A single-crystal blade casting process comprehensive evaluation system and optimization method

Abstract

This invention belongs to the field of single-crystal high-temperature alloy casting, specifically a comprehensive evaluation system and optimization method for the casting processability of single-crystal blades. The method includes the following steps: constructing a module for evaluating casting processability, which comprises at least one single-crystal casting and its associated structures; simulating the directional solidification process of the casting module through numerical simulation to obtain the physical field distribution information of the casting process; identifying casting defect risk areas based on the physical field distribution information and evaluating the casting process performance; optimizing and adjusting the module structural parameters and process parameters based on the evaluation results; comparing the optimized scheme with experimental data to establish a correlation mechanism between module design, casting geometry, and alloy system, thereby achieving iterative optimization and solidification of the casting process. This invention improves the reliability of single-crystal blade casting processes and reduces trial-and-error costs through simplified module design, multi-physics simulation and defect identification, multi-objective optimization, and experimental feedback learning.

Description

Technical Field

[0001] This invention belongs to the field of single-crystal high-temperature alloy casting, specifically a comprehensive evaluation system and optimization method for the casting process of single-crystal blades. Background Technology

[0002] Nickel-based single-crystal superalloys, due to their excellent high-temperature service performance, have become key materials for the fabrication of advanced aero-engine turbine blades. To improve the high-temperature load-bearing capacity of the blades, the content of refractory elements (Re, Ru, W, Ta, etc.) has gradually increased, leading to a significant increase in blade fabrication costs and a corresponding decrease in yield. The fabrication of single-crystal blades involves dozens of complex processes, including wax pressing, assembly, shell forming, and directional solidification. Furthermore, the complex geometry and topology of their hollow structure, along with significant differences in wall thickness, makes them highly susceptible to typical process defects such as shrinkage porosity, undercasting, orientation deviation, impurities, and recrystallization during directional solidification. These defects exhibit multi-scale characteristics, ranging from micron-level pores to millimeter-level grain anomalies, and their formation involves melt flow, solidification heat transfer, dendrite competitive growth, and strong coupling effects of multiple physical fields. Existing process evaluation methods largely rely on empirical trial and error, lacking systematic quantitative evaluation tools. This makes it difficult to meet the engineering requirements of modern precision casting for the efficient fabrication of single-crystal blades, and also fails to provide reasonable guidance for controlling process defects in different blade casting modules, resulting in long development cycles, high trial and error costs, and low yields. Summary of the Invention

[0003] The purpose of this invention is to provide a comprehensive evaluation system and optimization method for the casting processability of single-crystal blades, addressing the technical problems of existing single-crystal blade casting process design relying on empirical trial and error, lacking systematic evaluation and optimization methods, disconnecting numerical simulation from experimental verification, and struggling to achieve multi-objective collaborative optimization. This invention constructs a casting module containing multiple simplified geometric features, combining multi-physics numerical simulation and automatic defect identification technology to achieve quantitative evaluation of casting process performance. Based on the evaluation results, an optimization algorithm is used to perform multi-objective optimization of module structural parameters and process parameters. Furthermore, experimental data feedback is used to correct the simulation model, establishing a mapping relationship between the alloy system, casting geometry, and process parameters, forming a closed-loop mechanism of "simulation evaluation - optimization iteration - experimental verification - process solidification." Ultimately, this improves the scientific rigor and reliability of single-crystal blade casting process design, effectively controls casting defects such as impurities, orientation deviation, and shrinkage porosity, shortens the R&D cycle, and increases product yield.

[0004] The technical solution adopted by this invention to achieve the above objectives is: a method for comprehensive evaluation and optimization of the casting process of single-crystal blades, comprising the following steps:

[0005] Step S1: Construct a module for evaluating casting processability, the module comprising at least one single-crystal casting and its associated structures;

[0006] Step S2: Simulate the directional solidification process of the casting module through numerical simulation to obtain the physical field distribution information of the casting process;

[0007] Step S3: Based on the physical field distribution information, identify casting defect risk areas and evaluate casting process performance;

[0008] Step S4: Based on the evaluation results, optimize and adjust the module structure parameters and process parameters;

[0009] Step S5: Compare the optimized scheme with the experimental data to establish the correlation mechanism between module design, casting geometry and alloy system, and realize the iterative optimization and solidification of casting process.

[0010] Step S1 specifically includes:

[0011] Step S1-1: Design a simplified geometric model of the single crystal casting: Design a simplified geometric model of the single crystal casting, which consists of a multi-layer structure of blade and rim plate;

[0012] The leaf blade is a near-leaf-shaped solid structure with an outer diameter of 15–25 mm, an inner diameter of 10–20 mm, and a chord height of 5–15 mm, with the chord height being less than 60% of the inner diameter; the height of the leaf blade is 30–45 mm.

[0013] The blade edge is a polygonal solid structure with a length of 20–60 mm, a width of 20–40 mm, and a height of 3–5 mm. The chord direction of the blade is parallel to the length direction of the blade edge, and the center of the blade edge is collinear with the center of the blade edge. The width and length directions of the multiple blade edge layers are parallel to each other. The transition radius R between the blade edge and the blade edge is 0.5–2.5 mm, and the R value increases with the number of layers.

[0014] Step S1-2: Construct a multi-layer casting structure: Design each single crystal casting to contain 3 to 5 rim plates, with the width and length directions of each rim plate parallel to each other. The blade size and rim plate size increase layer by layer from bottom to top to simulate the geometric characteristics of different cross sections in the actual blade and to simulate the heat and mass transfer behavior of the blade rim plates to the greatest extent.

[0015] Step S1-3: Assemble multiple castings to form a module: Set 4 to 6 multi-layer castings in each module, arrange each casting at a predetermined angle relative to the center line of the module, and select one or more combinations of predetermined angles 0°, 60°, 120°, 180°, 240°, and 300° to simulate the influence of different module assembly methods on solidification heat transfer path and heat radiation shielding effect.

[0016] Step S1-4: Set up the module auxiliary structure: In the module, from top to bottom, set up risers, gating system, crystal selector and water-cooled base. The risers are used for feeding, the gating system is used to guide the molten metal into each casting, the crystal selector is used to control the grain orientation, and the water-cooled base is used to provide the cold source for directional solidification.

[0017] Step S1-5: Set the crystal selector structural parameters: Select a crystal selector with a helix angle of 30° to 50°, a helix diameter of 3 to 7 mm, and 1.5 to 3 turns. The angle between the extension direction of the helix section and the plane of the water-cooled base is 90°. Adjust the geometric parameters of the crystal selector to adapt to the crystal selection characteristics of different alloy systems.

[0018] Step S2 specifically includes:

[0019] Step S2-1: Mesh the constructed module geometry model using the finite element method or finite volume method; implement adaptive mesh refinement for key areas with drastic temperature gradient changes; the key areas include: the transition angle between the blade and the rim plate, the edge of the rim plate platform, the helical section of the crystal selector, and the root of the riser; wherein, the mesh size of the transition angle and the crystal selector is set to 0.3-0.8 mm, the mesh size of the main body of the blade is 1.2-2.0 mm, and the mesh size of the remaining parts is 2.0-3.0 mm, to ensure the simulation accuracy of the temperature field, the morphology of the mushy region, and the grain structure;

[0020] Step S2-2: Input the material parameters of the selected alloy system, including: density, specific heat capacity, thermal conductivity, solidus temperature, liquidus temperature, latent heat of solidification, and viscosity, and use phase diagram calculation or Scheil model to simulate the microsegregation behavior during solidification.

[0021] The alloy system is selected from one or more of the first-generation single-crystal alloy PWA1483, the second-generation single-crystal alloy CMSX-4, and the third-generation single-crystal alloy René N6.

[0022] Step S2-3: Define the interfacial heat transfer coefficients of the mold shell to the water-cooled copper plate, the alloy to the water-cooled base plate, and the alloy to the mold shell, and define the thermal radiation coefficient between the outer surface of the mold shell and the furnace environment as 0.4 to 0.8. At the same time, consider the mutual radiation shielding effect between multiple castings.

[0023] The heat transfer coefficient between the mold shell and the water-cooled copper plate is 400–600 W / m²·K, the heat transfer coefficient between the alloy and the water-cooled base plate is 800–1200 W / m²·K, and the heat transfer coefficient between the alloy and the mold shell is 400–600 W / m²·K.

