Einstein–Roscoe regression for the slag viscosity prediction problem in steelmaking

Saigo, Hiroto; Kc, Dukka B.; Saito, Noritaka

doi:10.1038/s41598-022-10278-w

Cited by 9 publications

(4 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is necessary to mention that property estimations by advanced machine-learning models have been applied to the slag system in recent years. [95,[137][138][139][140][141][142] Density is one of most commonly seen quantity and is relatively easy to be measured. Particularly, it has been found that the slag density varies little at different temperatures, which means the room-temperature measurements can extended for high-temperature applications with fair accuracy.…”

Section: Machine-learning Modelmentioning

confidence: 99%

“…For example, Chen et al [138,139] developed a structure-informed artificial neural network model to predict the viscosity of molten slags with the contribution of composition, temperature, and structural features well included. More recently, Saigo et al [142] applied the Einstein-Roscoe regression to predict slag viscosity even outside the training domain. Note that the machine-learning models are often published or reported while the associated codes are not open source, although these models feature fast and accurate viscosity prediction.…”

Section: Machine-learning Modelmentioning

confidence: 99%

See 1 more Smart Citation

High‐Temperature Thermophysical Property Characterization of Molten Blast Furnace Slag: A Critical Reviews

Liu

et al. 2023

steel research int.

View full text Add to dashboard Cite

Centrifugal granulation‐assisted thermal energy recovery (CGATER) remains the leading technology for the iron and steel industry to enable the low‐carbon treatment of high‐temperature blast furnace (BF) slag. Accurate knowledge of the thermophysical properties of BF slag is pivotal for the commercialization of the CGATER. However, it presents a formidable task to obtain these quantities due to the extremely high temperature of molten BF slag (≈1500 °C). To date, the thermophysical properties of BF slag have not been well characterized yet. Herein, five typical thermophysical properties including density, viscosity, surface tension, thermal conductivity, and emissivity are scrutinized with emphasis on the experimental measurements and empirical estimates. The working principles of these measurement methods are briefly overviewed and their merits and demerits are highlighted. It is noted that density, viscosity, and surface tension have been intensively studied; the quantities can be estimated with moderate uncertainty. However, the thermal conductivity and emissivity of BF slag are less tackled, mainly because of the scarcity of measurement tools at high temperatures. In the future, innovative measurement tools tailored for high‐temperature materials are essential, not only for enriching the thermophysical property database of BF slag but also for paving the way toward characterizing other complex high‐temperature melts.

show abstract

Section: Machine-learning Modelmentioning

confidence: 99%

Section: Machine-learning Modelmentioning

confidence: 99%

High‐Temperature Thermophysical Property Characterization of Molten Blast Furnace Slag: A Critical Reviews

Liu

et al. 2023

steel research int.

View full text Add to dashboard Cite

show abstract

“…On this basis, a dynamic prediction model for the later stages of steelmaking was established, and the effectiveness of the algorithm was demonstrated through trials. Saigo et al [8] proposed the Einstein-Roscoe expression and a transfer learning framework based on the Gaussian process to assist in measuring and estimating regression parameters for the viscosity prediction of the steelmaking process. Experiments showed the suggested technique was effective and that it outperformed other machine learning algorithms in clay prediction.…”

Section: Related Literaturementioning

confidence: 99%

Support Vector Machine with Robust Low-Rank Learning for Multi-Label Classification Problems in the Steelmaking Process

Li¹,

Liu

Guo

2022

Mathematics

View full text Add to dashboard Cite

In this paper, we present a novel support vector machine learning method for multi-label classification in the steelmaking process. The steelmaking process involves complicated physicochemical reactions. The end-point temperature is the key to the steelmaking process. According to the initial furnace condition information, the end-point temperature can be predicted using a data-driven method. Based on the setting value of the temperature before tapping, multi-scale predicted errors of the end-point temperature can be calculated and divided into different ranges. The quality evaluation problem can be attributed to the multi-label classification problem of molten steel quality. To solve the classification problem, considering that it is difficult to capture nonlinear relationships between the input and output in linear models, we propose a novel support vector machine with robust low-rank learning, which has the characteristics of class imbalance without label correlations; a low-rank constraint is used to deal with high-order label correlations in low-dimensional space. Furthermore, we derive an accelerated proximal gradient algorithm and then extend it to handle the nonlinear multi-label classifiers. To validate the proposed model, experiments are conducted with real data from a practical steelmaking problem. The results show that the proposed model can effectively solve the multi-label classification problem in industrial production. To evaluate the proposed approach as a general classification approach, we test it on multi-label classification benchmark datasets. The results illustrate that the proposed approach performs better than other state-of-the-art approaches across different scenarios.

show abstract

“…Concerning the presence of second-phase solid particles on slags, the paper of Saigo et al 15 focuses on the viscosity prediction problem in steelmaking and proposes Einstein–Roscoe regression (ERR), which learns the coefficients of the Einstein–Roscoe equation and is able to extrapolate to unseen domains. In experiments using the viscosity measurements in a high-temperature slag suspension system, ERR is compared favorably with various machine learning approaches.…”

Section: Introductionmentioning

confidence: 99%

Thermodynamic evaluation of viscosity behavior for CaO–SiO2–Al2O3–MgO slag systems examined at the temperatures range from 1500 to 1700 °C

Pereira,

Morales Pereira,

Bielefeldt

et al. 2023

Sci Rep

View full text Add to dashboard Cite

This work showed an application of computational tools to understand systematically the behavior of viscosity on CSAM systems relevant to industrial uses. Consequently in this study, the viscosity experimental data obtained from the literature were compared with the thermodynamic calculated results via the software FactSage v.7.3 for melts in CaO–SiO2–Al2O3–MgO slag system with the range of compositions slags cover 0–100 wt% CaO, 0–100 wt% SiO2, 0–100 wt% Al2O3 and 0–15 wt% MgO at temperature ranges of 1500–1700 °C. Using open-source software in Python, the results of viscosity, liquid, and solid fraction of the slag, as a function of composition and temperature, are represented by multiple color maps and by iso-viscosity contours. The results of the viscosity values indicated that the effect of all the oxides in the CSAM slag system follows the well-known behavior trend observed in the literature. Viscosities of the slag were found to increase with increasing SiO2 contents and decrease with increasing basicities (high CaO). The increase in Al2O3 content increases the viscosity values. An increase of 0–15% MgO depolymerized the slag melt and decreases the viscosity. However, above 5% MgO content occur a decrease in the liquid zone (single phase) and a liquid fraction (two-phase region) of the slag. For a constant MgO concentration, the increase in temperature generates an expansion of low-viscosity zones associated with an increase in the liquid phase of the slag. From the comparison between the calculated and experimental viscosities data keeps up within 30% average relative deviation (Δ), the predictions are considered acceptable for viscosity in the CSAM slag system at high temperatures.

show abstract

Einstein–Roscoe regression for the slag viscosity prediction problem in steelmaking

Cited by 9 publications

References 11 publications

High‐Temperature Thermophysical Property Characterization of Molten Blast Furnace Slag: A Critical Reviews

High‐Temperature Thermophysical Property Characterization of Molten Blast Furnace Slag: A Critical Reviews

Support Vector Machine with Robust Low-Rank Learning for Multi-Label Classification Problems in the Steelmaking Process

Thermodynamic evaluation of viscosity behavior for CaO–SiO2–Al2O3–MgO slag systems examined at the temperatures range from 1500 to 1700 °C

Contact Info

Product

Resources

About