In order to control the surface quality of continuous casting slabs, the changes in the composition and physical properties of mold flux during continuous casting were studied. In this study, three mold fluxes with Al2O3 content between 1.3 and 6.0mass% and three different basicities were used. Continuous casting steel with aluminum content of 0.7 mass% Al and conventional Al-killed steel. The research results are summarized as follows: (1) During the continuous casting process, the content of Al2O3 in the mold flux was increased to 30 mass %. Using the Equilibrium Effective Reaction Zone Model (EERZM), referring to the continuous casting results, the change of the composition of the mold flux can be reproduced through parameter fitting. (2) The analysis of the equilibrium effective reaction zone model EERZM shows that the viscosity of mold flux and the flow rate of molten steel affect the increase rate of Al2O3 content in mold flux. (3) According to the change of mold powder composition, the change of physical properties of mold powder is estimated. Analysis by FactSage software can estimate mold flux crystallization temperature and changes of main crystals. (4) Considering the influence of Al2O3 as an amphoteric oxide, the viscosity of the mold flux liquid slag can be estimated by modifying the Iida equation.
Key words: continuous casting; high aluminum steel; mold flux; composition change; modeling; crystallization temperature; viscosity
Introduction
Mold mold flux plays several important roles in molten steel continuous casting, including (1) heat insulation of molten steel, (2) prevention of secondary oxidation of molten steel, (3) adsorption of inclusions, (4) slowing down of primary Solid shell cooling, (5) to ensure the lubricating performance between the crystallizer and the solidified shell. It is extremely important to design and use the optimum mold flux that can achieve these effects according to the molten steel composition and continuous casting conditions. Although mold fluxes with appropriate properties are usually designed and used according to molten steel composition and continuous casting conditions, even mold fluxes with appropriate properties often cause problems in continuous casting operations, for example, in continuous casting high For alloy steels, such as high aluminum steel, high manganese steel and titanium-containing steel, the composition change of mold powder is an example that affects the quality of casting slab production. It has been reported that the compositional change of mold flux during continuous casting of high alloy steel is caused by the redox reaction between molten steel and mold flux. Mold powder generally contains a certain amount of SiO2, but elements such as Al, Mn, and Ti in high-alloy steel reduce the SiO2 in mold powder, resulting in significant changes in the composition of mold powder, resulting in continuous casting process protection Significant changes in slag properties make continuous casting process conditions unstable, causing various production problems and quality defects in slabs. Therefore, in order to improve the quality of high-alloy steel (high Al steel, high Mn steel, high Ti steel) and the stability of continuous casting, it is very important to grasp the composition change of mold powder during continuous casting. This paper will take high-aluminum steel as an example to illustrate the instability mechanism of the above-mentioned high-alloy steel continuous casting process. In the continuous casting of high aluminum steel, the redox reaction of Si and Al occurs between molten steel and liquid mold slag, as shown in formula (1). This reaction makes the concentration of SiO2 in the liquid mold flux greatly reduced, while the concentration of Al2O3 is greatly increased, and the composition and alkalinity of the mold flux are significantly changed. According to reports, in the process of continuous casting of high aluminum steel, the content of Al2O3 in the liquid mold flux of the mold increases from less than 5mass% to a maximum of about 35mass%, while the content of SiO2 decreases from more than 40mass% to about 20mass%. Since SiO2 is the main component of mold flux and a key factor in the molten oxide network, it has a great influence on the performance of mold flux. The change of SiO2 content will cause significant changes in the crystallization temperature and viscosity of mold flux, while the crystallization temperature And viscosity are important characteristics affecting the stability of continuous casting. In this case, the liquid mold slag absorbed between the primary solidified slab shell and the inner wall of the copper plate becomes uneven, forming a large depression, which may cause cracks in the slab and breakout accidents (BO). This is a very serious problem for casting production.
In this study, in order to prevent the occurrence of instability during continuous casting of high alloy steel, continuous casting tests and measurements were carried out on the properties of mold flux, and the validity of the following items was verified. •Analysis model for mold flux composition changes during continuous casting •A model for analyzing mold flux crystallization temperature changes due to composition changes •A model for analyzing viscosity changes caused by composition changes Mold flux, the composition and properties of the flux vary greatly.
