Prediction model for mechanical properties of Al-Si alloy, construction method of prediction model and method for predicting mechanical properties of Al-Si alloy by using model

Abstract

The invention belongs to the technical field of metallurgical materials, and particularly relates to a prediction model for mechanical properties of an Al-Si alloy, a construction method of the prediction model and a method for predicting the Al-Si alloy by using the model. According to the method, a machine learning method is combined, the research and development period of the Al-Si alloy is shortened, the Al-Si alloy with good performance can be obtained more easily by using the machine learning model to screen the optimal combination of the alloy components and the process, and a new thought is provided for development and design of the high-mechanical-performance Al-Si alloy based on machine learning. Moreover, in the process of establishing the data set, the maximum utilization of the disclosed experimental data is realized. In addition, in combination with knowledge increase features in the related material field, the prediction performance of the model is greatly improved, and the interpretability of the model is also improved.

Description

Technical Field

[0001] This invention belongs to the field of metallurgical materials technology, specifically relating to a predictive model for the mechanical properties of Al-Si alloys, its construction method, and a method for predicting Al-Si alloys using this model. Background Technology

[0002] Aluminum-silicon (Al-Si) alloys are widely used in automotive manufacturing, aerospace, and electronic components due to their excellent comprehensive properties, such as good casting performance, high strength, and hardness. At the same time, the low density and low cost of Al-Si alloys also make them an important material choice for achieving energy conservation and emission reduction. In recent decades, researchers have been trying to improve the mechanical properties of Al-Si alloys to enhance their load-bearing capacity, extend their service life, and ultimately reduce energy consumption.

[0003] However, traditional Al-Si alloys generally suffer from limitations in mechanical properties, requiring a trade-off between strength and toughness. For example, a high silicon content can increase the proportion of eutectic Si, and the fiberization of eutectic Si can improve casting fluidity, but it may introduce porosity and reduce strength. Adding Fe can easily lead to the formation of α-Fe or β-Fe phases. While the β-Fe phase can enhance strength and improve high-temperature stability, it can also increase brittleness and affect elongation. In summary, how to precisely control the alloy composition and heat treatment process to achieve excellent comprehensive mechanical properties in Al-Si alloys is one of the most pressing problems to be solved in the field of materials science.

[0004] However, the main process for discovering new alloys in existing technologies involves alloy composition design, smelting and preparation, structural characterization and performance testing, and then determining the alloy composition and preparation process based on structural characterization and performance testing. This method is not only costly to develop but also has a long development cycle. Summary of the Invention

[0005] The purpose of this invention is to provide a predictive model for the mechanical properties of Al-Si alloys, a method for constructing the model, and a method for predicting Al-Si alloys using the model. The method provided by this invention can effectively save time and cost, improve R&D efficiency, and make it easier to achieve a balance between the strength and toughness of Al-Si alloys, resulting in Al-Si alloys with good properties in all aspects.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a method for constructing a predictive model for the mechanical properties of Al-Si alloys, comprising the following steps: (1) Based on the known raw data of Al-Si alloys, perform data processing to construct an Al-Si alloy dataset; the raw data includes the elemental composition, preparation process conditions and mechanical property data of each Al-Si alloy; (2) The Al-Si alloy dataset is normalized and standardized. The preprocessed dataset is based on the cross-validation recursive feature elimination method of the adaptive enhancement algorithm. The root mean square error is used as the evaluation index to calculate the importance of each feature in the original dataset and sort them. Then, the number and types of each feature are screened, and the screened dataset is divided into training set and test set. Each feature is elemental composition, preparation process conditions and mechanical property data. (3) Based on the training set, different algorithms are used to train the model. During the training process, five-fold cross-validation is used to improve the generalization ability of the model and obtain the coefficient of determination. The coefficient of determination is used as the evaluation index to determine the model with the highest prediction ability and the model with the strongest generalization ability. (4) The model with the highest predictive ability and the model with the strongest generalization ability are fused to obtain the Al-Si alloy mechanical property prediction model; the mechanical properties include tensile strength, yield strength or elongation.

[0007] Preferably, the algorithm includes one or more of the following: linear regression, K-nearest neighbor algorithm, decision tree, random forest, gradient boosting decision tree, neural network, lightweight gradient boosting machine, and extreme gradient boosting algorithm.

