A method and system for simulation of micro porosity in castings and process window optimization

Abstract

The application discloses a cast micro-porosity simulation and process window optimization method and system, which voxelizes the cast and shell geometry, solves the solidification heat transfer process by using the apparent specific heat method, and records the liquidus and solidus crossing time, temperature gradient, cooling rate and local solidification time; calculates the Niyama criterion, dimensionless Niyama criterion and shrinkage shrinkage porosity volume fraction, and combines the dissolved gas supersaturation relationship to distinguish shrinkage, gas hole and mixed defects; according to the user's point selection position or the key position automatically selected by the risk field, the microcosmic simulation of the cellular automaton-finite volume method coupling is executed; through Latin hypercube sampling, Gaussian process regression and worst case optimization, the anti-vibration process window that still meets the safety threshold within the process parameter tolerance is obtained. The system adopts the architecture of decoupling the interface layer and the calculation engine layer, and can output risk positioning, cause explanation, parameter sensitivity and process adjustment suggestions.

Description

Technical Field

[0001] This invention belongs to the field of numerical simulation and process optimization technology of casting process, specifically involving a method and system for simulating micro porosity of castings and optimizing process windows. Background Technology

[0002] In high-temperature alloy precision casting, microporousness and subcutaneous porosity reduce the density and service reliability of castings. For castings with significant thickness differences, such as transition sections, guide vanes, and casings, or those with ring-spoke intersecting structures, defects often appear in localized hot spots and areas where end-feeding channels are easily blocked. These defects are small in size and scattered in distribution, and usually require comprehensive judgment based on fluorescence penetrant testing, radiographic testing, or metallographic results. Once out-of-tolerance defects are detected, rework is often necessary, and in severe cases, scrapping may result.

[0003] The existing technology has at least the following shortcomings: Firstly, existing commercial casting simulation tools can calculate temperature fields, solid fraction, and Niyama criteria, and can compare parameters through multi-condition calculations. However, conventional applications still mainly focus on positive verification of a given nominal process, and usually do not unify production tolerances, worst-case safety margins, and parameter sensitivity into a directly solvable process window problem. Therefore, good results from a single nominal condition calculation do not necessarily indicate that the process remains safe when there are normal fluctuations in pouring temperature, shell preheating, shell thickness, or local cooling.

[0004] Secondly, the Niyama criterion primarily reflects the temperature gradient, cooling rate, and difficulty of feeding at the end of solidification, making it suitable for evaluating the risk of shrinkage porosity. However, it cannot independently identify porosity caused by dissolved gas precipitation, gas evolution at the mold-shell interface, or the superposition of both with insufficient shrinkage feeding. If process measures are given based on only a single criterion, a gas-dominated problem may be misjudged as a feeding problem, or vice versa.

[0005] Third, while phase-field, micro-flow, or high-resolution dendrite models can provide strong mechanistic explanations, their computational domain and time scale are limited, making it difficult to directly cover the entire casting. Engineering-level macroscopic models have a large coverage area, but they cannot naturally provide information on interdendritic solute enrichment, final solidification channels, and pore morphology. Existing methods still do not adequately transfer boundary conditions between key macroscopic locations and local microscopic representative volume elements.

[0006] Fourth, the Niyama safety threshold, shrinkage volume fraction relationship, material thermophysical properties, and interfacial heat transfer parameters are all related to alloys, mold shells, and production lines. If uncalibrated general constants are used directly, the results are more suitable for relative trend comparisons and not for absolute compliance judgments. Existing solutions still lack sufficient support for reverse calibration of physical test results and blind testing verification after calibration.

[0007] Furthermore, purely data-driven methods require a large number of high-quality samples covering geometric, material, and process variations, while casting production samples are typically limited in number, have incomplete inspection labels, and are unevenly distributed in terms of operating conditions. Without physical constraints, surrogate models may produce unreliable extrapolations outside the sample range. Therefore, a more suitable approach is to generate constrained samples using a physical solver, reduce the cost of repeated solutions using machine learning models, and explicitly manage uncertainty and applicability.

[0008] Therefore, a method is needed that integrates macroscopic solidification criteria, local microscopic cause simulation, and anti-fluctuation optimization: it should both identify risks and distinguish possible formation mechanisms; it should both provide nominal operating conditions and evaluate the worst-case margin within the production tolerance range; and it should retain interfaces for calibration using material data, thermal history, and physical test results. Summary of the Invention

[0009] To address the aforementioned problems, this invention provides a method and system for simulating micro-porosity in castings and optimizing the process window. This approach uses macroscopic solidification solutions to determine the risk field, employs local cellular automata-finite volume methods to simulate and explain the formation of interdendritic dendrites at critical locations, and trains a surrogate model using physical simulation samples to solve for the worst-case Niyama margin under production tolerance constraints.

[0010] The first aspect of this invention proposes a method for simulating microporousness in castings and optimizing the process window, comprising six steps from S1 to S6: S1: Obtain the three-dimensional geometric data of the casting and the shell and establish a solid voxel model. Solve the solidification heat transfer process based on the apparent specific heat method. Calculate the equivalent thermal properties based on the shell formula and record the liquidus and solidus crossing times and solidification field quantities. S2: Calculate the Niyama criterion value based on the temperature gradient and cooling rate at the solidus line crossing time, further calculate the dimensionless Niyama criterion value and shrinkage-induced porosity volume fraction, calculate the porosity risk index based on the dissolved gas supersaturation relationship, and distinguish between shrinkage-induced, porosity-induced, and mixed defects. S3: Compare the local solidification time, Niyama criterion value and dissolved gas supersaturation index with the corresponding thresholds to generate hot spot markers, insufficient feeding markers and gas evolution markers. Based on the markers, determine the risk level, key location and formation mechanism of shrinkage-induced, porosity-induced or mixed type. S4: Receive the suspected location selected by the user, extract the macroscopic solidification conditions of the location, and perform a microscopic genetic simulation coupled with cellular automata and finite volume method on the local conformal representative volume element; S5: Sample the process parameter space, respond with the minimum Niyama criterion value in the preset evaluation area, train the Gaussian process regression surrogate model, and obtain the anti-fluctuation process window by maximizing the worst-case Niyama margin in the parameter fluctuation set. S6: When no user selection is received, automatically select one or more key locations based on the risk field, call the micro-genetic simulation, and summarize and output the macro-risk field, micro-organism and pore evolution results and process window. Among them, the microgenetic simulation is a coupled simulation of cellular automata and finite volume method, and the microporous simulation of castings and the process window optimization method are executed by the interface layer calling the calculation engine layer. The interface layer is used for parameter input, task scheduling and result display, and the calculation engine layer is used to perform geometric processing, solidification solution, criterion calculation, causal diagnosis, micro-simulation and process window optimization.