[0024] Step S2-4: Set directional solidification process parameters: Define the drawing speed as 2-8 mm / min and implement a segmented speed variation strategy, that is, use a higher speed at the blade body and a lower speed at the edge plate to optimize local solidification conditions; define the furnace body temperature distribution, including the heating zone temperature of 1500-1600℃, the transition zone temperature of 700-950℃, and the cooling zone temperature of 20-60℃; define the casting temperature as 1500-1580℃ and the mold shell preheating temperature as 1450-1550℃.

[0025] Step S2-5: Use a cellular automata-finite element coupled model to simulate grain nucleation and competitive growth during solidification. Input nucleation parameters and dendrite growth kinetic coefficients to predict grain orientation, primary dendrite spacing and the probability of impurity formation.

[0026] Step S2-1 specifically involves the following steps:

[0027] Step S2-1-1: Model preprocessing: Import the 3D geometric model of the module, complete model repair, chamfer simplification, and void removal, eliminate process features that have no impact on temperature field calculation, and ensure the convergence of the calculation model;

[0028] Step S2-1-2: Basic mesh initialization: Set the size of the global basic mesh, and use a structured mesh + unstructured hybrid mesh method to divide the area. Use a uniform coarse mesh in the normal area to reduce the overall computing power consumption.

[0029] Step S2-1-3: Temperature gradient determination: Perform preliminary temperature field iteration and rough calculation, solve the global temperature gradient field ΔT, extract the temperature change rate threshold, and determine key sensitive areas such as the gating system, abrupt changes in wall thickness, and solidification interface.

[0030] Step S2-1-4: Adaptive mesh refinement: The high-temperature gradient region after identification is automatically refined by mesh subdivision, and the mesh cell size is gradually reduced. The refinement ratio is controlled at 2 to 5 times to ensure the accuracy of temperature solution at solidification interface and hot spot.

[0031] Step S2-1-5: Mesh quality check: Check the mesh distortion rate, aspect ratio, and Jacobian matrix, remove inferior elements, ensure that the mesh quality meets the requirements of finite element / finite volume discretization solution, and complete the mesh independence verification.

[0032] Step S2-5 is specifically performed as follows:

[0033] Step S2-5-1: Multiphysics coupling initialization: Map the temperature field and flow field data obtained from the finite element solution to the cellular automaton computational domain, complete the spatial interpolation matching between the finite element mesh and the cellular mesh, and establish the time step coupling iteration relationship between the two models;

[0034] Step S2-5-2: Assign nucleation parameters: Input the nucleation parameters such as heterogeneous nucleation density, critical nucleation undercooling, and nucleation base radius. Combine the Gaussian distribution function to randomly generate grain nucleation positions and determine the triggering conditions for homogeneous and heterogeneous nucleation.

[0035] Step S2-5-3: Dendrite kinetics solution: Import dendrite growth kinetic coefficients, calculate the dendrite tip growth rate based on the KGT model, solve for the preferred grain growth orientation under different temperatures and degrees of undercooling, and establish the dendrite growth anisotropy function;

[0036] Step S2-5-4: Grain competition growth iteration: Update the cell state according to the solidification time step, simulate the grain growth, grain boundary engulfment and orientation competition process, and capture the solid-liquid interface evolution morphology in real time.

[0037] Step S2-5-5: Feature parameter extraction and calculation: After the solidification calculation is completed, the grain orientation angle distribution is statistically analyzed, the average dendrite spacing is measured, and the impurity determination threshold is combined to calculate the impurity ratio and formation probability, thus completing the quantitative prediction of microstructure.

[0038] The physical field distribution information includes: temperature gradient distribution, width and morphology of the mushy region, thermal radiation flux, grain orientation and stress concentration region;

[0039] Key indicators were extracted through post-processing analysis: Based on the extracted key indicators, independent cold zones, stress concentration zones, and abrupt morphological regions of the mushy area that cause defects such as impurities, orientation deviations, shrinkage porosity, hot cracks, and recrystallization were identified.

[0040] The independent cold zone refers to the area where the temperature is below the critical value, and the stress concentration zone refers to the area where the equivalent stress exceeds the alloy yield strength.

[0041] Key indicators include: axial temperature gradient at different heights of the casting, transverse temperature gradient, length of the mushy region, solid fraction distribution, grain orientation deviation angle, and equivalent stress distribution.

[0042] In step S3, the evaluation of casting process performance includes:

[0043] Step S3-1: Construction of evaluation index system: Determine the four major evaluation dimensions of casting defects, solidification uniformity, microstructure density and grain uniformity, and clarify the defect judgment criteria of shrinkage cavities, shrinkage porosity, hot cracks and inclusions;

[0044] Step S3-2: Simulation data extraction: Extract key simulation data such as solidification time, temperature gradient, cooling rate, solid fraction, porosity, and grain size to establish the original dataset of process performance;

[0045] Step S3-3: Quantitative calculation of individual indicators: Calculate the thermal section coefficient, solidification sequence consistency coefficient, grain inhomogeneity coefficient, and defect volume ratio respectively to complete the individual process performance score;

[0046] Step S3-4: Weighted comprehensive evaluation: The weight of each indicator is determined by the analytic hierarchy process, a comprehensive evaluation function for casting process is constructed, and the comprehensive process performance score is calculated.

[0047] Step S3-5: Process level determination: Based on the comprehensive score, classify the process into good and bad levels, locate the weak areas of the process, and mark the unreasonable process parameters such as pouring temperature, pouring speed, and mold temperature.

[0048] Step S3-6: Quantitatively classify the casting feasibility under the current module design and process parameters using weighted scoring or fuzzy comprehensive evaluation methods, and rate it as excellent, good, medium, or poor; if the rating is lower than the preset threshold, it is determined that optimization is needed.

[0049] Step S4 specifically includes:

[0050] Step S4-1: Optimization Variables and Constraints Setting: Divide the optimization variables into two categories. The first is the module structure parameters, including the number of castings, the casting arrangement angle, the relative position of the castings, and the crystal selector geometric parameters. The second is the casting process parameters, covering the pulling rate curve, the heating zone temperature, the transition zone temperature gradient, the cooling zone temperature, the mold shell thickness, and the mold shell material. At the same time, combined with the production boundary of the single crystal blade casting project, clarify the upper and lower limit constraint ranges of all optimization variables and limit the reasonable value range of the variables.

[0051] Step S4-2: Construction of multi-objective optimization function: The core optimization objectives are to eliminate impurity defects, eliminate grain orientation deviation defects, control the grain orientation deviation angle within 10°, constrain the width of the mushy region, and improve the uniformity of the temperature gradient. At the same time, auxiliary optimization requirements such as minimizing shrinkage defect rate, minimizing grain size, and minimizing temperature stress are also taken into account. The multi-dimensional optimization objectives are dimensionless normalized to eliminate the influence of dimensional differences and construct a unified multi-objective optimization evaluation function.

[0052] Step S4-3: Algorithm population initialization: Select genetic algorithm, particle swarm optimization algorithm or sparrow algorithm as the optimization algorithm, randomly generate the initial population based on the constraint interval, encode each group of structural parameters and process parameters combination, and generate multiple initial candidate parameter schemes;

[0053] Step S4-4: Iterative optimization calculation: Import the encoded parameter combinations into the numerical simulation model, perform batch simulation calculations and obtain defect index data such as impurity index, orientation deviation angle, mushy region width, temperature gradient, and defect rate, and solve the fitness value of the objective function; rely on the selection, crossover, and mutation operations of the algorithm to iteratively update the population, eliminate inferior parameter combinations, and continuously approach the optimal parameter range;

[0054] Step S4-5: Convergence Judgment and Verification Screening: Set the maximum number of iterations and convergence accuracy threshold for the algorithm. When the population fitness value tends to stabilize and the calculation error is less than the set threshold, terminate the iteration and output the candidate parameter combination that meets the optimization objective. Perform secondary numerical simulation verification on the candidate combination. If the defect level is not lower than the preset threshold, return to the parameter sampling or algorithm optimization stage for re-optimization. If the defect meets the standard, the combination is temporarily designated as the optimization scheme.

[0055] Steps S4-6: Closed-loop evaluation of the process scheme: Repeat steps S1-S3 to verify the provisional optimized process scheme, and iterate and optimize until the casting defects, grain structure and temperature field uniformity reach the optimal state. Finally, determine the optimal casting process scheme and complete the engineering verification.

