Tank Degasser Research Articles

Predicting the Liquid Steel End-Point Temperature during the Vacuum Tank Degassing Process Using Machine Learning Modeling

The present work focuses on predicting the steel melt temperature following the vacuum treatment step in a vacuum tank degasser (VTD). The primary objective is to establish a comprehensive methodology for developing and validating machine learning (ML) models within this context. Another objective is to evaluate the model by analyzing the alignment of the SHAP values with metallurgical domain expectations, thereby validating the model’s predictions from a metallurgical perspective. The proposed methodology employs a Random Forest model, incorporating a grid search with domain-informed variables grouped into batches, and a robust model-selection criterion that ensures optimal predictive performance, while keeping the model as simple and stable as possible. Furthermore, the Shapley Additive Explanations (SHAP) algorithm is employed to interpret the model’s predictions. The selected model achieved a mean adjusted R2 of 0.631 and a hit ratio of 75.3% for a prediction error within ±5 °C. Despite the moderate predictive performance, SHAP highlighted several aspects consistent with metallurgical domain expertise, emphasizing the importance of domain knowledge in interpreting ML models. Improving data quality and refining the model framework could enhance predictive performance.

Studies on Parameters Affecting Temperature of Liquid Steel and Prediction Using Modified AdaBoost.RT Algorithm Ensemble Extreme Learning Machine

The present work aimed to develop a predictive model for the end temperature of liquid steel in advance to support the smooth functioning of a vacuum tank degasser (VTD). An ensemble model that combines extreme learning machine (ELM) with a self-adaptive AdaBoost.RT algorithm was established for the regression problem. Based on analyzing the energy equilibrium of the VTD system, the factors were determined for predicting the end temperature of liquid steel. To establish a hybrid ensemble prediction model, an ELM algorithm was selected as the ensemble predictor due to its strong performance and robustness, and a modification of the AdaBoost.RT algorithm is proposed to overcome the drawback of the original AdaBoost.RT by embedding statistical theory to dynamically self-adjust the threshold value. For efficient VTD operations, an ensemble model that combines ELM with the self-adaptive AdaBoost.RT algorithm was established to model the end temperature of liquid steel. The proposed approach was analyzed and validated on actual production data derived from a steelmaking workshop in Baosteel. The experimental results reveal that the proposed model can improve the generalization performance, and the accuracy of the model is feasible for the secondary steel refining process. In addition, a polynomial equation is obtained from the ensemble predictive model for calculating the value of the end temperature. The predicted results are in good agreement with the actual data with <1.7% error.

The Degassing Laws for Railway Wheel Steel in a Vacuum Tank Degasser

The research presents the results of data analysis on degassing of wheel grades of steels in a tank degasser with a capacity of 120 tons, operated at the JSC “Ural Steel”. The volume of the analyzed sample included 754 steels for railway wheels (steel grades “2” and “T” according to State standard GOST 10791-2011) weighing more than 80 thousand tons received in November-December 2019.It was established that in order to guarantee the production of hydrogen content of less than 1.5 ppm and nitrogen before 0.007%, it is necessary to carry out vacuum treatment of metal with overheating of 110-130°C at the residual pressure of up to 3 mbar for 20-25 minutes and argon flow rate of at least 0.05 m3/ton. The regression equation was obtained, which allows to predict the results of degassing, as well as select the values of vacuum treatment parameters in order to achieve a given content of dissolved gases - hydrogen and nitrogen.

Objective Laws of Nitrogen Removal in the Vacuum Tank Degasser

This article discusses the problems of efficient removal of nitrogen in the vacuum tank degasser operating at the electric arc furnace shop of Uralskaya Stal (Ural Steel) OJSC in the course of manufacturing of high-quality low carbon steel grades by means of steel grades «2» and «T». In order to determine the reasonable and balanced treatment parameters that ensure the required level of nitrogen content in the above steel grades, an analysis of production data for the period of November-December 2016 has been carried out. This analysis is the basis for identifying the vacuum degassing parameters in compliance with the technological capabilities and well-balanced levels, which allow predicting the level of nitrogen content in steel. To assess the cumulative quantitative effect of the main parameters of vacuum degassing on nitrogen removal, there has been performed a regression analysis. As a result, there have been obtained multiple regression equations describing a rational combination of steel treatment parameters for achieving the required nitrogen removal level.