[0008] Preferably, the data processing includes missing value imputation, data filtering, and cleaning performed sequentially; The specific steps for filling in the missing values ​​are as follows: (1) Based on the elemental composition in the initial dataset, calculate the physical properties that may affect the mechanical properties of Al-Si alloy; (2) The elemental composition and preparation process conditions in the initial dataset are used as features, and tensile strength, yield strength and elongation are used as labels. The random forest algorithm is used to fill in the missing label data in the initial dataset.

[0009] Preferably, the data screening and cleaning involves filtering out data in the original data where the Si content is below 6 wt.% or above 12 wt.%.

[0010] Preferably, the number of samples in the training set accounts for 70-80% of the number of samples in the dataset.

[0011] Preferably, the model fusion is performed based on the Stacking method or the Blending method.

[0012] Preferably, the model fusion step is as follows: The model with the highest predictive ability and the model with the strongest generalization ability are used as base learners; The prediction results of the base learner are used as the input parameters of the meta learner and the model is trained to obtain the prediction model of the mechanical properties of Al-Si alloy.

[0013] This invention also provides a method for predicting the mechanical properties of Al-Si alloys using a predictive model, comprising the following steps: The virtual database is input into the Al-Si alloy mechanical property prediction model constructed by the construction method described above, and the mechanical properties of the alloy under different element ratios are output. Based on the predicted mechanical properties, the composition and preparation process conditions of the target Al-Si alloy are obtained; The steps for obtaining the virtual database are as follows: (a) Determine the element types in the Al-Si alloy; (b) The element types determined in step (a) are processed to generate multiple data sets with different element proportions; (c) Randomly match the multiple data with different element ratios to the preparation process conditions in the Al-Si alloy dataset to obtain a virtual database.

[0014] Preferably, the data processing in step (b) involves randomly generating multiple data sets with different element ratios using the Dirichlet function in Python.

[0015] The present invention also provides a computer-readable storage medium, including an input module, a running module, and an output module. The running module is an Al-Si alloy mechanical property prediction model constructed by the construction method described in the above technical solution; the input module is a virtual database; and the output module is the mechanical properties of the alloy.

[0016] Compared with existing aluminum alloy design technologies, this invention has the following advantages: (1) The main process for discovering new alloys in the prior art involves alloy composition design, smelting and preparation, structural characterization and performance testing, which is costly and time-consuming. The method provided by this invention can effectively save time and cost and improve R&D efficiency in the discovery of new alloys.

[0017] (2) It is difficult to achieve a balance between strength and toughness in Al-Si alloys using traditional alloy design methods. However, it is easier to obtain Al-Si alloys with good performance by using machine learning models to screen the optimal combination of alloy composition and process. This provides a new idea for the development and design of Al-Si alloys with high mechanical properties based on machine learning.

[0018] (3) In the process of building the dataset, the publicly available experimental data were utilized to the maximum extent. In addition, by adding features based on relevant materials science knowledge, the predictive performance of the model was not only greatly improved, but also the interpretability of the model was enhanced. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a simplified flowchart of the method designed in this invention; Figure 2 The graphs are: (a) the RMSE of tensile strength prediction as a function of the number of features, analyzed using the recursive feature elimination method; (b) the RMSE of yield strength prediction as a function of the number of features, analyzed using the recursive feature elimination method; and (c) the RMSE of elongation prediction as a function of the number of features, analyzed using the recursive feature elimination method. Figure 3 R for 8 models 2 Compare the bar charts; Figure 4 The integrated model fitting curves are shown for (a) tensile strength, (b) yield strength, and (c) elongation. Figure 5 Photograph of an Al-Si alloy tensile specimen; Figure 6 XRD pattern of Al-Si alloy material; Figure 7 SEM image of Al-Si alloy material; Figure 8 This is a stress-strain curve of an Al-Si alloy. Detailed Implementation

[0021] This invention provides a method for constructing a predictive model for the mechanical properties of Al-Si alloys, comprising the following steps: (1) Based on the known raw data of Al-Si alloys, perform data processing to construct an Al-Si alloy dataset; the raw data includes the elemental composition, preparation process conditions and mechanical properties of each Al-Si alloy; (2) The Al-Si alloy dataset is normalized and standardized. The preprocessed dataset is based on the cross-validation recursive feature elimination method of the adaptive enhancement algorithm. The root mean square error is used as the evaluation index to calculate the importance of each feature in the original dataset and sort them. Then, the number and types of each feature are screened, and the screened dataset is divided into training set and test set. Each feature is elemental composition, preparation process conditions and mechanical property data. (3) Based on the training set, different algorithms are used to train the model. During the training process, five-fold cross-validation is used to improve the generalization ability of the model and obtain the coefficient of determination. The coefficient of determination is used as the evaluation index to determine the model with the highest prediction ability and the model with the strongest generalization ability. (4) The model with the highest predictive ability and the model with the strongest generalization ability are fused to obtain the Al-Si alloy mechanical property prediction model; the mechanical properties include tensile strength, yield strength or elongation.