[0011] A second aspect of the present invention provides a system for simulating micro-porosity in castings and optimizing the process window, comprising: Macroscopic solidification solution module: used to acquire three-dimensional geometric data of castings and shells and establish solid voxel models; solve the solidification heat transfer process based on the apparent specific heat method, calculate equivalent thermal properties based on the shell formula, and record the liquidus and solidus crossing times and solidification field quantities. The criteria and defect quantification module is used to calculate the Niyama criterion value based on the temperature gradient and cooling rate at the solidus line crossing time, and further calculate the dimensionless Niyama criterion value and shrinkage-induced porosity volume fraction; it calculates the porosity risk index based on the dissolved gas supersaturation relationship and distinguishes between shrinkage-induced, porosity-induced and mixed defects. The risk and mechanism determination module is used to compare the local solidification time, Niyama criterion value and dissolved gas supersaturation index with the corresponding thresholds to generate hot spot flags, insufficient feeding flags and gas evolution flags, and determine the risk level, key location and formation mechanism of shrinkage, porosity or mixed type based on the flags. The microgenetic simulation module is used to receive the suspected location selected by the user, extract the macroscopic solidification conditions of the location, and perform a microgenetic simulation coupled with cellular automata and finite volume method on the local conformal representative volume element. The anti-fluctuation process window optimization module is used to sample the process parameter space, respond with the minimum Niyama criterion value in the preset evaluation region, train a Gaussian process regression surrogate model, and obtain the anti-fluctuation process window by maximizing the worst-case Niyama margin in the parameter fluctuation set. The analysis, arrangement, and output module is used to automatically select one or more key locations based on the risk field when no user selection is received, call the micro-genetic simulation, and summarize and output the macro-risk field, micro-organism and pore evolution results, and process window. Among them, the microgenetic simulation is a coupled simulation of cellular automata and finite volume method, and the microporous simulation of castings and the process window optimization method are executed by the interface layer calling the calculation engine layer. The interface layer is used for parameter input, task scheduling and result display, and the calculation engine layer is used to perform geometric processing, solidification solution, criterion calculation, causal diagnosis, micro-simulation and process window optimization.

[0012] A third aspect of the present invention also provides an electronic device comprising: a processor and a memory; The processor executes any of the above methods for simulation of microporousness in castings and optimization of process windows by calling programs or instructions stored in memory.

[0013] In a fourth aspect, the present invention also provides a computer-readable storage medium that stores a program or instructions that cause a computer to execute any of the preceding methods for simulation of microporousness in castings and optimization of process windows.

[0014] The present invention has at least the following beneficial effects: 1) Establish a one-way coupling path between macroscopic risk positioning and local microscopic interpretation. The macroscopic model is responsible for screening key locations, while the local cellular automata-finite volume method model is responsible for displaying the regions where dendrite growth, solute enrichment, and porosity preferentially occur, thereby avoiding direct micrometer-level calculations across the entire casting.

[0015] 2) Insufficient shrinkage and supersaturation of dissolved gas are modeled separately, and their superposition is allowed. The classification results provide a basis for selecting measures such as local cooling, adjustment of the feeding channel, melt degassing, or control of shell permeability.

[0016] 3) Supports both user-selected and automatic point selection. Users can verify known detection locations; if no selection location is provided, the system can automatically select key locations based on the minimum Niyama criterion value within the effective computational domain and spatial interval constraints.

[0017] 4) Explicitly incorporate production tolerances into the optimization problem. The optimization objective is not simply to increase the criterion value of the nominal operating condition, but to increase the worst-case Niyama margin within the fluctuation set, and output the range and sensitivity of safety parameters.

[0018] 5) The calibration constants in the relationships between material thermal properties, interfacial heat transfer, critical feeding pressure drop, dendrite arm spacing, Niyama safety threshold, and volume fraction can all be configured, facilitating calibration using thermal analysis, thermocouple records, and physical test results.

[0019] 6) The interface layer and the calculation engine layer are decoupled. The calculation engine can run and be tested independently of the graphical interface, which facilitates access to external temperature history, replacement of solvers, and consistent criteria and optimization processing for different accuracy levels. Attached Figure Description

[0020] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. It is obvious that the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings.

[0021] Figure 1 A flowchart of the method for simulating micro-porosity in castings and optimizing the process window provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a macroscopic multiphysics order reduction solution method provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a shrinkage porosity criterion and defect quantification method provided in an embodiment of the present invention; Figure 4 A flowchart of microgenetic simulation of cellular automata-finite volume method driven by point selection or automatic point selection provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of an anti-fluctuation process window optimization method provided in an embodiment of the present invention; Figure 6 This is a block diagram of the multi-scale simulation and anti-fluctuation process window optimization system for casting micro-porosity provided in an embodiment of the present invention; Figure 7 This is a schematic block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0023] Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts disclosed in this invention.

[0024] In the description of this invention, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The terms "installed," "connected," and "linked" should be interpreted broadly; for example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0025] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of methods and systems consistent with some aspects of the invention as detailed in the appended claims.

[0026] This invention provides a method, system, electronic device, and storage medium for simulating micro-porosity in castings and optimizing the process window. The following embodiments are used to illustrate the technical solution and its boundaries, and should not be construed as limiting the scope of protection or specific qualification thresholds.

[0027] Method Implementation Examples In the method embodiments, combined with Figure 1 The methods include S1 to S6: S1: Obtain the three-dimensional geometric data of the casting and the shell and establish a solid voxel model. Solve the solidification heat transfer process based on the apparent specific heat method. Calculate the equivalent thermal properties based on the shell formula and record the liquidus and solidus crossing times and solidification field quantities.

[0028] Specifically, in this embodiment of the invention, S1 is used to establish the solidification temperature field and the basic field quantities required for subsequent criteria at the whole part or representative cross-sectional scale, which can be specifically according to... Figure 2 And S21 to S24 will be implemented.

[0029] S2: Calculate the Niyama criterion value based on the temperature gradient and cooling rate at the solidus line crossing time, further calculate the dimensionless Niyama criterion value and shrinkage-induced porosity volume fraction, calculate the porosity risk index based on the dissolved gas supersaturation relationship, and distinguish between shrinkage-induced, porosity-induced, and mixed defects.

[0030] Specifically, in this embodiment of the invention, S2 is used to characterize the risk of insufficient shrinkage compensation and the risk of dissolved gas precipitation, respectively; the dimensionless Niyama relation is used to estimate the amount of shrinkage-induced loosening; and the gas supersaturation relation is used to estimate the risk of porosity. These two should not be confused. Specifically, they can be defined as follows: Figure 3 And S31 to S34 will be implemented.

[0031] S3: Compare the local solidification time, Niyama criterion value, and dissolved gas supersaturation index with the corresponding thresholds to generate hot spot markers, insufficient shrinkage markers, and gas evolution markers. Based on the markers, determine the risk level, key location, and formation mechanism of shrinkage-induced, porosity-induced, or mixed-type formation.

[0032] Specifically, in this embodiment of the invention, S3 converts the field quantity into a risk level and mechanism description. The threshold is preferentially calibrated by the measured data of the same alloy, shell and production line. If not calibrated, statistical quantiles can be used for relative screening, but absolute qualification cannot be directly claimed based on this. The specific generation of the mark and the method for determining the risk level, key position and shrinkage-induced, porosity-induced or mixed-type formation mechanism based on the mark are described in detail below.

[0033] S4: Receive the suspected location selected by the user, extract the macroscopic solidification conditions of the location, and perform a microscopic genetic simulation coupled with cellular automata and finite volume method on the local conformal representative volume element.

[0034] Specifically, in this embodiment of the invention, S4 corresponds to manually selected paths, suitable for verifying known detection locations or locations of interest to engineers. Specifically, it can be configured as follows: Figure 4 And S41 to S45 will be implemented.

[0035] S5: Sample the process parameter space, respond with the minimum Niyama criterion value in the preset evaluation region, train the Gaussian process regression surrogate model, and obtain the anti-fluctuation process window by maximizing the worst-case Niyama margin in the parameter fluctuation set.

[0036] Specifically, in this embodiment of the invention, S5 corresponds to the anti-fluctuation optimization path, which can be specifically described as follows: Figure 5 As implemented in S51 to S54, the surrogate model is only used within the parameter range covered by the training samples; if it exceeds this range, physical simulation samples should be added or the model should be retrained.

[0037] S6: When no user selection is received, automatically select one or more key locations based on the risk field, call the micro-genetic simulation, and summarize and output the macro-risk field, micro-organism and pore evolution results, and process window.

