Design method of Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy, storage medium and equipment

Abstract

The invention relates to the technical field of die-casting aluminum alloys, in particular to a design method of an Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy, a storage medium and equipment. The method comprises the following steps: establishing a thermodynamic database of an Al-Si-Mg-Cu quaternary system rich in an Al end; the solidification structure information of the Al-Si-Mg-Cu alloy is obtained; data corresponding to the input features and the output features are adopted to train a Gaussian regression model, and when R2 is larger than or equal to 0.9, training is completed; and the optimal component point in the preset alloy component content range is screened out. A computer program instruction is stored in the storage medium; the apparatus includes at least one processor, at least one memory, and computer program instructions stored in the memory; the method is implemented when computer program instructions are executed by a processor. According to the method, the optimal component point within the preset alloy component content range can be quickly screened out, and the design of the novel high-performance heat-treatment-free die-casting aluminum alloy is accelerated.

Description

Technical Field

[0001] The present invention relates to the technical field of die-cast aluminum alloys, and in particular to a design method, storage medium and equipment for an Al-Si-Mg-Cu heat-treatment-free die-cast aluminum alloy. Background Art

[0002] In automobile manufacturing, automobile lightweighting has become an important direction for energy conservation and emission reduction. Aluminum alloy has become an ideal material for achieving automobile lightweighting due to its excellent performance, low density, good processability and cost-effectiveness. As the proportion of aluminum alloy used in automobiles increases, the difficulty of the splicing process of body structural parts continues to increase, and the splicing efficiency is low; at the same time, automobiles have higher and higher requirements for the ultimate tensile strength, yield strength, elongation, hardness and corrosion resistance of aluminum alloy materials. When automobile structural parts collide or are subjected to force, aluminum alloy materials need to have extremely high ultimate tensile strength to ensure safety; the yield strength of aluminum alloy materials needs to be high enough to avoid permanent deformation during the use of the car. Aluminum alloys with high yield strength can ensure that the car maintains its structural stability under normal driving and collision conditions; in order to prevent brittle fracture of automobile structural parts when they encounter impact, aluminum alloy materials need to have good elongation.

[0003] The ultimate tensile strength and yield strength of traditional aluminum alloy materials are relatively low, which makes it difficult to meet the needs of high-strength and high-load automotive structural parts. Especially in the event of a collision, traditional aluminum alloys may not be able to effectively withstand external impact forces, affecting vehicle safety. In addition, traditional aluminum alloy materials have low elongation, are prone to brittle fracture when subjected to stress, lack toughness, and cannot effectively absorb collision energy, reducing the impact resistance of the car. Therefore, traditional aluminum alloy materials have certain deficiencies in meeting the high-performance requirements of modern automobiles for structural parts. High-performance requirements can only be met by optimizing alloy composition, improving processing technology, and increasing surface treatment.

[0004] Developing high-performance heat-treatment-free die-cast aluminum alloys for integrated die-casting of body structural parts can effectively break through this bottleneck. Tesla added Cu to the Al-Si-Mg-based alloy, developed a new alloy and applied it to the manufacture of the heat-treatment-free integrated die-casting rear floor of the Model Y model, reducing the number of parts and the overall body weight.

[0005] However, the optimal amount of Cu added to Al-Si-Mg based alloys is difficult to determine. In addition, the traditional Cu-modified Al-Si-Mg alloy development method usually has problems of low efficiency, high cost and long design cycle in determining the optimal content of alloy components. Therefore, it is urgent to develop a design method, storage medium and equipment for Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy by combining computational thermodynamics and active learning methods to accelerate the design of new high-performance heat-treatment-free die-casting aluminum alloys. Summary of the invention

