Al-Si-Mg-Yb alloy design method, storage medium and equipment

Abstract

The invention relates to the technical field of aluminum alloys, in particular to an Al-Si-Mg-Yb alloy design method, a storage medium and equipment. The method comprises the following steps: acquiring microstructure data and mechanical property data; processing the microstructure data and the mechanical property data to obtain an optimal Gaussian regression prediction model; obtaining prediction data from the model; processing the prediction data to obtain experimental data; and an optimal Gaussian regression prediction model is trained, and a component point with the optimal comprehensive mechanical property in the Al-Si-Mg-Yb alloy system is found by means of the optimal Gaussian regression prediction model. Computer program instructions are stored on the storage medium, and the method is implemented when the instructions are executed by the processor. The apparatus includes at least one processor, at least one memory, and computer program instructions stored in the memory that implement the method when executed by the processor. According to the method, the Al-Si-Mg-Yb alloy composition with the optimal comprehensive mechanical property can be rapidly screened out.

Description

Technical Field

[0001] The present invention relates to the technical field of aluminum alloys, and particularly to a design method, a storage medium, and a device for an Al-Si-Mg-Yb alloy. Background Art

[0002] Aluminum alloys have characteristics such as high specific strength, high specific stiffness, good electrical and thermal conductivity, good corrosion resistance, and easy processing. They are ideal lightweight materials and are widely used in fields such as automobiles, aerospace, and mechanical manufacturing.

[0003] Traditional hypoeutectic Al-Si-Mg has a small alloy density, good die-casting fluidity, strong plasticity, high tensile and compressive properties, fast heat dissipation, and good corrosion resistance, making it a good material for automotive structural parts. However, hypoeutectic Al-Si-Mg alloys often contain coarse primary crystals and lamellar eutectics. These coarse primary crystal structures and lamellar eutectic structures will cause stress concentration and become crack sources during service, seriously affecting the life of the castings.

[0004] In industrial production, rare earth elements are often added to refine the microstructure of Al-Si-Mg alloys and improve their mechanical properties. The rare earth element Yb easily forms ternary compounds with Al and Si, inhibits the growth of eutectic silicon, transforms the lamellar eutectic silicon into a fibrous or fine lamellar morphology, improves the comprehensive mechanical properties, and is affordable.

[0005] Currently, the industry mainly uses the "trial and error method" to study the modification effect of Yb content on alloys. However, due to the large number of alloying elements and the wide composition space, the R & D efficiency using the "trial and error method" is low and cannot meet the needs of today's rapidly developing society.

[0006] With the development of computer technology, various theoretical calculation-assisted material design methods have emerged. Among them, for multi-component alloy systems, computational thermodynamics and machine learning technologies have shown their unique design advantages. Through the computational thermodynamics method based on a thermodynamic database, the relationship between alloy composition and solidification structure can be efficiently obtained. Through the data-driven machine learning technology, the mutual relationship between alloy structure and properties can be connected. Coupling computational thermodynamics and machine learning technologies can obtain the quantitative relationship between alloy composition-structure-properties and accelerate the screening of alloy compositions.

[0007] In summary, it is necessary to develop a design method, a storage medium, and a device for an Al-Si-Mg-Yb alloy to quickly screen out the Al-Si-Mg-Yb alloy composition with the optimal comprehensive mechanical properties. Summary of the Invention

[0008] The object of the present invention is to provide a design method, a storage medium, and a device for an Al-Si-Mg-Yb alloy. The specific technical solutions are as follows:

[0009] In a first aspect, the present invention provides an Al-Si-Mg-Yb alloy design method, comprising:

[0010] Step S1, collecting thermodynamic descriptions of all boundary binary and ternary systems in the Al-Si-Mg-Yb alloy system reported in existing literature, and establishing a thermodynamic database of the Al-Si-Mg-Yb alloy system by thermodynamic extrapolation methods in existing literature; performing high-throughput Hill solidification simulation on the thermodynamic database, calculating solidification curves of all component points in the Al-Si-Mg-Yb alloy system, and obtaining its solidification microstructure data; collecting mechanical property data of the Al-Si-Mg-Yb alloy system in existing literature;

