对方大特钢 60Si2Mn 连铸坯中心疏松和中心偏析缺陷进行预测,并与 BP 和遗传算 法优化 BP 神经网络预测模型的预测结果进行分析对比. 结果表明:BP 及 GA鄄鄄BP 神经网络预测模型对连铸坯中心疏松和中 心偏析缺陷的预测准确率分别为 50% 、57郾 5% 、70% 和 72郾 5% ;而基于极限学习机的连铸坯预测模型预测准确率更高,对连铸 坯中心疏松和中心偏析缺陷的预测准确率分别为 85% 和 82郾 5% ,且该模型具有极快的运算时间,仅需 0郾 1 s. 该模型可对连铸 坯质量进行迅速准确地分析,为连铸坯质量预测的在线应用提供了一种新的方法. 关键词 连铸坯; BP 神经网络; 遗传算法; 极限学习机; 质量预测 分类号 TF777郾 2 收稿日期: 2017鄄鄄06鄄鄄12 基金项目: 国家自然科学基金资助项目(50874014) Quality prediction of the continuous casting bloom based on the extreme learning machine CHEN Heng鄄zhi 1) , YANG Jian鄄ping 1) , LU Xin鄄chun 1,2) , YU Xiang鄄zhuo 1) , LIU Qing 1) 苣 1)State Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, Beijing 100083, China 2)Fangda Special Steel Technology Co. , Ltd, Nanchang 330000, China 苣Corresponding author,E鄄mail: qliu@ ustb. edu. cn ABSTRACT To solve the problems of slow training, weak generalization ability, and low prediction accuracy in the traditional pre鄄 diction model established in terms of the BP neural network, a method of the quality prediction of the continuous casting bloom based on the extreme learning machine ( ELM) was proposed to predict the degree of the center porosity and the central segregation of 60Si2Mn continuous casting bloom produced by Fangda Special Steel. Comparing the prediction models of the BP neural network and the GA鄄鄄BP neural network, the results show that the prediction accuracy of the model based on ELM is improved to 85% and 82郾 5% in the center loose and central segregation, respectively, and the operation time is reduced to 0郾 1 s. The model can rapidly and accu鄄 rately analyze the quality of a continuous casting billet, thus providing a new method for the online application of continuous casting bil鄄 let quality prediction. KEY WORDS continuous casting bloom; BP neural network; genetic algorithm; extreme learning machine; quality prediction 连铸坯生产过程中,钢水洁净度较低或工艺操 作不稳定,往往造成连铸坯质量缺陷,如中心偏析、 横纵裂纹等,致使产品质量下降,废品率增加,难以 及时满足客户的订单需求. 目前,大多数应用于实 际生产的连铸坯质量缺陷检测模型,由于检验时间 的滞后性,难以对铸坯质量进行实时精准预测,不利 于钢铁企业高品质钢的生产. 如何及时准确地预报 和检测铸坯质量是钢铁企业可持续发展过程中亟待
向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有