Research and Application of a Rolling Gap Prediction Model in Continuous Casting

• Published in: Metals (Basel ) Vol. 9; no. 3; p. 380
• Main Authors: Lei, Zhufeng, Su, Wenbin
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 25.03.2019
• Subjects: Artificial intelligence; Automation; Continuous casting; Data acquisition systems; data stream; global optimization; Mathematical models; multi-source information fusion; Neural networks; Optimization techniques; Particle swarm optimization; Phase transitions; prediction; Prediction models; Principal components analysis; Process controls; Process parameters; Roll gap; roll gap value; Sensors; Slab casting; Steel; Support vector machines; support vector regression
• ISSN: 2075-4701; 2075-4701
• DOI: 10.3390/met9030380

Control of the roll gap of the caster segment is one of the key parameters for ensuring the quality of a slab in continuous casting. In order to improve the precision and timeliness of the roll gap value control, we proposed a rolling gap value prediction (RGVP) method based on the continuous casting process parameters. The process parameters collected from the continuous casting production site were first dimension-reduced using principal component analysis (PCA); 15 process parameters were chosen for reduction. Second, a support vector machine (SVM) model using particle swarm optimization (PSO) was proposed to optimize the parameters and perform roll gap prediction. The experimental results and practical application of the models has indicated that the method proposed in this paper provides a new approach for the prediction of roll gap value.