[0011] Step S2: pre-process the microstructure data and the mechanical properties data to form a data set required for machine learning, and then construct multiple Gaussian regression prediction models after processing them using a machine learning Gaussian regression algorithm. The coefficient of determination R is calculated by K-fold cross validation. 2 Determine the optimal Gaussian regression prediction model; use the optimal Gaussian regression prediction model to predict the Al-Si-Mg-Yb alloy system under the initially set wide composition range to obtain prediction data;

[0012] Step S3, finding the Pareto front in the predicted data to obtain multiple recommended composition points of the Al-Si-Mg-Yb alloy system, performing cluster analysis on each of the recommended composition points, and determining the optimal number of clusters using the Davies-Bouldin index, finding the recommended composition point with the highest EI value in each cluster and performing an ingot casting experiment based on the recommended composition point, and obtaining the mechanical property experimental data of the ingot;

[0013] Step S4: if the experimental data is within the predicted data range and the EI value is less than or equal to 0.003, it means that the optimal Gaussian regression prediction model has been trained; if the experimental data exceeds the predicted data range, the experimental data is added to the mechanical property data in step S1 and input into the optimal Gaussian regression prediction model to continue training until the training is completed;

[0014] Step S5: using the trained optimal Gaussian regression prediction model to find the component point with the best comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system.

[0015] Optionally, the microstructure data includes the phase fraction of each phase in the Al-Si-Mg-Yb alloy system;

[0016] The mechanical properties data include ultimate tensile strength, yield strength and elongation.

[0017] Optionally, the prediction data includes the ultimate tensile strength ± error range, the yield strength ± error range, and the elongation ± error range.

[0018] Optionally, the mass percentages of the raw material components of the Al-Si-Mg-Yb alloy system under the initially set broad composition range are as follows: silicon 4% - 10%, magnesium 0 - 0.7%, ytterbium 0 - 1.2%, the total content of each impurity element is less than or equal to 0.1%, the content of a single impurity element is less than or equal to 0.03%, and the balance is aluminum.

[0019] Optionally, the coefficient of determination R 2 for K-fold cross-validation is calculated as follows:

[0020]

[0021] where y i represents the true value of the i-th mechanical property data; represents the predicted value of the i-th mechanical property data; represents the mean value of all the mechanical property data; i represents the i-th mechanical property data among all the mechanical property data;

[0022] If R 2 is greater than or equal to 0.8, it indicates that the corresponding Gaussian regression prediction model is the optimal Gaussian regression prediction model and is retained;

[0023] If R 2 is less than 0.8, the corresponding Gaussian regression prediction model is discarded.

[0024] Optionally, the calculation formula for the EI value is as follows:

[0025] EI = EI(UTS) * EI(YS) * EI(EL);

[0026] where EI(UTS) represents the EI value of the ultimate tensile strength; EI(YS) represents the EI value of the yield strength; EI(EL) represents the EI value of the elongation;

[0027] The EI(UTS), the EI(YS), and the EI(EL) are all calculated using the following formula:

[0028]

[0029] Where, when EI(x) represents the EI value of the ultimate tensile strength; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of the Al-Si-Mg-Yb alloy reported in the existing literature; n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature; n + 1 represents the ingot number formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at the position numbered n + 1; f n+1 represents the ultimate tensile strength of the Al-Si-Mg-Yb alloy system at the position numbered n + 1 predicted by using the optimal Gaussian regression prediction model; represents the maximum value among all the ultimate tensile strength experimental data; D n represents the ultimate tensile strength experimental data corresponding to n; p(f n+1 |x n+1 ,D n ) represents the probability distribution of f n at x n+1 under D n+1 ;

[0030] When EI(x) represents the EI value of the yield strength; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of the Al-Si-Mg-Yb alloy reported in the existing literature; n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature; n + 1 represents the ingot number formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at the position numbered n + 1; f n+1 represents the yield strength of the Al-Si-Mg-Yb alloy system at the position numbered n + 1 predicted by using the optimal Gaussian regression prediction model; represents the maximum value among all the yield strength experimental data; D n represents the yield strength experimental data corresponding to n; p(f n+1 |x n+1 ,D n ) represents the probability distribution of f n at x n+1 under D n+1 ;

[0031] When EI(x) represents the EI value of the elongation; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of the Al-Si-Mg-Yb alloy reported in the existing literature; n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature; n + 1 represents the ingot number formed in the next ingot experiment after the current ingot experiment; x n+1Represents the microstructure information of the Al-Si-Mg-Yb alloy at the (n + 1)-th position; f n+1 Represents the elongation of the Al-Si-Mg-Yb alloy system at the (n + 1)-th position predicted using the optimal Gaussian regression prediction model; Represents the maximum value among all the elongation experimental data; D n Represents the elongation experimental data corresponding to n; p(f n+1 |x n+1 ,D n ) represents that under D n , the probability distribution of f n+1 at x n+1 .