[0006] The object of the present invention is to provide a design method, storage medium and device for Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy. The specific technical scheme is as follows: In a first aspect, the present invention provides a method for designing an Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy, comprising: Step S1, collecting thermodynamic data of pure elements, binary systems, ternary systems and quaternary systems in the Al-rich end Al-Si-Mg-Cu quaternary system according to existing literature, and establishing a thermodynamic database of the Al-rich end Al-Si-Mg-Cu quaternary system based on the obtained thermodynamic data; Step S2: Based on the thermodynamic database, using the Scheil-Gulliver model to simulate the effect of different Cu addition amounts on the solidification structure of the Al-Si-Mg-Cu alloy, thereby obtaining the solidification structure information of the Al-Si-Mg-Cu alloy; Step S3, using the solidification structure information as input features, using the ultimate tensile strength, yield strength and elongation of the initial experimental performance data of the Al-Si-Mg-Cu alloy as output features; using the data corresponding to the input features and the output features to train a Gaussian regression model until the fitting coefficient of the Gaussian regression model for the initial experimental performance data and the predicted performance data reaches R 2 When it is greater than or equal to 0.9, the training is completed; Step S4, using the trained Gaussian regression model to predict within the preset alloy component content range of yAl-8wt.%Si-0.4wt.%Mg-xCu, to obtain predicted performance data that changes with the increase of Cu mass percentage and the decrease of Al mass percentage; wherein the value range of y is 90wt.%~91.6wt.%; the value range of x is 0~1.6wt.%; the predicted performance data is converted into a comprehensive mechanical property factor Q DJR According to Q DJR and the optimal composition point within the preset alloy composition content range is screened out in combination with the maximum value of the EI value.

[0007] Optionally, if the optimal composition point corresponds to the Q DJR The maximum value within its confidence interval is greater than the Q of the alloy corresponding to any other component point within the preset alloy component content range. DJR The maximum value within its confidence interval indicates that the optimal component point screening is successful; If the optimal composition point corresponds to the alloy's Q DJR The maximum value within its confidence interval is less than the Q value of the alloy corresponding to any other component point within the preset alloy component content range.DJR If the value is the maximum value within its confidence interval, it means that the optimal component point needs to be reselected.

[0008] Optionally, when reselecting the optimal composition point, each of the measured data is first added to the initial experimental performance data, and then the training of the Gaussian regression model is completed in combination with step S3. Finally, the optimal composition point within the preset alloy component content range is screened out using step S4.

[0009] Optionally, the initial experimental performance data includes measured data of the ultimate tensile strength, yield strength and elongation of the alloy corresponding to 5 to 10 component points selected within the preset alloy component content range.

[0010] Optionally, during the process of training the Gaussian regression model, the kernel function used by the Gaussian regression model includes a kernel function for training ultimate tensile strength, a kernel function for training yield strength, and a kernel function for training elongation.

[0011] Optionally, a Calphad method is used to establish a thermodynamic database for the Al-rich end Al-Si-Mg-Cu quaternary system for the obtained thermodynamic data.

[0012] Optionally, the predicted performance data includes ultimate tensile strength, yield strength and elongation.

[0013] Optionally, the solidification structure information includes (Al) phase structure information, (Si) phase structure information, Mg2Si phase structure information, Al2Cu phase structure information and Q-Al5Cu2Mg8Si6 phase structure information.

[0014] In a second aspect, the present invention provides a computer storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method.

[0015] In a third aspect, the present invention provides an electronic device comprising: at least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method.

[0016] The application of the technical solution of the present invention has at least the following beneficial effects: (1) The present invention provides a design method for an Al-Si-Mg-Cu heat-treatment-free die-cast aluminum alloy, which can quickly screen out the optimal composition point within a preset alloy component content range, thereby accelerating the design of a new high-performance heat-treatment-free die-cast aluminum alloy. Specifically, the present invention uses the Scheil-Gulliver model to simulate the effect of different Cu addition amounts on the solidification structure of the Al-Si-Mg-Cu alloy based on the Al-rich end Al-Si-Mg-Cu quaternary system thermodynamic database, thereby obtaining the solidification structure information of the Al-Si-Mg-Cu alloy; uses the solidification structure information as input features, uses the ultimate tensile strength, yield strength and elongation of the initial experimental performance data of the Al-Si-Mg-Cu alloy as output features, and uses the data corresponding to the input features and the output features to train a Gaussian regression model; uses the trained Gaussian regression model to make predictions within the preset alloy component content range of yAl-8wt.%Si-0.4wt.%Mg-xCu, and according to Q DJR The changes in the EI value are combined with the maximum value of the EI value to quickly screen out the optimal composition point within the preset alloy component content range. In addition, the alloy corresponding to the optimal composition point is verified by experiments to have the best comprehensive mechanical properties, ensuring the accuracy and reliability of the design method. Therefore, the present invention can combine computational thermodynamics and active learning methods to shorten the development cycle of the optimal composition point of the Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy, reduce development costs, and effectively improve the comprehensive performance of the alloy, meeting the requirements of lightweight and structural strength of automobiles.

