Article
Metallurgy & Metallurgical Engineering
Ji Ya-feng, Song Le-bao, Sun Jie, Peng Wen, Li Hua-ying, Ma Li-feng
Summary: The article introduces the application of an optimized model based on support vector machine to improve the quality of hot strip rolling products. The experimental results show that the PCA-CS-SVM model has the best prediction accuracy and fastest convergence speed, meeting the production requirements.
JOURNAL OF CENTRAL SOUTH UNIVERSITY
(2021)
Article
Computer Science, Artificial Intelligence
Zhenhua Wang, Yuanming Liu, Tao Wang, Dianyao Gong, Dianhua Zhang
Summary: A new hybrid strip crown forecasting model is proposed in this study by combining extreme learning machine (ELM) and industrial data. The accuracy of the model is improved by using principal component analysis (PCA) and an improved particle swarm optimization algorithm based on S-curve decreasing inertia weight (SDWPSO). The performance of the proposed model is evaluated using MAE, MAPE, and RMSE, and compared with three other comparison models. The research shows that the hybrid PCA-SDWPSO-ELM method is suitable for parameter prediction and optimization in the iron and steel manufacturing industry, especially in the process of shape control in hot strip rolling.
Article
Chemistry, Multidisciplinary
Fengwei Jing, Junliang Li, Shimeng Hao, Jie Li, Jing Wang
Summary: This study focuses on the hot strip rolling process and proposes optimal rolling suggestions using neural networks and genetic algorithms. The research shows that the optimized process parameters can improve the rolling stability and meet the limit specifications.
APPLIED SCIENCES-BASEL
(2022)
Article
Engineering, Multidisciplinary
Pengfei Wang, Haifeng Wang, Xu Li, Dianhua Zhang, Wentian Li, Yulin Yao
Summary: A feedforward-feedback coordination control strategy based on double-layer optimization is proposed in this study for the cooperative control of roll bending in the flatness control system. The global optimization model and optimal allocation model have been proven effective in achieving optimal adjustments of actuators between feedforward control and feedback control.
APPLIED MATHEMATICAL MODELLING
(2021)
Article
Metallurgy & Metallurgical Engineering
Qiang Dong, Zhaoyang Li, Huihui Liu, Zhongxue Wang, Li Ji, Lecheng Zhang, Zheng Qi, Guangzhou Gao
Summary: The transversal temperature distribution and phase transformation of strip steel in the hot rolling process were analyzed using thermographic observations and thermal simulation tests. Finite element simulations were used to investigate the effects of temperature distribution and phase state on strip profile. The results showed that the temperature decrease at the edge led to smaller strip crowns with dual-phase state and edge bulges with single-phase state. Measures were proposed and applied to industrial production to prevent edge bulges and control strip crown.
STEEL RESEARCH INTERNATIONAL
(2023)
Article
Metallurgy & Metallurgical Engineering
Wu-quan Yang, Zhi-ting Zhao, Liang-yu Zhu, Xun-yang Gao, Li Wang
Summary: In this study, aiming at the problem of insufficient prediction accuracy of strip flatness at the outlet of cold tandem rolling, different ensemble methods were studied and a high-precision prediction ensemble model of strip flatness at the outlet was established. The results showed that bagging, boosting, and stacking ensemble methods had the most significant improvement in the prediction accuracy of the regression trees model. Among them, boosting method achieved the best performance. The stacking ensemble method improved both simple and complex models, with the greatest improvement effect on the simple base model.
JOURNAL OF IRON AND STEEL RESEARCH INTERNATIONAL
(2023)
Article
Automation & Control Systems
Derrez Mimoune, Mohamed Zaaf, Abdelaziz Amirat
Summary: This work contributes to the improvement of the hydrodynamic method used to predict pressures and rolling speeds during hot rolling of aluminum strips. A critical analysis of the existing method was conducted and an improved approach, which eliminates empirical coefficients and considers the variation of viscosity with pressure, was presented. Finite element simulations validated the improved method, producing reliable results that agree well with experimental data. The proposed approach is rapid and more user-friendly for industrial applications.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2022)
Article
Metallurgy & Metallurgical Engineering
Yu Liu, Xiaojun Wang, Jie Sun, Guangming Liu, Huaying Li, Yafeng Ji
Summary: This study focuses on the complex characteristics of quality control in the hot strip rolling process. A new prediction model for strip thickness, profile, and flatness is developed using machine learning guided by a rolling mechanism model. The fusion of rolling mechanism and process data, combined with optimization algorithms, improves the accuracy of the model. The results show a significant improvement in prediction accuracy, providing theoretical guidance for controlling strip quality and improving the quality of hot-rolled products.