Experimental method and calculation method
Composition change prediction model
The prediction model of the composition change of the mold flux during the silicon-aluminum redox reaction between the molten steel and the mold flux was studied, and various models for predicting the composition change of the mold flux during the continuous casting of high-alumina steel were proposed. In this study, the Equilibrium Effective Reaction Zone Model (EERZM), which combines thermodynamic calculations (FactSage) and hybrid models (ERZM: Effective Reaction Zone Model), was used. EERZM is a model that combines thermodynamic calculation software (FactSage) and hybrid model (ERZM) proposed by Van Ende et al. The purpose is to reproduce the change of mold powder composition in the continuous casting process of high alloy steel. FactSage is a software for calculating the thermodynamic equilibrium state of a multi-component system. It has the functions of calculating the equilibrium phase diagram and alloy phase diagram of the oxide system, preparing the potential-pH diagram, predicting the viscosity of slag glass, and preparing the thermodynamic database. The mixing model (ERZM) is a simple model for simulating interzone mixing reaction rates and mass transfer. The model divides the reaction region near the two-phase interface into multiple regions, calculates the equilibrium reaction between each region, and then simulates the reaction rate and mass transfer through mixing. At this time, the reaction rate and mass transfer degree of each phase interface reaction are simulated by changing the thickness (volume) of each region. Although the model only considers equilibrium reactions, changes in composition can be reproduced by taking into account reaction rates and mass transfer. Fig. 1 shows a schematic diagram of the prediction model for the composition change of mold flux used in this study. In this model, the mold slag liquid slag layer is divided into liquid slag layer 1, liquid slag layer 2, and liquid slag layer 3, and the reaction between molten steel and liquid slag layer 1 is regarded as the molten steel layer. As mentioned above, the reaction rate of the reaction between molten steel and mold slag and the mass transfer in the mold slag are simulated by changing the thickness of these layers, that is to say, the thickness of each liquid slag layer is a parameter of this model.
Fig.1 Schematic diagram of mold flux equilibrium effective reaction zone model (EERZM)
The model assumptions are as follows.
• The temperature of the liquid slag layer 1 in contact with the liquid steel layer is the same as that of the liquid steel layer.
•All mold fluxes entering the system flow in molten state.
•The mold flux flowing out from the system is the same as the consumed flux, flowing out from each layer according to the thickness of each layer.
• The size of each slag and steel layer is the same as the thickness and width of the mold.
•The thickness of each slag layer and steel layer is constant during the continuous casting process.
• Both solid and liquid phases are considered in equilibrium calculations.
• The overall thickness of the slag layer in the model is the same as the thickness of the liquid slag layer in the continuous casting process.
The calculation process will be briefly described here. Firstly, the equilibrium reaction between the molten steel and the liquid slag layer adjacent to the molten steel (ie, between the molten steel layer and the liquid slag layer 1 ) is calculated. Then calculate the inflow and outflow from liquid slag layer 1 to liquid slag layer 3 in sequence. At this time, the outflow of each layer is obtained by multiplying the consumption of mold slag by the ratio of the thickness of each layer to the thickness of the entire liquid slag layer, and the corresponding outflow of each layer is the same as the inflow of the previous layer. In the case of the liquid slag layer 3, the amount of mold flux inflow to the liquid slag layer 3 is the same as the amount of mold flux consumption. Finally, the mixing ratio of each layer of the adjacent liquid slag layer is 50%, and the average composition of the mold slag in the whole mold is the output mold slag result. These calculations are performed repeatedly with one step length, assuming that the completion time of one step is 1 min, and the calculation conditions are shown in Table 1. Assume that the temperature of each layer is constant, all inclusions are Al2O3, and assume that the flow rate from the molten steel into the mold powder is constant.
Table 1 Calculation conditions
Table 2 shows the test continuous casting conditions. The continuous casting steel types are medium carbon-0.7mass% Al high aluminum steel, aluminum-killed ultra-low carbon steel and low carbon steel. Select three kinds of mold fluxes with different basicity and properties (mold flux A, mold flux B and mold flux E), the alkalinity and performance of each case are different, and the ITP setting of each state measurement index is different. Under the molten steel flow rate RSteel-Mold flux and the mold powder flow rate get the index, define the case 1 as 1. The molten steel flow RSteel-Mold flux is expressed by the formula (2).