[0038] Specifically, in this embodiment of the invention, S6 corresponds to automatic point selection and end-to-end path arrangement. Automatic point selection uses the same cellular automaton-finite volume method calculation module as S4. The difference is that the key positions are determined by the risk field rather than by user selection.

[0039] Among them, the simulation of micro porosity in castings and the optimization of process windows are executed by calling the calculation engine layer from the interface layer. The interface layer is used for parameter input, task scheduling and result display, while the calculation engine layer is used to perform geometric processing, solidification solution, criterion calculation, cause diagnosis, microscopic simulation and process window optimization.

[0040] Specifically, in this embodiment of the invention, the interface layer and the computing engine layer adopt a decoupled architecture; the interface layer handles input, task status and result display, and the computing engine layer organizes geometric, solidification, criterion, microscopic and optimization calculations through a unified interface.

[0041] Furthermore, in the above-mentioned method for simulating microporousness in castings and optimizing the process window, combined with... Figure 2 S1 includes four steps: S21 to S24. S21: Read the 3D geometry in STL format, or read the 3D geometry in STP or STEP format obtained by format conversion or geometry kernel parsing, determine whether the center of each mesh unit is located inside the casting entity or the shell entity, and establish the entity voxel model on the structured Cartesian mesh.

[0042] In one implementation, the STL surface mesh can be directly voxelized using ray casting or spatial inclusion algorithms, and acceleration libraries such as Embree can be used to improve the judgment speed; STP and STEP files can be parsed or converted into triangular meshes by a compatible geometric kernel. The voxel side length can be selected according to the part size, allowable error, and computational resources, for example, 2 to 8 mm. For approximately rotating structures, axisymmetric sections or sector models with periodic boundaries can be used to reduce the computational load.

[0043] It should be noted that smaller voxel side lengths generally provide a better description of thin walls and local hot spots, but computational and memory usage increase with the number of meshes. Voxel accuracy should be determined by checking geometric volume errors or mesh convergence, rather than simply by subjectively stating that finer is more accurate.

[0044] S22: Numerically solve the transient heat conduction-solidification equation on a representative cross section or a three-dimensional voxel mesh to obtain the temperature field, and calculate the temperature gradient, cooling rate and local solidification time from the temperature field.

[0045] In one embodiment, the finite difference method or the finite volume method is used to solve the problem. The unknown quantity is the temperature field. Apparent specific heat is It is used to account for the latent heat of solidification in the mushy region; in spatial discretization, the surface thermal conductivity of the interface between adjacent heterogeneous materials can be harmonic averaged.

[0046] The convective-radiative boundary heat flux density on the outer surface of the shell can be written as: .in, and The ambient temperature and shell surface temperature are expressed in Kelvin; if the internal storage temperature is calculated in degrees Celsius, 273.15 should be added before the radiation term is calculated for conversion.

[0047] In the formula, ρ is the material density. For apparent specific heat capacity, For isobaric specific heat capacity, For latent heat of solidification, f s k represents the solid volume fraction. th Thermal conductivity, Here, q is the gradient or divergence operator, h is the boundary heat flux density pointing into the computational domain, and q is the gradient or divergence operator. amb ε is the convective heat transfer coefficient, and ε is the surface emissivity. is the Stefan-Boltzmann constant.

[0048] S23: Calculate the equivalent thermal conductivity, equivalent isobaric specific heat capacity, and equivalent density of the shell based on the proportions, porosity, pore size, number of layers, and thickness of each constituent phase. The number of layers and thickness are also used to establish the shell geometry and heat transfer path.

[0049] In one embodiment, the mass fraction of each solid component is first converted to a volume fraction, and the thermal conductivity k of the solid framework is calculated. s The equivalent thermal conductivity of the porous shell is then calculated using the Maxwell-Eucken two-phase effective medium model, with pore radiation equivalent heat conduction added in the high-temperature range. The number of layers and thickness are used to determine the shell geometry and heat transfer distance, while porosity and pore size are used for equivalent thermophysical property calculations.

[0050] Shell With castings (T) has the same physical dimension, both representing isobaric specific heat capacity, but corresponding to the shell and casting materials respectively; to avoid confusion with the solute partition coefficient, this specification uses k. th Thermal conductivity is expressed in kJ. seg This represents the solute equilibrium partition coefficient.

[0051] ; In the formula, The equivalent thermal conductivity of the shell is For the thermal conductivity of the solid-state framework, φ represents the thermal conductivity of the pore gas, and φ represents the porosity. =4ε d pore TK 3 This is the equivalent heat conduction term for pore radiation. d is the Stefan-Boltzmann constant. pore Where TK is the equivalent aperture, TK is the absolute temperature, and w is the equivalent isobaric specific heat capacity by mass fraction. i Weighted: ,eff =Σw i The equivalent density can be calculated based on the density and porosity of the solid phase skeleton.

[0052] The Maxwell-Eucken relation treats pores as a dispersed phase within a continuous solid matrix, representing an equivalent medium approximation. For shells with high porosity, through-cracks, significant interlayer thermal resistance, or strong anisotropy, calibration should be performed using measured thermal diffusivity or temperature history. Alternatively, a layered, anisotropic model could be used; one cannot claim to obtain absolutely accurate shell thermal properties solely based on formula calculations.

[0053] S24: Record the moment when the temperature of each casting voxel crosses the liquidus and solidus lines, and record the temperature gradient G and cooling rate at the moment of solidus line crossing. And the local solidification time tLST.

[0054] Specifically, temperature gradient Cooling rate Local solidification time ;in, and These are the moments when the same voxel crosses the liquidus and solidus, respectively. The local solidification time is the difference in time, not the difference between the solidus temperature and the liquidus temperature.

[0055] Furthermore, in the above-mentioned method for simulating microporousness in castings and optimizing the process window, combined with... Figure 3 S2 includes four steps: S31 to S34. S31: Calculate the Niyama criterion value at the moment when each voxel of the casting crosses the solidus line; S32: Based on the Darcy feeding resistance in the mushy region, critical feeding pressure drop, secondary dendrite arm spacing, liquid phase viscosity, total solidification shrinkage and solidification temperature range, convert the Niyama criterion value into a dimensionless Niyama criterion value, and calculate the shrinkage-induced porosity volume fraction. S33: Calculate the supersaturation index of dissolved gas based on the concentration of dissolved gas in the liquid phase at the end of solidification, temperature, ambient pressure, metal static pressure, pressure drop of the pasty region, and surface tension pressure of bubbles. S34: Based on the shrinkage-induced porosity volume fraction and dissolved gas supersaturation index, defects are classified into shrinkage-induced, porosity-induced, and mixed types, and the risk index of gas evolution at the shell interface is calculated based on the shell adhesive content, air permeability, and local solidification time of the surface layer. The Niyama criterion value is calculated using the following formula: The unit is ; The dimensionless Niyama criterion value is calculated using the following formula: ; The volume fraction of shrinkage-induced loosening is calculated using the following formula: ; In the formula, G represents the temperature gradient at the moment the solidus crosses, with units of K / mm. This represents the absolute value of the cooling rate at any given moment, expressed in K / s. In the unit expression, K and s represent Kelvin and seconds, respectively, and are not variables to be determined. This represents the Niyama criterion value converted to a unit of length per meter, when N y When expressed in millimeters, =103N y , This indicates the critical compensation pressure drop. Indicates the spacing between secondary dendrite arms. This represents the Kozeny-Carman constant. β represents the dynamic viscosity of the liquid phase, and β represents the total solidification shrinkage. Indicates the solidification temperature range. =∫(1-fl) 2 / fl 3 dfs is the integral of the shrinkage resistance in the mushy region, f l f represents the liquid volume fraction. l = 1 - f s f s df is the solid volume fraction. s This represents the solid fraction infinitesimal element, i.e., the integral variable; N represents the calibration coefficient. y,0 and m represent the calibration scale and shape index of the volume fraction relationship, respectively; The dissolved gas supersaturation index can be calculated based on the following relationship: Pg = [Cg, l / KS(T)] 2 When Pg is not less than P amb +P met -ΔP sh At +2γ / rp, a risk of bubble nucleation is identified; In the formula, Cg,0 and Cg,l represent the initial and final liquid phase dissolved gas concentrations, respectively; kg represents the gas equilibrium distribution coefficient; KS(T) represents the Sievert solubility coefficient; and P... amb P represents environmental pressure. met ΔP represents the static pressure of the metal. sh γ represents the pressure drop during the filling of the mushy region, γ represents the liquid-gas interfacial tension, and rp represents the initial bubble radius.

