Scholarlink

<p>To enhance the prediction accuracy of rolling force in cold tandem rolling processes, a deformation resistance-rolling force (DR-RF) coupled model is proposed based on dynamic deformation zone length iteration. This DR-RF model comprehensively accounts for the influence of material parameters, hot rolling, and cold rolling processes on deformation resistance, establishing a robust framework for cold rolling. To ensure the model's generalization capability across diverse stands and steel grades, a hierarchical progressive optimization strategy is introduced, leveraging a differential evolution-particle swarm optimization (DE-PSO) hybrid algorithm to effectively mitigate local optima. Experimental validation using industrial data demonstrates significant performance improvements. The unoptimized DR-RF model already exhibits superior accuracy compared to the conventional Hill model. Furthermore, DE-PSO model optimized DR-RF model achieves an overall rolling force prediction accuracy of 94.71%, representing a 7.11% improvement over Hill model. Notably, the first stand shows a 13.4% improvement in accuracy, and the third stand achieves the highest accuracy of 96.68%. The average prediction accuracy for 15 steel grades consistently remains within the range of 90.5%–97.2%. DR-RF model coupled with DE-PSO framework provides a robust theoretical foundation and practical solution for steel enterprises to achieve efficient and intelligent rolling across multiple stands and steel grades.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction model of rolling force in cold tandem rolling based on dynamic deformation zone coupling and DE-PSO hybrid optimization

  • 期刊论文

  • 发表日期：

    2026年5月23日

查看全文
  • Ji Zhang,
  • Zhi-Xuan Wang,
  • Qi Lu,
  • Zhen-Hua Bai

摘要

To enhance the prediction accuracy of rolling force in cold tandem rolling processes, a deformation resistance-rolling force (DR-RF) coupled model is proposed based on dynamic deformation zone length iteration. This DR-RF model comprehensively accounts for the influence of material parameters, hot rolling, and cold rolling processes on deformation resistance, establishing a robust framework for cold rolling. To ensure the model's generalization capability across diverse stands and steel grades, a hierarchical progressive optimization strategy is introduced, leveraging a differential evolution-particle swarm optimization (DE-PSO) hybrid algorithm to effectively mitigate local optima. Experimental validation using industrial data demonstrates significant performance improvements. The unoptimized DR-RF model already exhibits superior accuracy compared to the conventional Hill model. Furthermore, DE-PSO model optimized DR-RF model achieves an overall rolling force prediction accuracy of 94.71%, representing a 7.11% improvement over Hill model. Notably, the first stand shows a 13.4% improvement in accuracy, and the third stand achieves the highest accuracy of 96.68%. The average prediction accuracy for 15 steel grades consistently remains within the range of 90.5%–97.2%. DR-RF model coupled with DE-PSO framework provides a robust theoretical foundation and practical solution for steel enterprises to achieve efficient and intelligent rolling across multiple stands and steel grades.