Design method, system, medium and equipment for semi-solid cast-forged aluminum alloy material

Abstract

The invention discloses a design method, system, medium and equipment for a semi-solid cast-forged aluminum alloy material, and relates to the technical field of metal material design. The method comprises the following steps: constructing a data set of various element contents and various process parameters; constructing a first relation equation between the data in the data set and the performance parameters of the semi-solid cast and forged aluminum alloy; according to the data set, a plurality of candidate semi-solid cast-forged aluminum alloy component points are determined; evaluating the plurality of candidate semi-solid cast and forged aluminum alloy component points by using the first relation equation, correcting the first relation equation according to a first evaluation result and preset experimental data to obtain a second relation equation, and evaluating the plurality of candidate semi-solid cast and forged aluminum alloy component points by using the second relation equation to obtain a plurality of candidate semi-solid cast and forged aluminum alloy component points. And the aluminum alloy material meeting the preset performance is obtained according to the second evaluation result. The method is not limited by a specific alloy system, the aluminum alloy design efficiency and the performance prediction accuracy are improved, and the aluminum alloy material meeting the specified performance parameters can be automatically found.

Description

Technical Field

[0001] The present invention relates to the technical field of metal material design, and particularly to a design method, system, medium, and device for semi-solid casting and forging aluminum alloy materials. Background Art

[0002] Materials with suitable properties are an important cornerstone of human industrial progress. With the continuous exploration of new material applications, researchers need to conduct multiple semi-solid casting and forging experiments to discover material structures with ideal properties. However, the relevant operations of semi-solid casting and forging experiments (such as pouring operations, etc.) are usually carried out by researchers in an environment full of toxicity and potential hazards, posing great safety risks. In addition, there are numerous variables in semi-solid casting and forging experiments, making the experimental process complex and cumbersome. In the process of repeating experiments, researchers not only have to spend a lot of time dealing with these complex steps, but may also be distracted by long-term monotonous operations, increasing the risk of experimental errors. Such repetitive labor not only affects the work efficiency of researchers, but also makes it difficult for them to conduct more creative research, inhibiting the potential of scientific exploration and innovation.

[0003] In order to reduce the danger and complexity of experiments, the existing technology usually uses advanced automation technology and intelligent devices, which can help optimize the experimental process, enabling them to focus more on innovation and exploring the broad possibilities of new materials.

[0004] However, in the existing technology, the target materials for semi-solid casting and forging experiments are set by researchers, and researchers often rely on their own experience to make judgments, which makes it difficult for them to break through the limitations of inherent thinking and reduces the possibility of discovering new materials. Moreover, the control systems of existing semi-solid casting and forging machine tools often cannot adaptively adjust the process parameters of the machine tool according to the alloy composition, resulting in poor adaptability during processing in the laboratory and difficulty in performing complex operations. In addition, the judgment of the synthesized products under thermodynamics in the existing technology still needs to be carried out by researchers.

[0005] Therefore, the existing technology cannot accurately predict the properties of semi-solid casting and forging aluminum alloy materials, resulting in the inability to automatically find aluminum alloy materials that meet the specified performance parameters. Summary of the Invention

[0006] Based on this, in view of the problem that the existing technology cannot accurately predict the properties of semi-solid casting and forging aluminum alloy materials, resulting in the inability to automatically find aluminum alloy materials that meet the specified performance parameters, it is necessary to provide a design method, system, medium, and device for semi-solid casting and forging aluminum alloy materials.

[0007] The present invention adopts the following technical solutions:

[0008] In the first aspect, the present invention provides a design method for semi-solid casting and forging aluminum alloy materials, and the method includes:

[0009] Construct a dataset containing the contents of multiple metal elements, non-metal elements, and multiple process parameters;

[0010] Construct a first relationship equation between the data in the dataset and the thermal conductivity and tensile strength of semi-solid cast and forged aluminum alloy through a regression method with collective decision-making of decision trees; according to the dataset, determine multiple candidate semi-solid cast and forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength;

[0011] Evaluate the multiple candidate semi-solid cast and forged aluminum alloy composition points using the first relationship equation to obtain a first evaluation result, and correct the first relationship equation according to the first evaluation result and preset experimental data to obtain a second relationship equation, and use the second relationship equation to evaluate the multiple candidate semi-solid cast and forged aluminum alloy composition points again to obtain a second evaluation result;

[0012] Obtain an aluminum alloy material that meets the preset performance according to the second evaluation result.