[0032] Optionally, in the step S4, for the continued training, it is necessary to use the optimal Gaussian regression prediction model to re-predict the Al-Si-Mg-Yb alloy system under the initially set wide composition range to obtain updated prediction data; the updated prediction data is processed through the step S3 to obtain updated experimental data;

[0033] If the updated experimental data is within the range of the updated prediction data and the EI value is less than or equal to 0.003, it indicates that the training of the optimal Gaussian regression prediction model is completed;

[0034] If the updated experimental data exceeds the range of the updated prediction data, the updated experimental data is supplemented to the mechanical property data in the step S1 and input into the optimal Gaussian regression prediction model for continued training until the training is completed.

[0035] Optionally, in the step S5, the comprehensive mechanical property is denoted as Q, and its calculation formula is as follows:

[0036] Q = UTS + YS * Log(EL);

[0037] Wherein, UTS represents the ultimate tensile strength; YS represents the yield strength; EL represents the elongation.

[0038] In a second aspect, the present invention provides a computer storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the Al-Si-Mg-Yb alloy design method is implemented.

[0039] In a third aspect, the present invention provides an electronic device, including: at least one processor, at least one memory, and computer program instructions stored in the memory, and when the computer program instructions are executed by the processor, the Al-Si-Mg-Yb alloy design method is implemented.

[0040] Applying the technical solution of the present invention has at least the following beneficial effects:

[0041] The Al-Si-Mg-Yb alloy design method provided by the present invention couples computational thermodynamics and machine learning technologies, deeply excavates existing data, establishes an Al-Si-Mg-Yb alloy system, can quickly screen out the Al-Si-Mg-Yb alloy composition with the optimal comprehensive mechanical properties, realizes the efficient design of the Al-Si-Mg-Yb alloy system, significantly accelerates the research and development of new materials, and meets the needs of various fields for high-performance aluminum alloys. Specifically, the present invention obtains the microstructure data and mechanical property data of existing literatures by using step S1; processes the microstructure data and mechanical property data to obtain the optimal Gaussian regression prediction model by using step S2; predicts the Al-Si-Mg-Yb alloy system under a wide initial set composition range by the optimal Gaussian regression prediction model to obtain prediction data; processes the prediction data to obtain experimental data by using step S3; obtains the optimal Gaussian regression prediction model after training by using step S4; and finds out the composition point with the optimal comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system by means of the optimal Gaussian regression prediction model after training by using step S5. In addition, the present invention modifies the traditional hypoeutectic Al-Si-Mg alloy with the inexpensive rare earth element Yb, refines the eutectic silicon, generates intermediate compounds, and realizes the coordinated improvement of the strength and toughness of the Al-Si-Mg alloy.

[0042] In addition to the purposes, features and advantages described above, the present invention has other purposes, features and advantages. The following will refer to the drawings for a further detailed description of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0044] Figure 1 is a schematic flow chart of the Al-Si-Mg-Yb alloy design method in Embodiment 1;

[0045] Figure 2 is a schematic diagram of the verification result of the vertical section thermodynamics calculation from the pure ternary composition of 92.65% Al 7% Si 0.35% Mg to the pure quaternary composition of 91.15% Al 7% Si 0.35% Mg 1.5% Yb of the Al-Si-Mg-Yb alloy system;

[0046] Figure 3Schematic diagram of the verification result of the thermodynamic calculation of the vertical section from the pure ternary composition of 90.75% Al 9% Si 0.25% Mg to the pure quaternary composition of 89.25% Al 9% Si 0.25% Mg 1.5% Yb in the Al-Si-Mg-Yb alloy system;

[0047] Figure 4 It is the prediction result diagram of the ultimate tensile strength UTS in Example 1;

[0048] Figure 5 It is the prediction result diagram of the yield strength YS in Example 1;

[0049] Figure 6 It is the prediction result diagram of the elongation EL in Example 1;

[0050] Figure 7 It is the tensile curve diagram of aluminum alloy A356 and Examples 1 to 3. Specific implementation manners

[0051] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present invention.