[0056] Specifically, in this embodiment of the invention, Ny The classic Niyama criterion value; Darcy pressure drop in the mushy region and Proportional to 1 / N y 2 It is directly proportional. Therefore, all other things being equal, N y A smaller value usually indicates a larger pressure drop due to shrinkage and a higher risk of shrinkage-induced porosity. This relationship describes the shrinkage-compensation mechanism and cannot prove what kind of gas or interfacial reaction forms the pores.

[0057] It should be noted that the above includes , , The relationship between and m is the engineering normalization and calibration form adopted by existing software, which retains the physical trend of the Darcy pressure drop changing monotonically with the Niyama criterion, but it is not a universal porosity formula that does not require calibration. , and m should be determined from test data of the same alloy and production line. In another embodiment, a standard dimensionless form can be used. Then, the critical liquid volume fraction is inverted from the complete solid fraction - temperature integral. , and according to Calculate the porosity, where, .

[0058] For gas-induced porosity, the liquid-phase gas concentration Cg,l at the end of solidification can be calculated first according to the segregation relationship, and then Pg=[Cg,l / KS(T)] can be calculated according to the Sievert relationship. 2 Calculate the equilibrium precipitation pressure. When Pg exceeds the local nucleation conditions determined by ambient pressure, metal static pressure, pressure drop in the mushy region, and surface tension pressure, the risk of porosity increases. Lowering the ambient pressure may increase the precipitation tendency at a given gas content, but the vacuum during production may simultaneously promote melt degassing. Therefore, ambient pressure and initial gas content should be treated as independent inputs or calibration values, and it is inappropriate to simply assert that the degree of vacuum monotonically determines the amount of porosity.

[0059] To avoid ambiguity of symbols, this instruction manual uses... This represents the Kozeny-Carman constant with a value of 180, where C0 represents the nominal initial solute concentration of the alloy; Let represent the Stefan-Boltzmann constant, and γ represent the liquid-gas interfacial tension.

[0060] Furthermore, in the above-mentioned method for simulating microporousness in castings and optimizing the process window, S3 includes: The local solidification time tLST is compared with the first threshold t1; when tLST is not less than t1, a hot spot flag H=1 is generated, indicating that the corresponding voxel belongs to the last solidification hot spot; when tLST is less than t1, a hot spot flag H=0 is generated; where the first threshold t1 is the local solidification time quantile of the casting voxel with effective liquidus and solidus line crossing records in the evaluation area, or the threshold obtained by calibration from the test results of similar castings; Niyama criterion value N y With safety threshold N safe Comparison; when N y Less than N safe If the insufficient padding flag S=1 is generated, otherwise the insufficient padding flag S=0 is generated. The dissolved gas supersaturation index Ig and the porosity precipitation threshold Ig were compared. crit Comparison; when Ig is greater than I crit When gas evolution occurs, the gas evolution flag B=1; otherwise, the gas evolution flag B=0. The risk level, critical location, and formation mechanism are determined based on the hot spot indicator H, insufficient feeding indicator S, and gas evolution indicator B: When S=1 and B=0, it is determined to be a shrinkage-induced mechanism, indicating that the feeding pressure drop in the mushy region is too large and the residual liquid phase between dendrites cannot be replenished; when S=0 and B=1, it is determined to be a porosity-induced mechanism, indicating that the gas equilibrium partial pressure at the end of solidification exceeds the sum of the local pressure and the surface tension pressure and the conditions for bubble nucleation are met; when S=1 and B=1, it is determined to be a mixed mechanism; when S=0 and B=0, it is determined to be low risk, and H=1 is used to characterize the final solidification hot spot and increase the risk priority of the corresponding location, but does not determine the defect type alone; voxels that meet at least one of H=1, S=1, or B=1 are included in the candidate risk location set, and critical locations are determined from the candidate risk location set.

[0061] Specifically, in this embodiment of the invention, the first threshold t1 can be taken as the high quantile of the local solidification time of the evaluation area, such as P90; N safe Icrit prioritizes calibration based on the inspection results of qualified / defective parts from the same alloy and production line. When uncalibrated, it outputs a relative risk ranking and explicitly states that it cannot replace physical acceptance criteria.

[0062] Furthermore, combined Figure 4 S4 includes five steps: S41 to S45. S41: Receive one or more suspicious locations selected by the user in the 3D viewport, map the world coordinates of the suspicious locations to voxel indices, extract local conformal representative voxels at the corresponding locations, and extract at least one of the following as micro boundary conditions: cooling rate, temperature gradient, local solidification time, Niyama criterion value, melt gas content, and ambient pressure.

[0063] Users can select one or more suspicious locations in the 3D viewport; in one embodiment, a maximum of three locations can be selected, and if no locations are selected, S6 will automatically select points. World coordinates are mapped to voxel indices, and a local conformal section or volume of approximately 1–4 mm is extracted at that location. The microgrid step size can be 8–14 μm. Input microboundary conditions include… G, tLST, N y At least one of the following: melt gas content and ambient pressure; different input quantities can be selected according to the complexity of the microscopic model, and all of them are not required to participate simultaneously.

[0064] S42: Cellular automata are used to simulate equiaxed crystal nuclei, random orientation, and dendrite capture and growth with fourfold symmetry anisotropy.

[0065] In one implementation, local supercooling The critical supercooling degree exceeding the preset nucleation site Time-activated nucleation and assigned random orientation The dendrite tip velocity is ,in This is the dendrite growth kinetic coefficient, distinct from the liquid phase viscosity in the dimensionless Niyama relation. ; The interface cell captures liquid phase neighbors according to the anisotropic rule containing fourfold symmetry terms and forms fourfold symmetric preferred dendrite arms along the coordinate axis direction.

[0066] S43: The finite volume method is used to solve the solute transport in the liquid phase. The interface dissolution and solute enrichment in the liquid phase are calculated according to the solute equilibrium distribution coefficient, and the final solidification region between dendrites is determined accordingly.

[0067] In one embodiment, the solute mixing concentration satisfies: Solute enrichment alters the local liquidus line, which in turn changes the supercooling and growth rate, causing the residual liquid phase between dendrites to become the final solidification region.

[0068] In the formula, C is the intracellular mixed solute concentration. This refers to the concentration of the solute in the liquid phase. The solute diffusion coefficient in the liquid phase is... and They are the volume fractions of solid phase and liquid phase, respectively. =1- , This is the solute equilibrium partition coefficient. For reference ingredients The liquidus temperature at that point The slope of the liquidus line.