Table 2 Continuous casting conditions
RSteel-Mold flux [t/kg]: molten steel flow relative to the mold flow, QSteel [t/min]: molten steel flow (= the amount of molten steel entering the mold), QMold flux [kg/min]: the amount of mold slag added , ρ [t/m3]: molten steel density, Vc [m/min]: casting speed, tslab [m]: slab thickness, wslab [m]: slab width, CMold flux [kg/m2]: crystallization Mold slag consumption For the results of Al2O3 content changes obtained in the test casting under these conditions, the thickness of each layer (tSlag layer1, tSlag layer2, tSlag layer3, tSteel layer) of the model parameters was changed and adjusted.
Performance Estimation Analysis Model
The method of estimating the crystallization temperature and viscosity of mold powder which affects the stability of continuous casting is studied. The composition and properties of the mold fluxes used in this study are shown in Table 3. The basicities of the fluxes B to D are different, and the increase of Al2O3 content in each flux is simulated. The FactSage software method is used to estimate the characteristics of the mold flux. The crystallization temperature is the highest temperature at which the solid phase exists, and the main crystal type is the solid phase that exists at the crystallization temperature. The mold flux viscosity of each component at 1573K was calculated using the FactSage module, and then the various properties of the mold flux were measured. The samples obtained by quenching and cooling mold flux on water-cooled copper plates were analyzed by X-ray diffraction (XRD). Under the condition of cooling rate of 1300℃~℃/min, the crystallization temperature was measured by differential scanning calorimeter (DSC), and the viscosity of mold powder was measured by rotational viscometer at 1573K.
Table 3 Composition and measured properties of mold flux
Experimental and calculation results
Composition change prediction model
Figure 2 shows the results of Al2O3 concentration changes in the mold flux during the continuous casting test. In Case1 and Case2, the Al content in molten steel is relatively high, and the content of Al2O3 in mold slag changes greatly during continuous casting. In the case of using high-basic mold flux B, the concentration of Al2O3 in the mold flux is the highest, and its concentration can reach about 30 mass %. In Case3 and Case4 with lower Al content in molten steel, the concentration of Al2O3 in mold flux increased by about 5 mass %.
Fig.2 Changes of Al2O3 content in mold flux during continuous casting test
Figure 3 shows the results of adjusting the model parameters according to the actual changes in the Al2O3 content in Figure 2. In each case, by setting specific parameters, it was possible to reproduce the variation of the actual Al2O3 content during the experimental casting process. Table 4 shows the model parameters that reproduce the actual variation in Al2O3 content in each case. Among them, the model parameters rsteel.layer, rSlag.layer1, rSlag.layer2, and rSlag.layer3 are the ratios of the thickness of each layer to the rtotal slag.layer thickness of the entire mold flux (formulas (3), (4)).
Figure 3 The measured and calculated Al2O3 content changes in the mold powder during the continuous casting test
Table 4 Continuous casting conditions and model parameters
i = liquid steel layer slag layer 1, liquid slag layer 2 and liquid slag layer3
ri[-]: the ratio of the thickness of each layer to the thickness of the entire mold flux layer, ti [mm]: the corresponding thickness of each layer, tTotal.layer[mm]: the thickness of the entire liquid mold flux layer. As shown in Table 4, the model parameters are different in each case. The difference in model parameters in the two cases is considered to be due to the difference in mold powder properties and molten steel in and out. The relationship between the initial component viscosity of the mold powder and the rSlag layer1 and the relationship between the input index ITP and the rSteel layer are shown in Figure 4. It can be understood that the thickness of the liquid slag layer 1 decreases with the increase of the initial component viscosity, and the thickness of the molten steel layer increases with the increase of the input and output index of molten steel.
Figure 4 (a) The relationship between the mold powder viscosity and rSlag layer1; (b) the relationship between ITP and rSteel layer
Performance Estimation Model
The results of estimating the main crystal types and crystallization temperature with FactSage are shown in Table 5, Figure 5 and Figure 6. The main crystal species estimated by FactSage are basically consistent with the main crystal species observed. In terms of crystallization temperature, although the average value of the absolute value of the difference between the measured value and the estimated value of the mold flux in the whole experiment is as high as about 50°C, the trend of the crystallization temperature changing with the Al2O3 content is basically consistent with the measured values of all mold fluxes.