[0052] Example 1:

[0053] Refer to Figure 1 , the Al-Si-Mg-Yb alloy design method includes:

[0054] Step S1, collect existing literature (Literature 1: GUO C, DU Z. Thermodynamic optimization of the Mg–Tb and Mg–Yb systems [J]. Journal of Alloys and Compounds, 2006, 422(1–2): 102–108; Literature 2: BORZONE G, PARODI N, FERRO R. The magnesium-ytterbium system: A contribution to the thermodynamics of solid alloys [J]. International Journal of Materials Research, 2006, 97(4): 417–421; Literature 3: PREDEL B. Mg-Yb (Magnesium-Ytterbium) [M]. MADELUNG O. / / Li-Mg–Nd-Zr. Berlin / Heidelberg: Springer-Verlag, 1997: 1–2 [2023-02-24]; Literature 4: MCMASTERS O D, GSCHNEIDNER K A. Ytterbium-magnesium system [J]. Journal of the Less Common Metals, 1965, 8(5): 289–298; Literature 5: 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; Literature 6: ZHANG M, GAO J, YI W, et al. Thermodynamic descriptions of ternary Al–Si–Yb system and their application to understand solidification behaviors of Yb-modified Al–Si alloys [J]. Calphad, 2023, 83: 102625.)Thermodynamic descriptions of all the boundary binary and ternary systems in the reported Al-Si-Mg-Yb alloy system are used to establish a thermodynamic database for the Al-Si-Mg-Yb alloy system through the thermodynamic extrapolation method in the existing literature (Reference 7: MUGGIANU Y-M, GAMBINO M, BROS J-P. Enthalpies de formation des alliages liquides bismuth-étain-gallium à 723k. Choix d’une représentation analytique des grandeurs d’excès intégrales et partielles de mélange[J]. Journal de Chimie Physique, 1975, 72:83–88.); see. Figure 2 and Figure 3, the verification results of the thermodynamic calculations of the vertical section from the pure ternary composition of 92.65% Al 7% Si 0.35% Mg to the pure quaternary composition of 91.15% Al 7% Si 0.35% Mg 1.5% Yb in the Al-Si-Mg-Yb alloy system and the verification results of the thermodynamic calculations of the vertical section from the pure ternary composition of 92.65% Al 7% Si 0.35% Mg to the pure quaternary composition of 91.15% Al 7% Si 0.35% Mg 1.5% Yb in the Al-Si-Mg-Yb alloy system both conform to the said thermodynamic database; perform high-throughput Scheil solidification simulation on the said thermodynamic database, calculate the solidification curves of all composition points in the Al-Si-Mg-Yb alloy system, and obtain the solidification microstructure data; collect existing literature (Literature 8: XUH. Effects of Yb addition on microstructure and mechanical properties of A356 aluminum alloys[J]. Special Casting & Nonferrous Alloys, 2017, 37(8): 827–830; Literature 9: LI B, WANG H, JIE J, et al. Microstructure evolution and modification mechanism of the ytterbium modified Al–7.5%Si–0.45%Mg alloys[J]. Journal of Alloys and Compounds, 2011, 509(7): 3387–3392; Literature 10: ZHANG S, LENG J, LI C, et al. Influence of Yb modification on the microstructure and mechanical properties of A356.2 aluminum alloy[J]. Materials Science Forum, 2017, 898: 259–264; Literature 11: ZHANG S, WANG Z, DONG Y, et al. Aging behavior of A356.2 with Yb modified[J]. Materials Science Forum, 2018, 913: 90–95; Literature 12: GAO J, ZHONG J, LIU G, et al.Mechanical property data of the Al-Si-Mg-Yb alloy system in (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; Literature 13: YI W,LIU G,LU Z,etal.Efficient alloy design of Sr-modified A356 alloys driven by computational thermodynamics and machine learning[J].Journal of Materials Science & Technology,2022,112:277–290; Literature 14: SHEN Q,YIN Q,ZHAO H,et al.Inversely optimized design of Al-Mg-Si alloys using machine learning methods[J].Computational Materials Science,2024,242:113107.).