[0069] The dissolved gas field can adopt the same finite volume discretization form as the solute field, but uses the gas equilibrium distribution coefficient kg and the liquid-gas diffusion coefficient Dg. The gas field and the solute field share the solid fraction and interface position, but do not share the physical property parameters.

[0070] S44: The shrinkage porosity obtained from macroscopic calculations is distributed to the final solidification region between dendrites, and pores are generated at gas enrichment locations based on the local dissolved gas supersaturation index, thereby obtaining the spatial distribution of shrinkage-induced, porosity-induced, or mixed-type pores.

[0071] In one embodiment, the shrinkage-induced loosening target volume fraction and the permeability of the pasty region are calculated using the following formula: Shrinkage is preferentially distributed to interdendritic regions where solidification occurs later and connectivity is poor; This indicates the volume fraction of shrinkage-induced loosening. This is the calibrable upper limit of contraction.

[0072] Pore ​​locations are preferentially assigned to interdendritic regions where the liquid phase gas concentration is high and local nucleation conditions are met. Shrinkage porosity and porosity are calculated separately and then spatially superimposed; if the same mesh satisfies both conditions simultaneously, it can be marked as a mixed type.

[0073] S45: The single dendrite morphology is generated using the thermal diffusion phase field model, and the dendrite morphology and anisotropy of the phase field model are compared and verified with the simulation results of the cellular automata-finite volume method.

[0074] Here, phase-field comparison can only indicate the consistency of numerical morphology and anisotropy, and cannot replace the actual metallographic, compositional segregation or porosity measurement.

[0075] Furthermore, in the above-mentioned method for simulating microporousness in castings and optimizing the process window, combined with... Figure 5 S5 includes four steps: S51 to S54. S51: The process parameter space is sampled using the Latin hypercube method. The solidification heat transfer solver is run for each sample, and the minimum Niyama criterion value in the preset evaluation area or multiple preset probe positions is used as the response y.

[0076] The process parameters include at least one of the following: pouring temperature, shell preheating temperature, shell thickness, ambient heat transfer intensity, and local chilling intensity at critical locations; the preset evaluation area is a preset area of ​​interest or a set of multiple defect probe locations; the minimum value is the minimum value among spatial evaluation points under the same working condition, rather than arbitrarily taking the minimum value over the entire solidification time process, because N in this embodiment... y Calculated at the solidus line crossing time of each voxel.

[0077] S52: Training a Gaussian process regression surrogate model with a Matérn kernel function using sample process parameters x and response y. .

[0078] In one implementation, the Matérn kernel function is used, and the input parameters are normalized. The Gaussian process can simultaneously output the predicted mean and uncertainty; when the candidate process is located in a sparse region of the sample or the predicted uncertainty is too high, the physical solution sample should be increased instead of directly using the extrapolation results.

[0079] S53: For each candidate nominal process x, calculate the worst-case Niyama margin predicted by the regression surrogate model within the parameter fluctuation box U(x). And maximize that margin.

[0080] Specifically, This represents the Niyama criterion value predicted by the surrogate model under fluctuating process ξ. The safety threshold is determined by the test data. M(x) > 0 indicates that all parameters within the parameter fluctuation box are higher than the safety threshold. The larger M(x) is, the more resistant the process window is to fluctuations. ; By maximizing Obtain anti-fluctuation nominal process, This indicates that the surrogate model's prediction is still above the safety threshold within a given tolerance box. This indicates that there exists at least one combination of volatility that is unpredictable. Let be the production tolerance half-width of the j-th process parameter. The Niyama margin represents the difference between the predicted minimum Niyama criterion value and the safety threshold within the fluctuation set. A positive value indicates that the safety threshold is still met within the set fluctuation range.

[0081] S54: Output the set of process parameters with M(x)≥0, the safety margin of each parameter, and the parameter sensitivity based on the surrogate model, and generate corresponding process adjustment suggestions.

[0082] Specifically, in this embodiment of the invention, the process window can be represented as a set of parameters M(x)≥0 or a two-dimensional slice contour line; the parameter sensitivity can be estimated using the main effect variance proportion based on a surrogate model. Sensitivity is used for ranking and is not equivalent to a strict global causal contribution. Process measures may include local chilling, local shell thickness adjustment, directional air cooling, riser or ingate adjustment, but the specific structure and strength still need to be confirmed through physical recalculation and field tests.

[0083] Furthermore, the automatic key location selection in S6 includes: A set of candidate risk locations is established among the voxels of castings that meet the conditions for effective solidification recording. For candidate locations with insufficient feeding indicator S=1, they are sorted in ascending order of Niyama criterion value, and the voxel with the smallest Niyama criterion value is determined as the first critical location. For candidate locations with gas evolution indicator B=1 and insufficient feeding indicator S=0, they are sorted in descending order of dissolved gas supersaturation index to determine the porosity-induced critical location. When multiple critical locations are required, they are selected sequentially according to the sorting and spatial interval constraints are applied. The safety threshold is used to determine the absolute risk, and the field quantity sorting is used to determine the relatively most unfavorable location. The two do not need to be equal in value. When a suspicious location selected by the user is received, the suspicious location is designated as a critical location; when no suspicious location selected by the user is received, the critical location is automatically determined according to the aforementioned set of candidate risk locations and field quantity sorting; the interface layer calls the calculation engine layer through a unified analysis interface and receives the calculation results as a background task; the calculation engine layer performs at least one of the following: geometric voxelization, two-dimensional or three-dimensional solidification solution, criterion calculation, risk and mechanism diagnosis, cellular automata-finite volume method micro-simulation, and process window optimization.

[0084] Specifically, in this embodiment of the invention, when having effective Recorded in-voxel extraction of castings and will The corresponding position is determined as the first key position; when selecting multiple positions, press... The algorithm sorts the data in ascending order and combines this with the minimum spatial interval to remove duplicate points within the same hot spot.

[0085] Safety threshold N safe The parameter arg min is used to determine whether a critical location meets the absolute risk standard, while the field quantity sorting arg min is used to find the relatively most unfavorable location. The two are not required to be equal. The example values ​​are only specific to the material, geometry, mesh, and calibration parameters and should not be pre-fixed to a general most dangerous point threshold.

[0086] The simulation system adopts an architecture that decouples the interface layer from the computation engine layer. The interface layer is responsible for collecting inputs, creating background tasks, and displaying results; the computation engine layer is responsible for geometry voxelization, solidification solution, criterion calculation, risk diagnosis, microscopic simulation, and process window optimization.

[0087] In one implementation, the unified analysis interface run_analysis() sequentially organizes geometric processing, three-dimensional solidification solution, criteria and diagnosis, optional cellular automata-finite volume method microscopic simulation, and optional process window optimization; this interface can be called by a graphical interface or run independently in a batch processing environment without a graphical interface.

[0088] The computational engine allows for different precision levels: L1 uses a built-in 2D or 3D reduced-order solver for rapid parameter screening; L2 reads nodal temperature history, solidification time, temperature gradient, and cooling rate obtained from external commercial simulations or experiments for high-fidelity criterion calculation and calibration; L3 performs cellular automata-finite volume method microscopic simulations at key locations. Each level shares the same criterion, diagnostic, and optimization interfaces, but the absolute values ​​for different levels should be calibrated separately for mesh, material, and boundary conditions.

[0089] Example Explanation Taking a representative ring-spoke section of the transition section of K423A nickel-based cast superalloy as an example, the nominal process parameters include a pouring temperature of 1450 ℃, a shell preheating temperature of 1020 ℃, a shell thickness of 7 mm, and air cooling conditions. Latent heat of solidification, solidification shrinkage, solid fraction curves, and some shell parameters not provided in the material handbook are estimated using configurable literature values. Therefore, the following values ​​are used to illustrate the process and relative changes and cannot be directly used as product acceptance criteria without calibration.