[0056] Step S5 specifically includes:

[0057] Step S5-1: In the actual directional solidification experiment, thermocouples are placed at the crystal selector, blade root, edge plate center, and transition angle, and the measured temperature-time curves are recorded.

[0058] Step S5-2: Compare the measured temperature-time curve from step S5-1 with the simulated temperature-time curve from step S2, and calculate the average relative error and the maximum deviation; if the average relative error exceeds 10% or the maximum deviation exceeds 15℃, the simulation model needs to be corrected.

[0059] Step S5-3: When it is determined that the simulation model needs to be corrected, adjust the mesh generation strategy, interface heat transfer coefficient, thermal radiation coefficient or boundary conditions in step S2, and re-perform the numerical simulation until the error between the simulation results and the measured data is within the allowable range.

[0060] Step S5-4: Establish optimized module combination methods and process parameter databases for different alloy systems, including the first-generation single-crystal alloy PWA1483, the second-generation single-crystal alloy CMSX-4, and the third-generation single-crystal alloy René N6.

[0061] Step S5-5: Use a self-learning method to train the database from step S5-4 and establish a mapping model between alloy composition, casting geometry and optimal process parameters;

[0062] Steps S5-6: Solidify the optimized process scheme into a standardized process specification, and feed back the detection data from actual production to the simulation system to update the database and mapping model, so as to realize continuous optimization and self-learning of the casting process.

[0063] A comprehensive evaluation system for the casting processability of single-crystal blades, used to execute the aforementioned comprehensive evaluation and optimization method for the casting processability of single-crystal blades, includes:

[0064] The module construction module is used to generate a 3D module model containing multiple castings and their arrangement angles. It supports parameterized input of casting geometry, number of layers, transition angle radius, number of castings and angle distribution, and automatically generates assemblies of risers, gating system, crystal selector and water-cooled chassis. It is also used for parameterized settings of crystal selector helix angle, helix diameter and number of helix turns, where the diameter of water-cooled chassis is 180-400 mm and the mold shell thickness is 5-10 mm.

[0065] The simulation module integrates a numerical simulation engine, which can perform multi-physics simulation of the directional solidification process based on the input module model and process parameters, including temperature field, flow field, stress field and grain structure evolution, and has built-in databases of the thermal properties of various single crystal alloys and CAFE grain growth model.

[0066] The simulation module is equipped with a database of various single-crystal alloy thermal properties, including: thermal property parameters of the first-generation single-crystal alloy PWA1483, the second-generation single-crystal alloy CMSX-4, and the third-generation single-crystal alloy René N6.

[0067] The thermophysical parameters include: density, specific heat capacity, thermal conductivity, solidus temperature, liquidus temperature, latent heat of solidification, and viscosity.

[0068] The simulation module is equipped with nucleation parameters required for the CAFE grain growth model, including nucleation undercooling and nucleation density, and dendrite growth kinetic coefficients are preset according to the alloy system.

[0069] The evaluation and analysis module is used to extract key features such as temperature gradient, mushy region characteristics, grain orientation, and stress distribution from simulation results. Based on preset defect criteria, it automatically identifies defect risk areas and outputs a visual report, including temperature field cloud map, mushy region isosurface, grain morphology map, and defect probability distribution map.

[0070] The optimization iteration module automatically or assistedly adjusts the module structure and process parameters based on the evaluation results. It uses optimization algorithms to perform multi-objective optimization and drives the simulation module to perform iterative calculations until the optimization objectives are met, outputting the optimal combination of process parameters and the corresponding module configuration.

[0071] The optimal combination of process parameters output by the optimization iteration module includes: the drawing rate curve, the heating zone temperature, the transition zone temperature, the cooling zone temperature, the casting temperature, and the mold shell preheating temperature.

[0072] The feedback learning module is used to compare actual experimental data with simulation data, calculate errors and automatically correct simulation model parameters. At the same time, it stores the verified data into the optimization database and updates the alloy-geometry-process mapping model.

[0073] The present invention has the following beneficial effects and advantages:

[0074] 1. This invention constructs a casting module containing multiple simplified geometric features and various combined angles, which can systematically simulate the influence of different module structures on solidification heat transfer path, thermal radiation shielding effect and grain growth, providing a quantitative basis for module optimization design and overcoming the limitations of traditional empirical trial and error methods.

[0075] 2. This invention employs adaptive mesh refinement technology to perform detailed simulations of key components such as transition angles and platform regions. Combined with multiphysics field analysis of temperature field, morphology of mushy regions, and grain structure, it can accurately identify defect risk areas such as independent cold zones and stress concentration zones, and deeply analyze their coupling relationship with the morphology of mushy regions, providing a clear direction for process optimization.

[0076] 3. This invention incorporates parameters such as mold angle, number of castings, pulling speed, and holding furnace temperature into the scope of collaborative optimization. Through numerical simulation and iterative optimization, it can quickly obtain the optimal combination of process parameters that takes into account the characteristics of different alloy systems, effectively eliminate impurity defects, control grain orientation, and significantly improve the reliability and adaptability of the casting process.

[0077] 4. This invention constructs a chain-like evaluation and optimization mechanism among "module combination method - three-dimensional geometric configuration of casting - alloy system" by comparing and self-learning actual experimental data and simulated prediction data of directional solidification, thereby solidifying and inheriting process knowledge. A modular evaluation system is established for different generations of single-crystal alloys (such as PWA1483, CMSX-4, and René N6), making the optimized process scheme portable and scalable, significantly shortening the R&D cycle of new single-crystal blades, reducing manufacturing costs, and improving product qualification rate. Attached Figure Description

[0078] Figure 1 This is a flowchart illustrating the implementation of the present invention.

[0079] Figure 2 These are schematic diagrams of single-crystal castings and casting modules. (a) is a schematic diagram of the structure of a single-crystal casting, and (b) is a schematic diagram of the structure of a casting module composed of multiple single-crystal castings.

[0080] 1 is the crystal selector; 2 is the blade; 3 is the rim plate; 4 is the central column tube; 5 is the single crystal casting; 6 is the gating system; 7 is the riser.

[0081] Figure 3This is a 1 / 4 structural schematic diagram of the provided single-crystal blade process module and directional solidification system. Among them: (a) is a 1 / 4 structural cross-sectional view of the directional solidification system, (b) is a schematic diagram of local mesh refinement in the rim plate and blade area, and (c) is a schematic diagram of local mesh refinement in the crystal selector area;

[0082] 8 is the cooling zone; 9 is the transition zone; 10 is the heating zone of the directional solidification furnace; 11 is the water-cooled copper plate; 12 is the central column tube; 13 is the casting; 14 is the local grid of the crystal selector; 15 is the local grid of the rim plate; 16 is the local grid of the blade.

[0083] Figure 4 This is a schematic diagram of the mold assembly angle for a single-crystal casting. (a) is a top view of the mold assembly; (b) shows the mold assembly angle at 0°; (c) shows the mold assembly angle at 60°; (d) shows the mold assembly angle at 120°; (e) shows the mold assembly angle at 180°; (f) shows the mold assembly angle at 240°; and (g) shows the mold assembly angle at 300°.

[0084] Figure 5 The simulation results of PWA1483 alloy at a mold angle of 0° and a pulling rate of 3 mm / min are shown in the figure. Among them: (a) is the distribution of the directional solidification temperature field; (b) is the morphology of the mushy region; (c) is the distribution of the grain structure at the edge plate; 17 in the figure is the single crystal region and 18 is the impurity region at the edge plate.

[0085] Figure 6 The simulation results for PWA1483 alloy with a mold angle of 120° and a variable pulling rate are shown in the figure.

[0086] Figure 7 The simulation results of René N6 alloy after optimization of the model angle are shown in the figure. Detailed Implementation

[0087] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The present invention provides a comprehensive evaluation system and optimization method for the casting process performance of single-crystal blades. By constructing a casting module containing simplified geometric features and multiple combined angles, and combining multiphysics numerical simulation and automatic defect identification technology, the system achieves quantitative evaluation and multi-objective optimization of casting process performance. Furthermore, through experimental feedback, it establishes a mapping relationship between the alloy system, casting geometry, and process parameters, forming a closed-loop process iteration and knowledge solidification mechanism.

