A method for evaluating the inclusion absorption capacity of molten slag based on molecular dynamics

Abstract

The present application belongs to the technical field of molecular dynamics simulation and metallurgical slag design, and discloses a method for evaluating the inclusion absorption capacity of a slag based on molecular dynamics. An initial structure model of the slag and an initial structure model of the inclusion substrate are established. The initial structure model of the slag and the initial structure model of the inclusion substrate are pre-relaxed respectively, and the models are merged. The file obtained after the end of the wetting simulation is post-processed using a local tangent fitting method based on two-dimensional number density distribution to obtain the average contact angle in the equilibrium state. The pre-relaxed initial structure model of the slag is placed in an isolated vacuum system for continued simulation, and the adhesion work representing the interfacial bonding strength is calculated. The effective diffusion coefficient is obtained from the self-diffusion coefficient. The adhesion work and the effective diffusion coefficient are substituted into the interfacial mass transfer driving power model to calculate the comprehensive absorption evaluation index. The present application combines the thermodynamic interfacial wetting phenomenon and the kinetic component diffusion mechanism to improve the robustness and precision of the contact angle calculation of a complex slag system.

Description

Technical Field

[0001] This invention relates to the field of molecular dynamics simulation and metallurgical slag design technology, specifically to a method for evaluating the ability of slag to absorb inclusions based on molecular dynamics. Background Technology

[0002] In metallurgical processes such as electroslag remelting and ladle refining, the absorption of non-metallic inclusions in molten steel by the slag is a crucial step in improving the cleanliness of the steel. The inclusion absorption process typically involves two consecutive stages: First, after the inclusion comes into contact with the slag, the slag wets and spreads on the inclusion surface, providing interfacial conditions for subsequent interfacial reactions; second, the inclusion reacts and dissolves at the interface, and further migrates and diffuses into the interior of the slag, ultimately achieving inclusion removal. Therefore, a high-quality refining slag should possess both good wettability and strong mass transfer and migration capabilities to effectively remove inclusions.

[0003] The dissolution and absorption process of inclusions by slag can be subdivided into several continuous stages, including: diffusion of slag components across the concentration boundary layer to the solid surface, diffusion across the reaction product layer to the inclusion surface, chemical reaction at the interface, diffusion of reaction products into the concentration boundary layer, and further migration of reaction products into the depth of the slag. These stages are coupled with each other and jointly determine the inclusion removal efficiency. Some scholars have pointed out that the limiting stage of inclusion removal is often controlled by the characteristics of the intermediate product layer at the interface (formation and stripping of the intermediate phase) and the ion diffusion rate in the slag phase.

[0004] In summary, the absorption of inclusions by slag is the result of the combined effects of interfacial wetting, interfacial reaction, and dissolution-diffusion. The migration abilities of various ions differ in different metallurgical slag systems, leading to varying wetting behaviors towards inclusions. Good wettability helps increase the effective contact area and improve interfacial mass transfer conditions, while dissolution products such as Al... 3+ O 2- The continuous migration of inclusions into the depths of the slag helps maintain the concentration gradient at the interface, thereby promoting further dissolution and absorption of inclusions.

[0005] From a microscopic perspective, the diffusion capacity of different ions in molten slag is closely related to the melt structure, especially to the network aggregation units such as silicon-oxygen tetrahedra [SiO4]. 4- Aluminum-oxygen tetrahedron [AlO4] 5- The degree of polymerization and Ca 2+ Mg 2+ The distribution state of cations has an impact. Studies have shown that reducing slag viscosity can significantly promote the dissolution of inclusions such as Al2O3, which is essentially due to changes in the melt microstructure and reduction of diffusion resistance. The lower the degree of network aggregation, the more loose and open the melt structure, the smaller the migration resistance of solute ions, the higher the diffusion coefficient, and the more favorable it is for the migration of inclusion dissolution products into the interior of the slag.

[0006] Currently, research on the dissolution and absorption behavior of inclusions in molten slag largely relies on high-temperature experimental methods. These experiments are typically complex, requiring sample preparation and repeated high-temperature tests after composition adjustments. This is especially true for slags containing volatile or easily oxidized components, resulting in long experimental cycles, high costs, and significant difficulties in maintaining stable control over temperature, atmosphere, and interface states. Furthermore, macroscopic experiments cannot directly characterize microscopic mechanisms such as interface structure evolution, particle trajectories, and diffusion coefficients, thus hindering the revelation of the essential laws governing inclusion absorption in slag at the atomic scale.

[0007] With the development of computer simulation technology, molecular dynamics simulation has become an important tool for studying high-temperature metallurgical melts and their interfacial behavior. Based on Newton's equations of motion, this method describes the evolution of a system over time at the atomic scale, offering advantages such as low cost, high efficiency, controllable conditions, and good repeatability. Existing simulation studies have shown that in solid-liquid interfacial systems, wettability is usually closely related to interfacial adhesion strength; the greater the adhesion work, the easier it is for the liquid to wet and spread on the solid surface.

[0008] Therefore, molecular dynamics simulations have unique advantages in revealing the microscopic mechanisms of interfacial wetting, structural evolution, and component migration in slag-inclusion systems. To more efficiently guide the design of metallurgical slag systems, this invention proposes a molecular dynamics-based method for evaluating the inclusion absorption capacity of slag. Summary of the Invention

[0009] To address the aforementioned problems, this invention provides a molecular dynamics-based method for evaluating the inclusion absorption capacity of slag. This method can simultaneously predict the wetting behavior of slag on inclusions, interfacial bonding characteristics, and the diffusion capacity of related ions in the slag at the atomic scale. This enables rapid assessment of the inclusion absorption capacity of different slag systems and provides computational methods and theoretical basis for the design of high-performance metallurgical slag systems. This method is used to simultaneously characterize the wetting behavior between slag and inclusions, as well as the migration and diffusion potential of inclusion-related components in the slag, thereby evaluating whether the target slag system design meets smelting requirements.