[0017] (2) In the present invention, when reselecting the optimal composition point, each of the measured data is first added to the initial experimental performance data, and then the training of the Gaussian regression model is completed in combination with step S3. Finally, the optimal composition point within the preset alloy component content range is screened out by step S4. Through this dynamic adjustment and optimization, it is ensured that the optimal composition point within the preset alloy component content range is screened out, so that the finally obtained Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy has excellent mechanical properties and meets the high requirements in practical applications.

[0018] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1It is a schematic flow chart of a method for designing an Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy in Example 1; Figure 2 It is the comparison result of the calculated vertical cross-section phase diagram of the quaternary alloy system of 70wt.%Al-0.2wt.%Si-0.16wt.%Mg-29.64wt.%Cu to 70wt.%Al-0.2wt.%Si-0.28wt.%Mg-29.52wt.%Cu calculated by the thermodynamic database in Example 1 and the experimental data; Figure 3 is the non-equilibrium solidification phase diagram of the preset alloy at different Cu contents in Example 1; Figure 4 is a solidification microstructure diagram of the preset alloy at different Cu contents in Example 1; Figure 5 The Gaussian regression model in Example 1 predicts the ultimate tensile strength of the preset alloy compositions with different Cu contents; Figure 6 The Gaussian regression model in Example 1 predicts the yield strength of the preset alloy compositions with different Cu contents; Figure 7 The Gaussian regression model in Example 1 predicts the elongation of the preset alloy components with different Cu contents; Figure 8 The calculation result of the comprehensive mechanical properties obtained by calculating the predicted performance data in Example 1; Fig. 9 Calculate the expected improved EI value of the preset alloy composition with different Cu content in Example 1; Fig.10 It is a comparison chart of the experimental verification results of the optimal composition point with a Cu content of 0.62wt.% in Example 1 and the mechanical properties of the A356 alloy; Fig.11 The Gaussian regression model in comparative example 1 predicts the ultimate tensile strength of the preset alloy compositions with different Cu contents; Fig.12 The Gaussian regression model in comparative example 1 predicts the yield strength of the preset alloy compositions with different Cu contents; Fig.13 The prediction results of the Gaussian regression model in comparative example 1 for the elongation of preset alloy compositions with different Cu contents; Fig.14 The calculation results of the comprehensive mechanical properties obtained by calculating the predicted performance data in Comparative Example 1; Fig.15 The expected improved EI values ​​of the preset alloy compositions with different Cu contents calculated in Comparative Example 1; Among them, Figure 5~Figure 8 and Figure 11~Figure 14 The shaded areas near the curves in are the 95% confidence intervals; Figure 5~Figure 8 and Figure 10~Figure 14 In the table, UTS stands for ultimate tensile strength, YS stands for yield strength, and EL stands for elongation. DETAILED DESCRIPTION

[0020] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention. Embodiment 1:

[0021] See also Figure 1 , a design method for Al-Si-Mg-Cu heat treatment-free die-casting aluminum alloy, comprising: Step S1, according to existing literature, thermodynamic data of four pure elements (Al, Si, Mg and Cu) in the Al-rich end Al-Si-Mg-Cu quaternary system, six binary systems (Al-Si, Al-Mg, Al-Cu, Si-Mg, Si-Cu and Mg-Cu), four ternary systems (Al-Si-Mg, Al-Si-Cu, Al-Mg-Cu and Si-Mg-Cu) and one quaternary system (Al-Si-Mg-Cu) are collected, and a thermodynamic database of the Al-rich end Al-Si-Mg-Cu quaternary system is established using the Calphad method for the obtained thermodynamic data; wherein, the Al-rich end defined in this embodiment 1 means that the Al-Si-Mg-Cu quaternary system has a high mass percentage of Al, and the mass percentage of Al is 90wt.%~91.6wt.%; Specifically, the thermodynamic data of pure element Al, pure element Si, pure element Mg and pure element Cu are from the existing literature: DINSDALE A T. SGTE data for pure elements[J]. Calphad, 1991, 15(4): 317-425; The thermodynamic data of binary Al-Si, Al-Mg, Si-Mg and ternary Al-Si-Mg systems are from existing literature: TANG Y, DU Y, ZHANG L, et al. Thermodynamic description of the Al-Mg-Si system using a new formulation for the temperature dependence of the excess Gibbs energy[J]. Thermochimica Acta, 2012, 527: 131-142; Thermodynamic data of binary Al-Cu system comes from existing literature: LIANG SM, SCHMID-FETZER R. Thermodynamic assessment of the Al-Cu-Zn system, part II: Al-Cu binary system[J]. Calphad, 2015, 51: 252-260; The thermodynamic data of binary Si-Cu and ternary Al-Si-Cu are from existing literature: HALLSTEDT B, GRÖBNER J, HAMPL M, et al. Calorimetric measurements and assessment of the binary Cu-Si and ternary Al-Cu-Si phase diagrams[J]. Calphad, 2016, 53: 25-38; The thermodynamic data of binary Mg-Cu and ternary Al-Mg-Cu are from existing literature: BUHLER T, FRIESS G, SPENCER PJ, et al. A thermodynamic assessment of the Al-Cu-Mg ternary system[J]. Journal of Phase Equilibria, 1998, 19(4): 317-333; The thermodynamic data of the ternary Si-Mg-Cu system comes from the existing literature: ZHAO J, ZHOU J, LIU S, et al. Phase diagram determination and thermodynamic modeling of the Cu-Mg-Sisystem[J]. Journal of Mining and Metallurgy, Section B: Metallurgy, 2016, 52(1): 99-112; The thermodynamic data of the Q-Al5Cu2Mg8Si6 phase in the quaternary Al-Si-Mg-Cu system comes from the existing literature: PAN X, MORRAL JE, BRODY H D. Predicting the Q-phase in Al-Cu-Mg-Si alloys[J]. Journal of Phase Equilibria and Diffusion, 2010, 31(2): 144-148; The calculated vertical cross-section phase diagram (mass percentage, %) of the quaternary alloy system of 70wt.%Al-0.2wt.%Si-0.16wt.%Mg-29.64wt.%Cu to 70wt.%Al-0.2wt.%Si-0.28wt.%Mg-29.52wt.%Cu calculated using the thermodynamic database is compared with the experimental data. Figure 2 As shown (due to the large amount of data, only part of the results are shown here), Figure 2 This shows that the experimental data are basically consistent with the calculation results, indicating that the thermodynamic database is reliable.

[0022] Step S2, based on the thermodynamic database, using the Scheil-Gulliver model (from Pandat software) to simulate the effect of different Cu addition amounts on the solidification structure of the Al-Si-Mg-Cu alloy, so as to obtain the solidification structure information of the Al-Si-Mg-Cu alloy; Specifically, the Scheil-Gulliver model in the Pandat software is used to simulate the solidification behavior of the preset alloy of yAl-8wt.%Si-0.4wt.%Mg-xCu in the range of Cu content (i.e., mass percentage) of 0-1.6wt.%, and several composition points are selected to simulate the solidification behavior of the preset alloy of yAl-8wt.%Si-0.4wt.%Mg-xCu, and the non-equilibrium solidification phase diagram of the preset alloy is constructed, such as Figure 3As shown; where the value range of y is 90wt.%~91.6wt.%; the value range of x is 0~1.6wt.%; within the preset alloy component content range, the mass percentages of Mg and Si remain unchanged, the mass percentage of Cu increases (specifically the increase range is 0~1.6wt.%), and the mass percentage of Al decreases (specifically the decrease range is 91.6wt.%~90wt.%). Figure 3 It can be seen that the Cu content significantly affects the solidification path of the preset alloy: when the Cu content is less than 0.61wt.%, as the temperature decreases, the ternary eutectic reaction Liquid→(Al)+(Si)+Mg2Si and the ternary eutectic reaction Liquid→(Al)+(Si)+Q-Al5Cu2Mg8Si6 exist in the preset alloy solidification path at the same time. When the Cu content is higher than 0.61wt.%, the ternary eutectic reaction Liquid→(Al)+(Si)+Mg2Si in the preset alloy completely disappears, and only the ternary eutectic reaction Liquid→(Al)+(Si)+ Q-Al5Cu2Mg8Si6 exists, and the Mg2Si phase in the alloy is completely replaced by the Q-Al5Cu2Mg8Si6 phase.