Vibration‐Based Monitoring of Gas‐Stirring Intensity in Vacuum Tank Degassing

Liquid steel is typically stirred in a vacuum tank using argon gas injection to achieve a homogeneous composition and high‐purity steel. The aim of this work is to study the effect of vessel vibration on the operational state monitoring of the gas stirring in a vacuum tank degasser. Following an extensive analysis of vibration features, the root mean square (RMS) of vertical velocity is found to be the best feature for the measurement of the stirring intensity caused by the volumetric gas injection rate into the ladle. Smoothing is conducted using a centered median filter with a window length of 21 s. In this work, the operational state monitoring of gas stirring is described using a ladle responsiveness value (LRV). This describes the ability of a ladle to generate the maximum amount of vibration with the minimum amount of argon gas. The LRV summarized for each ladle reveals significant differences between them. Correspondingly, a rolling ladle responsiveness value (rLRV) is used for online monitoring of possible gas leakages. The rLRV can also be used for the online monitoring of the stirring efficiency and as its comparison with the overall efficiency of a specific ladle or all ladles.

An Integrated Methodology for Rule Extraction from ELM-Based Vacuum Tank Degasser Multiclassifier for Decision-Making

The present work proposes an integrated methodology for rule extraction in a vacuum tank degasser (VTD) for decision-making purposes. An extreme learning machine (ELM) algorithm is established for a three-class classification problem according to an end temperature of liquid steel that is higher than its operating restriction, within the operation restriction and lower than the operating restriction. Based on these black-box model results, an integrated three-step approach for rule extraction is constructed to interpret the understandability of the proposed ELM classifier. First, the irrelevant attributes are pruned without decreasing the classification accuracy. Second, fuzzy rules are generated in the form of discrete input attributes and the target classification. Last but not the least, the rules are refined by generating rules with continuous attributes. The novelty of the proposed rule extraction approach lies in the generation of rules using the discrete and continuous attributes at different stages. The proposed method is analyzed and validated on actual production data derived from a No.2 steelmaking workshop in Baosteel. The experimental results revealed that the extracted rules are effective for the VTD system in classifying the end temperature of liquid steel into high, normal, and low ranges. In addition, much fewer input attributes are needed to implement the rules for the manufacturing process of VTD. The extracted rules serve explicit instructions for decision-making for the VTD operators.

Denitrogenation model for vacuum tank degasser

Nitrogen in steel is both beneficial and detrimental depending on grade of steel and its application. To get desired low nitrogen during vacuum degassing process, VD parameters namely vacuum level, argon flow rate and holding time has to optimized depending upon initial nitrogen level. In this work a mathematical model to simulate nitrogen removal in tank degasser is developed and how various VD parameters affects nitrogen removal is studied. Ladle water model studies with bottom purging have shown two distinct flow regions, namely the plume region and the outside plume region. The two regions are treated as two separate reactors exchanging mass between them and complete mixing is assumed in both the reactors. In the plume region, transfer of nitrogen to single bubble is simulated. At the gas-liquid metal interface (bubble interface) thermodynamic equilibrium is assumed and the transfer of nitrogen from bulk liquid metal in the plume region to the gas-metal interface is obtained using mass transport principles. The model predicts variation of Nitrogen content in both the reactors with time. The model is validated with industrial process and the predicted results were found to have fair agreement with the measured results.

Nitrogen Injection in Molten Aluminum in a Tank Degasser

The injection of nitrogen in molten aluminum through a static impeller in a tank degassing unit is studied. Using basic principles of fluid mechanics, it is analyzed the influence of the nozzle diameter on the bubble diameter and the mean residence time of the bubbles in the molten aluminum. By means of transient isothermal 2D Computational Fluid Dynamics (CFD) simulations, the influence of the nitrogen volumetric flow rate on the phase distributions and the tank degasser dynamics is studied. Finally, an adiabatic CFD simulation is carried out in order to elucidate the changes in the molten aluminum temperature due to the injection of nitrogen at ambient temperature. This simulation shows that molten aluminum does not suffer drastic temperature reductions given that, in spite that the nitrogen is fed at ambient temperature, the mass of nitrogen is relatively small compared with the mass of aluminum.

