检索规则说明:AND代表“并且”;OR代表“或者”;NOT代表“不包含”;(注意必须大写,运算符两边需空一格)
检 索 范 例 :范例一: (K=图书馆学 OR K=情报学) AND A=范并思 范例二:J=计算机应用与软件 AND (U=C++ OR U=Basic) NOT M=Visual
作 者:李飞飞 宋勇 刘超 贾仁君 李博 LI F ei-fei;SONG Yong;LIU Chao;JIA Ren-jun;LI Bo(Institute of Engineering Technology,University of Science and Technology Beijing,Beijing 100083,China)
机构地区:[1]北京科技大学工程技术研究院
出 处:《冶金自动化》2019年第6期28-33,共6页Metallurgical Industry Automation
基 金:创新方法工作专项资助项目(2016IM010300)
摘 要:目前热轧带钢力学性能预报模型在实际应用过程中可靠性不高,不能给出预测结果的误差范围。为此,以某钢厂薄板坯连铸连轧CSP生产线的屈服强度预报模型为例,通过建立误差分布预测模型优化和提高原机理模型的准确率和稳定性。首先采用度量学习方法对生产过程数据中隐藏的工况进行划分,并通过可视化方法进行验证;然后通过主成分分析对度量学习后的数据样本进行降维,针对不同工况分别使用逻辑回归算法建立误差分布预测模型。利用该误差分布预测模型对原模型进行补偿后,采用十折十次交叉验证,屈服强度在±30 MPa误差区间下的预测准确率达到96%以上,且具有较高的预测稳定性。At present,the prediction model of the mechanical properties of hot-rolled strip steel is not highly reliable in practical application,and the error range of the prediction results cannot be given.For this reason,taking the yield strength prediction model of the thin slab continuous casting and rolling CSP production line of a steel plant as an example,the error distribution prediction model is established to optimize and improve the accuracy and stability of the original mechanism model. Firstly,the metric learning method is used to divide the hidden working conditions in the production process data,and the visualization method is used for verification. Then,the data samples of the metric learning are dimension-reduced by principal component analysis,and the logistic regression algorithm is used for different working conditions. Error distribution prediction model. After the original model is compensated by the error distribution prediction model,the prediction accuracy of the yield strength under the ± 30 MPa error interval is over 96%,and the prediction stability is high.
关 键 词:热轧带钢 力学性能预报 误差分布 误差补偿 机器学习 度量学习
分 类 号:TG3[金属学及工艺—金属压力加工]
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在链接到云南高校图书馆文献保障联盟下载...
云南高校图书馆联盟文献共享服务平台 版权所有©
您的IP:216.73.216.249