[0088] like Figure 1 The diagram shown is a flowchart of the implementation of this invention, and its core steps include: module construction, numerical simulation, process evaluation, optimization iteration, experimental verification, and feedback learning. Specifically, the method of this invention includes the following steps:

[0089] Step S1: Construct a module for evaluating casting processability;

[0090] The module includes at least one single-crystal casting and its associated structures. For example... Figure 2 As shown, the single crystal casting 5 includes a crystal selector 1, a blade 2, and a rim plate 3 forming a multi-layered simplified structure; the single crystal casting 5 is provided with a crystal selector assembly and a gating 6 from bottom to top, and multiple single crystal castings 5 ​​are evenly arranged along the circumferential direction of the central column tube 8, wherein the riser 7 is placed on the central injection tube 8, and the gating 6 is connected to the riser.

[0091] Specifically, the crystal selector 1 is located at the bottom of the single crystal casting to control the grain orientation. It adopts a spiral structure design with a spiral helix angle of 30° to 50°, a spiral diameter of 3 to 7 mm, and 1.5 to 3 turns. The spiral section extends at an angle of 90° to the plane of the water-cooled chassis.

[0092] The blade 2 constitutes the main body of the single crystal casting. It is a near-blade-shaped solid structure with an outer diameter of 15–25 mm, an inner diameter of 10–20 mm, a chord height of 5–15 mm, and a height of 30–45 mm.

[0093] The blade 3 is a polygonal solid structure with a length of 20–60 mm, a width of 20–40 mm, and a height of 3–5 mm. The chord direction of the blade is parallel to the length direction of the blade, and the center of the blade is collinear with the center of the blade.

[0094] The central column tube 4 serves as the central support structure of the module, with multiple single crystal castings 5 ​​evenly arranged along its circumference, and risers placed on the central column tube 4.

[0095] The single crystal casting 5 is a multi-layered simplified structure consisting of a crystal selector 1, a blade 2, and a rim plate 3. Each casting contains 3 to 5 rim plates, with the dimensions increasing layer by layer from bottom to top.

[0096] The gating system 6 connects to the riser 7, guiding the molten metal into each casting. The riser 7 is located at the top of the mold assembly for feeding purposes. Figure 3 The diagram shown is a 1 / 4 structural schematic of the directional solidification system, illustrating the arrangement of the modules within the furnace body. The directional solidification system includes a cooling zone 8, a transition zone 9, a heating zone 10, a water-cooled copper plate 11, a central column tube 12, and a casting 13. Within the directional solidification system, local mesh refinement is applied to key areas. Specifically, the refined mesh includes a local mesh 14 for the crystal selector, a local mesh 15 for the rim plate, and a local mesh 16 for the blade.

[0097] Cooling zone 8 is the lower part of the directional solidification furnace, with a temperature of 20-60℃, used to provide a solidification cooling environment.

[0098] The transition zone 9 is located between the heating zone and the cooling zone, with a temperature of 700-950℃, forming a temperature gradient.

[0099] The upper heating zone of the directional solidification furnace (zone 10) has a temperature of 1500–1600℃ and is used to maintain the temperature of the alloy melt.

[0100] A water-cooled copper plate 11 is placed at the bottom of the module to provide a cold source for directional solidification.

[0101] 12-piece central column tube Figure 2 4. It supports the entire module structure.

[0102] Casting 13, also known as single-crystal casting 5, is placed in the module.

[0103] Local mesh of the crystal selector (14): Adaptive mesh refinement in the crystal selector region to ensure the accuracy of temperature field and grain structure simulation. Local mesh of the rim plate (15): Adaptive mesh refinement in the rim plate region, focusing on the transition angle and plate edge. Local mesh of the blade body (16): Mesh of the main body of the blade.

[0104] Specifically, step S1 further includes:

[0105] S1-1: Simplified geometric model for designing single-crystal castings: The blade is a near-blade-shaped solid structure with an outer diameter of 15–25 mm, an inner diameter of 10–20 mm, a chord height of 5–15 mm (the chord height is less than 60% of the inner diameter), and a height of 30–45 mm; the rim plate is a polygonal solid structure with a length of 20–60 mm, a width of 20–40 mm, and a height of 3–5 mm; the chord direction of the blade is parallel to the length direction of the rim plate, the center of the blade circle is collinear with the center of the rim plate, and the width and length directions of the multi-layer rim plates are parallel to each other; the transition radius R between the blade and the rim plate is 0.5–2.5 mm, and the R value increases with the number of layers.

[0106] S1-2: Constructing a multi-layer casting structure: Each single-crystal casting contains 3 to 5 rim plates, with the blade size and rim plate size increasing layer by layer from bottom to top, to simulate the geometric characteristics and thermal mass distribution of different cross sections of actual blades.

[0107] S1-3: Assemble multiple castings to form a module: Each module consists of 4-6 multi-layer castings, arranged at predetermined angles relative to the module's centerline, such as... Figure 4 As shown, one or more combinations of 0° (b), 60° (c), 120° (d), 180° (e), 240° (f), and 300° (g) are selected to simulate the effects of different modeling methods on the solidification heat transfer path and thermal radiation shielding effect.

[0108] S1-4: Setting up the module's auxiliary structures: From top to bottom, risers, gating system, crystal selector, and water-cooled chassis are set up in sequence. The risers are used for feeding, the gating system guides the molten metal to flow in, the crystal selector controls the grain orientation, and the water-cooled chassis provides a cold source for directional solidification.

[0109] S1-5: Set the crystal selector structural parameters: Select a crystal selector with a helix angle of 30° to 50°, a helix diameter of 3 to 7 mm, and 1.5 to 3 turns. The angle between the extension direction of the helix section and the plane of the water-cooled chassis is 90° to adapt to the crystal selection characteristics of different alloy systems.

[0110] Step S2: Simulate the directional solidification process of the casting module using numerical simulation.

[0111] The constructed module geometry model is imported into numerical simulation software (such as Pro-CAST) for mesh generation, material parameter input, and boundary condition setting. A cellular automata-finite element (CAFE) model is then used to simulate grain nucleation and competitive growth. Key settings include:

[0112] S2-1 Mesh Generation: Adaptive mesh refinement is implemented in key regions with drastic temperature gradient changes (the transition angle between the blade and the rim, the edge of the rim platform, the selector spiral section, and the riser root) using the finite element method or finite volume method. Figure 3 As shown in the local mesh, the mesh size for the transition angle and selector is 0.3–0.8 mm, for the main body of the blade it is 1.2–2.0 mm, and for the remaining parts it is 2.0–3.0 mm, to ensure simulation accuracy.

[0113] Step S2-1 involves the following steps:

[0114] Step S2-1-1: Model preprocessing: Import the 3D geometric model of the module, complete model repair, chamfer simplification, and void removal, eliminate process features that have no impact on temperature field calculation, and ensure the convergence of the calculation model;

[0115] Step S2-1-2: Basic mesh initialization: Set the size of the global basic mesh, and use a structured mesh + unstructured hybrid mesh method to divide the area. Use a uniform coarse mesh in the normal area to reduce the overall computing power consumption.

[0116] Step S2-1-3: Temperature gradient determination: Perform preliminary temperature field iteration and rough calculation, solve the global temperature gradient field ΔT, extract the temperature change rate threshold, and determine key sensitive areas such as the gating system, abrupt changes in wall thickness, and solidification interface.

[0117] Step S2-1-4: Adaptive mesh refinement: The high-temperature gradient region after identification is automatically refined by mesh subdivision, and the mesh cell size is gradually reduced. The refinement ratio is controlled at 2 to 5 times to ensure the accuracy of temperature solution at solidification interface and hot spot.

[0118] Step S2-1-5: Mesh quality check: Check the mesh distortion rate, aspect ratio, and Jacobian matrix, remove inferior elements, ensure that the mesh quality meets the requirements of finite element / finite volume discretization solution, and complete the mesh independence verification.

[0119] S2-2 Material Parameter Input: Input the density, specific heat capacity, thermal conductivity, solid-liquid phase temperature, latent heat of solidification, viscosity, etc. of the selected alloy system (such as PWA1483, CMSX-4, René N6), and use the Scheil model to simulate microsegregation.

[0120] S2-3 Interface Heat Transfer and Radiation Settings: Define the heat transfer coefficients of the mold shell to the water-cooled copper plate (400~600 W / m²·K), the alloy to the water-cooled base plate (800~1200 W / m²·K), and the alloy to the mold shell (400~600 W / m²·K). The thermal radiation coefficient between the outer surface of the mold shell and the furnace environment is 0.4~0.8, and the mutual radiation shielding between multiple castings is considered.