[0010] To achieve the above objectives, the technical solution of the present invention is as follows: A method for evaluating the ability of slag to absorb inclusions based on molecular dynamics, comprising the following steps: S1. Determine the initial composition of the simulation system, including the composition of the slag system and the types of inclusions; S2. Based on the composition and density of the selected slag system, calculate the number of each atom and establish an initial slag structure model; S3. Determine the crystallographic surface orientation based on the selected inclusion type, select the target crystal plane as the substrate surface in contact with the slag, expand the cell of the inclusion, and construct the initial structure model of the inclusion substrate. S4. Determine the potential function and corresponding potential parameters that describe the interaction between atoms in the simulation system; S5. Based on the potential function and corresponding potential parameters, write the input file for the molecular dynamics simulation command. Pre-relax the initial structure model of the slag and the initial structure model of the inclusion substrate respectively. After pre-relaxation, merge the models. The merged model includes the slag system, the inclusion system, and the vacuum region set around the slag and inclusions. In the merged model, the initial distance between the slag system and the inclusion system is greater than the cutoff distance of the potential function used. S6. Adjust the distance between the slag system and the inclusion system in the merged model to 2Å-4Å, and run the wetting simulation process to enable the merged model to spontaneously wet and continue running until wetting equilibrium is reached; continuously record the energy, temperature, pressure, atomic trajectory and mean square displacement of the slag system in the merged model. S7. After the wetting simulation is completed, the atomic trajectory file obtained is post-processed using a local tangent fitting method based on two-dimensional number density distribution to analyze the wetting state of the slag system on the inclusion system and obtain the average contact angle under equilibrium conditions. i ; S8. The slag structure model obtained by pre-relaxation treatment of the initial slag structure model in step S5 is placed in an isolated vacuum system for further relaxation. The Laplace pressure difference and the true radius of curvature inside the droplet are calculated to obtain the surface tension of the slag, and combined with the average contact angle. i The adhesion work characterizing the interfacial bonding strength was calculated. W ad ; S9. Extract the mean square displacement versus time curves of each component in the slag system after the wetting simulation operation, fit the curves within the stable diffusion range where a linear relationship exists, and obtain the self-diffusion coefficient of each component based on the Einstein relation. Then, determine the effective diffusion coefficient based on the self-diffusion coefficient. D eff It is used to evaluate the mass transfer and migration potential of slag systems for specific inclusion types; S10, the adhesion function W ad With the effective diffusion coefficient D eff Substituting into the interface mass transfer driven power model, the calculation dimension is J·s -1 Comprehensive absorption evaluation index P abs The strength of the synergistic effect of slag system driving interface wetting and component migration is used as an evaluation of the ability of different slags to absorb inclusions.

[0011] In step S5, after the initial slag structure model has completed the pre-relaxation process, the boundary size of the simulation calculation domain is adjusted, and the size of other directions is adjusted by compensation to keep the total volume and number density of the slag system unchanged, so that the size of the slag structure model and the initial structure model of the inclusion substrate after pre-relaxation process are matched. At the same time, a vacuum region is added to facilitate the merging of the two models, and finally the merged model is output.

[0012] In step S7, the process of calculating the average contact angle under equilibrium conditions includes: Determine the solid-liquid contact reference surface: Perform time-averaged processing on the atomic trajectory after wetting equilibrium, statistically analyze the atomic number density along the Z-axis, calculate the spatial gradient after Gaussian smoothing, and locate the last significant negative peak; find the first positive peak of the atomic number density gradient in the transition region to the right of the significant negative peak. The inflection point of the component density curve of the inclusion substrate corresponding to the positive peak is the position where the component density of the slag system increases the fastest, thereby determining the solid-liquid contact reference surface. Constructing the liquid phase profile based on the solid-liquid contact reference surface: The droplet centroid is corrected based on periodic boundary conditions, and the statistical time-averaged number density spectrum is processed by two-dimensional spatial gridding. The isodensity line extraction method is used to identify the liquid phase interface boundary. Defining the adsorption layer range: Based on the number density distribution curve of atoms in the slag system along the direction perpendicular to the inclusion substrate, the first density peak adjacent to the inclusion substrate is taken as the characteristic of the interface adsorption layer, and the height coordinate corresponding to the first density minimum value that appears after the first density peak is taken as the upper boundary of the adsorption layer. Local tangent fitting: A set of boundary points located above the upper boundary of the adsorption layer is selected. Principal component analysis is used to perform orthogonal distance linear fitting on the boundary point set. The direction of the first principal component of the boundary point set is extracted as the local tangent of the liquid interface. The slope of the local tangent at the intersection of the three phases is analyzed, and the average contact angle is calculated. i .

[0013] The acquisition of slag surface tension in step S8 specifically involves: first, placing the pre-relaxed slag structure model in an isolated vacuum system, maintaining the same composition, number of atoms, and temperature conditions as the merged model during wetting, and continuing relaxation; using the droplet centroid as the reference origin, statistically analyzing the radial time-averaged density distribution curve, identifying the surface density peak that appears before entering the vacuum transition region where the density monotonically decreases to zero, and extracting the radial coordinate position corresponding to the surface density peak as the droplet's true radius of curvature. Simultaneously, the pressure tensor of the steady-state region where the droplet's internal radius is smaller than its true radius of curvature is extracted to obtain the Laplace pressure difference. P in And calculate the surface tension of the molten slag. s lv : ; Based on average contact angle i and slag surface tension s lv Solve for adhesion work W ad : .

[0014] Given the average contact angle i and slag surface tension s lv Then, combined with the solid surface energy of the target inclusions s sv Using formula c sl =s sv -s lv cos i Solving for solid-liquid interfacial energy c sl It is used to characterize the bonding strength at the solid-liquid interface.

[0015] In step S9, the self-diffusion coefficients of each component in the slag system are... D Solving using Einstein's relation, the formula is: ; In the formula, D Indicates the self-diffusion coefficient; t is the simulation time; It is the mean square displacement of the component within time t; The components that characterize the absorption process of target inclusions by the slag and can obtain a self-diffusion coefficient in the slag system are defined as characteristic mass transfer components. The effective diffusion coefficient is constructed by geometrically averaging the self-diffusion coefficients of each characteristic mass transfer component. D eff The formula is: ; In the formula, D eff It reflects the mass transfer and migration potential of the slag system to the relevant components of the target inclusions; D j For the first The self-diffusion coefficient of each characteristic mass transfer component; v j The stoichiometric coefficient is the chemical formula of the target inclusion corresponding to the characteristic mass transfer component. m This represents the total number of types of characteristic mass transfer components.

[0016] In step S10, the interface mass transfer driving power model, without considering interfacial chemical reactions, comprehensively considers the diffusion and migration characteristics of the characteristic mass transfer components of the slag system on the target inclusions and the interfacial wetting and bonding characteristics. Its expression is as follows: ; In the formula, P abs This is an evaluation index of the slag system's ability to dissolve and absorb target inclusions, expressed in units of 10. -12 J·s -1 This characterizes the synergistic effect of the slag system on interfacial wetting and component dissolution and migration per unit time.

[0017] Compared with the prior art, the present invention has the following advantages: Under high-temperature conditions, direct experimental observation of the atomic-scale structure, dynamic wetting process, and interfacial evolution of molten slag is easily constrained by experimental conditions, limiting a deeper understanding of the mechanism by which molten slag absorbs inclusions. This invention employs molecular dynamics simulations to reproduce the contact, wetting, spreading, interfacial interactions, and component diffusion and migration processes between molten slag and inclusions at the atomic scale. It not only obtains quantitative information such as molten slag structure, interfacial energy, and diffusion behavior, but also reveals interfacial configuration changes and particle trajectories that are difficult to observe directly under experimental conditions. Therefore, it provides an efficient, intuitive, and reproducible technical means for studying the behavior of molten slag in absorbing inclusions.