Study of Low Flow Rate Ladle Bottom Gas Stirring Using Triaxial Vibration Signals

Secondary steelmaking plays a great role in enhancing the quality of the final steel product. The metal quality is a function of metal bath stirring in ladles. The metal bath is often stirred by an inert gas to achieve maximum compositional and thermal uniformity throughout the melt. Ladle operators often observe the top surface phenomena, such as level of meniscus disturbance, to evaluate the status of stirring. However, this type of monitoring has significant limitations in assessing the process accurately especially at low gas flow rate bubbling. The present study investigates stirring phenomena using ladle wall triaxial vibration at a low flow rate on a steel-made laboratory model and plant scale for the case of the vacuum tank degasser. Cold model and plant data were successfully modeled by partial least-squares regression to predict the amount of stirring. In the cold model, it was found that the combined vibration signal could predict the stirring power and recirculation speed effectively in specific frequency ranges. Plant trials also revealed that there is a high structure in each data set and in the same frequency ranges at the water model. In the case of industrial data, the degree of linear relationship was strong for data taken from a single heat.

Mathematical Modeling of Nitrogen Removal from the Vacuum Tank Degasser

The removal of nitrogen from an industrial vacuum tank degasser depends on a series of operational parameters, steel composition, and contents of surface‐active elements in liquid steel, e.g., oxygen and sulfur. The effect of some specific elements on nitrogen removal in the vacuum degasser has been (well) examined. Still, it is quite challenging to assess the overall effect of the whole steel composition on the process. The focus of the present work was to predict nitrogen removal from the vacuum degasser specially taking into account the multi‐component effect of steel composition. An integrated computational fluid dynamics (CFD) model for simulating nitrogen (and hydrogen) removal in industrial vacuum tank degassers was therefore developed based on theories and methods that are relatively separate in the literature. In order to include the multi‐component effect, the model is coupled with an in‐house thermodynamics code that can be used to determine the activity coefficient of nitrogen, oxygen, or sulfur as a function of steel composition and temperature. The code was verified by comparing the calculated activity coefficients against experimental measurements from various sources. Efforts were also put into developing an on‐line use concept to control the nitrogen removal. The gas plume, flow field, and evolutions of nitrogen and hydrogen during vacuum treatment were predicted and the on‐line concept was demonstrated by presenting two operating diagrams. The results showed that the final nitrogen content decreases with an increase in the chemical reaction rate constant and a decrease in the initial nitrogen content. By contrary, the final nitrogen content increases with a decrease in the chemical reaction rate constant and an increase in the initial nitrogen content. Finally, the operating diagrams were validated by industrial data and observations.

Modeling Study of Nitrogen Removal from the Vacuum Tank Degasser

Low nitrogen content in liquid steel is required for most of the steelmaking companies, where vacuum degassing of liquid steel is usually carried out to remove nitrogen as well as other impurity elements during ladle treatment. This paper presents an integrated computational fluid dynamics (CFD) model for simulating the nitrogen removal in an industrial vacuum tank degasser (VTD). In the CFD model, oxygen and sulfur are considered as surface‐active elements decreasing the denitrogenation rate at the gas–steel interface. An activity coefficient and a desulfurization sub‐model are implemented to compute the effective activity coefficients of the related elements and the evolution of sulfur content in liquid steel during the process, respectively. The effect of various elements on denitrogenation rate in the VTD is studied with the developed model. For the used steel compositions the denitrogenation rate decreases with an increase in the content of [S], [O], [Cr], [Nb], [V], [Ti], or [Mn], whereas it increases with an increase in the content of [C], [Si], [P], [Ni], or [Al]. The model is validated by comparing final nitrogen contents measured from the industrial plant with the ones predicted by the model and good agreements were found. The model has proven to be a useful and accurate tool for predicting final nitrogen content of liquid steel in the plant.

Numerical Study on the Removal of Hydrogen and Nitrogen from the Melt of Medium Carbon Steel in Vacuum Tank Degasser

The steelmaking field has been seeing an increased demand of reducing hydrogen and nitrogen in liquid steel before casting. This is often accomplished by vacuum treatment. This paper focuses on developing a numerical model to investigate the removal of hydrogen and nitrogen from the melt of medium carbon steel in a commercial vacuum tank degasser. An activity coefficient model and the eddy-cell expression are implemented in the ANSYS FLUENT code to compute the activities of related elements and mass transfer coefficients of hydrogen and nitrogen in liquid steel. Several cases are simulated to assess the effect of gas flow rate and initial nitrogen content in liquid steel on degassing process and the calculated results are compared with industrial measured data.

