为改善轧机控制系统中变形抗力的预报精度，提高产品质量，在分析原单机架轧机控制系统数学模型的基础上，选定温度影响项的系数作为修正系数，并将原长期自学习的算法改为渐消记忆的自适应算法。实现了以实测轧制力数据动态校正变形抗力，并运用VC60++编程实现。实际应用表明，给出的快速自学习策略优于原来的变形抗力工程计算方法，有效减少了产品的厚度误差，提高了板材的成材率和经济效益。
To improve deformation resistance prediction accuracy of the mill control system and quality of products, on basis of analysis of the mathematical model of the single steckel mill control system, the coefficient of temperature was corrected, and the adaptive algorithm of gradually disappeared the memory was adopted to take the place of original long-term learning algorithm. Dynamic correcting deformation resistance was realized with the measured rolling force data by using VC60++. The practical application show that the rapid learning strategies is better than the original deformation resistance engineering calculation method, product thickness error is effectively reduced and the products yield and economic benefit are improved.
参考文献
| [1] | |
| [2] | 刘玠.热轧生产自动化技术 [M].北京:冶金工业出版社，2006：23-37 [2] 费静，李文斌等.热连轧精轧设定模型及其优化方法研究[J].鞍钢技术，2010,1:48-50.[3] 韩朝晖.轧机变形抗力自学习数学模型研究[D].上海：华东师范大学硕士论文，2008.10：31-33. |
- 下载量()
- 访问量()
- 您的评分：
-
10%
-
20%
-
30%
-
40%
-
50%