Table 5 Comparison of main crystals estimated by FactSage and measured main crystals
Fig.5 XRD pattern of mold powder B (top) and B-20 (bottom)
Figure 6 Comparison of FactSage crystallization temperature estimates and measured values
The results of viscosity estimation using FactSage are shown in Fig. 7. It can be seen from the figure that the trend of viscosity change with Al2O3 content is inconsistent with the measured value, especially in the area where the Al2O3 concentration is 25 mass % or above, the measured value deviates greatly from the estimated value calculated by FactSage. It is believed that this difference arises from the contribution of the amphoteric oxide Al2O3 to the basicity in the high Al2O3 content and high basicity region, and FactSage is not able to adequately reproduce the high basicity phenomenon of mold flux.
Figure 7 Comparison of FactSage estimated viscosity and measured viscosity
Discussion
The relationship between mold flux characteristics, continuous casting conditions and model parameters
As shown in Fig. 4, there is a correlation between the viscosity of the original composition of the mold powder and the rSlag layer1, and there is a correlation between the molten steel inflow and outflow index ITP and the rSteel layer. Firstly, the relationship between the viscosity of the initial composition of the mold flux and the thickness of the first slag layer is discussed, which may be related to the mass transfer of Al2O3 in the mold flux. Fig. 8 shows the estimation mechanism of the influence of the initial component viscosity of mold slag on the thickness of slag layer 1, taking case 1 and case 2, where the difference between the input and output index ITP of molten steel is small, as an example. Since the initial viscosity of mold flux B is lower, it is more easily affected by thermal convection and stirring, so the mass transfer rate of Al2O3 in mold flux B is qualitatively faster than that of mold flux A. Therefore, in the B-type mold flux, Al2O3 is more likely to move upward in the mold flux, and the Al2O3 content near the mold flux interface of the molten steel mold decreases. As a result, the oxidation-reduction reaction at the mold flux interface of the molten steel mold is promoted (the reaction is shown in formula (1)), and the content of Al2O3 in the mold flux is increased. In this model, the upward movement of Al2O3 corresponds to the increase in the thickness of the first slag layer near the mold flux interface. Therefore, in case 1 of using mold flux B, since the initial component viscosity of mold flux B is relatively low, it is considered that the thickness of liquid slag layer 1 is relatively large.
Fig. 8 Effect of viscosity on slag layer thickness
Next, the relationship between the molten steel inflow and outflow index ITP and the thickness of the molten steel layer will be considered, which is considered to be related to the renewal rate of molten steel. The estimation mechanism of the influence of the molten steel inflow and outflow index ITP on the thickness of the molten steel layer is shown in Figure 9. Since the larger the molten steel inflow and outflow index ITP is, the faster the molten steel update rate is, the greater the amount of molten steel that reacts with mold slag per unit time is. This situation corresponds to the larger thickness of the molten steel layer in the model. Therefore, under the condition of continuous casting where the molten steel inlet and outlet index ITP of the mold is larger, it is considered that the thickness of the molten steel layer in this model is larger.
Figure 9 The effect of ITP on the thickness of molten steel layer
From the above discussion, the correlation between the model parameters is considered to exist in the liquid slag layer 1 and the viscosity of the initial composition of the mold flux, and the relationship between the initial viscosity of the mold flux and the molten steel in-out index ITP, the initial composition of the mold flux The viscosity and the amount of molten steel in and out (the flow of molten steel entering the mold) have a great influence, and the alumina content in the mold powder of the accelerated material is accelerated. It can also be considered to use the relationship in Figure 4 to determine these model parameters from the continuous casting conditions and the properties of mold flux.