[0013] Further, after constructing the dataset containing the contents of multiple metal elements, non-metal elements, and multiple process parameters, it further includes:

[0014] Use a machine learning model to perform feature screening on the dataset to obtain feature data;

[0015] Among them, the feature data includes: element composition, heat treatment state, thermal conductivity, phase composition, and density.

[0016] Further, after using the machine learning model to perform feature screening on the dataset to obtain feature data, it further includes:

[0017] Adjust the hyperparameters of the machine learning model according to different alloy systems included in the dataset to obtain an optimized machine learning model;

[0018] Use the optimized machine learning model to screen the feature data again to obtain optimized feature data.

[0019] Further, determining multiple candidate semi-solid cast and forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength according to the dataset specifically includes:

[0020] Use a tabular conditional adversarial neural network to perform dimensionality reduction processing on the dataset, and identify a first performance region in the latent space after dimensionality reduction where the thermal conductivity and tensile strength are improved synergistically;

[0021] Reconstruct the first performance region through a decoder to obtain multiple candidate semi-solid cast and forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength.

[0022] Further, modifying the first relationship equation according to the first evaluation result and preset experimental data to obtain a second relationship equation specifically includes:

[0023] Improve the collective decision-making of the decision tree and the tabular conditional adversarial neural network according to the first evaluation result and preset experimental data to obtain an improved collective decision-making of the decision tree and an improved tabular conditional adversarial neural network;

[0024] Use a regression method with the improved collective decision-making of the decision tree to construct a second relationship equation between the feature data, the thermal conductivity of the semi-solid cast and forged aluminum alloy, and the tensile strength.

[0025] Further, when the machine learning model is a random forest model, the hyperparameters specifically include:

[0026] The number of decision trees, the maximum depth of the decision tree, the maximum number of leaf nodes, the minimum number of samples in the leaf nodes, and the number of features to be considered when finding the best split.

[0027] Further, the machine learning model integrates a grid search tuning algorithm and a Bayesian optimization algorithm.

[0028] In a second aspect, the present invention provides a design system for semi-solid cast and forged aluminum alloy materials, including:

[0029] A construction module for constructing a data set including the contents of various metal elements, the contents of non-metal elements, and various process parameters;

[0030] A determination module for constructing a first relationship equation between the data in the data set and the thermal conductivity and tensile strength of the semi-solid cast and forged aluminum alloy through a regression method with collective decision-making of the decision tree; determining multiple candidate semi-solid cast and forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength according to the data set;

[0031] An evaluation module for evaluating the multiple candidate semi-solid cast and forged aluminum alloy composition points using the first relationship equation to obtain a first evaluation result, and modifying the first relationship equation according to the first evaluation result and preset experimental data to obtain a second relationship equation, and using the second relationship equation to evaluate the multiple candidate semi-solid cast and forged aluminum alloy composition points again to obtain a second evaluation result;

[0032] A manufacturing module for obtaining an aluminum alloy material that meets the preset performance according to the second evaluation result.

[0033] The present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, a method for semi-solid casting-forging aluminum alloy materials is implemented.

[0034] The present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, a method for semi-solid casting-forging aluminum alloy materials is implemented.

[0035] At least one technical solution adopted by the present invention can achieve the following beneficial effects:

[0036] The present invention constructs a data set including the contents of various metal elements, non-metal elements, and various process parameters through a regression method with collective decision-making of decision trees, and constructs a first relationship equation between the data in the data set and the thermal conductivity and tensile strength of semi-solid casting-forging aluminum alloy. And according to the data set, multiple candidate semi-solid casting-forging aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength are determined; then the first relationship equation is used to evaluate the multiple candidate semi-solid casting-forging aluminum alloy composition points to obtain a first evaluation result, and the first relationship equation is corrected according to the first evaluation result and preset experimental data to obtain a second relationship equation; and the second relationship equation is used to evaluate the multiple candidate semi-solid casting-forging aluminum alloy composition points again to obtain a second evaluation result, and finally an aluminum alloy material meeting the preset performance is obtained according to the second evaluation result. Through the above solution, the present invention can be unrestricted by a specific alloy system, improves the design efficiency of aluminum alloy and the accuracy of performance prediction, and thus can automatically find an aluminum alloy material meeting specified performance parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings described herein are used to provide a further understanding of the present invention and constitute a part 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:

