中国机械工程

• • 上一篇    下一篇

[成形质量的闭环与自适应控制技术]热轧带钢出口凸度数据驱动建模及智能化预测分析

刘元铭1,2,3,4;王振华2,3;王涛1,2,3;刘文礼1;熊晓燕1   

  1. 1.太原理工大学机械与运载工程学院,太原,030024
    2.太原理工大学先进成形与智能装备研究院,太原,030024
    3.先进金属复合材料成形技术与装备教育部工程研究中心,太原,030024
    4.山西太钢不锈钢精密带钢有限公司,太原,030006
  • 出版日期:2020-11-25 发布日期:2020-11-27
  • 基金资助:
    国家自然科学基金资助项目(51904206,51974196);
    中国博士后科学基金资助项目(2020M670705);
    国家自然科学基金资助重点项目(U1710254);
    山西省科技重大专项(20181102015,MC2016-01);
    山西省应用基础面上青年基金资助项目(201801D221130,201901D211011);
    山西省高等学校科技创新项目(2019L0258,2019L0176)

Data-driven Modeling and Intelligent Prediction Analysis for Hot Strip Outlet Crowns

LIU Yuanming1,2,3,4;WANG Zhenhua2,3;WANG Tao1,2,3;LIU Wenli1;XIONG Xiaoyan1   

  1. 1.College of Mechanical and Vehicle Engineering,Taiyuan University of Technology,Taiyuan,030024
    2.Advanced Forming and Intelligent Equipment Research Institute,Taiyuan University of Technology,Taiyuan,030024
    3.Engineering Research Center of Advanced Metal Composites Forming Technology and Equipment,Ministry of Education,Taiyuan,030024
    4.Shanxi Taigang Stainless Precision Strip Co.,Ltd.,Taiyuan,030006
  • Online:2020-11-25 Published:2020-11-27

摘要: 提出一种基于热轧现场生产数据和智能算法的新型带钢出口凸度预测模型,该模型采用差分进化算法对支持向量机的惩罚因子和核函数宽度进行优化。确定了支持向量回归模型的最佳参数组合,采用大量实际生产数据对模型进行训练并将其用于带钢出口凸度预测。该模型结构简单、容易实现,其整体性能用平均绝对误差(MAE)、平均绝对百分误差(MAPE)、均方根误差(RMSE)和决定系数R2来评价。预测值和实际值的比较验证了所提出模型的可行性。

关键词: 凸度预测, 差分进化算法, 支持向量机, 生产数据, 热轧

Abstract: A new prediction model of strip outlet crowns was proposed based on hot rolling actual production data and intelligent algorithm. This model used differential evolution algorithm to optimize the penalty factor and kernel function width of SVM, and the optimal parameters combinations of support vector regression model were determined. The model was trained with a lot of actual production data and was used to predict the strip outlet crowns. The model structure was simple and easy to implement, and the overall performance was evaluated by mean absolute error, mean absolute percentage error, root mean square error and determination coefficient R2. The feasibility of the proposed model was verified by comparing the predicted values with the actual ones.

Key words: crown prediction, differential evolution algorithm, support vector machine(SVM), production data, hot rolling

中图分类号: