Multi-field coupling grid division method, device and product for liquid phase growth of silicon carbide

Abstract

The present application provides a multi-field coupling grid division method, device and product for growing silicon carbide by a liquid phase method, and relates to the field of single crystal growth technology. The method comprises: establishing a geometric model according to structural information of a silicon carbide growing system by a liquid phase method; performing grid division according to a maximum total amount of calculation grids to obtain a numerical calculation model; simulating electromagnetic induction heating processes, heat transfer flow processes and component transport processes under multiple different electromagnetic induction heating frequencies based on the numerical calculation model to obtain multiple groups of simulation results; calculating and obtaining target thickness fitting curves according to the multiple groups of simulation results, including: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve; determining the grid thickness of the multi-physical field coupling grid from the thickness fitting curve according to the current electromagnetic induction heating frequency, performing grid division and obtaining a target numerical calculation model.

Description

Technical Field

[0001] The present application relates to the technical field of single crystal growth, and in particular to a multi-field coupling grid division method, device and product for growing silicon carbide by a liquid phase method. Background Art

[0002] The liquid phase method is also called the solution growth method, which can be used to prepare large-sized silicon carbide crystals, etc. In the liquid phase method, electromagnetic induction heating is mainly used to provide heat for the crystal growth process, that is, the electromagnetic induction method is used to generate current inside the heated material (crucible, side graphite support and other parts), and the energy of these eddy currents is used to achieve the heating purpose. Induction heating as a heating method is of great significance in the liquid phase growth system. The temperature field and thermal field of the liquid phase growth system determine the quality of the crystal.

[0003] In the numerical calculation process of liquid phase growth of silicon carbide, the coupled solution process of multiple physical fields such as magnetic field, temperature field, velocity field and concentration field is involved. In the numerical solution process, the number of divided computational grids will affect the accuracy of the final numerical calculation results. If the computational grids are divided too little (especially the computational grids of the physical field boundary layer), it is difficult to obtain accurate solution results. If the computational grids are divided too much, it is easy to waste computing resources. Therefore, it is urgent to propose a multi-field coupled grid division method, device and product for liquid phase growth of silicon carbide to achieve accurate division of multi-physical field coupled grids, thereby achieving a balance between computing resources and the accuracy of numerical calculation results. Summary of the invention

[0004] In view of the above problems, embodiments of the present application provide a multi-field coupling grid division method, device and product for liquid phase growth of silicon carbide, so as to overcome the above problems or at least partially solve the above problems.

[0005] In a first aspect of an embodiment of the present application, a multi-field coupling grid division method for liquid phase growth of silicon carbide is provided, the method comprising:

[0006] According to the structural information of the liquid phase growth silicon carbide system, a geometric model is established;

[0007] According to the maximum total amount of computational grids, the geometric model is divided into computational grids, boundary conditions of the corresponding computational grids in the geometric model are set, and electromagnetic induction control equations, heat transfer flow control equations and component transport control equations are set to obtain a numerical calculation model;

[0008] Based on the numerical calculation model, the electromagnetic induction heating process, heat transfer flow process and component transport process under multiple different electromagnetic induction heating frequencies are simulated to obtain multiple groups of simulation results;

[0009] According to the plurality of sets of simulation results, a target thickness fitting curve is calculated, wherein the target thickness fitting curve includes: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve;

[0010] According to the current electromagnetic induction heating frequency, the grid thickness of the multi-physical field coupling grid of each region in the liquid phase method silicon carbide growth system is determined from the thickness fitting curve, and grid division is performed to obtain a target numerical calculation model.

[0011] The second aspect of the embodiment of the present application further provides a multi-field coupling meshing device for growing silicon carbide by a liquid phase method, which is applied to execute the multi-field coupling meshing method described in the first aspect of the embodiment of the present application, and the device comprises:

[0012] A model building module is used to build a geometric model based on the structural information of the liquid phase growth silicon carbide system;

[0013] A numerical calculation module, used to divide the geometric model into calculation grids according to the maximum total amount of calculation grids, set boundary conditions of the corresponding calculation grids in the geometric model, set electromagnetic induction control equations, heat transfer flow control equations and component transport control equations, and obtain a numerical calculation model;

[0014] A simulation module, for simulating the electromagnetic induction heating process, heat transfer flow process and component transport process under multiple different electromagnetic induction heating frequencies based on the numerical calculation model, and obtaining multiple groups of simulation results;

[0015] A fitting module, used for calculating a target thickness fitting curve according to the plurality of simulation results, wherein the target thickness fitting curve includes: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve;

[0016] The grid division module is used to determine the grid thickness of the multi-physical field coupling grid in each area of ​​the liquid phase method silicon carbide growth system from the thickness fitting curve according to the current electromagnetic induction heating frequency, perform grid division, and obtain the target numerical calculation model.

[0017] A third aspect of an embodiment of the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored on the memory, wherein the processor executes the computer program to implement the steps in the multi-field coupled meshing method for growing silicon carbide by the liquid phase method described in the first aspect of the embodiment of the present application.

[0018] The fourth aspect of the embodiments of the present application further provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the steps in the multi-field coupled meshing method for growing silicon carbide by the liquid phase method described in the first aspect of the embodiments of the present application.

[0019] The fifth aspect of the embodiments of the present application further provides a computer program product, which, when executed on an electronic device, enables a processor to implement the steps in the multi-field coupling grid division method for growing silicon carbide by the liquid phase method as described in the first aspect of the embodiments of the present application.

[0020] The embodiment of the present application provides a multi-field coupling grid division method for liquid phase growth of silicon carbide, the method comprising: establishing a geometric model according to structural information of a liquid phase growth of silicon carbide system; performing computational grid division on the geometric model according to the maximum total amount of computational grids, setting boundary conditions of the corresponding computational grids in the geometric model, setting electromagnetic induction control equations, heat transfer flow control equations and component transport control equations, and obtaining a numerical calculation model; based on the numerical calculation model, simulating electromagnetic induction heating processes, heat transfer flow processes and component transport processes under multiple different electromagnetic induction heating frequencies to obtain multiple groups of simulation results; calculating a target thickness fitting curve according to the multiple groups of simulation results, the target thickness fitting curve comprising: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve; determining the grid thickness of the multi-physical field coupling grid of each region in the liquid phase growth of silicon carbide system from the thickness fitting curve according to the current electromagnetic induction heating frequency, performing grid division, and obtaining a target numerical calculation model.

[0021] The embodiment of the present application establishes a numerical calculation model according to the maximum total amount of calculation grids, thereby simulating the electromagnetic induction heating process, heat transfer flow process and component transport process under different electromagnetic induction heating frequencies, and obtaining multiple groups of simulation results. Then, according to the information in the simulation results (such as carbon concentration information at the calculation grid, Lorentz force magnitude and current value, etc.), the most suitable concentration boundary layer thickness, heat boundary layer thickness and Lorentz force boundary layer thickness are determined, so as to generate a relationship curve between the electromagnetic induction heating frequency and the boundary layer thickness (i.e., the target thickness fitting curve) according to the multiple groups of simulation results. Therefore, in the actual application process, the grid thickness of the multi-physics field coupling grid (including the concentration boundary layer grid thickness, the heat boundary layer grid thickness and the Lorentz force boundary layer grid thickness) can be determined according to the size of the electromagnetic induction heating frequency adopted, so as to achieve accurate calculation grid division, and reduce unnecessary waste of computing resources on the premise of ensuring the accuracy of the results of numerical calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0023] Figure 1 It is a flowchart of the steps of the multi-field coupling grid division method for growing silicon carbide by the liquid phase method provided in an embodiment of the present application;

[0024] Figure 2 It is a schematic structural diagram of a system for producing silicon carbide by a liquid phase method provided in an embodiment of the present application;

[0025] Figure 3 is a schematic diagram of a concentration boundary layer thickness fitting curve provided in an embodiment of the present application;

[0026] Figure 4 is a schematic diagram of a thermal boundary layer thickness fitting curve provided in an embodiment of the present application;

[0027] Figure 5 is a schematic diagram of a Lorentz force boundary layer thickness fitting curve provided in an embodiment of the present application;

[0028] Figure 6 It is a structural schematic diagram of a multi-field coupling grid division device provided in an embodiment of the present application;

[0029] Figure 7 is a schematic diagram of an electronic device provided in an embodiment of the present application;

[0030] Figure symbols: 1. Silicon melt; 2. Graphite crucible; 3. Side graphite support; 4. Carbon felt; 5. Copper coil. DETAILED DESCRIPTION

[0031] The exemplary embodiments of the present application will be described in more detail below in conjunction with the accompanying drawings in the embodiments of the present application. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to enable the scope of the present application to be fully communicated to those skilled in the art.

[0032] As a third-generation semiconductor material, SiC has excellent properties such as large bandgap, high thermal conductivity, and high chemical stability. It is a semiconductor material with great application prospects. At present, there are three mainstream growth methods for SiC, namely physical vapor transport, high-temperature chemical vapor deposition, and liquid phase method. Among them, SiC grown by liquid phase method has been proven to have advantages such as high crystal quality, easy diameter expansion, and easy p-type doping. Induction heating is usually used to provide heat in the liquid phase method to grow SiC single crystal system.

[0033] The liquid phase method is also called the solution growth method. In the liquid phase method, the graphite crucible serves not only as a container but also as a carbon source for the Si melt. Specifically, the crucible is heated to a sufficiently high temperature by an induction coil, where the carbon dissolves into the Si melt due to the high temperature and local unsaturation. Then, the carbon is transported to the bottom of the seed crystal by convection and diffusion, where the temperature near the seed crystal is relatively low, and the carbon here becomes supersaturated, and finally a SiC crystal grows. Among them, this supersaturation provides a driving force for the growth of SiC.

[0034] Induction heating is the use of electromagnetic induction to generate current inside the heated material (such as crucible, side graphite support, etc.), relying on the energy of these eddy currents to achieve the purpose of heating. Induction heating as a heating method is of great significance in liquid phase growth systems. The temperature field and thermal field of the growth system determine the quality of the crystal. However, the relevant parameters of the induction heating system and the structural layout of the components will affect the distribution of heat during the induction heating process, and thus affect the growth quality and growth rate of the crystal. Therefore, for a specific induction heating system, the induction heating frequency, the geometry and relative position of the coil and crucible in the system should be carefully determined.

[0035] In the numerical calculation process of liquid phase growth of silicon carbide, the coupled solution process of multiple physical fields such as magnetic field, temperature field, velocity field and concentration field is involved. First, the magnetic field is solved to obtain the heat distribution on each component, as well as the Lorentz force distribution on each component, including the Lorentz force distribution in the silicon melt area; the parameters obtained from the induction heating process (solving the magnetic field) are loaded into the heat transfer flow process. This process is a one-way coupling, which is equivalent to the one-way coupling of the magnetic field to the temperature field and velocity field; next, the temperature field and velocity field distribution in the crystal growth furnace are solved, and the temperature field and velocity field are bidirectionally coupled; finally, the component transport process is solved. This process is to obtain the concentration field based on the temperature field and velocity field. This process is a one-way coupling solution.

[0036] In the process of numerical solution, the number of divided computational grids will affect the accuracy of the final numerical calculation results. The number of grids that save the most computational resources is different in different solution processes. For example, how to divide the grid to accurately solve the magnetic field; how to divide the grid to accurately solve the concentration field, etc. If the computational grid is divided too little (especially the computational grid of the physical field boundary layer), it is difficult to obtain an accurate solution result. If the computational grid is divided too much, it is easy to waste computational resources.

[0037] In view of the above problems, the embodiments of the present application propose a multi-field coupled meshing method, device and product for growing silicon carbide by a liquid phase method, so as to achieve accurate division of multi-physical field coupled meshes, thereby achieving a balance between computing resources and the accuracy of numerical calculation results. In conjunction with the accompanying drawings, the multi-field coupled meshing method, device and product for growing silicon carbide by a liquid phase method provided by the embodiments of the present application are described in detail through some embodiments and their application scenarios.

[0038] The first aspect of the present application provides a multi-field coupling grid division method for growing silicon carbide by liquid phase method, referring to Figure 1 , Figure 1 A flow chart of the steps of a multi-field coupling grid division method for growing silicon carbide by a liquid phase method provided in an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0039] Step S101, establishing a geometric model according to the structural information of the liquid phase method growth silicon carbide system.

[0040] In the embodiment of the present application, silicon carbide is prepared by liquid phase method. In the present embodiment, the liquid phase method for growing silicon carbide system refers to any system that can be used to prepare silicon carbide by liquid phase method (such as a 6-inch SiC growth system). The embodiment of the present application does not limit the specific structure and size parameters of the liquid phase method for growing silicon carbide system. Among them, the structural information of the liquid phase method for growing silicon carbide system can be information such as the size and position of each component in the system. For example, refer to Figure 2 , Figure 2 A schematic diagram of the structure of a liquid phase method for producing silicon carbide is shown, Figure 2 As shown in FIG. 1 , the liquid phase method for growing silicon carbide system includes at least the following components: 1. silicon melt; 2. graphite crucible; 3. side graphite support; 4. carbon felt; 5. copper coil. Figure 2 As shown, the graphite crucible 1 is located in the center of the system, and the crucible is filled with silicon melt 2. The side graphite support 3 is located on the side of the crucible, which serves as the main heating element in the system and transfers the heat obtained by electromagnetic induction to the crucible and the melt in the crucible through radiation. The carbon felt 4 is located outside the side graphite support 3, which plays a role of support and heat preservation. The copper coil 5 is located outside the carbon felt 4 for electromagnetic induction heating.

[0041] According to the structural information of the liquid phase method grown silicon carbide system (including the structure of each component and the size parameters and other information), the structure of the liquid phase method grown silicon carbide system is geometrically modeled using modeling software to obtain the corresponding geometric model.

[0042] Step S102, dividing the geometric model into computational grids according to the maximum total amount of computational grids, setting boundary conditions of the corresponding computational grids in the geometric model, setting electromagnetic induction control equations, heat transfer flow control equations and component transport control equations, and obtaining a numerical calculation model.

[0043] First, the geometric model is divided into computational grids according to the maximum total amount of computational grids. Specifically, the meshing software is used to divide the computational grids of each region in the liquid phase growth silicon carbide system (wherein the grid adopts a quadrilateral grid). Each component in the geometric model is divided into multiple computational grids. Among them, the maximum total amount of computational grids is determined based on the total amount of computing resources used for numerical calculations, because the more the number of divided computational grids, the more accurate the results obtained by numerical calculations, but the more computing resources are used for numerical calculations. Exemplarily, in order to ensure that each component uses a sufficiently dense computational grid, the maximum total amount of computational grids can be set to 4.8 million, that is, the number of grids in the entire computational area is 4.8 million. This includes meshing for the concentration boundary layer, meshing for the heat boundary layer, and meshing for the Lorentz force boundary layer. In addition, in the system of growing silicon carbide by the liquid phase method, the cross-sections of carbon felt, graphite rods, side graphite supports, copper coils, and SiC seed crystals are all regular shapes, which can be divided into structured grids to improve computational efficiency. In practice, the shape of the inner wall of the crucible may be a curved surface, so an unstructured grid is used when dividing the grids of the graphite crucible and the silicon melt (but the boundary layer grids in the silicon melt are all structured grids).

[0044] Set the boundary conditions of the corresponding computational grid in the geometric model. Specifically, define the boundary conditions and working pressures between the various components in the geometric model during the electromagnetic induction heating process, heat transfer flow process, and component transport process using the liquid phase method (i.e., the boundary conditions of each computational grid). For example: the boundary conditions of the contact surface between the inner wall of the crucible and the silicon melt, the temperature of the outer wall of the furnace in the geometric model, the pressure in the furnace, the boundary conditions of the gas-liquid interface (free surface) in the geometric model, and the boundary conditions of other solid-gas contact surfaces in the geometric model.

[0045] Specifically, the boundary conditions in the embodiments of the present application include:

[0046] Boundary conditions of electromagnetic induction heating process: The outer boundary of the region adopts a balloon boundary. On the edge of its boundary, the magnetic field is neither perpendicular to the edge nor parallel to the edge. This boundary condition is to simulate infinity and reduce the amount of calculation; the other boundaries adopt natural boundary conditions. The tangential component of the magnetic field intensity and the normal component of the magnetic induction intensity on the contact surface of the two contacting objects remain continuous ( , ). H t represents the tangential component of the magnetic field intensity, B n Represents the normal component of magnetic induction intensity.

[0047] Boundary conditions for heat transfer flow process: The inner wall of the coil is the temperature boundary condition (It is equivalent to giving this wall a fixed temperature C), the other walls are coupled walls (the temperature and heat flux density on both sides of the interface are equal); the surface tension gradient (Malaygo force) is set on the free surface of the silicon melt, and the rest are no-slip boundaries (i.e. )speed.

[0048] Boundary conditions at the interface (gas-liquid boundary): set the surface tension gradient ;in, represents the surface tension coefficient in N / m, and T represents the temperature in °C.

[0049] In this embodiment, setting the boundary conditions of the corresponding computational grid in the geometric model includes:

[0050] adding a dissolution boundary condition to the interface between the crucible and the silicon melt in the geometric model;

[0051] Add a precipitation boundary condition to the interface between the silicon melt and the crystal in the geometric model; wherein the dissolution boundary condition and the precipitation boundary condition are the same, and are both expressed as the following formula:

[0052] ;

[0053] in, is the molar mass of silicon; is the density of silicon; e is the mathematical constant Euler number; T is the temperature at the grid.

[0054] Specifically, before numerical solution, it is necessary to set corresponding boundary conditions to simulate the real situation. This includes adding dissolution boundary conditions to the interface between the crucible and the silicon melt to simulate the dissolution of the graphite crucible; adding precipitation boundary conditions to the interface between the silicon melt and the crystal to simulate the precipitation of silicon carbide crystals. Both of these boundaries are equilibrium concentration boundaries of carbon. C eq .

[0055] In addition, at the free surface of the melt (interface between the melt and the gas), the normal gradient of the concentration is set to ,in, C c represents the molar concentration of carbon, x Indicates the direction perpendicular to the free surface. In order to simulate the real situation, silicon carbide crystals will not precipitate on the free surface. In actual operation, water cooling is usually passed through the inner area of ​​the coil to keep the inner wall of the coil at room temperature. In order to save computing resources, the water cooling inside the coil is simplified, and the temperature of the inner wall of the coil is given as room temperature during the solution. The remaining contact surfaces are set as temperature continuous walls, and the temperature on the left and right sides of the contact surfaces with different structures is continuous.

[0056] In this embodiment, the geometric model is initialized, and the control equations corresponding to each component and each region in the geometric model during electromagnetic induction heating using the liquid phase method, the control equations in the heat transfer flow process, and the control equations in the component transport process are set. For example: the electromagnetic induction control equation in the geometric model, the heat transfer flow control equation satisfied by the melt-crystal region in the geometric model, the heat transfer flow control equation satisfied by the gas region in the geometric model, etc. Among them, the electromagnetic induction control equation is used to solve the distribution of induced heat in the geometric model when electromagnetic induction heating is performed under specific conditions (electromagnetic induction heating frequency), that is, the induced power of each calculation grid in the geometric model. The heat transfer flow control equation is used to solve the thermal field information (such as temperature) and flow field information of each calculation grid in the geometric model under specific conditions (electromagnetic induction heating frequency). In the embodiment of the present application, the specific calculation method of the heat transfer flow control equation is not limited. In addition, when defining the control equation, the control equation corresponding to the self-inductance phenomenon in the copper coil (i.e., the above-mentioned electromagnetic induction control equation) is set, so that when simulating the electromagnetic induction heating process, the influence of the self-inductance phenomenon in the copper coil on the thermal field is added, thereby further improving the accuracy of the obtained simulation results.

[0057] Specifically, the control equation of the electromagnetic induction heating process (i.e., the above-mentioned electromagnetic induction control equation) includes the following formula:

[0058] ;

[0059] ;

[0060] in, Indicates the electric field strength v / m, is the induced current A / m 2 , is the magnetic potential vector (dimensionless quantity, no unit), t is the current time s, r is the distance m, μ is the magnetic permeability H / m; during electromagnetic induction heating, an alternating current is passed through the copper coil ;in, J 0 is the current density amplitude (maximum value) A / m 2 ; is the angular frequency rad / s; t is the current time; , f is the electromagnetic induction frequency kHz;

[0061] Substitute the physical property parameters of each component into the control equations of the electromagnetic induction heating process and the heat transfer flow process, wherein the physical property parameters at least include the material conductivity and density of each component. In this embodiment, according to the materials used in the actual production process (i.e., the materials of each component in the liquid phase growth silicon carbide system), the physical property parameters of each component (the conductivity of each material, the density of the material, etc.) are substituted into the control equations of the electromagnetic induction heating process and the heat transfer flow process to establish the electromagnetic induction heating numerical model and the heat transfer flow numerical model.

[0062] The heat transfer flow governing equations include:

[0063] Continuity equation: ;

[0064] Momentum equation: ;

[0065] Energy equation: ;

[0066] in, It means to find the gradient (i.e., to solve the first-order derivative perpendicular to the crystal growth interface); Indicates speed m / s; Indicates density kg / m 3 ; p Indicates pressure Pa; Indicates dynamic viscosity Pa·s; Indicates the acceleration due to gravity m / s 2 ; represents the coefficient of thermal expansion; T Indicates the temperature at the grid in °C; Indicates reference temperature °C; is the Lorentz force density N / m 3 ; represents specific heat capacity; represents the thermal conductivity of silicon melt; Indicates heat source W / m 3 ; represents the transpose of the velocity gradient, t Indicates the current time; represents the induced current density; is the magnetic induction intensity.

[0067] Substitute the density of the solid material of each component into the governing equation of the heat transfer flow process in the solid material area:

[0068] ;

[0069] in, Indicates the density of solid materials kg / m 3 ; It represents the specific heat capacity of solid materials; Represents the thermal conductivity of solid materials; Indicates heat source W / m 3 ; The density of the solid material is respectively: the density of the crucible, the density of the graphite support component, the density of the copper coil or the density of the carbon felt. In this embodiment, the control equation of the solid material region includes: the control equation of the crucible region, the control equation of the graphite support component on the side of the crucible, the control equation of the copper coil and the control equation of the carbon felt. Therefore, when the density of the solid material of each component is substituted into the control equation, the density of the solid material is: the density of the crucible, the density of the graphite support component, the density of the copper coil or the density of the carbon felt.

[0070] The component transport control equations include:

[0071] ;

[0072] in, It means to find the gradient (i.e., to solve the first-order derivative perpendicular to the crystal growth interface); Indicates density; Indicates speed; represents the molar concentration of carbon; represents the mass diffusion coefficient of carbon in silicon melt; Indicates the second-order derivative (i.e., the second-order derivative perpendicular to the crystal growth interface). Specifically, in the subsequent numerical solution process, the component transport control equation is crucial. The key factor that ultimately leads to the quality of crystal growth and the amount of crystal precipitation is the carbon concentration distribution in the silicon melt. The initial component transport equation is:

[0073] ;

[0074] in , k=1,2……N ; Indicates density kg / m 3 ; t Indicates the current time. Indicates shared k The mass fraction of different components;x Indicates the distance m; u Indicates speed m / s; Indicates shared k Mass diffusion coefficient of different components in silicon melt kg / (m·s); express k The mass concentration source terms of different components in the silicon melt.

[0075] Combined with the numerical calculation process of the embodiment of the present application, the solution is reasonably simplified. This solution is for steady state, which is independent of time, that is, , this item can be ignored. In numerical calculation, only the transport and calculation of carbon are set, so k =1. In addition, when solving the problem, there will be no carbon atoms in the silicon melt in the initial state, so =0. The source of carbon is the carbon on the crucible wall dissolving into the silicon melt. Then convert the mass fraction into molar concentration, and the final component transport control equation in this numerical calculation process is obtained as follows: .

[0076] Step S103, based on the numerical calculation model, simulating the electromagnetic induction heating process, heat transfer flow process and component transport process under multiple different electromagnetic induction heating frequencies to obtain multiple groups of simulation results.

[0077] In this embodiment, each set of simulation results includes, at the electromagnetic induction heating frequency, the information of the electromagnetic induction heating process (the induction power of each calculation grid), the information of the heat transfer flow process (the temperature value, thermal field, flow field information, silicon melt velocity, etc. of each calculation grid), and the information of the component transport process (the carbon concentration of each calculation grid) obtained by solving the electromagnetic induction control equation, the heat transfer flow control equation and the component transport control equation. Among them, the commonly used frequency in the liquid phase growth of silicon carbide is 1kHz-5kHz, and multiple different electromagnetic induction heating frequencies can be taken within the frequency range of 1kHz-5kHz, such as 1kHz, 2kHz, 3kHz, 4kHz and 5kHz.

[0078] Specifically, for each electromagnetic induction heating frequency, the electromagnetic induction heating process under the electromagnetic induction heating frequency is simulated. Specifically, by solving the electromagnetic induction control equation in the electromagnetic induction heating numerical model, the electromagnetic induction heating process under the electromagnetic induction heating frequency (i.e., part of the simulation results) is obtained, including the current value on each component in the geometric model when electromagnetic induction heating is performed at the electromagnetic induction heating frequency, and the numerical value of the induced power of each calculation grid, i.e., the induced heat density (i.e., the induced power density) on each component. By solving the heat transfer flow control equation, the heat transfer flow process, i.e., the process of stable growth of silicon carbide single crystal (i.e., part of the simulation results) is simulated, including: the temperature value, thermal field, flow field information, silicon melt velocity, etc. of each calculation grid in the geometric model during the stable growth of silicon carbide single crystal. By solving the heat transfer flow control equation at the current time, the silicon melt area at the current time is obtained. Speed ​​information, and then the speed Substituting the component transport control equation into the equation and solving it, we can obtain information such as the carbon concentration corresponding to each computational grid at the current time (i.e., part of the simulation results).

[0079] Step S104, calculating and obtaining a target thickness fitting curve according to the multiple sets of simulation results, wherein the target thickness fitting curve includes: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve.

[0080] Among them, the concentration boundary layer thickness fitting curve refers to the average value of the thickness of the carbon concentration boundary layer on the side of the silicon melt at the solid-liquid interface, and the relationship curve between the electromagnetic induction heating frequency. When the temperature is high, the graphite crucible will be dissolved at the interface between the silicon melt and the graphite crucible. In the process of solving the heat transfer flow, the velocity field and temperature field of the silicon melt area will be obtained, and the carbon concentration distribution in the silicon melt will be obtained according to the velocity field and the temperature field, so as to solve the growth rate of the crystal through the carbon concentration distribution (that is, the component transport process is simulated). In addition, when solving the component transport control equation, the crucible dissolution process will be simulated near the interface between the graphite crucible and the silicon melt. There is also a concentration boundary layer in this area (the interface between the graphite crucible and the silicon melt), but in the numerical simulation of the growth of silicon carbide by the liquid phase method, only the carbon concentration distribution at the solid-liquid interface is ultimately concerned, so the embodiment of the present application does not impose special restrictions on the grid near the interface between the graphite crucible and the silicon melt.