[0121] S2-4 directional solidification process parameter settings: drawing speed 2~8 mm / min, implementing a segmented speed change strategy (high speed for blades, low speed for rim plates); furnace body temperature distribution: heating zone 1500~1600℃, transition zone 700~950℃, cooling zone 20~60℃; casting temperature 1500~1580℃, mold shell preheating temperature 1450~1550℃.

[0122] S2-5 Grain Growth Simulation: Using the CAFE model, nucleation parameters (nucleation undercooling, nucleation density) and dendrite growth kinetic coefficients are input to predict grain orientation, primary dendrite spacing, and the probability of impurity formation. Specifically:

[0123] Step S2-5-1: Multiphysics coupling initialization: Map the temperature field and flow field data obtained from the finite element solution to the cellular automaton computational domain, complete the spatial interpolation matching between the finite element mesh and the cellular mesh, and establish the time step coupling iteration relationship between the two models;

[0124] Step S2-5-2: Assign nucleation parameters: Input the nucleation parameters such as heterogeneous nucleation density, critical nucleation undercooling, and nucleation base radius. Combine the Gaussian distribution function to randomly generate grain nucleation positions and determine the triggering conditions for homogeneous and heterogeneous nucleation.

[0125] Step S2-5-3: Dendrite kinetics solution: Import dendrite growth kinetic coefficients, calculate the dendrite tip growth rate based on the KGT model, solve for the preferred grain growth orientation under different temperatures and degrees of undercooling, and establish the dendrite growth anisotropy function;

[0126] Step S2-5-4: Grain competition growth iteration: Update the cell state according to the solidification time step, simulate the grain growth, grain boundary engulfment and orientation competition process, and capture the solid-liquid interface evolution morphology in real time.

[0127] Step S2-5-5: Feature parameter extraction and calculation: After the solidification calculation is completed, the grain orientation angle distribution is statistically analyzed, the average dendrite spacing is measured, and the impurity determination threshold is combined to calculate the impurity ratio and formation probability, thus completing the quantitative prediction of microstructure.

[0128] Step S3: Identify defect risk areas based on physical field distribution information and evaluate casting process performance.

[0129] Step S3-1: Construction of evaluation index system: Determine the four major evaluation dimensions of casting defects, solidification uniformity, microstructure density and grain uniformity, and clarify the defect judgment criteria of shrinkage cavities, shrinkage porosity, hot cracks and inclusions;

[0130] Step S3-2: Simulation data extraction: Extract key simulation data such as solidification time, temperature gradient, cooling rate, solid fraction, porosity, and grain size to establish the original dataset of process performance;

[0131] Step S3-3: Quantitative calculation of individual indicators: Calculate the thermal section coefficient, solidification sequence consistency coefficient, grain inhomogeneity coefficient, and defect volume ratio respectively to complete the individual process performance score;

[0132] Step S3-4: Weighted comprehensive evaluation: The weight of each indicator is determined by the analytic hierarchy process, a comprehensive evaluation function for casting process is constructed, and the comprehensive process performance score is calculated.

[0133] Step S3-5: Process level determination: Based on the comprehensive score, classify the process into good and bad levels, locate the weak areas of the process, and mark the unreasonable process parameters such as pouring temperature, pouring speed, and mold temperature.

[0134] Step S3-6: Quantitatively classify the casting feasibility under the current module design and process parameters using weighted scoring or fuzzy comprehensive evaluation methods, and rate it as excellent, good, medium, or poor; if the rating is lower than the preset threshold, it is determined that optimization is needed.

[0135] Step S4: Optimize and adjust the module structure parameters and process parameters based on the evaluation results.

[0136] Step S4-1: Optimization Variables and Constraints Setting: Divide the optimization variables into two categories. The first is the module structure parameters, including the number of castings, the casting arrangement angle, the relative position of the castings, and the crystal selector geometric parameters. The second is the casting process parameters, covering the pulling rate curve, the heating zone temperature, the transition zone temperature gradient, the cooling zone temperature, the mold shell thickness, and the mold shell material. At the same time, combined with the production boundary of the single crystal blade casting project, clarify the upper and lower limit constraint ranges of all optimization variables and limit the reasonable value range of the variables.

[0137] Step S4-2: Construction of multi-objective optimization function: The core optimization objectives are to eliminate impurity defects, eliminate grain orientation deviation defects, control the grain orientation deviation angle within 10°, constrain the width of the mushy region, and improve the uniformity of the temperature gradient. At the same time, auxiliary optimization requirements such as minimizing shrinkage defect rate, minimizing grain size, and minimizing temperature stress are also taken into account. The multi-dimensional optimization objectives are dimensionless normalized to eliminate the influence of dimensional differences and construct a unified multi-objective optimization evaluation function.

[0138] Step S4-3: Algorithm population initialization: Select genetic algorithm, particle swarm optimization algorithm or sparrow algorithm as the optimization algorithm, randomly generate the initial population based on the constraint interval, encode each group of structural parameters and process parameters combination, and generate multiple initial candidate parameter schemes;

[0139] Step S4-4: Iterative optimization calculation: Import the encoded parameter combinations into the numerical simulation model, perform batch simulation calculations and obtain defect index data such as impurity index, orientation deviation angle, mushy region width, temperature gradient, and defect rate, and solve the fitness value of the objective function; rely on the selection, crossover, and mutation operations of the algorithm to iteratively update the population, eliminate inferior parameter combinations, and continuously approach the optimal parameter range;

[0140] Step S4-5: Convergence Judgment and Verification Screening: Set the maximum number of iterations and convergence accuracy threshold for the algorithm. When the population fitness value tends to stabilize and the calculation error is less than the set threshold, terminate the iteration and output the candidate parameter combination that meets the optimization objective. Perform secondary numerical simulation verification on the candidate combination. If the defect level is not lower than the preset threshold, return to the parameter sampling or algorithm optimization stage for re-optimization. If the defect meets the standard, the combination is temporarily designated as the optimization scheme.

[0141] Steps S4-6: Closed-loop evaluation of the process scheme: Repeat steps S1-S3 to verify the provisional optimized process scheme, and iterate and optimize until the casting defects, grain structure and temperature field uniformity reach the optimal state. Finally, determine the optimal casting process scheme and complete the engineering verification.

[0142] Step S5: Compare the optimized solution with the experimental data, establish a correlation mechanism, and realize process iterative optimization and solidification.

[0143] In the actual directional solidification experiment, thermocouples were set up (selector, blade root, edge plate center, and transition angle position) and the measured temperature-time curves were recorded.

[0144] S5-2 compares the measured curve with the simulated curve and calculates the average relative error and the maximum deviation. If the average relative error is greater than 10% or the maximum deviation is greater than 15℃, the simulation model needs to be corrected.

[0145] S5-3 Adjust the mesh generation strategy, interface heat transfer coefficient, thermal radiation coefficient, or boundary conditions, and resimulate until the error is within the allowable range.

[0146] S5-4 establishes optimized module combination methods and process parameter databases for different alloy systems (PWA1483, CMSX-4, René N6).

[0147] S5-5 uses self-learning methods (such as neural networks and regression analysis) to train the database and establish a mapping model between alloy composition, casting geometry and optimal process parameters.

[0148] S5-6 solidifies the optimized process scheme into a standardized process procedure and feeds back the detection data from actual production to the simulation system to update the database and mapping model, thereby achieving continuous optimization and self-learning.

[0149] This invention also provides a comprehensive evaluation system for the casting processability of single-crystal blades, comprising:

[0150] Module building module: Used to generate 3D module models containing multiple castings and their arrangement angles (e.g., Figure 2 , Figure 4 As shown, it supports parameterized input of casting geometry, number of layers, transition angle radius, number of castings, and angle distribution, and automatically generates assemblies of risers, gating systems, crystal selectors, and water-cooled bases. The crystal selector can be parameterized with helix angle, diameter, and number of turns, while the water-cooled base has a diameter of 180–400 mm and a mold shell thickness of 5–10 mm.

[0151] Simulation module: Integrates a numerical simulation engine, capable of performing multiphysics simulations of directional solidification processes based on input module models and process parameters, including temperature field, flow field, stress field, and grain structure evolution (e.g., ...). Figure 5-7 As shown in the figure, it includes a database of thermal properties for various single-crystal alloys (PWA1483, CMSX-4, René N6) and a CAFE grain growth model. The thermal properties include density, specific heat capacity, thermal conductivity, solid-liquid phase temperature, latent heat of solidification, and viscosity; the nucleation parameters (nucleation undercooling, nucleation density) and dendrite growth kinetic coefficients required by the CAFE model are preset according to the alloy system.