[0055] Step S2: After sequentially cleaning and normalizing the microstructure data and the mechanical property data, a dataset required for machine learning is formed. Then, multiple Gaussian regression prediction models are constructed through processing by the Gaussian regression algorithm of machine learning. The determination coefficient R of K-fold cross-validation (K takes the value of 4) is used to 2 determine the optimal Gaussian regression prediction model; the optimal Gaussian regression prediction model is used to predict the Al-Si-Mg-Yb alloy system under a wide initially set composition range to obtain prediction data;

[0056] Step S3: Find the Pareto front in the prediction data to obtain multiple recommended composition points of the Al-Si-Mg-Yb alloy system. Cluster analysis is performed on each of the recommended composition points, and the Davies-Bouldin index is used to determine the optimal number of clusters. The recommended composition point with the highest EI value in each cluster is found and an ingot experiment is carried out based on this to obtain the mechanical property experimental data of the ingot;

[0057] Step S4: Refer to Figures 4 to 6, if the experimental data is within the predicted data range and the EI value is less than or equal to 0.003, it indicates that the improvement space of the optimal Gaussian regression prediction model is small and the training is completed; if the experimental data exceeds the predicted data range, the experimental data will be supplemented to the mechanical property data in the step S1 and input into the optimal Gaussian regression prediction model for continuous training until the training is completed;

[0058] Step S5: Use the optimal Gaussian regression prediction model after training to find out the composition point with the optimal comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system.

[0059] The microstructure data includes the phase fractions of each phase in the Al-Si-Mg-Yb alloy system;

[0060] The mechanical property data includes ultimate tensile strength, yield strength and elongation;

[0061] The predicted data includes ultimate tensile strength ± error range, yield strength ± error range and elongation ± error range.

[0062] The mass percentages of the raw material components of the Al-Si-Mg-Yb alloy system under the initially set wide composition range are as follows: silicon 4% - 10%, magnesium 0 - 0.7%, ytterbium 0 - 1.2%, the total content of each impurity element is less than or equal to 0.1%, the content of a single impurity element is less than or equal to 0.03%, and the balance is aluminum.

[0063] The determination coefficient R 2 calculated by K-fold cross-validation is as follows:

[0064]

[0065] where, y i represents the true value of the i-th mechanical property data; represents the predicted value of the i-th mechanical property data; represents the mean value of all mechanical property data; i represents the i-th mechanical property data of all mechanical property data;

[0066] If R 2 is greater than or equal to 0.8, it indicates that the corresponding Gaussian regression prediction model is the optimal Gaussian regression prediction model and is retained;

[0067] If R 2 is less than 0.8, the corresponding Gaussian regression prediction model is discarded.

[0068] The calculation formula of the EI value is as follows:

[0069] EI = EI(UTS) * EI(YS) * EI(EL);

[0070] Among them, EI(UTS) represents the EI value of the ultimate tensile strength; EI(YS) represents the EI value of the yield strength; EI(EL) represents the EI value of the elongation;

[0071] The EI(UTS), EI(YS), and EI(EL) are all calculated by the following formula:

[0072]

[0073] Among them, when EI(x) represents the EI value of the ultimate tensile strength; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of the Al-Si-Mg-Yb alloy reported in the existing literature (i.e., Literature 8 - Literature 14); n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature (i.e., Literature 8 - Literature 14); n + 1 represents the ingot number formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at the position of n + 1; f n+1 represents the ultimate tensile strength of the Al-Si-Mg-Yb alloy system at the position of n + 1 predicted by using the optimal Gaussian regression prediction model; represents the maximum value among all the ultimate tensile strength experimental data; D n represents the ultimate tensile strength experimental data corresponding to n; p(f n+1 |x n+1 ,D n ) represents the probability distribution of f n at x n+1 under D n+1 ;

[0074] When EI(x) represents the EI value of the yield strength; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of the Al-Si-Mg-Yb alloy reported in the existing literature (i.e., Literature 8 - Literature 14); n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature (i.e., Literature 8 - Literature 14); n + 1 represents the ingot number formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at the position of n + 1; f n+1 represents the yield strength of the Al-Si-Mg-Yb alloy system at the position of n + 1 predicted by using the optimal Gaussian regression prediction model; represents the maximum value among all the yield strength experimental data; D n represents the yield strength experimental data corresponding to n; p(fn+1 |x n+1 ,D n ) represents the probability distribution of f at x under D; n Under D, x n+1 at the position of f n+1 probability distribution;