[0081] In a possible implementation, calculating a target thickness fitting curve according to the multiple sets of simulation results includes:

[0082] Step S201, determining the carbon concentration value of each calculation grid in the silicon melt region according to each group of simulation results.

[0083] Specifically, the simulation results are obtained by solving the control equation. Then, according to the component transport control equation in the simulation results, The solution result can be used to obtain the carbon concentration at each grid in the silicon melt area, that is, the proportion of carbon atoms and silicon atoms at the silicon carbide crystal growth interface can be known, that is, the carbon content and silicon content at the silicon carbide crystal growth interface.

[0084] Step S202 , calculating the supersaturation of each layer of the calculation grid, starting from the calculation grid on the side of the silicon melt at the solid-liquid interface, according to the carbon concentration value.

[0085] In a possible implementation manner, the calculating the supersaturation of each layer of the computing grid includes:

[0086] The supersaturation of each layer of the computational grid is calculated according to the following formula:

[0087] ;

[0088] in, S is the supersaturation, c Calculate the carbon concentration value mol / m at the grid for this layer 3 , c eq Calculate the carbon equilibrium concentration mol / m at the grid for this layer 3 .

[0089] Step S203, determining the distance between the calculation grid at the maximum supersaturation value and the solid-liquid interface as the target distance.

[0090] Step S204: determining the target concentration boundary layer grid thickness corresponding to the simulation result according to the target distance.

[0091] Specifically, the supersaturation of the solid-liquid interface is 0. This is because one side of the solid-liquid interface is solid, so the supersaturation S here is 0. When one side of the silicon melt is slightly away from the solid-liquid interface, and the temperature here is low, the carbon equilibrium concentration is low, and the carbon here will be supersaturated, causing silicon carbide crystals to precipitate; but when the distance from the solid-liquid interface is very far and the distance from the crucible is close, the temperature here is high, the equilibrium concentration is high, and the carbon here is unsaturated. Therefore, for one side of the silicon melt, the supersaturation first increases and then decreases as it moves away from the solid-liquid interface. There must be a maximum supersaturation Smax on the side of the silicon melt below the seed crystal in the silicon melt region. The distance between the maximum supersaturation and the solid-liquid interface is the thickness of the carbon concentration boundary layer on the silicon melt side of the solid-liquid interface. It should be noted that the thickness of the entire carbon concentration boundary layer is not exactly the same, and there is a certain fluctuation. The thickness of the carbon concentration boundary layer at the actual solid-liquid interface position is not uniform, and the carbon concentration boundary layer thickness at different positions of the solid-liquid interface is inconsistent. In order to make the grid division simpler and the calculation results more accurate, the embodiment of the present application can calculate the carbon concentration boundary layer thickness corresponding to multiple points on the solid-liquid interface, and then calculate the average thickness as the final carbon concentration boundary layer thickness (i.e., the target distance). Furthermore, in order to make the final calculation result more accurate, 5 layers of grids are divided on the basis of the average thickness (i.e., the obtained carbon concentration boundary layer thickness consists of 5 layers of grids, and the growth rate between each two adjacent grids is 1.2) as the target concentration boundary layer grid thickness, which can further improve the accuracy of the calculation results compared to the technical solution of taking 3 layers of grids for a given boundary layer thickness in the prior art.

[0092] Step S205, according to the target concentration boundary layer grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the concentration boundary layer thickness fitting curve, and the concentration boundary layer thickness fitting curve represents the relationship curve between the target concentration boundary layer grid thickness and the electromagnetic induction heating frequency.

[0093] Specifically, for each set of simulation results, a set of electromagnetic induction heating frequencies and corresponding target concentration boundary layer grid thicknesses can be obtained, and then multiple sets of electromagnetic induction heating frequencies and corresponding target concentration boundary layer grid thicknesses can be obtained based on multiple sets of simulation results, and then a fourth-order polynomial fitting is used to obtain the relationship curve between the two. Figure 3 , Figure 3 A schematic diagram of a concentration boundary layer thickness fitting curve is shown, such as Figure 3 As shown, the horizontal axis is the electromagnetic induction heating frequency, and the vertical axis is the most suitable concentration boundary layer grid thickness (ie, the target concentration boundary layer grid thickness). For example, when the electromagnetic induction heating frequency is 2 kHz, the corresponding most suitable concentration boundary layer grid thickness is 0.32 mm.

[0094] In this embodiment, the thermal boundary layer thickness fitting curve refers to the relationship curve between the average value of the thickness of the heat concentrated distribution area on each component and the electromagnetic induction heating frequency. In the liquid phase growth of silicon carbide system, the heat is mainly distributed on the graphite crucible, the side graphite support, the carbon felt and the copper coil. Therefore, in this embodiment, the components mentioned include: graphite crucible, side graphite support, carbon felt and copper coil, and the average thickness of the thermal boundary layer on these components at different frequencies is calculated. In the process of solving the electromagnetic induction heating, the final heat distribution is related to the layout between the components, the shape of the components, the electromagnetic shielding effect of the external components on the internal components, and the skin effect.

[0095] In a possible implementation, calculating and obtaining a target thickness fitting curve according to the multiple sets of simulation results includes:

[0096] Step S301, determining the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component according to each group of simulation results; the power density represents the value of the total heat induced by the corresponding calculation grid in an electromagnetic induction heating cycle divided by the cycle duration; the components include: crucible, side graphite support, carbon felt and copper coil.

[0097] Specifically, the simulation results are obtained by solving the relevant control equations. Then, according to the solution results of the heat transfer flow control equations in the simulation results (including the current values ​​of the calculation grids corresponding to each component in the simulation results), the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component can be obtained.

[0098] In a possible implementation, the step S301, determining the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component according to each group of simulation results, includes:

[0099] Step S3011, determining the current value of the outer surface of each component during the electromagnetic induction heating process and the current value of each calculation grid in each component according to the simulation results.

[0100] Specifically, when solving, the complex function method is used to solve the electromagnetic induction heating process. Taking into account the time harmonicity of the electromagnetic field, the physical quantities such as the magnetic potential vector and the current density are written in complex form. The field quantities and field sources in complex form are only functions of the spatial position and have nothing to do with time. In addition, when electromagnetic induction heating is performed, the frequency is also large, so its period is short, and the induced heat is defined as the average value within one period. The formula for calculating the magnetic potential vector is as follows:

[0101] ;

[0102] in, is the magnetic potential vector, is the imaginary part of the magnetic potential vector, is the real part of the magnetic potential vector, i is a complex unit. The currents in the copper coil and the conductor (i.e. the current values ​​of each component in the first simulation result) are finally solved as follows:

[0103] ;

[0104] in, is the current value A / m 2 , i is a plural unit, is the conductivity of the copper coil s / m, is the conductivity of other devices s / m, is the angular frequency rad / s, J 0 is the current density amplitude (maximum value) A / m 2 , is the magnetic potential vector.

[0105] According to the current value of the calculation grid on each component, the induced heat density (i.e., power density) on each component can be calculated. The induced heat is defined as the average value within an electromagnetic induction heating cycle, that is, the heat accumulated within an electromagnetic induction heating cycle, divided by the duration of the electromagnetic induction heating cycle, and the average heat per unit time. That is, the induced heat density represents the total heat sensed by each calculation grid (i.e., per unit area) of the component within an electromagnetic induction heating cycle divided by the cycle duration (the average heat sensed per time unit within the cycle).

[0106] Step S3012: According to the current value and the conductivity of each component, the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component are calculated according to the following formula:

[0107] ;

[0108] in, P is the induced heat density (induced power density) W / m 3 ; is the induced current density vector value A / m 2 ; is the induced electromotive force V / m; is the induced current density A / m 2 ; is the conjugate complex number of the induced current density; is the conductivity of the copper coil s / m (the conductivity of the copper coil is 59980000s / m); is the conductivity of other components s / m; A re is the imaginary part of the magnetic potential vector, A im is the real part of the magnetic potential vector; is the angular frequency, is the current density amplitude. The other components are: graphite crucible, side graphite support, carbon felt, silicon melt, and their respective conductivity They are: 70000s / m, 70000s / m, 450s / m, 1020000s / m respectively.

[0109] Step S302 : determining a calculation grid in each component where the power density of the calculation grid is 1 / e of the power density of the outer surface of the component as a target calculation grid.

[0110] Step S303, determining a target thermal boundary layer grid thickness of the region to which the component belongs according to the distance from the target calculation grid to the outer surface of the component.