[0023] Figure 4 is the solidification structure diagram of the preset alloy under different Cu contents. Figure 4It can be seen that the Cu content has a significant effect on the solidification structure of the preset alloy. As the Cu content increases, the content of the (Si) phase in the preset alloy decreases slightly, and the content of the (Al) phase remains basically stable. As the Cu content increases, the content of the (Al) phase remains stable, indicating that the Cu element has no significant inhibitory effect on the formation of the (Al) phase. The solubility of the Cu element is relatively low, so it will not have a significant effect on the stability of the (Al) phase. The content of the (Si) phase decreases slightly, which may be due to the eutectic or hypoeutectic reaction between the Cu element and the Si element, resulting in the conversion of part of Si into other compounds. In aluminum alloys, the Si element usually forms a eutectic structure with the Al matrix, so despite the addition of the Cu element, the (Si) phase remains relatively stable, and its content only decreases slightly. For the Mg2Si phase, its content decreases with the increase of the Cu content, and completely disappears when the Cu content is 0.61wt.%. For the Q-Al5Cu2Mg8Si6 phase, when the Cu content is less than 0.61wt.%, its content continues to increase with the Cu content; when the Cu content is greater than 0.61wt.%, the content of the Q-Al5Cu2Mg8Si6 phase tends to be stable. This phenomenon can be explained by the interaction between the preset alloy elements: when the Cu content is low, the Mg2Si phase in the preset alloy is relatively stable. As the Cu content increases, Cu reacts with Mg and Si to generate a new metal compound Q-Al5Cu2Mg8Si6 phase; in addition, the Cu element restricts the further combination of Mg and Si to form the Mg2Si phase by affecting the solid solution structure of the preset alloy; when the Cu content is greater than 0.61wt.%, the content of the Q-Al5Cu2Mg8Si6 phase tends to be stable. When the Cu content is high, there are not enough Mg atoms in the preset alloy to continue to generate the Q-Al5Cu2Mg8Si6 phase, so the phase fraction of the Q-Al5Cu2Mg8Si6 phase no longer increases significantly.

[0024] Step S3, using the solidification structure information as input features, the solidification structure information including (Al) phase structure information, (Si) phase structure information, Mg2Si phase structure information, Al2Cu phase structure information and Q-Al5Cu2Mg8Si6 phase structure information; using the ultimate tensile strength, yield strength and elongation of the initial experimental performance data of the Al-Si-Mg-Cu alloy as output features; using the data corresponding to the input features and the output features to train a Gaussian regression model until the fitting coefficient of the Gaussian regression model for the initial experimental performance data and the predicted performance data reaches R 2 When it is greater than or equal to 0.9, the training is completed; the ultimate tensile strength, yield strength, and elongation R of this embodiment 1 2They are: 0.9638706913760833, 0.9914384871024077, and 0.9864146081641261 respectively; in the process of training the Gaussian regression model, the kernel function and hyperparameters used by the Gaussian regression model are shown in Table 1.

[0025] Table 1 Kernel functions and hyperparameters of the Gaussian regression model used for each output feature in Example 1

[0026] Step S4, using the trained Gaussian regression model to predict within a preset alloy component content range, to obtain predicted performance data that changes with increasing Cu mass percentage and decreasing Al mass percentage; the predicted performance data includes ultimate tensile strength, yield strength and elongation; converting the predicted performance data into a comprehensive mechanical property factor Q DJR According to Q DJR and the optimal composition point within the preset alloy composition content range is screened out in combination with the maximum value of the EI value.

[0027] If the optimal composition point corresponds to the alloy's Q DJR The maximum value within its confidence interval is greater than the Q of the alloy corresponding to any other component point within the preset alloy component content range. DJR The maximum value within its confidence interval indicates that the optimal component point screening is successful; If the optimal composition point corresponds to the alloy's Q DJR The maximum value within its confidence interval is less than the Q value of the alloy corresponding to any other component point within the preset alloy component content range. DJR If the value is the maximum value within its confidence interval, it means that the optimal component point needs to be reselected.

