基于机器学习的滚筒式飞剪剪切质量预测  

Prediction of shear quality of drum flying shear based on machine learning

作　　者：辛岩莉 郜志英[1] 张勃洋[1] 王晓勇 XIN Yanli;GAO Zhiyang;ZHANG Boyang;WANG Xiaoyong(School of Mechanical Engineering,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]北京科技大学机械工程学院,北京100083

出　　处：《轧钢》2022年第3期91-96,共6页Steel Rolling

基　　金：国家自然科学基金项目(51775038)。

摘　　要：滚筒式飞剪剪切质量与剪刃重叠量、侧向间隙等参数的取值密切相关,参数匹配不当不仅会影响断面质量,也会使刀具磨损严重,甚至出现剪不断的问题。为了预测选定参数下飞剪的剪切质量以选择合理的工艺参数,通过DEFORM-3D仿真了不同工艺参数下的飞剪剪切过程,运用机器学习算法分别对剪切断面质量和整体剪切效果建立了预测模型。测试集在剪切断面质量的预测模型上准确率接近90%,在整体剪切效果预测模型上的误差小于15%,两类模型为现场选择剪切工艺参数和智能化生产提供了理论指导。The shear quality of drum flying shear is closely related to the values of parameters such as overlap and lateral clearance.Improper matching of parameters will not only affect the cross-section quality, but also cause serious tool wear, even the problem of continuous cutting. In order to predict the shear quality of flying shear under selected parameters and select reasonable process parameters, the shear process of flying shear under different process parameters was simulated by DEFORM-3D, and the prediction models of shear section quality and overall shear effect were established by machine learning algorithms. The accuracy of the test set in the prediction model of shear section quality was close to 90%, and the error in the prediction model of overall shear effect was less than 15%. The two models also provide theoretical guidance for selection of shear process parameters and intelligent production.

关 键 词：滚筒式飞剪 剪切质量 机器学习 DEFORM-3D 

分 类 号：TG335[金属学及工艺—金属压力加工] TP181[自动化与计算机技术—控制理论与控制工程]