[0090] In this representative cross-sectional example, N at the intersection point P1 of the outer ring and the spokes is... y The value is 2.76, the predicted shrinkage-induced porosity volume fraction is approximately 1.87%, and the local solidification time is approximately 156 s; the N at another intersection point P2 is... y The model assigns a value of 4.36, a predicted volume fraction of approximately 0.74%, and a local solidification time of approximately 130 s. The model places P1 at a high-risk position, indicating that this location has poor relative feeding conditions under the current parameters.

[0091] Sampling and Gaussian process regression were performed using casting temperature, shell preheating temperature, shell thickness, ambient heat transfer intensity, and local quench intensity as variables. The surrogate model yields the nominal N of the candidate robust process. y The value is approximately 3.70, and the worst-case prediction within the tolerance box is approximately 3.66; when this candidate process is re-entered into the physical solver for recalculation, the N value of P1 is... y The value increased from 2.76 to approximately 3.40. The surrogate prediction is not entirely identical to the physical recalculation; this difference reflects surrogate error based on a finite sample size. Therefore, the final process should be confirmed by physical recalculation and field testing.

[0092] Within this sample range, the main effect sensitivities estimated based on the surrogate model are approximately: local chill 88.0%, shell thickness 7.1%, ambient heat transfer intensity 4.1%, pouring temperature 0.5%, and shell preheating temperature 0.3%. These percentages describe the response surface variance distribution only within the parameter range of this example and do not represent the same proportions under other castings, other tolerance ranges, or other boundary conditions.

[0093] In another set of real STL 3D solutions and local microscopic self-tests, the key location N obtained by automatic point selection was...y The value is approximately 2.0. After passing the cooling conditions to the local cellular automata-finite volume method model, a demonstration result of approximately 17 μm of secondary dendrite arm spacing and approximately 1.68% of the area fraction of micropores due to shrinkage is obtained. The mesh, boundary conditions, and calibration parameters used in this 3D self-test are different from those in the aforementioned representative cross-section example, and the absolute values ​​of the two are not directly compared laterally.

[0094] The above examples show that this method can organize macro-risk screening, local micro-interpretation, and tolerance box optimization into a single process; its engineering effectiveness still depends on the calibration of actual geometry, material thermal properties, interfacial heat transfer, melt gas content, and physical defect data, and should be verified by blind testing with reserved samples.

[0095] System Implementation Examples A second aspect of the invention also proposes a system, in conjunction with Figure 6 ,include: Macroscopic solidification solution module 61: used to acquire three-dimensional geometric data of castings and shells and establish solid voxel models; solve the solidification heat transfer process based on the apparent specific heat method, calculate equivalent thermal properties based on the shell formula, and record the liquidus and solidus crossing times and solidification field quantities. The criterion and defect quantification module 62 is used to calculate the Niyama criterion value based on the temperature gradient and cooling rate at the solidus line crossing time, and further calculate the dimensionless Niyama criterion value and shrinkage-induced porosity volume fraction; calculate the porosity risk index based on the dissolved gas supersaturation relationship, and distinguish between shrinkage-induced, porosity-induced and mixed defects. The risk and mechanism determination module 63 is used to compare the local solidification time, Niyama criterion value and dissolved gas supersaturation index with the corresponding thresholds to generate hot spot flags, insufficient feeding flags and gas evolution flags, and determine the risk level, key location and formation mechanism of shrinkage, porosity or mixed type based on the flags. The microgenetic simulation module 64 is used to receive the suspected location selected by the user, extract the macroscopic solidification conditions of the location, and perform a microgenetic simulation coupled with cellular automata and finite volume method on the local conformal representative volume element. The anti-fluctuation process window optimization module 65 is used to sample the process parameter space, respond with the minimum Niyama criterion value in the preset evaluation area, train the Gaussian process regression surrogate model, and obtain the anti-fluctuation process window by maximizing the worst-case Niyama margin in the parameter fluctuation set. The analysis, arrangement and output module 66 is used to automatically select one or more key locations based on the risk field when no user selection is received, call the micro-genetic simulation, and summarize and output the macro-risk field, micro-organism and pore evolution results and process window. Among them, the microgenetic simulation is a coupled simulation of cellular automata and finite volume method. The microporous simulation and process window optimization system for castings is executed by the interface layer calling the calculation engine layer. The interface layer is used for parameter input, task scheduling and result display, while the calculation engine layer is used to perform geometric processing, solidification solution, criterion calculation, gene diagnosis, micro-simulation and process window optimization.

[0096] A third aspect of the present invention also provides an electronic device comprising: a processor and a memory; The processor executes any of the above methods for simulating microporousness in castings and optimizing process windows by calling programs or instructions stored in memory.

[0097] In a fourth aspect, the present invention also provides a computer-readable storage medium that stores a program or instructions that cause a computer to execute any of the preceding methods for simulation of microporousness in castings and optimization of process windows.

[0098] Figure 7 This is a schematic block diagram of an electronic device provided in an embodiment of the present invention.

[0099] like Figure 7 As shown, the electronic device includes at least one processor 701, at least one memory 702, and at least one communication interface 703. The various components of the electronic device are coupled together via a bus system 704. The communication interface 703 is used for information transmission with external devices. It is understood that the bus system 704 is used to implement communication between these components. In addition to a data bus, the bus system 704 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 7 The general designated all buses as Bus System 704.

[0100] It is understood that the memory 702 in this embodiment may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0101] In some implementations, memory 702 stores elements such as executable units or data structures, or subsets thereof, or extended sets thereof: operating systems and applications.

[0102] The operating system, comprising various system programs such as the framework layer, core library layer, and driver layer, is used to implement various basic business functions and handle hardware-based tasks. The application programs, including media players and browsers, are used to implement various application functions. A program implementing any method in the casting micro-porosity simulation and process window optimization method provided in this embodiment of the invention can be included in the application programs.

[0103] In this embodiment of the invention, the processor 701 executes the steps of various embodiments of the method for simulating micro porosity of castings and optimizing process windows provided by the present invention by calling the program or instructions stored in the memory 702, specifically, the program or instructions stored in the application program.

[0104] S1: Obtain the three-dimensional geometric data of the casting and the shell and establish a solid voxel model. Solve the solidification heat transfer process based on the apparent specific heat method. Calculate the equivalent thermal properties based on the shell formula and record the liquidus and solidus crossing times and solidification field quantities. S2: Calculate the Niyama criterion value based on the temperature gradient and cooling rate at the solidus line crossing time, further calculate the dimensionless Niyama criterion value and shrinkage-induced porosity volume fraction, calculate the porosity risk index based on the dissolved gas supersaturation relationship, and distinguish between shrinkage-induced, porosity-induced, and mixed defects. S3: Compare the local solidification time, Niyama criterion value and dissolved gas supersaturation index with the corresponding thresholds to generate hot spot markers, insufficient feeding markers and gas evolution markers. Based on the markers, determine the risk level, key location and formation mechanism of shrinkage-induced, porosity-induced or mixed type. S4: Receive the suspected location selected by the user, extract the macroscopic solidification conditions of the location, and perform a microscopic genetic simulation coupled with cellular automata and finite volume method on the local conformal representative volume element; S5: Sample the process parameter space, respond with the minimum Niyama criterion value in the preset evaluation area, train the Gaussian process regression surrogate model, and obtain the anti-fluctuation process window by maximizing the worst-case Niyama margin in the parameter fluctuation set. S6: When no user selection is received, automatically select one or more key locations based on the risk field, call the micro-genetic simulation, and summarize and output the macro-risk field, micro-organism and pore evolution results and process window. Among them, the microgenetic simulation is a coupled simulation of cellular automata and finite volume method, and the microporous simulation of castings and the process window optimization method are executed by the interface layer calling the calculation engine layer. The interface layer is used for parameter input, task scheduling and result display, and the calculation engine layer is used to perform geometric processing, solidification solution, criterion calculation, causal diagnosis, micro-simulation and process window optimization.