[0022] This invention processes known raw data of Al-Si alloys to construct an Al-Si alloy dataset; the raw data includes the elemental composition, preparation process conditions, and mechanical properties of each Al-Si alloy.

[0023] As one embodiment of the present invention, the data processing includes missing value imputation, data filtering, and data cleaning performed sequentially.

[0024] As one embodiment of the present invention, the specific steps for filling in missing values ​​are as follows: (1) Calculate the physical properties that may affect the mechanical properties of Al-Si alloy based on the atomic percentage of elements in the initial dataset; (2) The elemental composition and preparation process conditions in the initial dataset are used as features, and tensile strength, yield strength or elongation are used as labels. The random forest algorithm is used to fill in the missing label data in the initial dataset.

[0025] As one embodiment of the present invention, the physical properties are shown in Table 1; Table 1 Physical characteristics and their calculation formulas

[0026] Note: In the above formula, The total number of elements in the alloy is , and the is the in the alloy. Atomic concentration (proportion) of the elements, Indicates the first The corresponding property parameters of each element For the first atomic radius of an element The average atomic radius of the alloy, Indicates the first Electronegativity of elements For average electronegativity, Indicates the first Type of element and the first Enthalpy of mixing between elements It is the ideal gas constant (value 8.314).

[0027] As one embodiment of the present invention, the data screening and data cleaning are as follows: data with Si content below 6wt.% and above 12wt.% in the original data are screened out and removed.

[0028] This invention normalizes and standardizes the Al-Si alloy dataset. The resulting preprocessed dataset is based on the cross-validation recursive feature elimination method of the adaptive enhancement algorithm. The root mean square error is used as the evaluation index to calculate the importance of each feature in the original dataset and sort them. Then, the number and types of each feature are filtered, and the filtered dataset is divided into training set and test set. In one embodiment of the present invention, the number of samples in the training set accounts for 70-80% of the number of samples in the dataset.

[0029] This invention trains models using different algorithms based on a training set. During the training process, cross-validation is used to improve the generalization ability of the model and obtain the coefficient of determination. The coefficient of determination is used as an evaluation index to determine the model with the highest predictive ability and the model with the strongest generalization ability. As one embodiment of the present invention, the model with the highest predictive ability and the model with the strongest generalization ability are used as base learners.

[0030] As one embodiment of the present invention, the algorithm may include linear regression (LR), K-nearest neighbors (KNN), decision tree (DT), random forest (RF), gradient boosting decision tree (GBDT), neural network (NN), lightweight gradient boosting machine (LGBM), and extreme gradient boosting algorithm (XGB). Before training the model, the hyperparameter settings for different algorithms are also included. Preferably, LR, KNN, DT, RF, and GBDT use grid search to find the optimal hyperparameters, while NN, LGBM, and XGB use Bayesian optimization to find their hyperparameters.

[0031] This invention fuses the model with the highest predictive ability and the model with the strongest generalization ability to obtain a predictive model for the mechanical properties of Al-Si alloys; the mechanical properties include tensile strength, yield strength or elongation.

[0032] As one embodiment of the present invention, the model fusion is based on the Stacking method or the Blending method. The specific process is as follows: the prediction results of the base learner are used as the input parameters of the meta learner and the model is trained to obtain the Al-Si alloy mechanical property prediction model.

[0033] The present invention also provides a method for predicting the mechanical properties of Al-Si alloys using a predictive model, comprising the following steps: The virtual database is input into the Al-Si alloy mechanical property prediction model constructed by the construction method described above, and the mechanical properties of the alloy under different element ratios are output. Based on the predicted mechanical properties, the composition and preparation process conditions of the target Al-Si alloy are obtained; The steps for obtaining the virtual database are as follows: (a) Determine the element types in the Al-Si alloy; (b) The element types determined in step (a) are processed to generate multiple data sets with different element proportions; (c) Randomly match the multiple data with different element ratios to the preparation process conditions in the Al-Si alloy dataset to obtain a virtual database.