[0028] When reselecting the optimal composition point, first add the measured data to the initial experimental performance data, then complete the training of the Gaussian regression model in combination with step S3, and finally, use step S4 to screen out the optimal composition point within the preset alloy composition content range.

[0029] The initial experimental performance data includes the measured data of the ultimate tensile strength, yield strength and elongation of the alloy corresponding to 7 component points selected within the preset alloy component content range. Specifically, high-purity Al single substance, high-purity Si single substance, high-purity Mg single substance and high-purity Cu single substance with a purity of 99.99 wt.% are used as raw materials, and a total of seven samples with different Cu contents are designed with serial numbers A1~A7, as shown in Table 2. The total mass of all raw materials is designed to be 75g, and an electronic balance with model AUY120 is used for weighing. Subsequently, the alloy samples were prepared by melting in a vacuum melting furnace of model CXZG-0.5. High-purity argon was used as the protective gas during the melting process to prevent the influence of impurity gas on the quality of the sample. The melting parameters used were derived from the existing literature: ZHANG S, YI W, ZHONG J, et al. Computer alloy design of Ti modified Al-Si-Mg-Sr casting alloys for achieving simultaneous enhancement in strengthand ductility[J]. Materials, 2023, 16(1): 306.

[0030] The alloy samples were prepared into powder samples using diamond files, each with a mass of 0.2 g. The composition of the alloy samples was analyzed using inductively coupled plasma mass spectrometry (ICP-MS) and chemical analysis (CA) methods. Some of the results are shown in Table 3. Subsequently, the alloy samples were prepared into tensile parts for mechanical property testing using a DK7725 wire cutting machine. Three tensile parts were prepared for each alloy sample, for a total of 21 tensile parts. Subsequently, the surfaces of the tensile parts were polished using metallographic sandpapers of 240 mesh, 400 mesh, 600 mesh, 800 mesh, 1500 mesh and 2000 mesh in sequence to remove the oxide layer, oil stains and microcracks on the surface of the tensile parts. Finally, the mechanical properties of each tensile part were tested using a universal testing machine of model America instron-3369. The test results are shown in Table 4. Among them, the ultimate tensile strength, yield strength and elongation were tested in accordance with GB / T 228.1-2021 "Tensile Test of Metallic Materials Part 1: Room Temperature Test Method".

[0031] Table 2 Seven samples with different Cu contents, numbered A1 to A7

[0032] Table 3 Composition analysis results of some alloy samples

[0033] Table 4 Mechanical properties of each tensile component

[0034] In step S4, the trained Gaussian regression model is used to predict the preset alloy components with different Cu contents, and the following is obtained: Figure 5~Figure 7 In addition, the data in Table 4 are marked on Figure 5~Figure 7 In. By Figure 5~Figure 7 It can be seen that the predicted points are completely consistent with the experimental points, and the confidence interval can well include the error value of the experimental points, indicating that the Gaussian regression model after training has good prediction accuracy.

[0035] Using comprehensive mechanical properties factor Q DJR =UTS+YS•log 10 (EL) Convert multiple objectives (i.e., UTS, YS, EL) into a single objective representing comprehensive mechanical properties to facilitate efficient global design. For specific conversion results, see Figure 8 Among them, the comprehensive mechanical properties factor Q DJR From the existing literature: GAO J, ZHONG J, LIU G, et al. Accelerated discovery of high-performance Al-Si-Mg-Sc casting alloys by integrating active learning with high-throughput CALPHAD calculations[J]. Science and Technology of Advanced Materials, 2023, 24(1): 2196242. The expected improvement function (i.e., Expected Improvement, EI) is used to calculate the optimal composition point within the preset alloy composition content range. The calculation results are shown in Fig. 9 Among them, the expected improvement function comes from the existing literature: GAO J, ZHONG J, LIU G, et al. Accelerated discovery of high-performance Al-Si-Mg-Sc casting alloys by integrating active learning with high-throughput CALPHAD calculations[J]. Science and Technology of Advanced Materials, 2023, 24(1): 2196242. According to Q DJRThe optimal composition point within the preset alloy composition content range is selected by combining the maximum value of the EI value. Specifically, the optimal composition point is 90.98wt.%Al-8wt.%Si-0.4wt.%Mg-0.62wt.%Cu. At this time, Q DJR =390.21394±15.57331 MPa, Q DJR The maximum value is 405.78725 MPa.