[0105] When the above steps are executed by the processor, all or some of the functions in two-dimensional or three-dimensional macroscopic solution, manual or automatic point selection, microscopic simulation, and process window optimization can be selected according to the task configuration.

[0106] Any method in the casting micro-porosity simulation and process window optimization method provided in this embodiment of the invention can be applied to, or implemented by, the processor 701. The processor 701 can be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed through integrated logic circuits in the hardware of the processor 701 or through software instructions. The processor 701 can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor.

[0107] The steps of any method in the casting micro-porosity simulation and process window optimization method provided in this invention can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software units in the decoding processor. The software units can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 702, and processor 701 reads the information in memory 702 and combines it with hardware to complete the steps of the method.

[0108] Those skilled in the art will understand that although some embodiments described herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of the invention and form different embodiments.

[0109] Those skilled in the art will understand that the descriptions of the various embodiments have different focuses, and for parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0110] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and all such modifications and variations fall within the scope defined by the appended claims.

[0111] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for simulating microporousness in castings and optimizing the process window, characterized in that, It includes six steps, S1 to S6: S1: Obtain the three-dimensional geometric data of the casting and the shell and establish a solid voxel model. Solve the solidification heat transfer process based on the apparent specific heat method. Calculate the equivalent thermal properties based on the shell formula and record the liquidus and solidus crossing times and solidification field quantities. S2: Calculate the Niyama criterion value based on the temperature gradient and cooling rate at the solidus line crossing time, further calculate the dimensionless Niyama criterion value and shrinkage-induced porosity volume fraction, calculate the porosity risk index based on the dissolved gas supersaturation relationship, and distinguish between shrinkage-induced, porosity-induced, and mixed defects. S3: Compare the local solidification time, Niyama criterion value and dissolved gas supersaturation index with the corresponding thresholds to generate hot spot markers, insufficient feeding markers and gas evolution markers. Based on the markers, determine the risk level, key location and formation mechanism of shrinkage-induced, porosity-induced or mixed type. S4: Receive the suspected location selected by the user, extract the macroscopic solidification conditions of the location, and perform a microscopic genetic simulation coupled with cellular automata and finite volume method on the local conformal representative volume element; S5: Sample the process parameter space, respond with the minimum Niyama criterion value in the preset evaluation area, train the Gaussian process regression surrogate model, and obtain the anti-fluctuation process window by maximizing the worst-case Niyama margin in the parameter fluctuation set. S6: When no user selection is received, automatically select one or more key locations based on the risk field, call the micro-genetic simulation, and summarize and output the macro-risk field, micro-organism and pore evolution results and process window. The microscopic formation simulation is a coupled simulation of cellular automata and finite volume method. The microporous simulation and process window optimization method of casting is executed by the interface layer calling the calculation engine layer. The interface layer is used for parameter input, task scheduling and result display, and the calculation engine layer is used to perform geometric processing, solidification solution, criterion calculation, cause diagnosis, microscopic simulation and process window optimization.

2. The method according to claim 1, characterized in that, S1 includes four steps from S21 to S24: S21: Read the 3D geometry in STL format, or read the 3D geometry in STP or STEP format obtained by format conversion or geometry kernel parsing, determine whether the center of each mesh unit is located inside the casting entity or shell entity, and establish the entity voxel model on the structured Cartesian mesh. S22: Numerically solve the transient heat conduction-solidification equation on a representative cross section or a three-dimensional voxel mesh to obtain the temperature field, and calculate the temperature gradient, cooling rate and local solidification time from the temperature field; S23: Calculate the equivalent thermal conductivity, equivalent isobaric specific heat capacity, and equivalent density of the shell based on the proportions, porosity, pore size, number of layers, and thickness of each constituent phase. The number of layers and thickness are also used to establish the shell geometry and heat transfer path. S24: Record the moment when the temperature of each casting voxel crosses the liquidus and solidus lines, and record the temperature gradient G and cooling rate at the moment of solidus line crossing. And the local solidification time tLST.

3. The method according to claim 1, characterized in that, S2 includes four steps, S31 to S34: S31: Calculate the Niyama criterion value at the moment when each voxel of the casting crosses the solidus line; S32: Based on the Darcy feeding resistance in the mushy region, critical feeding pressure drop, secondary dendrite arm spacing, liquid phase viscosity, total solidification shrinkage and solidification temperature range, convert the Niyama criterion value into a dimensionless Niyama criterion value, and calculate the shrinkage-induced porosity volume fraction. S33: Calculate the supersaturation index of dissolved gas based on the concentration of dissolved gas in the liquid phase at the end of solidification, temperature, ambient pressure, metal static pressure, pressure drop of the pasty region, and surface tension pressure of bubbles. S34: Based on the shrinkage-induced porosity volume fraction and dissolved gas supersaturation index, the defects are classified into shrinkage-induced, porosity-induced, and mixed types, and the risk index of gas evolution at the shell interface is calculated based on the shell adhesive content, air permeability, and local solidification time of the surface layer. The Niyama criterion value is calculated using the following formula: The unit is ; The dimensionless Niyama criterion value is calculated using the following formula: ； The volume fraction of shrinkage-induced loosening is calculated using the following formula: ; In the formula, G represents the temperature gradient at the moment the solidus crosses, with units of K / mm. This represents the absolute value of the cooling rate at any given moment, expressed in K / s. In the unit expression, K and s represent Kelvin and seconds, respectively, and are not variables to be determined. This represents the Niyama criterion value converted to a unit of length per meter, when N y When expressed in millimeters, =103N y , This indicates the critical compensation pressure drop. Indicates the spacing between secondary dendrite arms. This represents the Kozeny-Carman constant. β represents the dynamic viscosity of the liquid phase, and β represents the total solidification shrinkage. Indicates the solidification temperature range. =∫(1-fl) 2 / fl 3 dfs is the integral of the shrinkage resistance in the mushy region, f l f represents the liquid volume fraction. l = 1 - f s f s df is the solid volume fraction. s This represents the solid fraction infinitesimal element, i.e., the integral variable; N represents the calibration coefficient. y,0 and m represent the calibration scale and shape index of the volume fraction relationship, respectively; The dissolved gas supersaturation index can be calculated based on the following relationship: Pg = [Cg, l / KS(T)]²; when Pg is not less than P amb +P met -ΔP sh At +2γ / rp, a risk of bubble nucleation is identified; In the formula, Cg,0 and Cg,l represent the initial and final liquid phase dissolved gas concentrations, respectively; kg represents the gas equilibrium distribution coefficient; KS(T) represents the Sievert solubility coefficient; and P... amb P represents environmental pressure. met ΔP represents the static pressure of the metal. sh γ represents the pressure drop during the filling of the mushy region, γ represents the liquid-gas interfacial tension, and rp represents the initial bubble radius.