[0111] For each component in the system, the power density gradually decreases from the outer surface of the component to the inside. In this embodiment, the thickness of the thermal boundary layer is the thickness at which the power density in the component (crucible, side graphite support, carbon felt or copper coil) drops to 1 / e (e is the natural base) of the power density on the outer surface of the component. Specifically, the target thermal boundary layer grid thickness includes: the calculated grid thickness of the thermal boundary layer belonging to the crucible surface, the calculated grid thickness of the thermal boundary layer belonging to the side graphite support surface, the calculated grid thickness of the thermal boundary layer belonging to the carbon felt surface, and the calculated grid thickness of the thermal boundary layer belonging to the copper coil surface. It should be noted that the thickness of the entire thermal boundary layer is not exactly the same, there is a certain fluctuation, the actual thermal boundary layer thickness is not uniform, and the thermal boundary layer thickness at different positions is inconsistent. In order to make the grid division simpler and the calculation result more accurate, the embodiment of the present application can calculate the thermal boundary layer thickness corresponding to multiple points, and then calculate the average thickness as the final thermal boundary layer thickness. Furthermore, in order to make the final calculation results more accurate, 5 layers of grids are divided on the basis of the average thickness (that is, the obtained thermal boundary layer thickness consists of 5 layers of grids, and the growth rate between every two adjacent grids is 1.2) as the target thermal boundary layer grid thickness. Compared with the technical solution of taking 3 layers of grids for a given boundary layer thickness in the prior art, the accuracy of the calculation results can be further improved.

[0112] Step S304, based on the target thermal boundary layer grid thickness and electromagnetic induction heating frequency of the area to which each component belongs corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the thermal boundary layer thickness fitting curve, and the thermal boundary layer thickness fitting curve represents a relationship curve between the thermal boundary layer grid thickness and the electromagnetic induction heating frequency of the area to which each component belongs.

[0113] Specifically, for each set of simulation results, a set of electromagnetic induction heating frequencies and corresponding target thermal boundary layer grid thicknesses can be obtained, and then multiple sets of electromagnetic induction heating frequencies and corresponding target thermal boundary layer grid thicknesses can be obtained based on multiple sets of simulation results, and then a fourth-order polynomial fitting is used to obtain the relationship curve between the two. Figure 4 , Figure 4 A schematic diagram of a thermal boundary layer thickness fitting curve is shown, such as Figure 4 As shown, the horizontal axis is the electromagnetic induction heating frequency, and the vertical axis is the most suitable thermal boundary layer grid thickness (i.e., the target thermal boundary layer grid thickness). Corresponding thermal boundary layer thickness fitting curves are generated for different components, so the thermal boundary layer thickness fitting curve finally obtained in this embodiment includes multiple curves, namely: crucible thermal boundary layer thickness fitting curve, side graphite support thermal boundary layer thickness fitting curve, carbon felt thermal boundary layer thickness fitting curve, and copper coil thermal boundary layer thickness fitting curve. For example, Figure 4 As shown, the solid line curve represents the crucible thermal boundary layer thickness fitting curve, and the dotted line curve represents the side graphite support thermal boundary layer thickness fitting curve. When the electromagnetic induction heating frequency is 2kHz, the corresponding crucible thermal boundary layer's most suitable target thermal boundary layer grid thickness is 32.3mm, and the corresponding side graphite support thermal boundary layer's most suitable target thermal boundary layer grid thickness is 35.8mm.

[0114] In this embodiment, the Lorentz force boundary layer thickness fitting curve is the relationship curve between the average thickness of the calculation grid in the Lorentz force concentrated distribution area in the silicon melt (near the edge of the silicon melt near the crucible) and the electromagnetic induction heating frequency. Because the rest of the components in the system are solid, although the Lorentz force will be generated, it will not affect it.

[0115] In a possible implementation, calculating a target thickness fitting curve according to the multiple sets of simulation results includes:

[0116] Step S401 : determining the current value of each calculation grid of the silicon melt during the electromagnetic induction heating process according to each group of simulation results.

[0117] Step S402: Calculate the Lorentz force of each calculation grid on one side of the silicon melt in the interface region between the silicon melt and the crucible according to the current value.

[0118] In this embodiment, it is necessary to couple the electromagnetic induction heating process with the heat transfer flow process, and consider the influence of the Lorentz force in the melt when solving the heat transfer flow process (i.e., the heat transfer flow control equation in the heat transfer flow numerical model), and consider the influence of the Malegonez force (surface tension) on the free surface of the silicon melt. Specifically, in the electromagnetic induction control equation, the Lorentz force control equation of the silicon melt region is set to solve the corresponding Lorentz force parameter value (i.e., the Lorentz force density). The electromagnetic induction control equation includes: the Lorentz force control equation of the silicon melt region; according to the Lorentz force control equation of the silicon melt region as described below, the Lorentz force parameter value of each calculation grid (on the silicon melt side of the interface between the silicon melt and the crucible) is calculated according to the current value:

[0119] ;

[0120] in, is the Lorentz force density N / m 3 , the Lorentz force density is defined as the time-averaged value (independent of time, equivalent to the average value in an electromagnetic induction heating cycle, that is, the average Lorentz force per time unit obtained by dividing the accumulated Lorentz force in a cycle by the cycle length); is the magnetic induction intensity T; is the real part of magnetic induction intensity T; is the imaginary part of magnetic induction intensity T; is the induced current density A / m 2 , is the real part of the current density A / m 2 , is the imaginary part of the current density A / m 2 Because the Lorentz force in each place (each computational grid) in the silicon melt is different, the Lorentz force is loaded on each computational grid during loading, which is equivalent to loading the Lorentz force density. It is meaningless to perform volume integral on it, so only the Lorentz force density is described.

[0121] Step S403 , determining the target Lorentz force boundary layer grid thickness corresponding to the simulation result according to the magnitude of the Lorentz force of each calculation grid of the silicon melt.

[0122] The computational grids at which the Lorentz force (i.e., the Lorentz force density) of each computational grid of the silicon melt is 1 / e of the Lorentz force of the outer surface of the silicon melt are determined as candidate computational grids. The target Lorentz force boundary layer grid thickness of the silicon melt is determined according to the distance from the candidate computational grid to the outer surface of the silicon melt.

[0123] For silicon melt, the Lorentz force gradually decreases from the outer surface of the silicon melt to the inside. In this embodiment, the thickness of the Lorentz force boundary layer is taken as the thickness at which the Lorentz force density in the silicon melt drops to 1 / e (e is the natural base) of the Lorentz force density on the outer surface of the silicon melt. It should be known that the thickness of the entire Lorentz force boundary layer is not exactly the same, there is a certain fluctuation, the actual Lorentz force boundary layer thickness is not uniform, and the Lorentz force boundary layer thickness at different positions is inconsistent. In order to make the grid division simpler and the calculation results more accurate, the embodiment of the present application can calculate the Lorentz force boundary layer thickness corresponding to multiple points (that is, the shortest distance from the candidate calculation grid to the outer surface of the silicon melt), and then calculate the average thickness as the final Lorentz force boundary layer thickness. Furthermore, in order to make the final calculation results more accurate, 5 layers of grids are divided on the basis of the average thickness (that is, the obtained Lorentz force boundary layer thickness consists of 5 layers of grids, and the growth rate between every two adjacent grids is 1.2) as the target Lorentz force boundary layer grid thickness. Compared with the technical solution of taking 3 layers of grids for a given boundary layer thickness in the prior art, the accuracy of the calculation results can be further improved.

[0124] Step S404, according to the target Lorentz force boundary layer grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the Lorentz force boundary layer thickness fitting curve, and the Lorentz force boundary layer thickness fitting curve represents the relationship curve between the target Lorentz force boundary layer grid thickness and the electromagnetic induction heating frequency.

[0125] Specifically, for each set of simulation results, a set of electromagnetic induction heating frequencies and corresponding target Lorentz force boundary layer grid thicknesses can be obtained, and then based on multiple sets of simulation results, multiple sets of electromagnetic induction heating frequencies and corresponding target Lorentz force boundary layer grid thicknesses can be obtained, and then a fourth-order polynomial fitting can be used to obtain the relationship curve between the two. Figure 5 , Figure 5 A schematic diagram of a Lorentz force boundary layer thickness fitting curve is shown, as Figure 5 As shown in FIG. 1 , the horizontal axis is the electromagnetic induction heating frequency, and the vertical axis is the most suitable Lorentz force boundary layer grid thickness (i.e., the target Lorentz force boundary layer grid thickness). For different components, corresponding Lorentz force boundary layer thickness fitting curves are generated respectively. For example, Figure 5 As shown, when the electromagnetic induction heating frequency is 2 kHz, the corresponding target Lorentz force boundary layer grid thickness is 10.3 mm.

[0126] Step S105, according to the current electromagnetic induction heating frequency, determining the mesh thickness of the multi-physical field coupling mesh of each region in the liquid phase method silicon carbide growth system from the thickness fitting curve, performing mesh division, and obtaining a target numerical calculation model.