[0018] The absorption and removal of inclusions by slag is not determined by a single factor, but rather by the combined influence of interfacial interaction and migration / mass transfer capabilities. This invention deeply couples parameters characterizing wetting and bonding capabilities with parameters characterizing migration and diffusion capabilities, and employs computer algorithms to uniformly identify contact angles, avoiding subjective differences in human judgment. Based on this, it efficiently analyzes, optimizes, and ranks different slag composition systems, thus providing a reliable theoretical basis and efficient preliminary research method for metallurgical slag system design, composition optimization, and process optimization. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the implementation of the molecular dynamics-based method for evaluating the ability of slag to absorb inclusions as described in this invention. Figure 2 This is a schematic diagram of the number density distribution and density gradient for determining the reference surface of solid-liquid contact and the starting point of the contact angle fitting in this invention; (a) is the number density distribution, and (b) is the density gradient distribution. Figure 3 This is a schematic diagram of the two-dimensional isodense profile projection and local fitting of the equilibrium contact angle of the molten slag droplet in this invention; Figure 4 This is a schematic diagram of the radial density distribution of molten slag droplets in this invention; Figure 5 This is a graph showing the relationship between the surface tension of the simulated slag system and MgO in this invention; Figure 6The diagram shows the evolution of the wetting and spreading morphology of molten slag droplets on the surface of a solid substrate over time in this invention; (a) is the initial contact state after 50 ps, ​​(b) is the initial spreading state after 200 ps, ​​(c) is the middle spreading state after 500 ps, ​​(d) is the late spreading state after 800 ps, ​​(e) is the near-equilibrium state after 1000 ps, ​​and (f) is the equilibrium state after 1200 ps. Figure 7 This is a graph showing the variation of the equilibrium contact angle of the slag-inclusion system with the MgO content in the slag in this invention. Figure 8 This is a graph showing the relationship between the adhesion work of slag and inclusion substrate and the MgO content in the slag in this invention. Figure 9 The graph shows the mean square displacement and time evolution curves of the characteristic mass transfer components in this invention, as well as the relationship between the effective diffusion coefficient and the MgO content in the slag; (a) shows the mean square displacement and ion diffusion coefficient of sample AM1, and (b) shows the effective diffusion coefficient of each sample. Figure 10 This is a comparison chart of the effective absorption power evaluation index of different slag systems for target inclusions in this invention. Detailed Implementation

[0020] To make the technical solution, implementation process, and beneficial effects of the present invention clearer, the specific embodiments of the present invention are further described below. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0021] like Figure 1 The diagram shows the flowchart of a method for evaluating the ability of slag to absorb inclusions based on molecular dynamics. The specific steps are as follows: S1. Determine the initial composition of the simulation system, including the composition of the slag system and the types of inclusions; S2. Based on the composition and target density of the selected slag system, calculate the number of each atom and establish an initial slag structure model; S3. Based on the selected inclusion type, determine the inclusion crystal surface as the wetting simulation substrate, expand the inclusion crystal structure to construct the initial structure model of the inclusion substrate. S4. Determine the potential function and corresponding potential parameters that describe the interaction between atoms in the simulation system; S5. Based on the potential function and corresponding potential parameters, write the input file for the molecular dynamics simulation command. Pre-relax the initial structure model of the slag and the initial structure model of the inclusion substrate respectively. After pre-relaxation, merge the models. The merged model includes the slag system, the inclusion system, and the vacuum region set around the slag and inclusions. In the merged model, the slag system and the inclusion system maintain a sufficient distance, which should be greater than the cutoff distance of the potential function used, to ensure that there is no interaction between the two systems after merging. S6. Adjust the distance between the slag system and the inclusion system in the merged model, which can be around 3 Å, for example, 2 Å-4 Å. Then start the simulation to enable the merged model to spontaneously wet and continue running until wetting equilibrium is reached. Continuously record the energy, temperature, pressure, atomic trajectories and mean square displacement of the slag system in the merged model. S7. After the wetting simulation is completed, the atomic trajectory file obtained is post-processed using a local tangent fitting method based on two-dimensional number density distribution to analyze the wetting state of the slag system on the inclusion system and obtain the average contact angle under equilibrium conditions. i ; S8. The slag structure model obtained after pre-relaxation processing of the initial slag structure model in step S5 is placed in an isolated vacuum system for further relaxation. The Laplace pressure difference and the true radius of curvature inside the droplet are calculated to obtain the surface tension of the slag, and combined with the average contact angle. i The adhesion work characterizing the interfacial bonding strength was calculated. W ad ; S9. After the wetting simulation in step S6 is completed, extract the mean square displacement versus time curves of each component in the slag system. Fit the curves within the stable diffusion range where a linear relationship exists. Calculate the self-diffusion coefficient of each component based on the Einstein relation, and then calculate the effective diffusion coefficient based on the self-diffusion coefficient. D eff It is used to evaluate the mass transfer and migration potential of slag systems for specific inclusion types; S10, the adhesion function W ad With the effective diffusion coefficient D eff Substituting into the interface mass transfer driven power model, the calculated power has the dimension of J·s. -1 Comprehensive absorption evaluation index P abs This serves as an evaluation of the ability of different slags to absorb inclusions.

[0022] In step S2, the initial slag structure model is established in a cubic simulation computational domain, which can be constructed using MaterialsStudio software, and the width (Y-axis direction) of the cubic simulation computational domain is made to match the initial structure model of the inclusion substrate as closely as possible.

[0023] In step S2, the initial structure model of the slag can also be constructed into a predetermined shape using the Packmol program; during the modeling process, the initial distance between each atom should be controlled within a reasonable range to avoid atomic overlap during relaxation, which could lead to atom loss.

[0024] In step S3, the initial structural model of the inclusion substrate can be expanded using Materials Studio software based on a standard crystal structure to obtain a substrate model that meets the simulation size requirements.

[0025] The data file corresponding to the initial structural model includes the total number of atoms, the number of atom types, the mass of each type of atom, the atom number, the atom type, and the x, y, and z coordinate information of each atom.

[0026] In step S5, the initial slag structure model is relaxed and reaches equilibrium using an NPT isothermal and isobaric ensemble in an independent simulation domain, and the initial structure model of the inclusion substrate is relaxed using an NVT canonical ensemble. The relaxed structure files are saved separately for subsequent simulation calls.

[0027] In step S5, after the molten slag droplets have completed the relaxation process, their simulation is switched from the NPT isothermal and isobaric ensemble to the NVT canonical ensemble to fix the total volume. Then, the boundary size of the simulation computation domain is adjusted to match the size of the inclusion substrate model. The size of other directions is adjusted by compensation to keep the total volume and number density of the molten slag system basically unchanged.

[0028] In step S5, after the size adjustment is completed, the molten slag droplets are relaxed to eliminate unreasonable atomic contacts; then a vacuum region is added around the system to make it compatible with the simulation calculation domain of the initial structural model of the inclusion substrate, and the structural file is output for subsequent merging calculations.

[0029] The input file in step S5 is used to set the simulation system type, atom type, boundary conditions, ensemble form, temperature control method, pressure control method, relaxation time, and output information; at the same time, it sets the initial distance between the slag droplets and the inclusion substrate so that the two can spontaneously wet each other in the subsequent simulation in step S6.

