基于CAHA-CatBoost的热轧硅钢凸度边降协同预测  

Collaborative Prediction for Profiles and Edge Reducings of Hot-rolled Silicon Steels Based on CAHA-CatBoost

作　　者：崔航 王笑辰 宋君 王奎越 CUI Hang;WANG Xiaochen;SONG Jun;WANG Kuiyue(Ansteel Beijing Research Institute Co.,Ltd.,Beijing 102200,China)

机构地区：[1]鞍钢集团北京研究院有限公司,北京102200

出　　处：《鞍钢技术》2025年第1期23-28,共6页Angang Technology

基　　金：国家重点研发计划项目(2022YFB3304800)。

摘　　要：由于热轧生产具有非线性、遗传性和强耦合性的特征,因此热轧硅钢断面轮廓控制过程中凸度和边降存在耦合关系。然而,现有研究均采用经验公式来计算热轧硅钢的凸度和边降,精度受限。提出了一种基于混沌人工蜂鸟算法和类别梯度提升算法结合的热轧硅钢凸度边降协同预测模型,结果表明,该预测模型的R2达到0.7932,模型的测试时间仅需0.015 s,大幅优于其他模型,具有更好的精度和更高的效率。Due to the nonlinear,genetic and strongly coupled characteristics in manufacturing hot-rolled products,there is a coupling relationship between profiles and edge reducings in the cross-sections of hot-rolled silicon steels during controlling of their profiles.However,the existing studies for controlling profiles were based on applications of all empirical formulas for calculating the profiles and edge reducings of hot-rolled silicon steels,resulting in limitation of calculation accuracy.Therefore,in order to overcome this limitation,the collaborative prediction model for profiles and edge reducings of hot-rolled silicon steels based on chaotic artificial hummingbird algorithm and categorical boosting(CAHA-CatBoost)was proposed.The study results showed that the R2 of this model was up to 0.7932,and its prediction time was only 0.015 seconds,which was significantly better than other models with better accuracy and higher efficiency.

关 键 词：热轧 硅钢 凸度边降协同预测 混沌人工蜂鸟算法 类别梯度提升算法 

分 类 号：TG337[金属学及工艺—金属压力加工]