Abstract:Central carbon segregation is a typical internal defect of continuous cast steel billets. Real-time and accurate carbon segregation prediction is of great significance for lean control of the production quality in continuous casting processes. In this paper, a data-driven regularized extreme learning machine (R-ELM) model is proposed for the prediction of carbon segregation index (CSI). To improve model performance, outliers in industrial data were eliminated by means of boxplot tool. Besides, Pearson correlat… Show more
“…The results indicated that the model was suitable for a complex hot-rolling process. Zou et al [13] established a carbon segregation prediction model for CC billets based on the regularized ELM (RELM) algorithm, by comparing it with multiple linear regression and the classical ELM model. The results showed that the RELM model had better prediction accuracy and generalization ability.…”
Section: Related Workmentioning
confidence: 99%
“…Effective quality prediction can improve the final slab quality, save energy consumption, and stabilize the production. Several machine learning algorithms have been widely used by scholars in the field of CC to construct excellent quality prediction models, such as: random forest (RF) [8], [9], support vector machine (SVM) [10], [11], and artificial neural networks, especially extreme learning machine (ELM) [12], [13]. Nevertheless, most studies ignore the natural class imbalance in CC datasets that the number of normal slabs is much more than that of abnormal slabs.…”
A slab quality prediction model based on machine learning plays an important role in improving final slab quality. However, the class imbalance of continuous casting datasets has a negative impact on the training of basic machine-learning models. In this study, weighted extreme learning machine (WELM) models are constructed to predict the slab quality of under different operation patterns by feeding millions of data. The results show that WELM models can achieve better prediction performance on the two types of continuous casting datasets than the basic algorithms. The superiority of WELM is demonstrated by the relatively high-precision identification of every kind of slab. The performance of WELM models with different weighting schemes is studied and the model with the golden section ratio weighting method is recommended for application as a quality prediction model. Meanwhile, WELM can still maintain a good predictive performance and generalization ability when training a large amount of data. This model can satisfy the demands for slab quality prediction and optimize the continuous casting process.INDEX TERMS Quality prediction, weighted extreme learning machine, continuous casting, class imbalance.
“…The results indicated that the model was suitable for a complex hot-rolling process. Zou et al [13] established a carbon segregation prediction model for CC billets based on the regularized ELM (RELM) algorithm, by comparing it with multiple linear regression and the classical ELM model. The results showed that the RELM model had better prediction accuracy and generalization ability.…”
Section: Related Workmentioning
confidence: 99%
“…Effective quality prediction can improve the final slab quality, save energy consumption, and stabilize the production. Several machine learning algorithms have been widely used by scholars in the field of CC to construct excellent quality prediction models, such as: random forest (RF) [8], [9], support vector machine (SVM) [10], [11], and artificial neural networks, especially extreme learning machine (ELM) [12], [13]. Nevertheless, most studies ignore the natural class imbalance in CC datasets that the number of normal slabs is much more than that of abnormal slabs.…”
A slab quality prediction model based on machine learning plays an important role in improving final slab quality. However, the class imbalance of continuous casting datasets has a negative impact on the training of basic machine-learning models. In this study, weighted extreme learning machine (WELM) models are constructed to predict the slab quality of under different operation patterns by feeding millions of data. The results show that WELM models can achieve better prediction performance on the two types of continuous casting datasets than the basic algorithms. The superiority of WELM is demonstrated by the relatively high-precision identification of every kind of slab. The performance of WELM models with different weighting schemes is studied and the model with the golden section ratio weighting method is recommended for application as a quality prediction model. Meanwhile, WELM can still maintain a good predictive performance and generalization ability when training a large amount of data. This model can satisfy the demands for slab quality prediction and optimize the continuous casting process.INDEX TERMS Quality prediction, weighted extreme learning machine, continuous casting, class imbalance.
“…Rejection of elements from the solid phase leads to an increase in the residual melt and consequently to a positive microsegregation. [ 31–33 ] The primary solidification in steel castings can be columnar dendritic, equiaxed dendritic, or globular, depending on the boundary conditions. [ 34–37 ] During dendritic solidification, the microsegregated melt gets captured between the dendritic arms.…”
Section: Reduction Of Cross‐sectional Area and Ductility Influencing ...mentioning
Continuous casting of premium steel grades requires a process with a high degree of precision and the knowledge about the mechanical behavior of the steel at temperatures above 800 C. Herein, several origins of effects which lead to unwanted impairment of the hot strand shell like segregations, size, amount, kind, and distribution of precipitates as well as porosities from a metallurgical point of view are dealt. The systematic description of potential defect reasons helps to predict harmful operation parameters in context with the chemical composition of steel grades. A compilation of results from experiments at Department of Ferrous Metallurgy of RWTH Aachen University is complemented by a literature review. It is focused on the high temperature ductility and the underlying mechanisms inside the solidifying steel. Finally, potential measures to adjust the continuous casting process to prevent defects are elaborated.
“…Varfolomeev et al [26] and Ye et al [27] used the random forest algorithm to predict crack occurrence in the continuous casting billets. Furthermore, researchers, including the author, have also applied AI algorithms to predict central carbon segregation in continuous casting billets [28][29][30]. However, the previous studies only assessed whether cracks would occur (two-category problem) or determined the probability of cracks.…”
The accurate prediction of internal cracks in steel billets is of great importance for the stable production of continuous casting. However, it is challenging, owing to the strong nonlinearity, and coupling among continuous casting process parameters. In this study, an internal crack prediction model based on the principal component analysis (PCA) and deep neural network (DNN) was proposed by collecting sufficient industrial data. PCA was used to reduce the dimensionality of the factors influencing the internal cracks, and the obtained principal components were used as DNN input variables. The 5-fold cross-validation results demonstrate that the prediction accuracy of the DNN model is 92.2%, which is higher than those of the decision tree (DT), extreme learning machine (ELM), and backpropagation (BP) neural network models. Moreover, the variance analysis showed that the prediction results of the DNN model were more stable. The PCA-DNN model can provide a useful reference for real production, owing to its strong learning ability and fault-tolerant ability.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.