[0034] In one embodiment of the present invention, the data processing in step (b) involves using the dirichlet function in Python to randomly generate multiple data sets with different element ratios. Specifically, 100,000 data sets with different element ratios can be generated.

[0035] The present invention also provides a computer-readable storage medium, including an input module, a running module, and an output module. The running module is an Al-Si alloy mechanical property prediction model constructed by the construction method described in the above technical solution; the input module is a virtual database; and the output module is the mechanical properties of the alloy.

[0036] This invention shortens the research and development cycle of Al-Si alloys by combining machine learning methods. The resulting alloy material overcomes the contradiction between strength and toughness compared with existing Al-Si alloys, and achieves a synergistic improvement in tensile strength, yield strength and elongation.

[0037] To further illustrate the present invention, the following detailed description of the invention's solutions, in conjunction with the accompanying drawings and embodiments, is provided, but should not be construed as limiting the scope of protection of the present invention.

[0038] Example 1 (1) Collect relevant data on the elemental composition, preparation process conditions, tensile strength, yield strength and elongation of Al-Si alloy from publicly published literature or public databases, and summarize them into an initial dataset; calculate the physical properties that may affect the mechanical properties of Al-Si alloy based on the elemental composition ratio and formula in the initial dataset (physical properties are: valence electron concentration, poor electronegativity, first ionization energy, second ionization energy and third ionization energy), fill in the missing values ​​in the order of the number of missing label values ​​from smallest to largest using the random forest algorithm, and remove the data in the dataset with Si content below 6wt.% and above 12wt.% to obtain the Al-Si alloy dataset.

[0039] (2) The Al-Si alloy dataset was normalized and standardized to obtain a preprocessed dataset. The preprocessed dataset was then used to screen the number and types of Al-Si alloy features based on the recursive feature elimination method of the adaptive enhancement algorithm with root mean square error as the evaluation index. The results are as follows: Figure 2 As shown, according to Figure 2 The screening results removed characteristic atomic radius (AR1), quantum number (QN), metal ion radius (RM), electron affinity (EA), and mixing entropy (...). ); The filtered dataset is divided into a training set and a test set, with the training set accounting for 80% and the test set accounting for 20% of the entire dataset.

[0040] (3) The model was trained using eight algorithms: LR, KNN, DT, RF, GBDT, NN, LGBM, and XGB, based on the training set. Among them, LR, KNN, DT, RF, and GBDT used grid search to find the optimal hyperparameters, while NN, LGBM, and XGB used Bayesian optimization to find their hyperparameters. The coefficient of determination (R) was used to train the model. 2 () is used as a model evaluation metric. For example... Figure 3 As shown, for tensile strength and yield strength, the R of the LGBM model... 2 The highest value indicates that the LGBM model has the best predictive ability; therefore, R is chosen. 2 The LGBM model with the highest value was chosen as the base learner. Regarding elongation, both the XGB and LGBM models showed relatively good predictive ability; therefore, R was selected. 2 The XGB and LGBM models with higher values ​​were used as base learners. To improve the generalization ability of the ensemble model, R0 values ​​from the training set were selected. 2 R on the test set 2 The KNN model with the least difference was used as the base learner. The LGBM, XGB, and KNN models were subjected to 10 Bayesian optimizations based on the tree-structured Parsons estimator (TPE), and five-fold cross-validation was performed on the training set to ensure that the models learned the data sufficiently.

[0041] (S4) An ensemble learning algorithm is used to fuse the single model with the strongest predictive ability and the single model with the best generalization ability. Using Stacking or Blending methods, the prediction results of the base learners are used as input parameters to train the meta-learners, thereby obtaining prediction models for the tensile strength, yield strength, and elongation of Al-Si alloys. The prediction results of the models are as follows: Figure 4 As shown, from Figure 4 It can be seen that the R-value of the tensile strength prediction model is... 2 The R-squared value of the predictive model for yield strength is 0.9308. 2 The R-value of the elongation prediction model is 0.9485. 2 The value of 0.7778 indicates that the three models have good fitting ability.

[0042] (S5) The element types are determined to be Al, Si, Mg, Fe, Ni, and Cu. 100,000 data points with different component ratios are randomly generated using the Dirichlet function in Python, and process parameters in the dataset are randomly matched to form a virtual database.