[0030] In step S7, after the wetting process reaches stability, visualization software such as OVITO and VMD can be used to perform autonomous analysis of the trajectory file. The trajectory file should at least contain atomic coordinate information. Alternatively, a self-developed post-processing program can be used to perform rapid quantitative analysis of the simulation results to obtain wetting behavior, contact angle, and related statistical parameters.

[0031] In step S8, the slag structure model must maintain the same composition, number of atoms, and temperature conditions as the merged model. The simulation continues in an isolated vacuum system using an NVT canonical ensemble, and the Laplace pressure difference inside the droplet is calculated. P in and the true radius of curvature of the droplet Used to calculate the surface tension of molten slag droplets s lv .

[0032] Before the molten slag droplets wet the inclusion substrate, the droplets can be kept at a certain distance from the substrate to allow for equilibrium relaxation. Simultaneously, the average total potential energy before wetting can be statistically analyzed. U sep When the liquid and inclusion substrates are far enough apart and do not interact with each other, U sep It can be considered as the sum of the potential energies of the droplet and the substrate; Once the wetting process of the molten slag droplets on the inclusion substrate reaches equilibrium, the trajectory data and average total potential energy of the equilibrium phase are captured. U wet ; Based on the above potential energy data, the potential energy difference per unit area of ​​solid-liquid interaction can be calculated. W inter The calculation principle is as follows: ; In the formula, W inter is the potential energy difference per unit area of ​​solid-liquid interaction. This parameter is usually positively correlated with the strength of interfacial bonding and can be used as a characterizing parameter of solid-liquid wetting and interfacial bonding strength; A is the calculation base area used to convert the total interaction energy into energy per unit area; the base area A should be the actual wetting contact area between the slag droplet and the inclusion substrate, but the base area A can also be the area of ​​the inclusion substrate in the calculation domain, i.e., A = L x × L y ,in L x For and L y These are the lengths of the inclusion substrate in the x and y directions, respectively, to avoid inconvenience and subjective errors when calculating the actual contact area of ​​the droplets; Furthermore, to characterize the minimum work required for the reversible separation of a unit area solid-liquid interface into independent solid and liquid surfaces, the work of adhesion needs to be calculated. W ad The principle is as follows: ; In the formula, W ad For adhesion work; s sv It is the surface energy of solids. s lv It is the surface energy of the liquid. c sl It is the solid-liquid interface energy.

[0033] The surface tension of the slag s lv Based on the Laplace pressure differential value P in and true radius of curvature Solve using the following formula: Surface tension of molten slag s lv The adhesion work can then be solved using the Young-Dupré relation: in, s lv For the surface tension of the molten slag, i To balance the contact angle.

[0034] The contact angle is the angle between the local tangent of the liquid-vacuum interface and the surface of the solid substrate on the liquid phase side at the three-phase contact line.

[0035] The trajectory file of the wetting process is post-processed by a program to calculate the average contact angle under equilibrium conditions, thereby minimizing human analysis errors. The specific post-processing process and program algorithm principles and steps are as follows: Step 1: First, statistically analyze the atomic number density along the Z-axis. r Its spatial gradient d is calculated after Gaussian smoothing. r / dz; Locate the last significant negative peak of the atomic number density gradient (the rightmost falling edge) to eliminate interference from layered density oscillations within the crystal. Then, search for the first positive peak of the atomic number density gradient within the transition region to the right of the significant negative peak; this positive peak corresponds to the matrix component density of the inclusions. r sub The inflection point of the curve is also the component density of the slag system. r drop The position where the rise is fastest is therefore used as a supplementary criterion to determine the solid-liquid contact reference surface.Z base ,like Figure 2 The number density distribution and density gradient curves are shown. This method is immune to the diffusion process of substrate atoms into the slag, determining the reference surface solely based on changes in the density field.

[0036] Step 2: Extracting the continuous material after wetting equilibrium N For frame trajectory data, a minimum mirror convention is used to eliminate the coordinate truncation effect caused by atoms crossing the boundary, especially for periodic boundary conditions. The geometric center coordinates of each frame droplet are calculated using the circumferential mean method, and the atoms in each frame are translated and aligned using this center as the origin to eliminate droplet centroid drift. Subsequently, a thickness of Δ is selected in the direction perpendicular to the interface. y The slices are statistically analyzed, and the projection distribution of atoms within the slices onto a two-dimensional plane is statistically analyzed. A time-averaged number density matrix is ​​constructed by dividing the spatial grid. At the same time, the density matrix is ​​smoothed by a two-dimensional Gaussian filter to eliminate high-frequency noise interference caused by atomic thermal motion, resulting in a clear and stable spatial density distribution map.

[0037] The third step involves statistical analysis of the density distribution in the liquid phase region. Given the density inhomogeneity caused by thermal fluctuations within the droplet, a high percentile statistic is used to characterize the droplet bulk density. r l To eliminate interference from low-density regions, the critical density of the liquid interface is defined as... r c = c · r l ,in c The value can range from 0.15 to 0.50, representing a dimensionless attenuation threshold coefficient, corresponding to the transition ratio of interface density attenuation from bulk to gas phase values. The contour tracing method is used to extract the density value from the density field. r c The closed contour line is the geometric representation of the liquid interface, and its coordinate point set is the basis for calculating the contact angle.

[0038] Step 4: Using the solid-liquid contact reference surface Z base Starting from the z-axis, establish a high-fit interval along the z-axis, with the lower limit set to... Z base +Δ Z min The upper limit is set to Z base +Δ Z max Among them, Δ Z minThe extent of the adsorption layer is determined by the first minimum point appearing to the right of the slag number density distribution curve on the reference plane. This point marks the end of the adsorption oscillation layer at the solid-liquid interface. By using this point as the starting point for fitting, the interference of the near-wall anisotropic layer on the slope of the interface tangent can be eliminated; Δ Z max Based on the wetting effect settings, multiple slag composition settings are kept unique to maintain the maximum height limit of local interface linearity. Exceeding this range is considered an approximate failure of the tangent. Interface contour point sets on the left and right sides within this interval are extracted. Principal component analysis (PCA), a method with coordinate rotation invariance, is used to extract the best-fit straight line direction for each interface point set through covariance matrix decomposition; this is the local interface tangent direction. Finally, the angles between the left and right tangents and the reference plane are calculated, and their average value is taken as the final equilibrium contact angle. i .