STEEL RESEARCH INTERNATIONAL
(2023)
Article
Engineering, Manufacturing
Qing-Long Wang, Xu Li, Jie Sun, Yuan-Ming Liu, Xin-Chun Zhang, Zhang-Qi Wang
Summary: The SmartCrown technology significantly enhances the cross-directional control capability of strip mill, and the mathematical modeling and 3D finite element analysis validate its effectiveness. The numerical model accuracy is confirmed through rolling force data, and the model is applied to investigate the performance of rolled strips with different widths.
JOURNAL OF MANUFACTURING PROCESSES
(2021)
Article
Materials Science, Multidisciplinary
Zhenhua Wang, Yu Huang, Yuanming Liu, Tao Wang
Summary: Based on industrial data, a strip steel prediction model was constructed using four machine learning algorithms. The research results showed that the XGBoost algorithm achieved a determination coefficient of 0.971 for strip crown prediction, with the lowest error indices, indicating optimal generalization performance. These results contribute to the application of industrial data and machine learning in strip shape control and have practical value for the intelligent preparation of the steel process.
Article
Computer Science, Artificial Intelligence
Qinglong Wang, Jie Sun, Yunjian Hu, Wenqiang Jiang, Xinchun Zhang, Zhangqi Wang
Summary: Flatness deviations in the tandem cold-rolling process have a direct impact on product quality and shape. Conventional physics-based numerical models are inadequate for accurately predicting flatness in this complex operating environment. This study proposes a novel approach using deep convolutional neural networks to effectively predict flatness profiles without additional data pre-processing, achieving remarkable predictive performance with fewer model parameters and lower computational complexity.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Engineering, Multidisciplinary
Hao Wu, Jie Sun, Wen Peng, Dianhua Zhang
Summary: This study investigates the formation mechanism of residual stress induced by incompatible deformation caused by temperature and phase transformation in the run-out table cooling stage of hot-rolled strip based on the fiber rods model. It also establishes an analytical model for calculating residual stresses at different run-out table cooling stages and introduces a strip buckling model to explore the relationship between residual stress and flatness defects. The accuracy of the analytical model is verified through a multi-field coupled finite element model based on the validation of measured data, and the distribution of residual stress calculated by the analytical model is consistent with the finite element model.
APPLIED MATHEMATICAL MODELLING
(2023)
Article
Engineering, Manufacturing
Yu Wang, Changsheng Li, Lianggui Peng, Ruida An, Xin Jin
Summary: This paper develops a prediction model based on CNN to accurately predict strip flatness under different conditions. Data preprocessing methods like the isolated forest algorithm and data folding technique were utilized. The model achieved high accuracy by modifying the loss function and using an Inception module as the basic network structure.
JOURNAL OF MANUFACTURING PROCESSES
(2021)
Article
Metallurgy & Metallurgical Engineering
Tianru Jiang, Kai Zhao, Wei Zhao, Zhimin Lv
Summary: This paper proposes a joint model based on Conditional Generative Adversarial Nets and Artificial Neural Network for process parameters generation and quality prediction. The model can predict the quality of products ahead of schedule using generated parameters, eliminating the need for actual process parameter inputs. Experimental results demonstrate that the generated process parameters align with actual production and the quality prediction accuracy meets production requirements, validating the applicability of the proposed model in simulating the actual rolling process and predicting strip quality ahead of hot rolling production, providing a reference for future planning and scheduling.
IRONMAKING & STEELMAKING
(2023)
Article
Engineering, Multidisciplinary
Hao Wu, Jie Sun, Wen Peng, Lei Jin, Dianhua Zhang
Summary: This study establishes an analytical model for the coupling of temperature, deformation, and residual stress to explore the mechanism of residual stress formation in hot-rolled strip and how to control it. The accuracy of the model is verified by comparing it with a finite element model, and a method to calculate the critical exit crown ratio to maintain strip flatness is proposed.
APPLIED MATHEMATICAL MODELLING
(2024)