[0043] (S6) The model is called to predict the mechanical properties of the alloys in the virtual database. Based on the target mechanical property range, the virtual database is screened to deduce the alloy composition and process with high mechanical properties. In this embodiment, the following mechanical properties are used as the criteria: tensile strength ≥300 MPa; yield strength ≥230 MPa; elongation ≥8%. The alloy composition is deduced as follows: silicon (Si): 6.0%-7.0%; copper (Cu): 3.0%-4.0%; nickel (Ni): 0.8%-1.2%; manganese (Mn): 0.15%-0.25%; iron (Fe): 0.15%-0.25%; aluminum (Al): balance. The preparation process is as follows: the alloy is melted in a vacuum arc melting furnace under inert gas protection and heat-treated by solution treatment (temperature 500℃, time 8 hours).

[0044] Specifically, taking an alloy with the following composition as an example: silicon (Si): 6.63%; copper (Cu): 3.53%; nickel (Ni): 0.67%; manganese (Mn): 0.22%; iron (Fe): 0.25%; aluminum (Al): balance; The specific preparation process is as follows: The required mass of each raw material is calculated according to the proportions of Al-50%Cu, Al-50%Si, Al-10%Fe, Al-30%Mn, Al-10%Ni master alloy and pure Al (99.99%) (by mass percentage), taking a total mass of 200g as an example: Al-50%Si: 26.51g; Al-50%Cu: 14.11g; Al-10%Fe: 4.90g; Al-30%Mn: 1.45g; Al-10%Ni: 13.47g; Al: 139.56g.

[0045] After weighing the required mass using an electronic balance, the mixture was combined. A vacuum arc melting furnace was used, with the mechanical pump evacuated for approximately 20 minutes, followed by argon gas introduction. Evacuation was then continued for another 10 minutes for gas purging, and this process was repeated three times. Cooling water was then turned on, and the molecular pump was activated to further evacuate the furnace, introducing argon gas again. Melting was carried out under argon protection at a temperature of 1450℃. After the surface was completely melted, the mixture was allowed to cool, flipped, and melted three times on each side to obtain a rectangular alloy ingot measuring 760mm × 340mm × 140mm and weighing approximately 200g. This ingot was then laser-cut into standard bone-shaped tensile test specimens measuring 20mm × 3mm × 2mm according to the drawings. The actual tensile test specimens are shown below. Figure 5 As shown. Then, the tensile specimens were heat-treated: the specimens were placed in a muffle furnace and heated to 500°C in air, held at 500°C for 8 hours, and then immediately removed and placed in warm water at 50°C for 30 seconds. After the quenching treatment was completed, the specimens were cooled to room temperature in air to obtain Al-Si alloy.

[0046] Before characterization and tensile testing, the specimens need to be coarsely ground, finely ground, and mechanically polished with sandpaper.

[0047] test: The Al-Si alloy prepared in this embodiment was characterized by X-ray diffraction (XRD) with a scanning range of 20°–90° and a scanning speed of 7° / min. Finally, the obtained diffraction patterns were analyzed using Jade 9.0 software for phase characterization and lattice constant calculation, as shown in the results. Figure 6 As shown in the figure. XRD analysis reveals that the main phase is an Al matrix, supplemented by a eutectic Si phase.

[0048] To demonstrate the microstructural characteristics of the Al-Si alloy obtained in this embodiment, the tensile fracture morphology of the Al-Si alloy was characterized using scanning electron microscopy (SEM). The results are as follows: Figure 7 As shown. By Figure 7 It can be seen that after heat treatment, the fracture structure exhibits typical ductile fracture characteristics, with fewer cleavage planes and cleavage steps, and small but deep dimples. Therefore, it can be concluded that heat treatment improves the toughness of the alloy.

[0049] The material was tested using an electronic universal testing machine at a speed of 1.0 × 10⁻⁶. -4 s -1The tensile strength, yield strength, and elongation of the Al-Si alloy prepared in this embodiment were tested at a constant strain rate, and the results are as follows: Figure 8 As shown. By Figure 8 As can be seen, the tensile strength of the Al-Si alloy obtained in this embodiment is 315.78 MPa, the yield strength is 240.4 MPa, and the elongation is 9.33%, which is basically consistent with the predicted results, confirming the usability of the model.

[0050] To verify that the Al-Si alloy obtained in this embodiment possesses superior comprehensive mechanical properties, a comparison of its mechanical properties with existing alloys is shown in Table 2. A380, 6061, and 4015 are all commonly used Al-Si alloys in industry. Under similar heat treatment processes, the Al-Si alloy obtained in this embodiment maintains good comprehensive mechanical properties, indicating that the composition ratio of this alloy is superior to that of existing Al-Si alloys.