[0075] When EI(x) represents the EI value of elongation; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of the Al-Si-Mg-Yb alloy reported in the existing literature (i.e., Literature 8 - Literature 14); n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature (i.e., Literature 8 - Literature 14); n + 1 represents the number of the ingot formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at the position of n + 1; f n+1 represents the elongation of the Al-Si-Mg-Yb alloy system at the position of n + 1 predicted by using the optimal Gaussian regression prediction model; represents the maximum value among all the elongation experimental data; D n represents the elongation experimental data corresponding to n; p(f n+1 |x n+1 ,D n ) represents the probability distribution of f at x under D; n Under D, x n+1 at the position of f n+1 probability distribution.

[0076] In the step S4, for the continued training, it is necessary to use the optimal Gaussian regression prediction model to re-predict the Al-Si-Mg-Yb alloy system under the initially set wide composition range to obtain updated prediction data; the updated prediction data is processed through the step S3 to obtain updated experimental data;

[0077] If the updated experimental data is within the range of the updated prediction data and the EI value is less than or equal to 0.003, it indicates that the training of the optimal Gaussian regression prediction model is completed;

[0078] If the updated experimental data exceeds the range of the updated prediction data, the updated experimental data is supplemented to the mechanical property data in the step S1 and input into the optimal Gaussian regression prediction model for continued training until the training is completed.

[0079] In the step S5, the comprehensive mechanical property is denoted as Q, and its calculation formula is as follows:

[0080] Q = UTS + YS * Log(EL);

[0081] Among them, UTS represents the ultimate tensile strength; YS represents the yield strength; EL represents the elongation rate.

[0082] The method of the ingot experiment is as follows:

[0083] Layer the Al-Si-Mg-Yb alloy raw materials at the recommended composition points into the crucible. First, lay half of the amount of pure aluminum evenly at the bottom of the crucible; second, lay pure magnesium and pure silicon evenly on the pure aluminum in sequence; then, lay the remaining half of the amount of pure aluminum on the pure silicon; finally, place the crucible in a melting furnace (specifically, a CXZG-0.5 type vacuum induction melting furnace), turn on the vacuum pump, and fill it with argon for gas washing until the air pressure reaches 500 Pa, and melt the alloy by heating in stages; during the staged heating and melting of the alloy, first, under a current of 200 - 210 A (specifically 210 A), heat for 200 - 300 s (specifically 240 s) to raise the temperature inside the melting furnace to 620 - 630 °C (specifically 625 °C), at this time the pure aluminum starts to melt and a melt appears; second, raise the current to 230 - 240 A (specifically 240 A), keep it for 300 s, heat the furnace temperature to 710 - 720 °C (specifically 720 °C), continue to raise the current to 245 - 255 A (specifically 250 A), and raise the furnace temperature to 730 - 740 °C (specifically 735 °C); then, shake the crucible at a frequency of 1 time per second to ensure the alloying of the melt; finally, turn off the power supply, and when the temperature inside the furnace drops to 665 - 675 °C (specifically 670 °C), pour the melt into a graphite mold to obtain a cylindrical ingot (the cylindrical ingot is 100 mm high and 20 mm in diameter).

[0084] Use wire cutting to cut out tensile test specimens from the center of the cylindrical ingot and use them for mechanical property testing to obtain the mechanical property experimental data of the cylindrical ingot.

[0085] Adopt step S5 to find that the composition point with the best comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system is 88.77% Al - 10% Si - 0.45% Mg - 0.78% Yb.

[0086] Example 2:

[0087] Different from Example 1, adopt the Al-Si-Mg-Yb alloy design method described in Example 1 to find that the composition point with the best comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system is 94.25% Al - 5% Si - 0.45% Mg - 0.3% Yb.

[0088] Example 3:

[0089] Different from Example 1, the component point with the optimal comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system was found by using the Al-Si-Mg-Yb alloy design method described in Example 1, which is 92.26% Al-7% Si-0.25% Mg-0.49% Yb.