[0152] Evaluation and Analysis Module: This module extracts key features such as temperature gradient, mushy region characteristics, grain orientation, and stress distribution from simulation results. Based on preset defect criteria, it automatically identifies defect risk areas and outputs a visual report, including temperature field cloud map, mushy region isosurface, grain morphology map, and defect probability distribution map (such as the probability distribution of impurities, orientation deviation, and shrinkage defects).

[0153] Optimization Iteration Module: Based on the evaluation results, the module structure and process parameters are automatically or assisted to be adjusted. The optimization algorithm is used to perform multi-objective optimization and drive the simulation module to perform iterative calculations until the optimization objectives are met. The optimal combination of process parameters (pulling rate curve, heating zone temperature, transition zone temperature, cooling zone temperature, pouring temperature, mold shell preheating temperature) and the corresponding module configuration are output.

[0154] Feedback learning module: It is used to compare actual experimental data with simulation data, calculate errors and automatically correct simulation model parameters, and store qualified data into the optimization database to update the alloy-geometry-process mapping model.

[0155] The present invention will be further described below with reference to specific embodiments.

[0156] Example 1: Process optimization of the first-generation single-crystal alloy PWA1483

[0157] This embodiment uses a typical first-generation single-crystal superalloy, PWA1483, with the following nominal composition (wt.%): C 0.07, Cr 12, Co 9, W 3.8, Mo 1.85, Al 3.6, Ti 4.1, Ta 5, Ni 8.5. See Table 1 for details.

[0158] Table 1 Nominal composition of the test alloy PWA1483

[0159] element C Cr Co W Mo Al Ti Ta Ni content 0.07 12 9 3.8 1.85 3.6 4.1 5 Bal.

[0160] The specific implementation steps are as follows:

[0161] (1) A single-crystal casting module and directional solidification system were constructed using 3D modeling software. To reduce the amount of computation, a 1 / 4 symmetric model was adopted. The specific structure is as follows: Figure 2 , Figure 3 As shown;

[0162] (2) Figure 4Single-crystal castings and directional solidification systems with different mold angles (0°, 60°, 120°, 180°, 240°, 300°) were imported into Pro-Cast numerical simulation software, and mesh generation, mold shell construction, and boundary parameter settings were performed sequentially. A variable-size mesh generation method was used: the mesh size at the blade location was 1.0 mm, while the mesh size at the rim and transition angle locations was refined to 0.5 mm; the mold shell thickness was uniformly set to 7.5 mm.

[0163] (3) Set boundary parameters: Input the thermal properties of PWA1483 alloy (density, thermal conductivity, solid-liquid phase temperature, latent heat of solidification, etc.); Set the interface heat transfer coefficient, where the heat transfer coefficient between the alloy and the water-cooled chassis is 1000 W / (m²·K), the heat transfer coefficient between the alloy and the mold shell is 500 W / (m²·K), and the heat transfer coefficient between the mold shell and the water-cooled chassis is 500 W / (m²·K); Set the process parameters: casting temperature 1540℃, heating zone temperature 1540℃, transition zone temperature 820℃, cooling zone temperature 30℃, thermal radiation coefficient 0.6, and pulling speed 3 mm / min;

[0164] (4) Directional solidification simulations were performed on single-crystal modules with different module angles to obtain temperature field distribution, morphology of the pasty region, and grain structure. Typical results are as follows: Figure 5 As shown. Figure 5 The simulation yielded the physical field distribution and grain structure, including:

[0165] (a) Distribution of directional solidification temperature field: shows the temperature distribution in different areas of the casting.

[0166] (b) Morphology of the pasty region: showing the morphology and distribution of the solid-liquid two-phase region.

[0167] (c) shows the grain structure distribution at the edge plate: In the figure, 17 is the single crystal region and 18 is the impurity crystal region of the edge plate, indicating that there are impurity crystal defects at the edge plate platform under the 0° module angle.

[0168] (5) The model angle was adjusted based on the simulation results. As the model angle increased, the area of ​​the isolated undercooled region decreased, and the area of ​​the impurity defect decreased accordingly. After multiple optimization iterations, the optimal model angle range was determined to be 105°~150°.

[0169] (6) Based on the optimal module angle, further optimize other process parameters. Adjust the pulling speed range to 2~5 mm / min and the holding furnace temperature range to 1500~1580℃. After multiple iterations of optimization, the optimal combination of process parameters was determined to be: holding furnace temperature 1540℃, a higher pulling speed of 3.0~4.0 mm / min at the blade body, and a lower pulling speed of 2.0~2.5 mm / min at the edge plate. A segmented speed change strategy was adopted to optimize local solidification conditions.

[0170] (7) Save the optimized single-crystal blade preparation parameters to form a standard process specification for PWA1483 alloy.

[0171] Example 2: Process adaptability optimization of the third-generation single-crystal alloy René N6

[0172] As the generation of single-crystal superalloys increases, the content of refractory elements Re and Ta increases, leading to significant differences in alloy thermophysical properties and potentially narrowing the process window. This example uses the typical third-generation single-crystal alloy René N6 as the test material to verify and optimize the adaptability of process parameters. The nominal composition (wt.%) of the René N6 alloy is: Cr 6.4, Co 9.7, W 6.4, Mo 0.6, Al 5.7, Ti 1.0, Ta 6.5, Hf 0.1, Re 2.9, Ni Bal. In Example 2, the typical third-generation single-crystal René N6 was used as the test alloy, and the alloy composition is shown in Table 2.

[0173] Table 2 Nominal composition of the test alloy René N6

[0174] element Cr Co W Mo Al Ti Ta Hf Re Ni content 6.4 9.7 6.4 0.6 5.7 1.0 6.5 0.1 2.9 Bal.

[0175] The specific implementation steps are as follows:

[0176] (1) The basic process parameters optimized in Example 1 were used as initial inputs: mold angle 105°~140°, blade pulling speed 3.0~4.0 mm / min, edge plate pulling speed 2.0~3.0 mm / min, and furnace temperature 1540℃; Figure 6 As shown, the optimized simulation results demonstrate that under the conditions of a model angle of 120°, a blade pulling rate of 3 mm / min, and a rim plate pulling rate of 2.5 mm / min, heterocrystalline defects in the rim plate region are effectively eliminated, and the grain structure is all single crystal.

[0177] (2) The casting processability of DD33 alloy was verified using Pro-Cast numerical simulation software. The simulation results showed that impurity defects still existed at the edge plate platform when the mold assembly angle was 105° or 140°;

[0178] (3) Based on the existing parameters, the model angle was adjusted in a targeted manner. Through multiple simulation iterations, the optimal model angle range applicable to DD33 alloy was determined to be 115°~135°.

[0179] (4) Maintaining the pulling rate and furnace temperature consistent with the optimization results of Example 1 (1540℃, 3.0~4.0 mm / min at the blade, 2.0~2.5 mm / min at the rim), the verification results show that under this parameter combination, impurity defects in the rim region are effectively eliminated; Figure 7 As shown, the optimization results for the third-generation single-crystal alloy René N6 are presented. Under the conditions of a module angle of 135°, a blade pulling rate of 3 mm / min, and a rim plate pulling rate of 2.5 mm / min, the impurity defects in the rim plate region are effectively controlled, verifying the adaptability of the process parameters to different alloy systems.

[0180] (5) Save the optimized preparation parameters for René N6 alloy to form a process specification suitable for third-generation single crystal alloys, and feed it back to the optimization database to update the mapping model of alloy system-geometric features-process parameters.

[0181] As can be seen from the two embodiments above, this invention systematically optimizes key parameters such as the module assembly angle, pulling rate, and temperature field distribution of single-crystal castings through a combination of 3D modeling and numerical simulation. This effectively controls casting defects such as impurities and significantly improves the casting process performance of single-crystal alloys of different generations. The established chain-like evaluation and optimization mechanism of "module assembly method - casting 3D geometry configuration - alloy system" provides theoretical guidance for the evaluation of module processability of different casting systems, while greatly reducing the production cycle and blade manufacturing cost.