[0039] Once the equilibrium contact angle is known, the solid-liquid interfacial energy can be further calculated by combining it with the solid surface energy of the target inclusion. c sl To characterize the bonding strength at the solid-liquid interface: ; Based on the mean square displacement data obtained from the slag equilibrium simulation of the system, the mean square displacement of the target ion versus time curve was plotted; a stable diffusion range exhibiting a linear relationship was selected, and linear fitting was performed on this range. The self-diffusion coefficient of the target ion was then obtained based on the Einstein relation. ; The absorption of inclusions by slag is essentially a non-equilibrium kinetic process controlled by the coupling of interfacial wetting and spreading with component migration and mass transfer. Based on this, without considering interfacial reactivity, this invention proposes an interfacial mass transfer-driven power model, derived from physical dimensions. This model integrates the interfacial wetting characteristics of the slag system for target inclusions with the microscopic diffusion potential of characteristic mass transfer components, thereby enabling a quantitative evaluation of the inclusion absorption capacity of different slags. ; In the formula, P abs This is an evaluation index of the slag system's ability to dissolve and absorb target inclusions, expressed in units of 10. -12 J·s -1 This value characterizes the intensity of the synergistic effect between interfacial wetting and component dissolution and migration in the slag system per unit time. A larger value indicates a stronger overall absorption capacity of the slag system for target inclusions. D eff The effective diffusion coefficient of the relevant characteristic mass transfer component in the slag is defined as the component that is related to the absorption process of the target inclusion and can obtain its self-diffusion coefficient in the bulk phase of the slag.

[0040] The interface mass transfer driven power model is based on interface interaction characteristic parameters and migration behavior characteristic parameters obtained from molecular dynamics simulations. Its purpose is to jointly determine, classify, and optimize the inclusion absorption capacity of different slag systems. Among these, adhesion work... W ad Used to characterize the wetting and interfacial bonding ability between slag and inclusions, its value is related to the interfacial interaction energy. W inter The correlation is good; while the diffusion coefficients of characteristic mass transfer components in slag show good correlation. D These parameters are used to reflect the kinetic mass transfer potential of slag to inclusions. By comprehensively analyzing the above parameters, the physical distortion and one-sidedness that may be caused by a single index evaluation can be effectively avoided, and the determination of the inclusion absorption capacity of different slag systems has a more rigorous scientific basis and better engineering applicability.

[0041] This method takes the study of the interfacial behavior of the CaF2-CaO-Al2O3-MgO slag system and the MgAl2O4 inclusion system as an example, and can be extended to other slag and inclusion systems.

[0042] The method includes steps such as model building, interface wetting simulation, potential energy analysis, contact angle calculation, surface tension characterization, adhesion work calculation, diffusion coefficient fitting, and comprehensive absorption capacity evaluation.

[0043] The composition of the target slag system is selected based on the scheme to be studied, and the type of target inclusions is determined.

[0044] Based on the composition of the slag system, the target density, and the size of the simulation domain, the number of atoms in each component is calculated, and an initial slag structural model is established. This initial slag structural model can be constructed in a cubic simulation domain, and preferably, the model's dimensions in the direction parallel to the substrate match the initial structural model of the inclusion substrate.

[0045] Select the crystal surface for wetting simulation based on the type of target inclusion, and perform cell expansion processing on the corresponding crystal structure to establish an initial structural model of the inclusion substrate. The initial structural model of the inclusion substrate should meet the planar size and thickness requirements required for interface simulation. The position of the substrate fixing layer should avoid significantly affecting the droplet wetting and spreading behavior and the relaxation of the interface atomic structure.

[0046] Determine the potential functions and corresponding potential parameters that describe all interatomic interaction forces contained in the simulation system; Write the input file for the molecular dynamics simulation command, and set the system boundary conditions, integration step size, simulation ensemble, temperature and pressure control mode, relaxation time and output parameters so that it can complete processes such as slag pre-melting, simulation domain size adjustment and merging of slag and inclusion substrate.

[0047] The initial structural model of the slag and the initial structural model of the inclusion substrate were pre-relaxed separately. After pre-relaxation, the models were merged. The slag structural model before merging after pre-relaxation was placed in an isolated vacuum system for further simulation. The Laplace pressure difference and the true radius of curvature inside the droplet were calculated to obtain the surface tension of the slag. Set the simulation parameters and run the molecular dynamics simulation of the merged model to output data such as atomic coordinates, velocity, temperature, pressure, mean square displacement, and total potential energy.

[0048] Post-processing analysis was performed on the trajectory file after wetting equilibrium to obtain the wetting state, contact angle and interface-related characteristic parameters of the molten slag droplets on the inclusion substrate.

[0049] The obtained mean square displacement data are fitted to obtain the self-diffusion coefficient of the characteristic mass transfer component, and the mass transfer and migration capability of the slag system for inclusion components is characterized accordingly.

[0050] In this embodiment, the merging model includes slag droplets, an inclusion substrate, and a vacuum region disposed around the slag droplets and the inclusion substrate.

[0051] In this embodiment, the slag and inclusions are preferably modeled and pre-relaxed separately. The slag droplets are preferably pre-melted and density balanced under an NPT isothermal and isobaric ensemble before being switched to an NVT canonical ensemble to maintain a basically stable overall volume; the inclusion substrate is preferably structurally relaxed under an NVT canonical ensemble.

[0052] In this embodiment, after the slag and inclusion models are combined, a reasonable initial spacing should be set to ensure that the droplets can spontaneously contact, wet and spread on the substrate surface, so that the slag can contact the inclusion substrate and gradually reach wetting equilibrium in the subsequent simulation process.

[0053] Furthermore, to evaluate the strength of interfacial interactions, the average total potential energy was statistically calculated under the same atomic number, temperature, and boundary size conditions as the composite system, before wetting the slag droplets and inclusion substrate. U sep After the composite system reaches wetting equilibrium, the trajectory data of the equilibrium interval are statistically analyzed to obtain the average total potential energy. U wet .

[0054] According to potential energy U sep and U wet Calculate the interaction energy per unit area of ​​the solid-liquid interface using the aforementioned formula. W inter The aforementioned W interThe potential energy difference between solid and liquid per unit area is usually positively correlated with the strength of interfacial bonding. It can be used as a characterization parameter for solid-liquid wetting and bonding strength and can be used for simple judgment.

[0055] Furthermore, the trajectory file of wetting equilibrium is analyzed through a post-processing program. Preferably, the post-processing program determines the solid-liquid contact reference surface based on the density gradient peak localization mechanism, extracts the smoothed two-dimensional number density matrix under time averaging, and extracts the interface contour; subsequently, it automatically defines the lower limit of the effective fitting interval based on the first minimum point in the slag number density distribution curve to eliminate the interference of the near-wall density oscillation layer; then, it applies principal component analysis to extract the eigenvector corresponding to the maximum singular value as the local tangent of the interface, thereby more robustly obtaining the left and right contact angles and the average contact angle. i ,like Figure 3 As shown.

[0056] In this embodiment, the simulation continues to run and the Laplace pressure difference inside the droplets is calculated after the droplets have relaxed and before they merge. P in and true radius of curvature The surface tension of the molten slag droplets is then solved through post-processing. s lv .

[0057] In this embodiment, after obtaining the equilibrium contact angle, the surface tension of the molten slag droplets is considered. s lv And the Young-Dupré relation for calculating adhesion work W ad The adhesion work is used to characterize the minimum work required for a unit area of ​​solid-liquid interface to reversibly separate into independent solid and liquid surfaces; adhesion work W ad The size directly reflects the interfacial wetting and bonding ability between the slag and inclusions. W ad The higher the value, the stronger the driving force for wetting and spreading.