[0090] The ingot experiments were respectively carried out on the component points with the optimal comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system found in Examples 1 to 3, and the mechanical property experimental data of the ingots are shown in Table 1. Among them, Table 1 also shows the aluminum alloy with the grade of A356. In Table 1, the mechanical properties were tested in accordance with GB / T 228.1-2021 "Metallic materials - Tensile testing - Part 1: Method of test at room temperature".

[0091] Table 1 Mechanical property test results of Examples 1 to 3

[0092] Number Ultimate tensile strength (MPa) Yield strength (MPa) Elongation (%) Example 1 242.54 134.34 7.92 Example 2 203.68 97.94 16.68 Example 3 224.17 112.05 14.79 Aluminum alloy A356 206.62 109.60 7.19

[0093] The ingot experiments were respectively carried out on the component points with the optimal comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system found in Examples 1 to 3 to obtain the ingots. The tensile curve tests were completed on the ingots obtained in Examples 1 to 3 and the aluminum alloy with the grade of A356 in accordance with GB / T228.1-2021 "Metallic materials - Tensile testing - Part 1: Method of test at room temperature". The test results are as Figure 7 shown.

[0094] From Figure 7 and Table 1, it can be seen that in Example 1, while ensuring that the elongation rate exceeds that of A356, the ultimate tensile strength and yield strength are increased by 35.92 MPa and 24.74 MPa respectively; in Example 2, while ensuring that the strength is equivalent to that of A356, the elongation rate is 9.49% higher than that of A356; the ultimate tensile strength, yield strength and elongation rate of Example 3 are 17.55 MPa, 2.45 MPa and 7.6% higher than those of A356 respectively.

[0095] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. 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. Al-Si-Mg-Yb alloy design method, characterized in that: include: Step S1, collecting thermodynamic descriptions of all boundary binary and ternary systems in the Al-Si-Mg-Yb alloy system reported in existing literature, and establishing a thermodynamic database of the Al-Si-Mg-Yb alloy system by thermodynamic extrapolation methods in existing literature; performing high-throughput Hill solidification simulation on the thermodynamic database, calculating solidification curves of all component points in the Al-Si-Mg-Yb alloy system, and obtaining its solidification microstructure data; Collect the mechanical properties data of Al-Si-Mg-Yb alloy system in existing literature; Step S2: pre-process the microstructure data and the mechanical properties data to form a data set required for machine learning, and then construct multiple Gaussian regression prediction models after processing them using a machine learning Gaussian regression algorithm. The coefficient of determination R is calculated by K-fold cross validation. 2 Determine the optimal Gaussian regression prediction model; use the optimal Gaussian regression prediction model to predict the Al-Si-Mg-Yb alloy system under the initially set wide composition range to obtain prediction data; Step S3, finding the Pareto front in the predicted data to obtain multiple recommended composition points of the Al-Si-Mg-Yb alloy system, performing cluster analysis on each of the recommended composition points, and determining the optimal number of clusters using the Davies-Bouldin index, finding the recommended composition point with the highest EI value in each cluster and performing an ingot casting experiment based on the recommended composition point, and obtaining the mechanical property experimental data of the ingot; Step S4: if the experimental data is within the predicted data range and the EI value is less than or equal to 0.003, it means that the optimal Gaussian regression prediction model has been trained; if the experimental data exceeds the predicted data range, the experimental data is added to the mechanical property data in step S1 and input into the optimal Gaussian regression prediction model to continue training until the training is completed; Step S5: using the trained optimal Gaussian regression prediction model to find the component point with the best comprehensive mechanical properties in the Al-Si-Mg-Yb alloy system.

2. The Al-Si-Mg-Yb alloy design method according to claim 1, characterized in that: The microstructure data includes the phase fraction of each phase in the Al-Si-Mg-Yb alloy system; The mechanical properties data include ultimate tensile strength, yield strength and elongation.

3. The Al-Si-Mg-Yb alloy design method according to claim 1, characterized in that: The predicted data include ultimate tensile strength ± error interval, yield strength ± error interval and elongation ± error interval.

4. The Al-Si-Mg-Yb alloy design method according to claim 1, characterized in that: The mass percentages of the raw material components of the Al-Si-Mg-Yb alloy system under the initially set wide composition range are as follows: silicon 4% to 10%, magnesium 0 to 0.7%, ytterbium 0 to 1.2%, the total content of each impurity element is less than or equal to 0.1%, the content of a single impurity element is less than or equal to 0.03%, and aluminum is the balance.