[0182] Those skilled in the art will understand that the above description is merely a preferred embodiment of the present invention, and the features described in the various embodiments of this disclosure can be combined or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. This is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for comprehensive evaluation and optimization of the casting process of single-crystal blades, characterized in that, Includes the following steps: Step S1: Construct a module for evaluating casting processability, the module comprising at least one single-crystal casting and its associated structures; Step S2: Simulate the directional solidification process of the casting module through numerical simulation to obtain the physical field distribution information of the casting process; Step S3: Based on the physical field distribution information, identify casting defect risk areas and evaluate casting process performance; Step S4: Based on the evaluation results, optimize and adjust the module structure parameters and process parameters; Step S5: Compare the optimized scheme with the experimental data to establish the correlation mechanism between module design, casting geometry and alloy system, and realize the iterative optimization and solidification of casting process.

2. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 1, characterized in that, Step S1 specifically includes: Step S1-1: Design a simplified geometric model of the single crystal casting: Design a simplified geometric model of the single crystal casting, which consists of a multi-layer structure of blade and rim plate; The leaf blade is a solid structure with a near-leaf shape, with an outer diameter of 15–25 mm, an inner diameter of 10–20 mm, and a chord height of 5–15 mm, the chord height being less than 60% of the inner diameter; the height of the leaf blade is 30–45 mm. The blade edge is a polygonal solid structure with a length of 20–60 mm, a width of 20–40 mm, and a height of 3–5 mm. The chord direction of the blade is parallel to the length direction of the blade edge, and the center of the blade edge is collinear with the center of the blade edge. The width and length directions of the multiple blade edge layers are parallel to each other. The transition radius R between the blade edge and the blade edge is 0.5–2.5 mm, and the R value increases with the number of layers. Step S1-2: Construct a multi-layer casting structure: Design each single crystal casting to contain 3 to 5 rim plates, with the width and length directions of each rim plate parallel to each other. The blade size and rim plate size increase layer by layer from bottom to top to simulate the geometric characteristics of different cross sections in the actual blade and to simulate the heat and mass transfer behavior of the blade rim plates to the greatest extent. Step S1-3: Assemble multiple castings to form a module: Set 4 to 6 multi-layer castings in each module, arrange each casting at a predetermined angle relative to the center line of the module, and select one or more combinations of predetermined angles 0°, 60°, 120°, 180°, 240°, and 300° to simulate the influence of different module assembly methods on solidification heat transfer path and heat radiation shielding effect. Step S1-4: Set up the module auxiliary structure: In the module, from top to bottom, set up risers, gating system, crystal selector and water-cooled base. The risers are used for feeding, the gating system is used to guide the molten metal into each casting, the crystal selector is used to control the grain orientation, and the water-cooled base is used to provide the cold source for directional solidification. Step S1-5: Set the crystal selector structural parameters: Select a crystal selector with a helix angle of 30° to 50°, a helix diameter of 3 to 7 mm, and 1.5 to 3 turns. The angle between the extension direction of the helix section and the plane of the water-cooled base is 90°. Adjust the geometric parameters of the crystal selector to adapt to the crystal selection characteristics of different alloy systems.

3. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 1, characterized in that, Step S2 specifically includes: Step S2-1: Mesh the constructed module geometry model using the finite element method or finite volume method; implement adaptive mesh refinement for key areas with drastic temperature gradient changes; The key areas include: the transition angle between the blade and the rim plate, the edge of the rim plate platform, the helical section of the crystal selector, and the root of the riser; wherein, the grid size of the transition angle and the crystal selector is set to 0.3-0.8 mm, the grid size of the main body of the blade is 1.2-2.0 mm, and the grid size of the remaining parts is 2.0-3.0 mm, to ensure the simulation accuracy of the temperature field, the morphology of the mushy region, and the grain structure; Step S2-2: Input the material parameters of the selected alloy system, including: density, specific heat capacity, thermal conductivity, solidus temperature, liquidus temperature, latent heat of solidification, and viscosity, and use phase diagram calculation or Scheil model to simulate the microsegregation behavior during solidification. The alloy system is selected from one or more of the first-generation single-crystal alloy PWA1483, the second-generation single-crystal alloy CMSX-4, and the third-generation single-crystal alloy René N6. Step S2-3: Define the interfacial heat transfer coefficients of the mold shell to the water-cooled copper plate, the alloy to the water-cooled base plate, and the alloy to the mold shell, and define the thermal radiation coefficient between the outer surface of the mold shell and the furnace environment as 0.4 to 0.

8. At the same time, consider the mutual radiation shielding effect between multiple castings. The heat transfer coefficient between the mold shell and the water-cooled copper plate is 400–600 W / m²·K, the heat transfer coefficient between the alloy and the water-cooled base plate is 800–1200 W / m²·K, and the heat transfer coefficient between the alloy and the mold shell is 400–600 W / m²·K. Step S2-4: Set directional solidification process parameters: Define the drawing speed as 2-8 mm / min and implement a segmented speed variation strategy, that is, use a higher speed at the blade body and a lower speed at the edge plate to optimize local solidification conditions; define the furnace body temperature distribution, including the heating zone temperature of 1500-1600℃, the transition zone temperature of 700-950℃, and the cooling zone temperature of 20-60℃; define the casting temperature as 1500-1580℃ and the mold shell preheating temperature as 1450-1550℃. Step S2-5: Use a cellular automata-finite element coupled model to simulate grain nucleation and competitive growth during solidification. Input nucleation parameters and dendrite growth kinetic coefficients to predict grain orientation, primary dendrite spacing and the probability of impurity formation.

4. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 3, characterized in that, Step S2-1 specifically involves the following steps: Step S2-1-1: Model preprocessing: Import the 3D geometric model of the module, complete model repair, chamfer simplification, and void removal, eliminate process features that have no impact on temperature field calculation, and ensure the convergence of the calculation model; Step S2-1-2: Basic mesh initialization: Set the size of the global basic mesh, and use a structured mesh + unstructured hybrid mesh method to divide the area. Use a uniform coarse mesh in the normal area to reduce the overall computing power consumption. Step S2-1-3: Temperature gradient determination: Perform preliminary temperature field iteration and rough calculation, solve the global temperature gradient field ΔT, extract the temperature change rate threshold, and determine the key sensitive areas of the gating system, abrupt wall thickness changes, and solidification interface. Step S2-1-4: Adaptive mesh refinement: The high-temperature gradient region after identification is automatically refined by mesh subdivision, and the mesh cell size is gradually reduced. The refinement ratio is controlled at 2 to 5 times to ensure the accuracy of temperature solution at solidification interface and hot spot. Step S2-1-5: Mesh quality check: Check the mesh distortion rate, aspect ratio, and Jacobian matrix, remove inferior elements, ensure that the mesh quality meets the requirements of finite element / finite volume discretization solution, and complete the mesh independence verification.

5. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 3, characterized in that, Step S2-5 is specifically performed as follows: Step S2-5-1: Multiphysics coupling initialization: Map the temperature field and flow field data obtained from the finite element solution to the cellular automaton computational domain, complete the spatial interpolation matching between the finite element mesh and the cellular mesh, and establish the time step coupling iteration relationship between the two models; Step S2-5-2: Assign nucleation parameters: Input the nucleation parameters such as heterogeneous nucleation density, critical nucleation undercooling, and nucleation base radius. Combine the Gaussian distribution function to randomly generate grain nucleation positions and determine the triggering conditions for homogeneous and heterogeneous nucleation. Step S2-5-3: Dendrite kinetics solution: Import dendrite growth kinetic coefficients, calculate the dendrite tip growth rate based on the KGT model, solve for the preferred grain growth orientation under different temperatures and degrees of undercooling, and establish the dendrite growth anisotropy function; Step S2-5-4: Grain competition growth iteration: Update the cell state according to the solidification time step, simulate the grain growth, grain boundary engulfment and orientation competition process, and capture the solid-liquid interface evolution morphology in real time. Step S2-5-5: Feature parameter extraction and calculation: After the solidification calculation is completed, the grain orientation angle distribution is statistically analyzed, the average dendrite spacing is measured, and the impurity determination threshold is combined to calculate the impurity ratio and formation probability, thus completing the quantitative prediction of microstructure.

6. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 1, characterized in that, The physical field distribution information includes: temperature gradient distribution, width and morphology of the mushy region, thermal radiation flux, grain orientation and stress concentration region; Key indicators were extracted through post-processing analysis: Based on the extracted key indicators, independent cold zones, stress concentration zones, and abrupt morphological regions of the mushy area that cause defects such as impurities, orientation deviations, shrinkage porosity, hot cracks, and recrystallization were identified. The independent cold zone refers to the area where the temperature is below the critical value, and the stress concentration zone refers to the area where the equivalent stress exceeds the alloy yield strength. Key indicators include: axial temperature gradient at different heights of the casting, transverse temperature gradient, length of the mushy region, solid fraction distribution, grain orientation deviation angle, and equivalent stress distribution.

7. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 1, characterized in that, In step S3, the evaluation of casting process performance includes: Step S3-1: Construction of evaluation index system: Determine the four major evaluation dimensions of casting defects, solidification uniformity, microstructure density and grain uniformity, and clarify the defect judgment criteria of shrinkage cavities, shrinkage porosity, hot cracks and inclusions; Step S3-2: Simulation data extraction: Extract key simulation data such as solidification time, temperature gradient, cooling rate, solid fraction, porosity, and grain size to establish the original dataset of process performance; Step S3-3: Quantitative calculation of individual indicators: Calculate the thermal section coefficient, solidification sequence consistency coefficient, grain inhomogeneity coefficient, and defect volume ratio respectively to complete the individual process performance score; Step S3-4: Weighted comprehensive evaluation: The weight of each indicator is determined by the analytic hierarchy process, a comprehensive evaluation function for casting process is constructed, and the comprehensive process performance score is calculated. Step S3-5: Process level determination: Based on the comprehensive score, classify the process into good and bad levels, locate the weak areas of the process, and mark the unreasonable process parameters such as pouring temperature, pouring speed, and mold temperature. Step S3-6: Quantitatively classify the casting feasibility under the current module design and process parameters using weighted scoring or fuzzy comprehensive evaluation methods, and rate it as excellent, good, medium, or poor; if the rating is lower than the preset threshold, it is determined that optimization is needed.

8. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 1, characterized in that, Step S4 specifically includes: Step S4-1: Optimization Variables and Constraints Setting: Divide the optimization variables into two categories. The first is the module structure parameters, including the number of castings, the casting arrangement angle, the relative position of the castings, and the crystal selector geometric parameters. The second is the casting process parameters, covering the pulling rate curve, the heating zone temperature, the transition zone temperature gradient, the cooling zone temperature, the mold shell thickness, and the mold shell material. At the same time, combined with the production boundary of the single crystal blade casting project, clarify the upper and lower limit constraint ranges of all optimization variables and limit the reasonable value range of the variables. Step S4-2: Construction of multi-objective optimization function: The core optimization objectives are to eliminate impurity defects, eliminate grain orientation deviation defects, control the grain orientation deviation angle within 10°, constrain the width of the mushy region, and improve the uniformity of the temperature gradient. At the same time, auxiliary optimization requirements such as minimizing shrinkage defect rate, minimizing grain size, and minimizing temperature stress are also taken into account. The multi-dimensional optimization objectives are dimensionless normalized to eliminate the influence of dimensional differences and construct a unified multi-objective optimization evaluation function. Step S4-3: Algorithm population initialization: Select genetic algorithm, particle swarm optimization algorithm or sparrow algorithm as the optimization algorithm, randomly generate the initial population based on the constraint interval, encode each group of structural parameters and process parameters combination, and generate multiple initial candidate parameter schemes; Step S4-4: Iterative optimization calculation: Import the encoded parameter combinations into the numerical simulation model, perform batch simulation calculations and obtain defect index data such as impurity index, orientation deviation angle, mushy region width, temperature gradient, and defect rate, and solve the fitness value of the objective function; rely on the selection, crossover, and mutation operations of the algorithm to iteratively update the population, eliminate inferior parameter combinations, and continuously approach the optimal parameter range; Step S4-5: Convergence Judgment and Verification Screening: Set the maximum number of iterations and convergence accuracy threshold for the algorithm. When the population fitness value tends to stabilize and the calculation error is less than the set threshold, terminate the iteration and output the candidate parameter combination that meets the optimization objective. Perform a second numerical simulation verification on the candidate combination. If the defect level is not lower than the preset threshold, return to the parameter sampling or algorithm optimization stage for re-optimization. If the defects meet the standards, then this combination is tentatively considered the optimal solution; Steps S4-6: Closed-loop evaluation of the process scheme: Repeat steps S1-S3 to verify the provisional optimized process scheme, and iterate and optimize until the casting defects, grain structure and temperature field uniformity reach the optimal state. Finally, determine the optimal casting process scheme and complete the engineering verification.

9. The method for comprehensive evaluation and optimization of the casting processability of single-crystal blades according to claim 1, characterized in that, Step S5 specifically includes: Step S5-1: In the actual directional solidification experiment, thermocouples are placed at the crystal selector, blade root, edge plate center, and transition angle, and the measured temperature-time curves are recorded. Step S5-2: Compare the measured temperature-time curve from step S5-1 with the simulated temperature-time curve from step S2, and calculate the average relative error and the maximum deviation; if the average relative error exceeds 10% or the maximum deviation exceeds 15℃, the simulation model needs to be corrected. Step S5-3: When it is determined that the simulation model needs to be corrected, adjust the mesh generation strategy, interface heat transfer coefficient, thermal radiation coefficient or boundary conditions in step S2, and re-perform the numerical simulation until the error between the simulation results and the measured data is within the allowable range. Step S5-4: Establish optimized module combination methods and process parameter databases for different alloy systems, including the first-generation single-crystal alloy PWA1483, the second-generation single-crystal alloy CMSX-4, and the third-generation single-crystal alloy René N6. Step S5-5: Use a self-learning method to train the database from step S5-4 and establish a mapping model between alloy composition, casting geometry and optimal process parameters; Steps S5-6: Solidify the optimized process scheme into a standardized process specification, and feed back the detection data from actual production to the simulation system to update the database and mapping model, so as to realize continuous optimization and self-learning of the casting process.

10. A comprehensive evaluation system for the casting processability of single-crystal blades, used to execute the comprehensive evaluation and optimization method for the casting processability of single-crystal blades as described in any one of claims 1-9, characterized in that, include: The module construction module is used to generate a 3D module model containing multiple castings and their arrangement angles. It supports parameterized input of casting geometry, number of layers, transition angle radius, number of castings and angle distribution, and automatically generates assemblies of risers, gating system, crystal selector and water-cooled chassis. It is also used for parameterized settings of crystal selector helix angle, helix diameter and number of helix turns, where the diameter of water-cooled chassis is 180-400 mm and the mold shell thickness is 5-10 mm. The simulation module integrates a numerical simulation engine, which can perform multi-physics simulation of the directional solidification process based on the input module model and process parameters, including temperature field, flow field, stress field and grain structure evolution, and has built-in databases of the thermal properties of various single crystal alloys and CAFE grain growth model. The simulation module is equipped with a database of various single-crystal alloy thermal properties, including: thermal property parameters of the first-generation single-crystal alloy PWA1483, the second-generation single-crystal alloy CMSX-4, and the third-generation single-crystal alloy René N6. The thermophysical parameters include: density, specific heat capacity, thermal conductivity, solidus temperature, liquidus temperature, latent heat of solidification, and viscosity. The simulation module is equipped with nucleation parameters required for the CAFE grain growth model, including nucleation undercooling and nucleation density, and dendrite growth kinetic coefficients are preset according to the alloy system. The evaluation and analysis module is used to extract key features such as temperature gradient, mushy region characteristics, grain orientation, and stress distribution from simulation results. Based on preset defect criteria, it automatically identifies defect risk areas and outputs a visual report, including temperature field cloud map, mushy region value surface, grain morphology map, and defect probability distribution map. The optimization iteration module automatically or assistedly adjusts the module structure and process parameters based on the evaluation results. It uses optimization algorithms to perform multi-objective optimization and drives the simulation module to perform iterative calculations until the optimization objectives are met, outputting the optimal combination of process parameters and the corresponding module configuration. The optimal combination of process parameters output by the optimization iteration module includes: the drawing rate curve, the heating zone temperature, the transition zone temperature, the cooling zone temperature, the casting temperature, and the mold shell preheating temperature. The feedback learning module is used to compare actual experimental data with simulation data, calculate errors and automatically correct simulation model parameters. At the same time, it stores the verified data into the optimization database and updates the alloy-geometry-process mapping model.

Images