Numerical Simulation of Dehydrogenation of Liquid Steel in the Vacuum Tank Degasser

Vacuum tank degassers are often utilized to remove hydrogen from liquid steel. A new comprehensive numerical model, which has been developed to simulate hydrogen removal in the vacuum degassers, is presented in this paper. The degassing model consists of two sub-models, which calculate the gas-steel flow field and the species transport of hydrogen. An extended k–e turbulence model is adopted to consider the effect of gas injection on the turbulent properties and an interfacial area concentration model is introduced to compute the interfacial area density between liquid steel and the bubbles. The fluid dynamic sub-model is validated with a physical gas stirred tank, which is believed to have similar flow phenomena as the studied vacuum degasser based on the modified Froude number. Two fundamental expressions for mass transfer coefficient, which have been paid little attention by the researchers concentrating on vacuum degassing, are evaluated with a simulation case corresponding to practical operation. The effect of vacuum pressure on the dehydrogenation process is investigated and, moreover, the integrated model is verified with industrial measurements. The predicted final hydrogen contents in liquid steel show good agreement with the measured ones. The model and the main results are presented.

Study of Fluid Flow and Mixing Behaviour of a Vacuum Degasser

In recent times the demand of ultra-low carbon steel (ULCS) with improved mechanical properties such as good ductility and good workability has been increased as it is used to produce cold-rolled steel sheets for automobiles. For producing ULCS efficiently, it is necessary to improve the productivity of the vacuum degassers such as RH, DH and tank degasser. Recently, it has been claimed that using a new process, called REDA (revolutionary degassing activator), one can achieve the carbon content below 10 ppm in less time. As such, REDA process has not been studied thoroughly in terms of fluid flow and mass transfer which is a necessary precursor to understand and design this process. Therefore, momentum and mass transfer of the process has been studied by solving momentum and species balance equations along with k–e turbulent model in two-dimension (2D) for REDA process. Similarly, computational fluid dynamic studies have been made in 2D for tank and RH degassers to compare them with REDA process. Computational results have been validated with published experimental and theoretical data. It is found that REDA process is the most efficient among all these processes in terms of mixing efficiency. Fluid flow phenomena have been studied in details for REDA process by varying gas flow rate, depth of immersed snorkel in the steel, diameter of the snorkel and change in vacuum pressure. It is found that design of snorkel affects the melt circulation in the bath significantly.

Enhancement in Bath Mixing and Plume area in a New Degassing Process - A Computational Fluid Dynamic Study

Reaction between the various species in slag and metal phase is usually mass transfer controlled. There have been continuous efforts to increase the reaction efficiency in slag-metal system, especially during decarburization of steel to produce the ultra low carbon steel (ULCS) in secondary steelmaking. It has been found that the surface reaction is a dominant factor in the final stage of decarburization. In the initial stage, the inner site reaction is major factor in the refining process. The mixing of bath affects the later reaction. However, the former reaction (surface reaction) is affected by the plume size area at the top of the metal surface. Therefore, a computational study has been made to understand the fluid dynamics of a new secondary steelmaking process called Revolutionary Degasser Activator (REDA) to study the bath mixing and plume area. REDA process has been considered as it is claimed that this process can reduce the carbon content in steel below 10ppm in a less time than the other existing processes such as RH and Tank degasser. This study shows that both bath mixing and plume area are increased in REDA process facilitating it to give the desired carbon content in less time. Qualitative comments are made on slag-metal reaction system based on this finding.

Computational Fluid Dynamics Study of New Vacuum Degassing Process

Recently, the demand of the steel having superior chemical and physical properties has increased for which the content of carbon must be in ultra low range. There are many processes which can produce low carbon steel such as tank degasser and RH (Rheinstahl-Heraeus) processes. It has been claimed that using a new process, called REDA (Revolutionary Degassing Activator), one can achieve the carbon content below 10ppm in less time. REDA process, in terms of installment cost, is in between the tank degasser and RH processes. As such, REDA process has not been studied thoroughly. Fluid flow phenomena affect the decarburization rate the most besides the chemical reaction rate. Therefore, momentum balance equations along with k-? turbulent model have been solved for gas and liquid phases in two-dimension (2D) for REDA process. The fluid flow phenomena have been studied in details for this process by varying gas flow rate, depth of immersed snorkel in the steel, diameter of the snorkel and change in vacuum pressure. It is found that the design of the snorkel affects the melt circulation of the bath significantly.