5. The Al-Si-Mg-Yb alloy design method according to claim 1, characterized in that: Coefficient of determination R using K-fold cross validation 2 The calculation formula is as follows: Among them, y i represents the true value of the i-th mechanical property data; represents the predicted value of the i-th mechanical property data; represents the mean value of all the mechanical property data; i represents the i-th mechanical property data of all the mechanical property data; If R 2 If it is greater than or equal to 0.8, it means that the corresponding Gaussian regression prediction model is the optimal Gaussian regression prediction model and is retained; If R 2 If it is less than 0.8, the corresponding Gaussian regression prediction model will be discarded.

6. The Al-Si-Mg-Yb alloy design method according to claim 1, characterized in that: The calculation formula of the EI value is as follows: EI=EI(UTS)*EI(YS)*EI(EL); Among them, EI(UTS) represents the EI value of ultimate tensile strength; EI(YS) represents the EI value of yield strength; EI(EL) represents the EI value of elongation; The EI (UTS), the EI (YS) and the EI (EL) are all calculated using the following formula: Wherein, when EI(x) represents the EI value of the ultimate tensile strength; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of the Al-Si-Mg-Yb alloy reported in the existing literature; n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature; n+1 represents the number of the ingot formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at position n+1; f n+1 represents the ultimate tensile strength of the Al-Si-Mg-Yb alloy system at position n+1 predicted by using the optimal Gaussian regression prediction model; Represents the maximum value of all ultimate tensile strength test data; D n represents the ultimate tensile strength experimental data corresponding to n; p(f n+1 |x n+1 ,D n ) means in D n Next, x n+1 F n+1 The probability distribution of When EI(x) represents the EI value of yield strength; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of Al-Si-Mg-Yb alloy reported in the existing literature; n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature; n+1 represents the number of the ingot formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at position n+1; f n+1 represents the yield strength of the Al-Si-Mg-Yb alloy system at position n+1 predicted by using the optimal Gaussian regression prediction model; Represents the maximum value of all yield strength test data; D n represents the yield strength experimental data corresponding to n; p(f n+1 |x n+1 ,D n ) means in D n Next, x n+1 F n+1 The probability distribution of When EI(x) represents the EI value of elongation; x represents the microstructure information corresponding to the alloy composition of the mechanical properties of Al-Si-Mg-Yb alloy reported in the existing literature; n represents the number of composition points of the Al-Si-Mg-Yb alloy system in the existing literature; n+1 represents the number of the ingot formed in the next ingot experiment after the current ingot experiment; x n+1 represents the microstructure information of the Al-Si-Mg-Yb alloy at position n+1; f n+1 It represents the elongation of the Al-Si-Mg-Yb alloy system at position n+1 predicted by using the optimal Gaussian regression prediction model; Indicates the maximum value of all elongation test data; D n represents the experimental data of elongation corresponding to n; p(f n+1 |x n+1 ,D n ) means in D n Next, x n+1 F n+1 The probability distribution of .

7. The Al-Si-Mg-Yb alloy design method according to claim 1, characterized in that: In step S4, the continued training requires using the optimal Gaussian regression prediction model to re-predict the Al-Si-Mg-Yb alloy system under the initially set wide composition range to obtain updated prediction data; the updated prediction data is processed through step S3 to obtain updated experimental data; If the updated experimental data is within the updated prediction data range, and the EI value is less than or equal to 0.003, it means that the training of the optimal Gaussian regression prediction model is completed; If the updated experimental data exceeds the updated prediction data range, the updated experimental data is added to the mechanical property data in step S1 and input into the optimal Gaussian regression prediction model to continue training until the training is completed.

8. The Al-Si-Mg-Yb alloy design method according to claim 1, characterized in that: The comprehensive mechanical properties in step S5 are recorded as Q, which is calculated using the following formula: Q = UTS + YS * Log (EL); Among them, UTS stands for ultimate tensile strength; YS stands for yield strength; EL stands for elongation.

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-Yb alloy design method according to 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, the Al-Si-Mg-Yb alloy design method according to any one of claims 1 to 8 is implemented.

Images