[0036] The method for preparing alloy samples and the mechanical properties testing method in the initial experimental performance data were used to obtain the measured mechanical properties data corresponding to the optimal composition point of 90.98wt.%Al-8wt.%Si-0.4wt.%Mg-0.62wt.%Cu. The results are shown in Fig.10 The new alloys in Table 5; Fig.10 The mechanical properties of the A356 alloy are also shown in the figure, which are measured using the mechanical properties test method in the initial experimental performance data; Table 5 also provides the method for preparing the alloy sample and the mechanical properties test method in the initial experimental performance data, and obtains the measured mechanical properties data corresponding to the 90.9wt.%Al-8wt.%Si-0.4wt.%Mg-0.7wt.%Cu composition point. In addition, the data in Table 5 are marked in Figure 5~Figure 8 middle.

[0037] Table 5 Experimental verification results of the preset alloys at different Cu contents

[0038] From Table 5 and Figure 5~Figure 8 It is known that the experimental verification results at two composition points with Cu content of 0.62wt.% and 0.7wt.% are consistent with the predicted performance data, and the error is negligible. This shows that the Gaussian regression model after training has good prediction accuracy. Therefore, the optimal composition alloy designed by the present invention within the preset alloy composition content range of yAl-8wt.%Si-0.4wt.%Mg-xCu is: 90.98wt.%Al-8wt.%Si-0.4wt.%Mg-0.62wt.%Cu.

[0039] See also Fig.10 It can be seen that the experimental verification results at the optimal composition point with a Cu content of 0.62wt.% are significantly better than the mechanical properties of A356 alloy.

[0040] Comparative Example 1: Different from Example 1, when training the Gaussian regression model in step S3, the preset alloy component content change is used as the input feature, wherein the mass percentages of Mg and Si remain unchanged, the mass percentage of Cu increases, specifically in the range of 0~1.6wt.%, and the mass percentage of Al decreases, specifically in the range of 91.6wt.%~90wt.%.

[0041] The Gaussian regression model is trained using the data corresponding to the input features and the output features. During the training of the Gaussian regression model, the kernel function and hyperparameters used by the Gaussian regression model are shown in Table 6.

[0042] Table 6 Kernel functions and hyperparameters of the Gaussian regression model used for each output feature in Comparative Example 1

[0043] The ultimate tensile strength, yield strength, and elongation of this comparative example 1 are R 2 They are 0.6128910488100194, 0.8055718322044677, and 0.6580101273505832 respectively; it can be seen that R 2 All of them are less than 0.9, and it is preliminarily judged that the Gaussian regression model after training in this comparative example 1 is not worthy of being adopted.

[0044] Further, the Gaussian regression model trained in this comparative example 1 is used to predict the preset alloy components with different Cu contents, and the following is obtained: Figure 11~Figure 13 In addition, the data in Table 2 are marked on Figure 11~Figure 13 In. By Figure 11~Figure 13 It can be seen that the predicted points are not completely consistent with the experimental points. The errors are large at Cu contents of 0, 0.5wt.% and 1.5wt.%, and the confidence interval cannot well include the error values ​​of the experimental points, indicating that the prediction accuracy of the Gaussian regression model after training is poor.

[0045] Using comprehensive mechanical properties factor Q DJR =UTS+YS•log 10 (EL) The multi-objective mechanical properties (i.e., UTS, YS, EL) are converted into a single objective representing the comprehensive mechanical properties to facilitate efficient global design. For specific conversion results, see Fig.14 The expected improvement function (Expected Improvement, EI) is calculated to determine the optimal composition point within the preset alloy composition content range. The calculation results are shown in Fig.15 According to Q DJRThe optimal composition point within the preset alloy composition content range is selected by combining the maximum value of the EI value. Specifically, the optimal composition point is 90wt.%Al-8wt.%Si-0.4wt.%Mg-1.6wt.%Cu. At this time, Q DJR The predictions do not describe the initial experimental performance data well, especially at 0.5 wt.% Cu. DJR The predicted results deviate seriously from the initial experimental performance data, further verifying that the results of comparative example 1 using the preset alloy component content change as the input feature are not worth considering.