Desulfurization practices in ArcelorMittal Flat Carbon Western Europe

The development of steel refining operations at ArcelorMittal FCWE for the production of ULS steels is reported. Several plants with different equipments produce grades with S <30 ppm, thus permitting to identify the most efficient process, from hot metal desulfurization to secondary metallurgy. For sites producing steels with S <30 ppm, four secondary metallurgy routes coexist in ArcelorMittal FCWE. The Vacuum Tank Degasser makes it possible to reach very low S and N contents simultaneously. To face the increase of the share of Ultra Low Sulfur products, it is necessary to further enhance the kinetics of desulfurization by steel - slag stirring at the atmospheric pressure.

Carbon Concentration in ULC Steels Numerically Tracked in Vacuum Processes

Ultra low carbon (ULC) steel grades for packaging or automotive industry are produced in ARCELOR group either by RH or Vacuum Tank Degasser (VTD) routes. Industrial Operations Research Center (IORC : ex IRSID) has developed kinetic models of decarburization, based on the physical description of the different reaction sites for each type of reactor. These models have been largely used in past years to adjust the operating conditions during decarburization. To go further, the computational fluid dynamics Fluent software, used to characterize the ladle hydrodynamics in terms of velocity and turbulence of the liquid steel flows, was coupled to the kinetic models in order to better assess the effect of metal liquid flow on the performances of decarburization. Numerical computations have been performed in such a way, from industrial configuration and running conditions. Computations show that, whatever the reactor-type, heterogeneity in carbon concentration is generated during the vacuum decarburization treatments of ULC steels. The highest differences become apparent at the beginning of the treatment and can reach up to 80 ppm in RH, then progressively decrease. Heterogeneity in concentration still exists after 15 or 20 minutes for a classical treatment time. The maximum differences between poor and rich areas in ladle vary from 5 ppm to 10 ppm, depending on stirring and fluid flow conditions. They are more pronounced in RH than in VTD and affect a larger volume of metal. But, in both cases, heterogeneity completely disappears 3 to 4 minutes after killing. These numerical results are consistent with specific measurements, carried out on an industrial installation. A finer description of the RH reactor from a chemical point of view, obtained from residence time distribution computations, largely improves the numerical results and gives some complementary explanations on the RH working.

Carbon concentration tracked by numerical simulation in vacuum tank degasser during decarburisation treatment

Two-phase flows in gas stirred steel bath and their effects on the carbon concentration distribution in the ladle during a decarburisation vacuum treatment are analysed. For this purpose, a numerical tool based on the coupling between the CFD Fluent software and kinetic model developed at IRSID has been used. The CFD Fluent software, equipped with a Lagrangian approach, provides a local description of the ladle hydrodynamics (molten steel flow and turbulence). The kinetic model is based on the physical description of the different reaction sites in the vacuum tank degasser. It is used to determine the evolution of the global carbon concentration in ladle. The coupling gives a complete mapping of the carbon concentration in the ladle according to time treatment. Numerical results applied to a water model configuration reveal good agreement in terms of velocity and turbulence profiles with experimental data provided from laser devices. In industrial configuration, the effect of the bath surface pressure and of the gas flow rate is analysed in terms of generated hydrodynamics and carbon distribution in the ladle.

Mechanisms and Modeling of Liquid Steel Decarburization below 10 ppm Carbon

Steel grades for packaging or automotive industry applications require ultra low carbon levels and carbon contents below 10 ppm have to be obtained in the industrial decarburization reactor (tank degasser or vacuum circulation degasser, known as RH). In the ultra low carbon range (below 10 ppm in the vacuum vessel), the decarburization kinetics cannot be obtained from the analysis of metal samples but can be deduced from continuous exhaust gas analysis and flow rate measurements. This was done on an industrial RH-OB degasser and a reliable and precise measurement of reaction rates in the ultra low carbon range was obtained. The rate was evaluated by an overall volumetric coefficient of decarburization ak⁄S. A model of decarburization kinetics in vacuum vessel has been developed at IRSID based on a physical description of three different reaction sites: spontaneous CO boiling in metal layer just beneath bath surface where carbon and oxygen contents are in excess with respect to equilibrium CO partial pressure, decarburization by purging gas in zones where metal and injected gas bubbles are in contact and decarburization of metal drops projected in vacuum vessel by bubble bursting. The experimental results indicate that CO boiling is responsible for the largest part of decarburization, but, the decarburization of metal drops projected in vacuum vessel is dominating below 10 ppm. The dynamic modeling of the process describes simultaneously the evolution of vessel pressure, oxygen and carbon contents. It explains with very good accuracy the experimental results and is used to define optimal process parameters for industrial treatments.