[0127] Specifically, in practical applications, the currently required computational grid thickness is determined from each thickness fitting curve according to the electromagnetic induction heating frequency (i.e., the current electromagnetic induction heating frequency) actually required for growing silicon carbide by the liquid phase method. Specifically, the multi-physics field coupling grids in each region include: concentration boundary layer grid thickness, heat boundary layer grid thickness, and Lorentz force boundary layer grid thickness. Thus, the constructed geometric model is meshed based on the grid thickness to perform the subsequent numerical simulation process. As a result, in the actual application process, accurate computational grid division can be achieved, and unnecessary waste of computing resources can be reduced while ensuring the accuracy of the results of numerical calculations.

[0128] In the embodiments of the present application, the relationship curve is mainly fitted by using the electromagnetic induction heating frequency and the target boundary layer grid thickness in multiple sets of simulation results. In this process, the more data used in the fitting curve, the more accurate the fitted relationship curve is, and correspondingly, the greater the amount of calculation required to be performed. Therefore, the present application evaluates the accuracy of the fitted curve, and if the accuracy does not meet the requirements, obtains a new set of simulation results (including the electromagnetic induction heating frequency and the target boundary layer grid thickness), combines the existing data, and re-fits to obtain a more accurate relationship curve, until the accuracy of the obtained relationship curve meets the requirements. The specific contents are as follows:

[0129] In a possible implementation, calculating a target thickness fitting curve according to the multiple sets of simulation results includes:

[0130] Step S501, determining a target grid thickness in each group of simulation results according to the multiple groups of simulation results, wherein the target grid thickness includes: a target concentration boundary layer grid thickness, a target heat boundary layer grid thickness, and a target Lorentz force boundary layer grid thickness.

[0131] Specifically, for the multiple groups of simulation results obtained, step S501 includes: executing steps S201-S204, generating a target concentration boundary layer grid thickness according to the multiple groups of simulation results (assuming the number is k); executing steps S301-S303, generating a target heat boundary layer grid thickness according to the multiple groups of simulation results (assuming the number is k); executing steps S401-S403, generating a target Lorentz force boundary layer grid thickness according to the multiple groups of simulation results (assuming the number is k).

[0132] Step S502 : generating an initial thickness fitting curve by using a fourth-order polynomial fitting method according to the target grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results.

[0133] The initial thickness fitting curve includes: an initial concentration boundary layer thickness fitting curve, an initial heat boundary layer thickness fitting curve and an initial Lorentz force boundary layer thickness fitting curve. Specifically, step S502 includes: executing step S205, according to the target concentration boundary layer grid thickness (obtained in step S501) and the electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, using a fourth-order polynomial fitting method to generate an initial concentration boundary layer thickness fitting curve; according to step S304, according to the target heat boundary layer grid thickness (obtained in step S501) and the electromagnetic induction heating frequency corresponding to the multiple groups of simulation results of the regions to which each component belongs, using a fourth-order polynomial fitting method to generate an initial heat boundary layer thickness fitting curve; according to step S404, according to the target Lorentz force boundary layer grid thickness (obtained in step S501) and the electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, using a fourth-order polynomial fitting method to generate an initial Lorentz force boundary layer thickness fitting curve.

[0134] Step S503, calculating the fitting degree of the initial thickness fitting curve.

[0135] The fitting degree of each initial concentration boundary layer thickness fitting curve, initial heat boundary layer thickness fitting curve and initial Lorentz force boundary layer thickness fitting curve obtained in step S502 is calculated respectively. The fitting degree represents the accuracy of the corresponding curve.

[0136] In a possible implementation manner, calculating the fitting degree of the initial thickness fitting curve includes:

[0137] Step S5031, obtaining the target grid thickness and electromagnetic induction heating frequency corresponding to a new set of simulation results as the actual grid thickness and the actual electromagnetic induction heating frequency.

[0138] Specifically, after generating an initial thickness fitting curve based on k groups of simulation results, a new group of simulation results, i.e., the k+1th group of simulation results, is obtained, and the target grid thickness and electromagnetic induction heating frequency in the k+1th group of simulation results are determined as the actual grid thickness and the actual electromagnetic induction heating frequency.

[0139] Step S5032: determining the predicted grid thickness corresponding to the actual electromagnetic induction heating frequency from the initial thickness fitting curve.

[0140] For example, if the actual electromagnetic induction heating frequency of the k+1th group of simulation results is 2 kHz, the predicted grid thickness corresponding to 2 kHz in the curve is determined from the initial thickness fitting curve (including: concentration boundary layer grid thickness, heat boundary layer grid thickness and Lorentz force boundary layer grid thickness).

[0141] Step S5033, calculating the residual sum of squares of the initial thickness fitting curve according to the predicted grid thickness and the actual grid thickness.

[0142] Specifically, the residual sum of squares is calculated as follows:

[0143] ;

[0144] in, is the residual sum of squares, y 实际 is the actual mesh thickness, y 预测 is the predicted grid thickness. The smaller the residual sum of squares is, the closer the fitted curve is to the actual value.

[0145] Step S5034, calculating the fitting degree of the initial thickness fitting curve according to the residual sum of squares according to the following formula:

[0146] ;

[0147] Among them, R 2 is the degree of fitting, is the residual sum of squares, y 实际 is the actual mesh thickness. 2 The value ranges from 0 to 1, and the closer to 1, the better the fit.

[0148] Step S504, when the degree of fitting is less than a preset threshold, obtain the target grid thickness and electromagnetic induction heating frequency corresponding to a new set of simulation results, regenerate the initial thickness fitting curve, until the degree of fitting of the initial thickness fitting curve is greater than or equal to the preset threshold, and determine the initial thickness fitting curve as the target thickness fitting curve.

[0149] Specifically, when the degree of fitting is less than the preset threshold, it means that the accuracy of the initial thickness fitting curve currently obtained according to the k group of simulation results is low and needs to be refitted. Then the k+1 group of simulation results (target grid thickness and electromagnetic induction heating frequency) is obtained, and the fourth-order polynomial fitting method is re-adopted to fit the data in the k+1 group of simulation results to obtain a new initial thickness fitting curve, and the degree of fitting is calculated again according to steps S501-S504 until the degree of fitting is greater than or equal to the preset threshold (for example, 0.999), and the qualified initial thickness fitting curve is determined as the target thickness fitting curve, and the target thickness fitting curve includes: concentration boundary layer thickness fitting curve, heat boundary layer thickness fitting curve and Lorentz force boundary layer thickness fitting curve.

[0150] Embodiment 1,

[0151] The first step is to mesh the silicon melt area. Figure 3 It can be seen that at 4kHz, the thickness of the solid-liquid interface concentration boundary layer D1 is 0.64mm. Five layers of grids need to be divided at a distance of 0.64mm on the silicon melt side of the solid-liquid interface. The thickness of the first layer of grids is x mm, the thickness of the second layer of grids is 1.2x mm, and the thickness of the third layer of grids is 1.2 2 x mm, the fourth layer mesh thickness is 1.2 3 x mm, the fifth layer mesh thickness is 1.2 4 x mm, so the first layer of grid thickness x=0.086mm is obtained. The reason why the boundary layer grid is set to gradually increase in thickness (growth rate of 1.2) is to avoid the boundary layer and other grid sizes being too different, resulting in a sudden change in a certain physical quantity. The first layer of grid thickness of the solid-liquid interface concentration boundary layer is 0.086mm, and the width of the five layers of grids at the boundary layer is 0.8mm (in the actual calculation of the liquid phase growth silicon carbide model, the grid and the aspect ratio must be greater than 0.1 and less than 10 to solve each process more accurately).

[0152] The second step is based on Figure 5 It can be seen that at 4kHz, the thickness of the Lorentz force boundary layer D3 of the silicon melt is 7.3mm (near the interface between the silicon melt and the graphite crucible), the thickness of the first layer of grid is x mm, the thickness of the second layer of grid is 1.2x mm, and the thickness of the third layer of grid is 1.2 2 x mm, the fourth layer mesh thickness is 1.2 3 x mm, the fifth layer mesh thickness is 1.2 4x mm, so the first layer grid thickness x=0.98mm, the width of the five layers of grids at the boundary layer is 2mm. The reason why the boundary layer grid is set to gradually increase in thickness (growth rate is 1.2) is to avoid a large difference between the boundary layer and other grid sizes, forming a sudden change in the value of a certain physical quantity.

[0153] In the third step, after the boundary layer grid size is determined, the grid size of the silicon melt area is set to 2mm according to the concentration boundary layer grid width of the solid-liquid interface is 0.8mm and the Lorentz force boundary layer grid width is 2mm, and it is an automatically divided unstructured grid.

[0154] Meshing of the side graphite support area: According to Figure 4 It can be seen that the side graphite support thermal boundary layer thickness D2 is 24.1mm at 4kHz. If the side graphite support thickness is only 22mm (less than the boundary layer thickness) in this embodiment, the thermal boundary layer can be ignored in this area, and five layers of structured grids can be automatically divided in the thickness direction. The grids in the height direction are all adaptive, and the aspect ratio of each grid is limited to be greater than 0.25 and less than 4.

