Research On The Thermal Conductivity Of Basic Slag System For Mold Fluxes Based On Molecular Dynamic Simulation

Mold fluxes are important functional materials used in the continuous casting process,which control the heat transfer of mold and influence the continuous casting process.The thermal conductivity of mold fluxes is one of the key parameters reflecting the heat transfer performance,and it is also an important indicator for the design of mold fluxes.Existing research on the thermal conductivity of mold fluxes is mainly based on the hot-wire method and transient laser method.However,the thermal conductivity test results of these two methods have a significant deviation when the temperature exceeds 1000K,which is not conducive to the research on thermal conductivity of mold fluxes.In order to avoid the problems in experiments,the equilibrium molecular dynamics(EMD)simulation was used to calculate the thermal conductivity of the basic slag system of mold fluxes.The research results provide data support and theoretical guidance for mold flux design.In this paper,the thermal conductivity of CaO-SiO2-Al2O3 slag system was calculated by EMD simulation,verifying the accuracy of Matsui potential,Hirao potential and EMD simulation parameters to study the thermal conductivity of mold fluxes.Then,based on the EMD simulation,the relationship between the temperature,composition,microstructure and thermal conductivity of molten and glassy basic slag system for mold fluxes was studied.The thermal conductivity of cuspidine,which is the main crystalline product in the film,was further studied by the EMD method.Finally,combined with the machine learning algorithm and EMD method,a thermal conductivity prediction model of the CaO-SiO2-Al2O3 slag system was constructed.The main research contents and results are as follows:①In view of the problem that the high-temperature thermal conductivity test methods are difficult to remove systematic errors,the EMD simulation was used to calculate the thermal conductivity of basic slag system for mold fluxes and the accuracy of the EMD method to calculate the thermal conductivity was verified.On this basis,the thermal conductivity of molten CaO-SiO2-Al2O3 ternary slag was calculated,and the relationship between composition,microstructure and thermal conductivity was initially established.The results show that the Ca-O structure is greatly affected by Al2O3 content in CaO-SiO2-Al2O3 system.The two peaks of the O-Ca-O bond angle distribution curve gradually overlap with increasing Al2O3 content.The aggregation of Si-O and Al-O networks and the enhancement of bridging role of Ca2+ are beneficial to increasing the thermal conductivity of molten CaO-SiO2-Al2O3 slags.②The influence of temperature on the thermal conductivity of CaO-SiO2-Al2O3 system was studied.The effect of temperature on the thermal conductivity of glassy and molten slags was clarified,and the relationship between the temperature,microstructure and thermal conductivity of the slags was established.The thermal conductivity of CaOSiO2-Al2O3 slags increases with increasing temperature,CaO/SiO2 ratio and CaO/Al2O3 ratio.The effect of temperature is the most significant.③Based on the study of the CaO-SiO2-Al2O3 ternary slag system,the thermal conductivity of the molten CaO-SiO2-Al2O3-MgO-Na2O five-element slag system was further studied.The thermal conductivity and microstructure of samples were clarified.The relationship between composition,microstructure and thermal conductivity was established.The results show that the thermal conductivity of molten high-SiO2 slags increases with increasing CaO/SiO2 ratio.When the CaO/SiO2 ratio is 0.25,the thermal conductivity of the molten slags decreases with increasing Al2O3/SiO2 ratio;when the CaO/SiO2 ratios are 0.43 and 0.67,the thermal conductivity of the melts first increases and then decreases with increasing Al2O3/SiO2 ratio.The aggregation of Si-O and Al-O networks and the bridging effect of Ca2+ promote phonon transport and increase the thermal conductivity of molten slags.The thermal conductivity of molten low-SiO2 slags first increases and then decreases with increasing CaO/Al2O3 ratio and is less affected by the CaO/SiO2 ratio.The average coordination number and the first sharp diffraction peak of the neutron diffraction curve(ND-FSDPA)of melts increase with increasing CaO/Al2O3 ratio and are less affected by the change of CaO/SiO2 ratio.The Al-O structure is the skeleton of the molten low-SiO2 slags and depolymerizes significantly with the increase of CaO/Al2O3 ratio.④The thermal conductivity of cuspidine at different temperatures was calculated,filling the gap in the thermal conductivity of cuspidine and clarifying the thermal conduction mechanism in cuspidine.The thermal conductivity of cuspidine decreases from 14.3W/(mK)to 3.4W/(mK)with increasing temperature(293K-1673K).The vibration density of states(VDOS)of the Ca-F structure is mainly concentrated in the low-frequency region(0-20THz),and the VDOS of the Si-O network is distributed in the full frequency range(0-60THz).⑤Combined with EMD simulation and machine learning algorithm,a prediction model for the thermal conductivity of the CaO-SiO2-Al2O3 slag system was established.The model based on random forest(RF)algorithm has the highest accuracy compared with the elastic net regularization(ENR)and artificial neural network(ANN)algorithms.The external prediction results of the RF model are similar to the experimental results of the front-heating front-detection initial lase flash(FFL)method,and the relative error is less than 15%.