[0058] After determining the equilibrium contact angle, the solid-liquid interfacial energy can be further calculated by combining the solid surface energy of the target inclusion. c sl The thermodynamic state of the bonding strength at the solid-liquid interface is used to characterize the solid-liquid two-phase interface.

[0059] In this embodiment, to characterize the migration and diffusion capacity, the mean square displacement data of the slag components obtained from the wetting simulation are post-processed to plot the relationship curve between the mean square displacement and time; then, a stable diffusion interval with a linear relationship is selected, and the diffusion coefficient of each component is obtained according to the Einstein relation.

[0060] The dissociation and exfoliation of complex multi-component inclusions at the interface and their migration into the slag bulk phase are essentially a mass transfer process involving the synergistic coupling and joint deduction of each component. This invention constructs an effective diffusion coefficient for characteristic mass transfer components, representing the overall migration potential of inclusions, by weighting the geometric average according to their stoichiometric proportions in the target inclusion. ; In the formula, D eff It reflects the migration and mass transfer potential of the slag system for the relevant components of the target inclusions; D j The self-diffusion coefficient of a certain characteristic mass transfer component; v j The stoichiometric coefficient is the chemical formula of the target inclusion corresponding to the characteristic mass transfer component. m This represents the total number of types of characteristic mass transfer components. D eff The larger the value, the easier it is for the components related to the absorption process of the target inclusions to overcome steric hindrance and diffuse and migrate in the bulk slag phase, thereby effectively avoiding mass transfer hindrance caused by the enrichment of dissolved products at the interface front.

[0061] Furthermore, the adhesion work calculated above... W ad With the effective diffusion coefficient of characteristic mass transfer components D eff Substituting into the aforementioned interface mass transfer driven power model, the comprehensive absorption capacity evaluation index of the slag system for specific target inclusions is calculated. P abs Based on this, a series of slag composition systems can be efficiently optimized.

[0062] In this embodiment, the adhesion force of electroslag slag on the surface of a MgAlO4 inclusion solid substrate is studied to explore the relationship between changes in slag composition and solid-liquid adhesion force. Furthermore, the strength of the slag's ability to continuously dissolve inclusions is explored through changes in the ion diffusion coefficient of the slag.

[0063] The specific process is as follows: (1) Based on the chemical composition of the metallurgical slag, an initial model of the slag is established. The specific components of the slag include: CaF2, CaO, Al2O3, and MgO. The specific components are shown in Table 1.

[0064] Table 1. Composition of the CaF2-CaO-Al2O3-MgO slag system (mass fraction, %)

[0065] The molar ratio is obtained based on the mass ratio of each component in the slag. Then, the number of different atoms (Si, Al, Ca, Mg and O) is determined based on the total number of atoms. Finally, the density of the slag is calculated using empirical formulas. The specific results are shown in Table 2.

[0066] Table 2. Atom number, density, and simulation domain side length of the CaF2-CaO-Al2O3-MgO slag system

[0067] Based on the aforementioned atomic types, quantities, and slag density, a randomly distributed amorphous structure model is constructed. The initial structure file and potential function of the slag to be tested are written into the .in file executed by the Lammps software. This invention uses a Born-Mayer-Huggins (BMH) type potential function to describe the interactions between ions. The potential parameters can be obtained through literature review or first-principles fitting, ensuring the reproduction of experimental properties such as slag density and structure factor. In the Lammps software, the initial slag structure is first minimized to eliminate unreasonable atomic configurations. Then, periodic boundary conditions are used, with a running time step of 1 fs and 100,000 relaxation steps, to bring the system to thermodynamic equilibrium.

[0068] (2) The relaxed slag structure model is placed in an extended vacuum layer computational domain, allowing it to spontaneously contract and aggregate at the target simulation temperature to form stable isolated droplets. The radial stratified statistical method is used to analyze the density distribution of the droplets to determine the radius of the Gibbs equivalent stress surface at the gas / liquid interface. ,like Figure 4 As shown in the figure, the arrows indicate that the density distribution of each layer is obtained by extending outward layer by layer from the center of the circle.

[0069] Based on the Young-Laplace equation, the average pressure difference Δ between the uniform region inside the droplet and the vacuum phase was calculated. P in The surface tension of the molten slag is solved by combining the formula. s lv The simulation employs periodic boundary conditions and a canonical ensemble throughout, ensuring the rigor of the statistical ensemble by maintaining constant atomic number, volume, and temperature.

[0070] Figure 5The calculated results of the surface tension of the molten slag as a function of MgO content at 2023 K are presented. Analysis shows that the surface tension of the molten slag exhibits a significant decreasing trend with increasing MgO mass fraction. This indicates that the addition of MgO plays a role in reducing surface tension in the electroslag system studied in this study. This is because the introduction of MgO releases additional free oxygen ions into the melt, promoting network depolymerization of the complex aluminum-oxygen tetrahedral network. Furthermore, Mg... 2+ It can act as a network modifier to disrupt the original long-range ordered structure, resulting in varying degrees of loosening of the overall network structure inside the melt, which is macroscopically manifested as a decrease in surface tension.

[0071] As the MgO content further increases, the surface tension tends to level off and then slightly rebounds. The core mechanism lies in the fact that, compared to the main cation Ca in the system, the surface tension... 2+ Mg 2+ It has a smaller ionic radius and a higher ionic field strength. As the MgO concentration continuously increases, a large amount of Mg... 2+ The aggregation alters the local coordination structure of the melt, and its stronger electrostatic attraction tends to attract surrounding O₂. 2- This leads to the formation of more compact Mg-O-Mg clusters; the formation of these dense micro-regions strengthens the local binding energy of the melt, offsetting the tension reduction effect caused by network deagglomeration, and ultimately resulting in a rebound in the surface tension of the system.

[0072] (3) Establish an initial model of the inclusion substrate according to the selected inclusion type. In different models, the size of the inclusion substrate is 161.6 Å × 24.2 Å. The wetting droplet is considered to be a cylindrical droplet, so the substrate is designed as a cuboid. The thickness is designed as three layers: the bottom is a fixed layer to constrain all atomic degrees of freedom and prevent the substrate from drifting as a whole; the middle is a thermostatic layer to maintain the simulation temperature through a Nosé-Hoover thermostat; and the top is a free layer to allow atoms to move freely and directly participate in the interface interaction.

[0073] In a three-dimensional axisymmetric droplet (cap-shaped), the three-phase contact line has curvature, and the line tension effect affects the contact angle measurement results as the droplet size changes. However, in a two-dimensional cylindrical droplet, due to the periodic boundary conditions, the three-phase contact line is an infinitely extending straight line in the third-dimensional direction with zero curvature, thus eliminating the interference of the line tension effect on the contact angle measurement. Therefore, the intrinsic interfacial wetting behavior can be directly studied.