[0155] Meshing of graphite crucible area: According to Figure 4 It can be seen that the graphite crucible thermal boundary layer thickness D2 is 19.0 mm at 4 kHz. The first layer grid thickness is x mm, the second layer grid thickness is 1.2x mm, and the third layer grid thickness is 1.2 2 xmm, the fourth layer grid thickness is 1.2 3 x mm, the fifth layer mesh thickness is 1.2 4 x mm, so the first layer grid thickness x=2.55mm, the five layers of grid width at the boundary layer are all 10mm. Assuming that the graphite crucible thickness is 26mm in this embodiment, the remaining area is automatically divided into calculation grids except the thermal boundary layer, and shares nodes with the network at the interface of silicon melt.

[0156] Grid division of carbon felt area: Assume that the thickness D2 of the carbon felt thermal boundary layer is 355.9 mm at 4 kHz. However, in this example or other systems where silicon carbide is grown by the liquid phase method, the thickness of the carbon felt cannot exceed 150 mm. Therefore, in order to ensure the accuracy of the solution of the heat transfer process, five layers of adaptive structured grids are automatically divided in the thickness direction of the carbon felt. Among them, the grids share nodes with the interface grids of the side graphite supports. The grids in the height direction of the carbon felt are all adaptive, and the aspect ratio of each grid is limited to be greater than 0.25 and less than 4.

[0157] Mesh division of the copper coil area: Assume that the thermal boundary layer thickness D2 of the copper coil is 1.02 mm at 4 kHz, and the thickness of the copper coil is 4 mm. Take the first layer of mesh thickness as x mm, the second layer of mesh thickness as 1.2x mm, and the third layer of mesh thickness as 1.22 x mm, the fourth layer mesh thickness is 1.2 3 x mm, the fifth layer mesh thickness is 1.2 4 x mm, so the first layer grid thickness x = 0.14 mm, the growth rate is 1.2, and five boundary layers are added (the computational grid in the rest of the copper coil area is adaptively divided).

[0158] The second aspect of the embodiment of the present application further provides a multi-field coupling meshing device for growing silicon carbide by a liquid phase method, which is used to execute the multi-field coupling meshing method for growing silicon carbide by a liquid phase method described in the first aspect, referring to Figure 6 , Figure 6 A schematic diagram of the structure of a multi-field coupling grid division method is shown in FIG. Figure 6 As shown, the device comprises:

[0159] A model building module is used to build a geometric model based on the structural information of the liquid phase growth silicon carbide system;

[0160] A numerical calculation module, used to divide the geometric model into calculation grids according to the maximum total amount of calculation grids, set boundary conditions of the corresponding calculation grids in the geometric model, set electromagnetic induction control equations, heat transfer flow control equations and component transport control equations, and obtain a numerical calculation model;

[0161] A simulation module, for simulating the electromagnetic induction heating process, heat transfer flow process and component transport process under multiple different electromagnetic induction heating frequencies based on the numerical calculation model, and obtaining multiple groups of simulation results;

[0162] A fitting module, used for calculating a target thickness fitting curve according to the plurality of simulation results, wherein the target thickness fitting curve includes: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve;

[0163] A grid division module is used to determine the grid thickness of the multi-physical field coupling grid in each area of ​​the liquid phase method silicon carbide growth system from the thickness fitting curve according to the current electromagnetic induction heating frequency, perform grid division, and obtain a target numerical calculation model.

[0164] In a possible implementation, the fitting module includes a fitting degree calculation submodule, which is used to perform the following steps:

[0165] Determine a target grid thickness in each group of simulation results according to the multiple groups of simulation results, wherein the target grid thickness includes: a target concentration boundary layer grid thickness, a target heat boundary layer grid thickness, and a target Lorentz force boundary layer grid thickness;

[0166] Generate an initial thickness fitting curve using a fourth-order polynomial fitting method according to the target grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results;

[0167] Calculating the fitting degree of the initial thickness fitting curve;

[0168] When the degree of fitting is less than a preset threshold, the target grid thickness and electromagnetic induction heating frequency corresponding to a new set of simulation results are obtained, and the initial thickness fitting curve is regenerated until the degree of fitting of the initial thickness fitting curve is greater than or equal to the preset threshold, and the initial thickness fitting curve is determined as the target thickness fitting curve.

[0169] In a possible implementation, the fitting module includes a concentration boundary layer thickness fitting curve fitting submodule, which is used to perform the following steps:

[0170] According to each group of simulation results, determining the carbon concentration value of each calculation grid in the silicon melt region;

[0171] According to the carbon concentration value, starting from the calculation grid on the side of the silicon melt at the solid-liquid interface, the supersaturation of each layer of the calculation grid is calculated;

[0172] The distance between the computational grid at the maximum value of supersaturation and the solid-liquid interface is determined as the target distance;

[0173] Determining the target concentration boundary layer grid thickness corresponding to the simulation result according to the target distance;

[0174] According to the target concentration boundary layer grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the concentration boundary layer thickness fitting curve, and the concentration boundary layer thickness fitting curve represents the relationship curve between the target concentration boundary layer grid thickness and the electromagnetic induction heating frequency.

[0175] In a possible implementation manner, the calculating the supersaturation of each layer of the computing grid includes:

[0176] The supersaturation of each layer of the computational grid is calculated according to the following formula:

[0177] ;

[0178] in, S is the supersaturation, c Calculate the carbon concentration value at the grid for this layer, c eq Calculate the carbon equilibrium concentration at the grid for this layer.

[0179] In a possible implementation, the fitting module includes a thermal boundary layer thickness fitting curve fitting submodule, which is used to perform the following steps:

[0180] According to each group of simulation results, the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component are determined; the power density represents the value of the total heat induced by the corresponding calculation grid in an electromagnetic induction heating cycle divided by the cycle length; the components include: crucible, side graphite support, carbon felt and copper coil;

[0181] Determine the computational grid in each component where the power density of the computational grid is 1 / e of the power density of the outer surface of the component as the target computational grid;

[0182] Determine the target thermal boundary layer grid thickness of the region to which the component belongs according to the distance from the target calculation grid to the outer surface of the component;

[0183] According to the target thermal boundary layer grid thickness and electromagnetic induction heating frequency of the area to which each component belongs corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the thermal boundary layer thickness fitting curve, and the thermal boundary layer thickness fitting curve represents the relationship curve between the target thermal boundary layer grid thickness and the electromagnetic induction heating frequency of the area to which each component belongs.

[0184] In a possible implementation, determining the power density of the outer surface of each component in the geometric model and the power density of each computational grid in each component according to each group of simulation results includes:

[0185] According to the simulation results, determining the current value of the outer surface of each component during the electromagnetic induction heating process, and the current value of each calculation grid in each component;

[0186] According to the current value and the conductivity of each component, the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component are calculated according to the following formula:

[0187] ;

[0188] in, P is the induced heat density (induced power density) W / m 3 ; is the induced current density vector A / m 2 ; is the induced electromotive force V / m; is the induced current density A / m 2 ; is the conjugate complex number of the induced current density; is the conductivity of the copper coil s / m; is the conductivity of other components s / m; A re is the imaginary part of the magnetic potential vector, A im is the real part of the magnetic potential vector; is the angular frequency, is the current density amplitude.

[0189] In a possible implementation, the fitting module includes a Lorentz force boundary layer thickness fitting curve fitting submodule, which is used to perform the following steps:

[0190] According to each group of simulation results, determining the current value of each calculation grid of the silicon melt during the electromagnetic induction heating process;

[0191] Calculating the Lorentz force of each calculation grid on one side of the silicon melt at the interface between the silicon melt and the crucible according to the current value;

[0192] Determining the target Lorentz force boundary layer grid thickness corresponding to the simulation result according to the magnitude of the Lorentz force of each calculation grid of the silicon melt;

[0193] According to the target Lorentz force boundary layer grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the Lorentz force boundary layer thickness fitting curve, and the Lorentz force boundary layer thickness fitting curve represents the relationship curve between the target Lorentz force boundary layer grid thickness and the electromagnetic induction heating frequency.

[0194] In a possible implementation manner, calculating the fitting degree of the initial thickness fitting curve includes:

[0195] Obtaining the target grid thickness and electromagnetic induction heating frequency corresponding to a new set of simulation results as actual grid thickness and actual electromagnetic induction heating frequency;

[0196] Determining a predicted grid thickness corresponding to the actual electromagnetic induction heating frequency from the initial thickness fitting curve;

[0197] Calculating the residual sum of squares of the initial thickness fitting curve according to the predicted grid thickness and the actual grid thickness;

[0198] According to the residual sum of squares, the fitting degree of the initial thickness fitting curve is calculated according to the following formula:

[0199] ;

[0200] Among them, R2 is the degree of fitting, is the residual sum of squares, y 实际 is the actual mesh thickness.