[0051] Table 2 Comparison of mechanical properties of the Al-Si alloy obtained in Example 1 with existing alloys

[0052] Note: T4 indicates that the alloy has undergone T4 heat treatment, and T5 indicates that the alloy has undergone T5 heat treatment.

[0053] Although the above embodiments have provided a detailed description of the present invention, they are only some embodiments of the present invention, and not all embodiments. Other embodiments can be obtained based on these embodiments without creative effort, and these embodiments all fall within the protection scope of the present invention.

Claims

1. A method for constructing a predictive model for the mechanical properties of Al-Si alloys, characterized in that, Includes the following steps: (1) Based on the known raw data of Al-Si alloys, perform data processing to construct an Al-Si alloy dataset; the raw data includes the elemental composition, preparation process conditions and mechanical property data of each Al-Si alloy; (2) The Al-Si alloy dataset is normalized and standardized. The preprocessed dataset is based on the cross-validation recursive feature elimination method of the adaptive enhancement algorithm. The root mean square error is used as the evaluation index to calculate the importance of each feature in the original dataset and sort them. Then, the number and types of each feature are screened, and the screened dataset is divided into training set and test set. Each characteristic includes elemental composition, preparation process conditions, and mechanical property data; (3) The model is trained using different algorithms based on the training set. During the training process, the model's generalization ability is improved and the coefficient of determination is obtained through five-fold cross-validation. Using the coefficient of determination as an evaluation metric, we determine the model with the highest predictive power and the model with the strongest generalization ability. (4) The model with the highest predictive ability and the model with the strongest generalization ability are fused to obtain the Al-Si alloy mechanical property prediction model; the mechanical properties include tensile strength, yield strength or elongation.

2. The construction method as described in claim 1, characterized in that, The algorithm includes one or more of the following: linear regression, K-nearest neighbors algorithm, decision tree, random forest, gradient boosting decision tree, neural network, lightweight gradient boosting machine, and extreme gradient boosting algorithm.

3. The construction method as described in claim 1, characterized in that, The data processing includes missing value imputation, data filtering, and cleaning performed sequentially. The specific steps for filling in the missing values ​​are as follows: (1) Based on the elemental composition in the initial dataset, calculate the physical properties that may affect the mechanical properties of Al-Si alloy; (2) The elemental composition and preparation process conditions in the initial dataset are used as features, and tensile strength, yield strength and elongation are used as labels. The random forest algorithm is used to fill in the missing label data in the initial dataset.

4. The construction method as described in claim 1, characterized in that, The data screening and cleaning process involves filtering out data where the Si content is below 6 wt.% or above 12 wt.%.

5. The construction method as described in claim 1, characterized in that, The number of samples in the training set accounts for 70-80% of the total number of samples in the dataset.

6. The construction method as described in claim 1, characterized in that, The model fusion is performed based on either the Stacking or Blending method.

7. The construction method as described in claim 1, characterized in that, The steps for model fusion are as follows: The model with the highest predictive ability and the model with the strongest generalization ability are used as base learners; The prediction results of the base learner are used as the input parameters of the meta learner and the model is trained to obtain the prediction model of the mechanical properties of Al-Si alloy.

8. A method for predicting the mechanical properties of Al-Si alloys using a predictive model, comprising the following steps: Input the virtual database into the Al-Si alloy mechanical property prediction model constructed by the construction method described in any one of claims 1 to 7, and output the mechanical properties of the alloy under different element ratios; Based on the predicted mechanical properties, the composition and preparation process conditions of the target Al-Si alloy are obtained; The steps for obtaining the virtual database are as follows: (a) Determine the elemental composition of the Al-Si alloy; (b) The element types determined in step (a) are processed to generate multiple data sets with different element proportions; (c) Randomly match the multiple data with different element ratios to the preparation process conditions in the Al-Si alloy dataset to obtain a virtual database.

9. The method as described in claim 8, characterized in that, The data processing in step (b) involves using the Dirichlet function in Python to randomly generate multiple data sets with different element ratios.

10. A computer-readable storage medium, characterized in that, It includes an input module, a running module, and an output module. The running module is an Al-Si alloy mechanical property prediction model constructed by the construction method according to any one of claims 1 to 7. The input module is a virtual database. The output module is the mechanical properties of the alloy.

Images