[0074] (4) The relaxed slag structure model, i.e., the droplet model, is embedded into the simulation domain of the relaxed inclusion substrate model, so that the two share a unified simulation domain boundary; the distance between the droplet and the substrate is adjusted to 3 Å, and long-term relaxation is performed under a canonical ensemble for 1,200,000 time steps, so that the slag droplet spontaneously spreads and wets the substrate surface. The wetting process of sample AM1 changes with time as follows: Figure 6 As shown. It is worth noting that during the wetting process, the slag droplets and the inclusion substrate are kept separate and the mean square displacement of each particle in the slag is continuously recorded to statistically analyze the changes in the diffusion coefficients of each component of the slag.

[0075] The graph showing the wetting process over time reveals a significant and continuous dynamic spreading process of the molten slag droplets on the solid substrate surface. In the initial wetting stage, the molten slag droplets, upon initial contact with the inclusion substrate, maintain a relatively high height and a small contact area, exhibiting a large initial contact angle. As the relaxation time increases, driven by interfacial interactions, the molten slag droplets rapidly deform and spread outwards, with the dynamic contact angle continuously decreasing. Finally, the spreading process becomes gradual, and the molten slag droplet morphology reaches a thermodynamically stable state, forming a flattened spherical cap with a small contact angle on the inclusion substrate. This indicates that the electroslag system possesses good wetting performance on the inclusion substrate at the simulated temperature.

[0076] (5) After the molten slag droplets have fully spread and reached equilibrium on the surface of the inclusion substrate, the atomic trajectory file is imported into the post-processing module. Through computer-executed steps such as spatial feature clustering, density gradient peak localization, Gaussian filtering smoothing, and principal component analysis, errors from human analysis are avoided, and the equilibrium average contact angle is obtained. The results are as follows: Figure 7 As shown, with the increase of MgO content, the contact angle generally shows a pattern of first decreasing and then slightly increasing. The initial decrease in contact angle indicates that the introduction of MgO reduces the spreading resistance of molten slag droplets to some extent, enhancing their wetting tendency on the inclusion substrate. However, when the MgO content continues to increase, the contact angle slightly increases, indicating that the effect of reducing interfacial energy gradually approaches saturation, and the local structural rearrangement of the interface exerts a certain constraint on the wetting behavior.

[0077] (6) To further explore the interfacial interaction mechanism, based on the surface tension of the molten slag s lv The average adhesion work of the equilibrium wetting process was calculated. W ad The result is as follows Figure 8 As shown in the figure, the analysis reveals that with increasing MgO content, the adhesion work does not increase monotonically with improved wetting and spreading, but rather shows an overall decreasing trend; moreover, when the MgO content further increases, the adhesion work shows a slight rebound. This is because the average contact angle... iThe fluctuations are relatively small, and the change in adhesion work is mainly influenced by the surface tension of the liquid phase itself. s lv This means that the cohesive energy dominates the process. Therefore, although the depolymerization of MgO leads to the breakage of the slag network and exhibits better macroscopic spreading properties, this depolymerization also weakens the interaction forces within the slag liquid phase, thus reducing the energy required to overcome the separation process, i.e., the adhesion work. It can be concluded that the change in slag wetting behavior is mainly the result of the combined effects of liquid-phase cohesion and interfacial structure regulation.

[0078] (7) Further, in order to explore the microscopic dynamic evolution law of the system, the output mean square displacement data is imported into the post-processing program to plot the evolution curve of the mean square displacement of each specified ion in the system with simulation time. At the same time, a stable diffusion segment with a good linear relationship is selected in the curve, and the data of this segment is linearly fitted to solve the diffusion coefficient. Figure 9 The solution results for sample AM1 at 2023 K and the effective diffusion coefficients of each sample for the target characteristic mass transfer components are presented. The results show that the diffusion and migration capabilities of each particle in the slag system differ significantly, with their self-diffusion coefficients decreasing sequentially, in the following order: D F > D Ca > D Mg > D O > D Al Al 3+ With O 2- The lowest diffusion coefficient indicates that the two form a complex aluminum-oxygen polyhedral anionic network structure; while F - It exists outside the complex network structure and possesses relatively strong diffusion and penetration capabilities, which is key to giving the molten slag its high fluidity; in addition, Mg 2+ Compared to Ca 2+ It possesses a higher ionic field strength and polarization ability, resulting in its ability to more strongly adsorb surrounding oxygen ions in the melt and form a higher concentration than Ca. 2+ The more tightly coordinated local cluster structure restricts its free thermal diffusion behavior.

[0079] (8) Based on the effective diffusion coefficient of the characteristic combination element obtained above, the adhesion work and effective diffusion coefficient of each slag system are substituted into the interface mass transfer driving power model to calculate the comprehensive absorption evaluation index of each slag system for the target inclusion. P abs The result is as follows Figure 10As shown in the figure, the results indicate that with the increase of MgO mass fraction, the overall absorption intensity of the system exhibits an evolution pattern of first significantly increasing and then slightly decreasing. Analysis suggests that the addition of MgO causes depolymerization of the network structure, thereby weakening the interfacial adhesion work to some extent. However, the depolymerization process simultaneously releases the migration ability of confined ions, thus significantly enhancing the overall absorption power. However, when the MgO content continues to increase... P abs The decline indicates that the high concentration of Mg 2+ This not only degrades interfacial affinity but also inhibits the synergistic diffusion of ions into the bulk phase, macroscopically reflecting that excessive MgO is not conducive to the continuous and efficient dissolution of inclusions.

[0080] The above results indicate that kinetic parameters exhibit a higher weight in the experimental slag with a stronger variation range. Although network deagglomeration leads to a certain decrease in interfacial adhesion work, its effect on improving component migration efficiency is more significant, indicating that the inclusion absorption process is limited by kinetic mass transfer. This result also indirectly verifies the scientific validity of the evaluation method described in this invention, aiming to provide research ideas for slag system optimization design focusing on improving slag phase fluidity and mass transfer rate.

[0081] In summary, based on the molecular dynamics-based method for evaluating the ability of slag to absorb inclusions provided by this invention, in the electroslag remelting slag system set in this embodiment, it is recommended that the mass fraction of MgO be controlled within the range of approximately 7% to 11%. At this point, the interfacial wettability and kinetic mass transfer potential of the slag system to inclusions achieve optimal coupling and matching, resulting in the best overall removal effect.

[0082] The technical solution provided by this invention has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this invention, aiming to help understand the method and core ideas of this invention. It should be noted that this invention is not limited to the specific forms disclosed herein, and the above embodiments should not be construed as excluding other possible solutions. It should be understood that any modifications and variations made by those skilled in the art within the spirit and scope of this invention should fall within the protection scope of the appended claims.