Application of viscosity estimation equation
As shown in Figure 7, the viscosity calculated by FactSage is inconsistent with the measured value. Therefore, the viscosity estimation equation proposed in the past literature was used to verify whether the trend of the measured values was reproducible. Although viscosity estimation equations for various molten oxides have been proposed, their use is limited to the compositional region used in the derivation, and outside this compositional region their accuracy decreases significantly. The viscosity estimation equations in the ROUND ROBIN project (1997-2000) and the report show that the revised Iida equation and Ribould equation have higher estimation accuracy for steel continuous casting mold flux. Among the equations studied in this project, the modified Iida equation is the most accurate and suitable for viscosity estimation over a wide range of Al2O3 content, because it can account for the change of amphoteric oxide behavior with composition and alkalinity. The revised Iida equation is represented by equations, formula (5) to formula (10).
μcalc [Pa·sec]: estimated viscosity, A, E[-]: constant, μ0 [Pa·sec]: simulated single-molecule molten mold flux viscosity, μ0i [Pa·sec]: simulated molten state of each oxide Viscosity, Bi[-]: Basicity index, α i [-]: Specific coefficient of each oxide, Wi [mass%]: Mass percentage of each oxide, Bi(j)[-]: Corrected base Sexual index, j[-]: Corrected ratio coefficient number, α*i[-]: Modified ratio coefficient, T [K]: Temperature. In the modified Iida equation, the relationship between the composition and viscosity of molten oxides can be corrected with measured data. Among them, the ratio coefficient αi of the amphoteric oxide is not a constant value, but a function α*i of the mass percentage Wi of the amphoteric oxide, and the basicity index Bi is fitted to minimize the error between the estimated value of viscosity μcalc and the measured value . Based on the viscosity measurement data in the range of Al2O3 concentration less than 12mass%, Iida et al. proposed α*Al2O3 as a linear function of continuous casting mold flux. In this study, an accurate estimation of the higher Al2O3 concentration range was required, so the functional form of α*Al2O3 was also investigated. The data in Table 3 is used here as the viscosity measurement data for correction, and μ0i and αi use the values in the literature. Through the test results, the ratio coefficient α*Al2O3 of Al2O3 is obtained as the following cubic function formula (11) of Wi and Bi.
Figure 10 shows the estimated results of the Iida equation corrected by formula (11). It can be seen from the figure that the estimated value obtained by the revised Iida equation is different from the estimated value of FactSage, and the trend of viscosity change with Al2O3 content can be reproduced . This is considered possible because the trend of the Al2O3 specific coefficient as a function of alkalinity and content can be reproduced. Therefore, formula (11) can be used to accurately estimate the viscosity of mold flux in high Al steels with large changes in Al2O3 content.
Figure 10 Comparison of the viscosity estimated by the modified Iida equation and the measured viscosity
Conclusion
Aiming at the stable continuous casting production of high alloy steel, a method for predicting the composition and performance changes of mold powder in the process of continuous casting of high aluminum steel was evaluated, and the following results were obtained. (1) Three kinds of mold fluxes with different Al2O3 concentrations (1.3 ~ 6.0mass%) and different basicities were applied to 0.7mass%Al steel and ordinary Al-killed steel, respectively. During the continuous casting process, the reduction of [Al] in the steel to the oxides in the mold flux increases the Al2O3 concentration in the mold flux to about 30mass%. The composition change in the continuous casting process is reproduced by the composition change model EERZM (Equilibrium Effective Reaction Zone Model). (2) Investigate the relationship between the obtained model parameters (thickness of liquid slag layer, thickness of steel layer, etc.), continuous casting conditions and physical properties of mold slag. The accelerated increase of the Al2O3 content in the mold flux has a great influence. (3) The main crystals and crystallization temperature were estimated by using the thermodynamic software Fact-Sage. The estimated value of the main crystal species is in good agreement with the measured value, and the error between the estimated value of the crystallization temperature and the measured value is about 50 °C. It is also possible to estimate the influence of changes in the content of Al2O3 in mold flux on the main crystals and crystallization temperature. (4) Although FactSage cannot estimate the change of viscosity with the change of Al2O3 content in the liquid mold flux of the mold, the modified Iida equation considering the basic behavior of amphoteric oxides with the change of composition can successfully estimate the change of viscosity with the change of Al2O3 content. The equations were corrected using the measured data of the viscosity of mold liquid mold flux in the range of Al2O3 concentration from 1.6mass% to 34.5mass%.