[0201] The present application also provides an electronic device, referring to Figure 7 , Figure 7 Schematic diagram of an electronic device proposed in an embodiment of the present application. Figure 7 As shown, the electronic device 100 includes: a memory 110 and a processor 120. The memory 110 and the processor 120 are connected via a bus communication. A computer program is stored in the memory 110. The computer program can be run on the processor 120 to implement the steps in the multi-field coupled meshing method for growing silicon carbide by a liquid phase method disclosed in the embodiment of the present application.

[0202] The embodiment of the present application also provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the steps in the multi-field coupled meshing method for growing silicon carbide by a liquid phase method as disclosed in the embodiment of the present application.

[0203] The embodiment of the present application also provides a computer program product. When the computer program product is run on an electronic device, the processor is enabled to implement the steps of the multi-field coupling grid division method for liquid phase growth of silicon carbide as disclosed in the embodiment of the present application.

[0204] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0205] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, devices, electronic devices, and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0206] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0207] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0208] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present application.

[0209] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0210] The above is a detailed introduction to the multi-field coupled grid division method, device and product for liquid phase growth of silicon carbide provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A multi-field coupling grid generation method for liquid phase growth of silicon carbide, characterized in that: The method comprises: According to the structural information of the liquid phase growth silicon carbide system, a geometric model is established; According to the maximum total amount of computational grids, the geometric model is divided into computational grids, boundary conditions of the corresponding computational grids in the geometric model are set, and electromagnetic induction control equations, heat transfer flow control equations and component transport control equations are set to obtain a numerical calculation model; Based on the numerical calculation model, the electromagnetic induction heating process, heat transfer flow process and component transport process under multiple different electromagnetic induction heating frequencies are simulated to obtain multiple groups of simulation results; According to the plurality of sets of simulation results, a target thickness fitting curve is calculated, wherein the target thickness fitting curve includes: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve; According to the current electromagnetic induction heating frequency, the grid thickness of the multi-physical field coupling grid of each region in the liquid phase method silicon carbide growth system is determined from the thickness fitting curve, and grid division is performed to obtain a target numerical calculation model.

2. The multi-field coupling grid generation method for liquid phase growth of silicon carbide according to claim 1, characterized in that: The step of calculating a target thickness fitting curve according to the plurality of sets of simulation results comprises: Determine a target grid thickness in each group of simulation results according to the multiple groups of simulation results, wherein the target grid thickness includes: a target concentration boundary layer grid thickness, a target heat boundary layer grid thickness, and a target Lorentz force boundary layer grid thickness; Generate an initial thickness fitting curve using a fourth-order polynomial fitting method according to the target grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results; Calculating the fitting degree of the initial thickness fitting curve; When the degree of fitting is less than a preset threshold, the target grid thickness and electromagnetic induction heating frequency corresponding to a new set of simulation results are obtained, and the initial thickness fitting curve is regenerated until the degree of fitting of the initial thickness fitting curve is greater than or equal to the preset threshold, and the initial thickness fitting curve is determined as the target thickness fitting curve.

3. The multi-field coupling grid generation method for liquid phase growth of silicon carbide according to claim 1, characterized in that: The step of calculating a target thickness fitting curve according to the plurality of sets of simulation results comprises: According to each group of simulation results, determining the carbon concentration value of each calculation grid in the silicon melt region; According to the carbon concentration value, starting from the calculation grid on the side of the silicon melt at the solid-liquid interface, the supersaturation of each layer of the calculation grid is calculated; The distance between the computational grid at the maximum value of supersaturation and the solid-liquid interface is determined as the target distance; Determining the target concentration boundary layer grid thickness corresponding to the simulation result according to the target distance; According to the target concentration boundary layer grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the concentration boundary layer thickness fitting curve, and the concentration boundary layer thickness fitting curve represents the relationship curve between the target concentration boundary layer grid thickness and the electromagnetic induction heating frequency.

4. The multi-field coupling grid generation method for liquid phase growth of silicon carbide according to claim 3, characterized in that: The calculating of the oversaturation of each layer of the computing grid includes: The supersaturation of each layer of the computational grid is calculated according to the following formula: ； in, S is the supersaturation, c Calculate the carbon concentration value at the grid for this layer, c eq Calculate the carbon equilibrium concentration at the grid for this layer.

5. The multi-field coupling grid generation method for liquid phase growth of silicon carbide according to claim 1, characterized in that: According to the multiple sets of simulation results, a target thickness fitting curve is calculated, including: According to each group of simulation results, the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component are determined; the power density represents the value of the total heat induced by the corresponding calculation grid in an electromagnetic induction heating cycle divided by the cycle length; the components include: crucible, side graphite support, carbon felt and copper coil; Determine the computational grid in each component where the power density of the computational grid is 1 / e of the power density of the outer surface of the component as the target computational grid; Determine the target thermal boundary layer grid thickness of the region to which the component belongs according to the distance from the target calculation grid to the outer surface of the component; According to the target thermal boundary layer grid thickness and electromagnetic induction heating frequency of the area to which each component belongs corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the thermal boundary layer thickness fitting curve, and the thermal boundary layer thickness fitting curve represents the relationship curve between the target thermal boundary layer grid thickness and the electromagnetic induction heating frequency of the area to which each component belongs.

6. The multi-field coupling grid generation method for liquid phase growth of silicon carbide according to claim 5, characterized in that: Determining the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component according to each group of simulation results includes: According to the simulation results, determining the current value of the outer surface of each component during the electromagnetic induction heating process, and the current value of each calculation grid in each component; According to the current value and the conductivity of each component, the power density of the outer surface of each component in the geometric model and the power density of each calculation grid in each component are calculated according to the following formula: ； in, P is the induced heat density (induced power density) W / m 3 ; is the induced current density vector A / m 2 ; is the induced electromotive force V / m; is the induced current density A / m 2 ; is the conjugate complex number of the induced current density; is the conductivity of the copper coil s / m; is the conductivity of other components s / m; A re is the imaginary part of the magnetic potential vector, A im is the real part of the magnetic potential vector; is the angular frequency, is the current density amplitude.

7. The multi-field coupling grid generation method for liquid phase growth of silicon carbide according to claim 1, characterized in that: The step of calculating a target thickness fitting curve according to the plurality of sets of simulation results comprises: According to each group of simulation results, determining the current value of each calculation grid of the silicon melt during the electromagnetic induction heating process; Calculating the Lorentz force of each calculation grid on one side of the silicon melt at the interface between the silicon melt and the crucible according to the current value; Determining the target Lorentz force boundary layer grid thickness corresponding to the simulation result according to the magnitude of the Lorentz force of each calculation grid of the silicon melt; According to the target Lorentz force boundary layer grid thickness and electromagnetic induction heating frequency corresponding to the multiple groups of simulation results, a fourth-order polynomial fitting method is used to generate the Lorentz force boundary layer thickness fitting curve, and the Lorentz force boundary layer thickness fitting curve represents the relationship curve between the target Lorentz force boundary layer grid thickness and the electromagnetic induction heating frequency.

8. The multi-field coupling grid generation method for liquid phase growth of silicon carbide according to claim 2, characterized in that: The calculating the fitting degree of the initial thickness fitting curve comprises: Obtaining the target grid thickness and electromagnetic induction heating frequency corresponding to a new set of simulation results as actual grid thickness and actual electromagnetic induction heating frequency; Determining a predicted grid thickness corresponding to the actual electromagnetic induction heating frequency from the initial thickness fitting curve; Calculating the residual sum of squares of the initial thickness fitting curve according to the predicted grid thickness and the actual grid thickness; According to the residual sum of squares, the fitting degree of the initial thickness fitting curve is calculated according to the following formula: ； Among them, R 2 is the degree of fitting, is the residual sum of squares, y 实际 is the actual mesh thickness.

9. A multi-field coupling grid division device for growing silicon carbide by liquid phase method, characterized in that: Applied to executing the multi-field coupling grid division method according to any one of claims 1 to 8, the device comprises: A model building module is used to build a geometric model based on the structural information of the liquid phase growth silicon carbide system; A numerical calculation module, used to divide the geometric model into calculation grids according to the maximum total amount of calculation grids, set boundary conditions of the corresponding calculation grids in the geometric model, set electromagnetic induction control equations, heat transfer flow control equations and component transport control equations, and obtain a numerical calculation model; A simulation module, for simulating the electromagnetic induction heating process, heat transfer flow process and component transport process under multiple different electromagnetic induction heating frequencies based on the numerical calculation model, and obtaining multiple groups of simulation results; A fitting module, used for calculating a target thickness fitting curve according to the plurality of simulation results, wherein the target thickness fitting curve includes: a concentration boundary layer thickness fitting curve, a heat boundary layer thickness fitting curve and a Lorentz force boundary layer thickness fitting curve; The grid division module is used to determine the grid thickness of the multi-physical field coupling grid in each area of ​​the liquid phase method silicon carbide growth system from the thickness fitting curve according to the current electromagnetic induction heating frequency, perform grid division, and obtain the target numerical calculation model.

10. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the multi-field coupling grid division method for growing silicon carbide by a liquid phase method according to any one of claims 1 to 8.

Images