Claims

1. A method for evaluating the ability of slag to absorb inclusions based on molecular dynamics, characterized in that, Includes the following steps: S1. Determine the initial composition of the simulation system, including the composition of the slag system and the types of inclusions; S2. Based on the composition and density of the selected slag system, calculate the number of each atom and establish an initial slag structure model; S3. Determine the crystallographic surface orientation based on the selected inclusion type, select the target crystal plane as the substrate surface in contact with the slag, expand the cell of the inclusion, and construct the initial structure model of the inclusion substrate. S4. Determine the potential function and corresponding potential parameters that describe the interaction between atoms in the simulation system; S5. Based on the potential function and corresponding potential parameters, write the input file for the molecular dynamics simulation command. Pre-relax the initial structure model of the slag and the initial structure model of the inclusion substrate respectively. After pre-relaxation, merge the models. The merged model includes the slag system, the inclusion system, and the vacuum region set around the slag and inclusions. In the merged model, the initial distance between the slag system and the inclusion system is greater than the cutoff distance of the potential function used. S6. Adjust the distance between the slag system and the inclusion system in the merged model to 2Å-4Å, and run the wetting simulation process to enable the merged model to spontaneously wet and continue running until wetting equilibrium is reached; continuously record the energy, temperature, pressure, atomic trajectory and mean square displacement of the slag system in the merged model. S7. After the wetting simulation is completed, the atomic trajectory file obtained is post-processed using a local tangent fitting method based on two-dimensional number density distribution to analyze the wetting state of the slag system on the inclusion system and obtain the average contact angle under equilibrium conditions. θ ; S8. The slag structure model obtained by pre-relaxation treatment of the initial slag structure model in step S5 is placed in an isolated vacuum system for further relaxation. The Laplace pressure difference and the true radius of curvature inside the droplet are calculated to obtain the surface tension of the slag, and combined with the average contact angle. θ The adhesion work characterizing the interfacial bonding strength was calculated. W ad ; S9. Extract the mean square displacement versus time curves of each component in the slag system after the wetting simulation operation, fit the curves within the stable diffusion range where a linear relationship exists, and obtain the self-diffusion coefficient of each component based on the Einstein relation. Then, determine the effective diffusion coefficient based on the self-diffusion coefficient. D eff It is used to evaluate the mass transfer and migration potential of slag systems for specific inclusion types; S10, the adhesion function W ad With the effective diffusion coefficient D eff Substituting into the interface mass transfer driven power model, the calculation dimension is J·s -1 Comprehensive absorption evaluation index P abs The strength of the synergistic effect of slag system driving interface wetting and component migration is used as an evaluation of the ability of different slags to absorb inclusions.

2. The method for evaluating the ability of slag to absorb inclusions based on molecular dynamics according to claim 1, characterized in that, In step S5, after the initial slag structure model has completed the pre-relaxation process, the boundary size of the simulation calculation domain is adjusted, and the size of other directions is adjusted by compensation to keep the total volume and number density of the slag system unchanged, so that the size of the slag structure model and the initial structure model of the inclusion substrate after pre-relaxation process are matched. At the same time, a vacuum region is added to facilitate the merging of the two models, and finally the merged model is output.

3. The method for evaluating the ability of slag to absorb inclusions based on molecular dynamics according to claim 1, characterized in that, In step S7, the process of calculating the average contact angle under equilibrium conditions includes: Determine the solid-liquid contact reference surface: Perform time-averaged processing on the atomic trajectory after wetting equilibrium, statistically analyze the atomic number density along the Z-axis, calculate the spatial gradient after Gaussian smoothing, and locate the last significant negative peak; find the first positive peak of the atomic number density gradient in the transition region to the right of the significant negative peak. The inflection point of the component density curve of the inclusion substrate corresponding to the positive peak is the position where the component density of the slag system increases the fastest, thereby determining the solid-liquid contact reference surface. Constructing the liquid phase profile based on the solid-liquid contact reference surface: The droplet centroid is corrected based on periodic boundary conditions, and the statistical time-averaged number density spectrum is processed by two-dimensional spatial gridding. The isodensity line extraction method is used to identify the liquid phase interface boundary. Defining the adsorption layer range: Based on the number density distribution curve of atoms in the slag system along the direction perpendicular to the inclusion substrate, the first density peak adjacent to the inclusion substrate is taken as the characteristic of the interface adsorption layer, and the height coordinate corresponding to the first density minimum value that appears after the first density peak is taken as the upper boundary of the adsorption layer. Local tangent fitting: A set of boundary points located above the upper boundary of the adsorption layer is selected. Principal component analysis is used to perform orthogonal distance linear fitting on the boundary point set. The direction of the first principal component of the boundary point set is extracted as the local tangent of the liquid interface. The slope of the local tangent at the intersection of the three phases is analyzed, and the average contact angle is calculated. θ .

4. The method for evaluating the ability of slag to absorb inclusions based on molecular dynamics according to claim 1, characterized in that, The acquisition of slag surface tension in step S8 specifically involves: first, placing the pre-relaxed slag structure model in an isolated vacuum system, maintaining the same composition, number of atoms, and temperature conditions as the merged model during wetting, and continuing relaxation; using the droplet centroid as the reference origin, statistically analyzing the radial time-averaged density distribution curve, identifying the surface density peak that appears before entering the vacuum transition region where the density monotonically decreases to zero, and extracting the radial coordinate position corresponding to the surface density peak as the droplet's true radius of curvature. Simultaneously, the pressure tensor of the steady-state region where the droplet's internal radius is smaller than its true radius of curvature is extracted to obtain the Laplace pressure difference. P in And calculate the surface tension of the molten slag. σ lv : ; Based on average contact angle θ and slag surface tension σ lv Solve for adhesion work W ad : .

5. The method for evaluating the ability of slag to absorb inclusions based on molecular dynamics according to claim 4, characterized in that, Given the average contact angle θ and slag surface tension σ lv Then, combined with the solid surface energy of the target inclusions σ sv Using formula γ sl =σ sv -σ lv cos θ Solving for solid-liquid interfacial energy γ sl It is used to characterize the bonding strength at the solid-liquid interface.

6. The method for evaluating the ability of slag to absorb inclusions based on molecular dynamics according to claim 1, characterized in that, In step S9, the self-diffusion coefficients of each component in the slag system are... D Solving using Einstein's relation, the formula is: ； In the formula, D Indicates the self-diffusion coefficient; t is the simulation time; It is the mean square displacement of the component within time t; The components that characterize the absorption process of target inclusions by the slag and can obtain a self-diffusion coefficient in the slag system are defined as characteristic mass transfer components. The effective diffusion coefficient is constructed by geometrically averaging the self-diffusion coefficients of each characteristic mass transfer component. D eff The formula is: ； In the formula, D eff It reflects the mass transfer and migration potential of the slag system to the relevant components of the target inclusions; D j For the first Self-diffusion coefficient of each characteristic mass transfer component; v j The stoichiometric coefficient is the chemical formula of the target inclusion corresponding to the characteristic mass transfer component. m This represents the total number of types of characteristic mass transfer components.

7. The method for evaluating the ability of slag to absorb inclusions based on molecular dynamics according to claim 1, characterized in that, In step S10, the interface mass transfer driving power model, without considering interfacial chemical reactions, comprehensively considers the diffusion and migration characteristics of the characteristic mass transfer components of the slag system on the target inclusions and the interfacial wetting and bonding characteristics. Its expression is as follows: ； In the formula, P abs This is an evaluation index of the slag system's ability to dissolve and absorb target inclusions, expressed in units of 10. -12 J·s -1 This characterizes the synergistic effect of the slag system on interfacial wetting and component dissolution and migration per unit time.

Images