[0046] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A design method for Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy, characterized in that: include: Step S1, collecting thermodynamic data of pure elements, binary systems, ternary systems and quaternary systems in the Al-rich end Al-Si-Mg-Cu quaternary system according to existing literature, and establishing a thermodynamic database of the Al-rich end Al-Si-Mg-Cu quaternary system based on the obtained thermodynamic data; Step S2: Based on the thermodynamic database, using the Scheil-Gulliver model to simulate the effect of different Cu addition amounts on the solidification structure of the Al-Si-Mg-Cu alloy, thereby obtaining the solidification structure information of the Al-Si-Mg-Cu alloy; Step S3, using the solidification structure information as input features, using the ultimate tensile strength, yield strength and elongation of the initial experimental performance data of the Al-Si-Mg-Cu alloy as output features; using the data corresponding to the input features and the output features to train a Gaussian regression model until the fitting coefficient of the Gaussian regression model for the initial experimental performance data and the predicted performance data reaches R 2 When it is greater than or equal to 0.9, the training is completed; Step S4, using the trained Gaussian regression model to predict within the preset alloy component content range of yAl-8wt.%Si-0.4wt.%Mg-xCu, to obtain predicted performance data that changes with the increase of Cu mass percentage and the decrease of Al mass percentage; wherein the value range of y is 90wt.%~91.6wt.%; the value range of x is 0~1.6wt.%; the predicted performance data is converted into a comprehensive mechanical property factor Q DJR According to Q DJR and the optimal composition point within the preset alloy composition content range is screened out in combination with the maximum value of the EI value.

2. The Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method according to claim 1, characterized in that: If the optimal composition point corresponds to the alloy's Q DJR The maximum value within its confidence interval is greater than the Q of the alloy corresponding to any other component point within the preset alloy component content range. DJR The maximum value within its confidence interval indicates that the optimal component point screening is successful; If the optimal composition point corresponds to the alloy's Q DJR The maximum value within its confidence interval is less than the Q value of the alloy corresponding to any other component point within the preset alloy component content range. DJR If the value is the maximum value within its confidence interval, it means that the optimal component point needs to be reselected.

3. The Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method according to claim 2, characterized in that: When reselecting the optimal composition point, first add the measured data to the initial experimental performance data, then complete the training of the Gaussian regression model in combination with step S3, and finally, use step S4 to screen out the optimal composition point within the preset alloy composition content range.

4. The Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method according to claim 2, characterized in that: The initial experimental performance data includes measured data of the ultimate tensile strength, yield strength and elongation of the alloy corresponding to 5 to 10 component points selected within the preset alloy component content range.

5. The Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method according to claim 1, characterized in that: In the process of training the Gaussian regression model, the kernel function used by the Gaussian regression model includes a kernel function for training ultimate tensile strength, a kernel function for training yield strength, and a kernel function for training elongation.

6. The Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method according to claim 1, characterized in that: The obtained thermodynamic data were used to establish a thermodynamic database for the Al-rich end Al-Si-Mg-Cu quaternary system using the Calphad method.

7. The Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method according to claim 1, characterized in that: The predicted performance data include ultimate tensile strength, yield strength and elongation.

8. The Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method according to claim 1, characterized in that: The solidification structure information includes (Al) phase structure information, (Si) phase structure information, Mg2Si phase structure information, Al2Cu phase structure information and Q-Al5Cu2Mg8Si6 phase structure information.

9. A computer storage medium, characterized in that Computer program instructions are stored thereon, and when the computer program instructions are executed by a processor, the Al-Si-Mg-Cu heat-treatment-free die-casting aluminum alloy design method as described in any one of claims 1 to 8 is implemented.

10. An electronic device, characterized in that: include: At least one processor, at least one memory, and computer program instructions stored in the memory, when the computer program instructions are executed by the processor, implement the Al-Si-Mg-Cu heat treatment-free die-casting aluminum alloy design method as described in any one of claims 1 to 8.

Images