4. The method according to claim 1, characterized in that, S3 includes: The local solidification time tLST is compared with the first threshold t1; when tLST is not less than t1, a hot spot flag H=1 is generated, indicating that the corresponding voxel belongs to the last solidification hot spot; when tLST is less than t1, a hot spot flag H=0 is generated; where the first threshold t1 is the local solidification time quantile of the casting voxel with effective liquidus and solidus line crossing records in the evaluation area, or the threshold obtained by calibration from the test results of similar castings; Niyama criterion value N y With safety threshold N safe Comparison; when N y Less than N safe If the insufficient padding flag S=1 is generated, otherwise the insufficient padding flag S=0 is generated. The dissolved gas supersaturation index Ig and the porosity precipitation threshold Ig were compared. crit Comparison; when Ig is greater than I crit When gas evolution occurs, the gas evolution flag B=1; otherwise, the gas evolution flag B=0. The risk level, critical location, and formation mechanism are determined based on the hot spot indicator H, insufficient feeding indicator S, and gas evolution indicator B: When S=1 and B=0, it is determined to be a shrinkage-induced mechanism, indicating that the feeding pressure drop in the mushy region is too large and the residual liquid phase between dendrites cannot be replenished; when S=0 and B=1, it is determined to be a porosity-induced mechanism, indicating that the gas equilibrium partial pressure at the end of solidification exceeds the sum of the local pressure and the surface tension pressure and the conditions for bubble nucleation are met; when S=1 and B=1, it is determined to be a mixed mechanism; when S=0 and B=0, it is determined to be low risk, and H=1 is used to characterize the final solidification hot spot and increase the risk priority of the corresponding location, but does not determine the defect type alone; voxels that meet at least one of H=1, S=1, or B=1 are included in the candidate risk location set, and critical locations are determined from the candidate risk location set.

5. The method according to claim 1, characterized in that, S4 includes five steps, S41 to S45: S41: Receive one or more suspicious locations selected by the user in the three-dimensional viewport, map the world coordinates of the suspicious locations to voxel indices, extract local conformal representative voxels at the corresponding locations, and extract at least one of the following as micro boundary conditions: cooling rate, temperature gradient, local solidification time, Niyama criterion value, melt gas content and ambient pressure. S42: Cellular automata are used to simulate equiaxed crystal nuclei, random orientation, and dendrite capture and growth with fourfold symmetry and anisotropy. S43: The finite volume method is used to solve the solute transport in the liquid phase. The interface dissolution and solute enrichment in the liquid phase are calculated according to the solute equilibrium distribution coefficient, and the final solidification region between dendrites is determined accordingly. S44: The shrinkage porosity obtained from macroscopic calculation is allocated to the final solidification region between dendrites, and pores are generated at the gas enrichment location according to the local dissolved gas supersaturation index, thereby obtaining the spatial distribution of shrinkage-induced, porosity-induced, or mixed-type pores. S45: The single dendrite morphology is generated using the thermal diffusion phase field model, and the dendrite morphology and anisotropy of the phase field model are compared and verified with the simulation results of the cellular automata-finite volume method.

6. The method according to claim 1, characterized in that, S5 includes four steps from S51 to S54: S51: The process parameter space is sampled using the Latin hypercube method. The solidification heat transfer solver is run for each sample, and the minimum Niyama criterion value in the preset evaluation area or multiple preset probe positions is used as the response y. S52: Train a Gaussian process regression surrogate model with a Matérn kernel function using the sample process parameters x and the response y. ; S53: For each candidate nominal process x, calculate the worst-case Niyama margin predicted by the regression surrogate model within the parameter fluctuation box U(x). And maximize that margin; S54: Output the set of process parameters with M(x)≥0, the safety margin of each parameter, and the parameter sensitivity based on the surrogate model, and generate corresponding process adjustment suggestions; The process parameters include at least one of the following: casting temperature, shell preheating temperature, shell thickness, ambient heat transfer intensity, and local quenching intensity at critical locations; This represents the Niyama criterion value predicted by the surrogate model under fluctuating process ξ. The safety threshold is determined by the test data. M(x)>0 indicates that all parameters in the parameter fluctuation box are higher than the safety threshold. The larger M(x) is, the more resistant the process window is to fluctuations. The Niyama margin represents the difference between the predicted minimum Niyama criterion value in the fluctuation set and the safety threshold. Its value is positive, indicating that the safety threshold is still met within the set fluctuation range.

7. The method according to claim 1, characterized in that, The automatic selection of key positions in S6 includes: A set of candidate risk locations is established among the voxels of castings that meet the conditions for effective solidification recording. For candidate locations with insufficient feeding indicator S=1, they are sorted in ascending order of Niyama criterion value, and the voxel with the smallest Niyama criterion value is determined as the first critical location. For candidate locations with gas evolution indicator B=1 and insufficient feeding indicator S=0, they are sorted in descending order of dissolved gas supersaturation index to determine porosity-induced critical locations. When multiple critical locations are required, they are selected sequentially according to the above sorting and spatial interval constraints are applied. The safety threshold is used to determine the absolute risk, and the field quantity sorting is used to determine the relatively most unfavorable location. The two do not need to be equal in value. When a suspicious location selected by the user is received, the suspicious location is designated as a critical location; when no suspicious location selected by the user is received, the critical location is automatically determined according to the aforementioned set of candidate risk locations and field quantity sorting; the interface layer calls the calculation engine layer through a unified analysis interface and receives the calculation results as a background task; the calculation engine layer performs at least one of the following: geometric voxelization, two-dimensional or three-dimensional solidification solution, criterion calculation, risk and mechanism diagnosis, cellular automata-finite volume method micro-simulation, and process window optimization.

8. A system for simulating micro-porosity in castings and optimizing the process window, characterized in that, include: Macroscopic solidification solution module: used to acquire three-dimensional geometric data of castings and shells and establish solid voxel models; solve the solidification heat transfer process based on the apparent specific heat method, calculate equivalent thermal properties based on the shell formula, and record the liquidus and solidus crossing times and solidification field quantities. The criteria and defect quantification module is used to calculate the Niyama criterion value based on the temperature gradient and cooling rate at the solidus line crossing time, and further calculate the dimensionless Niyama criterion value and shrinkage-induced porosity volume fraction; it calculates the porosity risk index based on the dissolved gas supersaturation relationship and distinguishes between shrinkage-induced, porosity-induced and mixed defects. The risk and mechanism determination module is used to compare the local solidification time, Niyama criterion value and dissolved gas supersaturation index with the corresponding thresholds to generate hot spot flags, insufficient feeding flags and gas evolution flags, and determine the risk level, key location and formation mechanism of shrinkage, porosity or mixed type based on the flags. The microgenetic simulation module is used to receive the suspected location selected by the user, extract the macroscopic solidification conditions of the location, and perform a microgenetic simulation coupled with cellular automata and finite volume method on the local conformal representative volume element. The anti-fluctuation process window optimization module is used to sample the process parameter space, respond with the minimum Niyama criterion value in the preset evaluation region, train a Gaussian process regression surrogate model, and obtain the anti-fluctuation process window by maximizing the worst-case Niyama margin in the parameter fluctuation set. The analysis, arrangement, and output module is used to automatically select one or more key locations based on the risk field when no user selection is received, call the micro-genetic simulation, and summarize and output the macro-risk field, micro-organism and pore evolution results, and process window. The microscopic formation simulation is a coupled simulation of cellular automata and finite volume method. The microporous simulation and process window optimization method of casting is executed by the interface layer calling the calculation engine layer. The interface layer is used for parameter input, task scheduling and result display, and the calculation engine layer is used to perform geometric processing, solidification solution, criterion calculation, cause diagnosis, microscopic simulation and process window optimization.

9. An electronic device, characterized in that, include: Processor and memory; The processor executes a method for simulating microporousness in castings and optimizing process windows as described in any one of claims 1 to 7 by calling programs or instructions stored in the memory.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program or instructions that cause a computer to execute a method for simulating microporousness in castings and optimizing process windows as described in any one of claims 1 to 7.

Images