[0038] Figure 1 is a schematic flow chart of a design method for semi-solid casting-forging aluminum alloy materials provided by the present invention;

[0039] Figure 2 is a schematic diagram of a prediction result after adjusting hyperparameters provided by the present invention;

[0040] Figure 3 is a generated result diagram of a low-performance area and a high-performance area provided by the present invention;

[0041] Figure 4 is a comparison diagram of original data and data after calibration and augmentation provided by the present invention;

[0042] Figure 5 Experimental result diagram of a candidate semi-solid cast and forged aluminum alloy composition point reconstructed by a decoder provided by the present invention;

[0043] Figure 6 Schematic diagram of a design system for semi-solid cast and forged aluminum alloy materials provided by the present invention;

[0044] Figure 7 Schematic diagram of a computer device for a method of realizing semi-solid cast and forged aluminum alloy materials provided by the present invention. Detailed implementation manners

[0045] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and the corresponding drawings. 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 without creative efforts fall within the protection scope of the present invention.

[0046] At present, the server mentioned in the present invention can be a server set up on a service platform, or a device such as a desktop computer or a laptop computer that can execute the solution of the present invention. The "first" and "second" mentioned in the present invention only serve to distinguish different objects and do not represent the order of precedence. For the convenience of description, the server is only used as the execution subject for description below. The technical solutions provided by each embodiment of the present invention are described in detail below with reference to the drawings.

[0047] The present invention proposes a design method for semi-solid cast and forged aluminum alloy materials, as Figure 1 shown, specifically including the following steps:

[0048] S10: Construct a data set including the contents of various metal elements, non-metal elements, and various process parameters.

[0049] In this embodiment, the data set includes, but is not limited to, a data set of semi-solid cast and forged aluminum alloy. The data set can be collected based on any alloy data. The constructed original data set contains 1021 pieces of data, and the number of features in the data set is 156. The contents of various metal elements and non-metal elements include, but are not limited to, the element contents of "Al", "Mg", "Si", "Ti", "Fe", "Cu", "Cr", and various process parameters include, but are not limited to, die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution time, solution temperature, aging time, aging temperature. These process parameters are used to construct the original data set for predicting the thermal conductivity and tensile strength of the alloy.

[0050] S20: Construct a first relationship equation between the data in the dataset and the thermal conductivity and tensile strength of semi-solid cast-forged aluminum alloy through a regression method with collective decision-making of decision trees; determine multiple candidate semi-solid cast-forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength according to the dataset.

[0051] In this embodiment, the dimensionality constraint of the regression method with collective decision-making of decision trees applies to binary operators, specifically including "+", "-", "×", and "÷". The preset thermal conductivity refers to a preset high thermal conductivity, and the preset tensile strength refers to a preset high tensile strength. The number of multiple candidate semi-solid cast-forged aluminum alloy composition points is 10% to 20% of the size of the dataset.

[0052] Optionally, the regression method with collective decision-making of decision trees is implemented based on a multi-objective jellyfish optimization algorithm. And in order to fully search the feature space, the population size is not less than 150, and the number of individuals in each population is not less than 80; in order to maintain search diversity, the complexity of the initial individuals is not less than 8.

[0053] S30: Evaluate multiple candidate semi-solid cast-forged aluminum alloy composition points using the first relationship equation to obtain a first evaluation result, and correct the first relationship equation according to the first evaluation result and preset experimental data to obtain a second relationship equation, and then use the second relationship equation to evaluate multiple candidate semi-solid cast-forged aluminum alloy composition points again to obtain a second evaluation result.

[0054] In this embodiment, evaluation refers to experimentally verifying the candidate semi-solid cast-forged aluminum alloy composition points with high thermal conductivity and high tensile strength using the relationship equation. The candidate semi-solid cast-forged aluminum alloy composition points refer to the potential composition points that can manufacture aluminum alloy materials with high thermal conductivity and high tensile strength.

[0055] S40: Obtain an aluminum alloy material that meets the preset performance according to the second evaluation result.

[0056] In this embodiment, the preset performance refers to the preset thermal conductivity and the preset tensile strength. Determine the composition points of the aluminum alloy material with high thermal conductivity and high tensile strength according to the second evaluation result, and then obtain the aluminum alloy material that meets the preset performance according to the composition points of the aluminum alloy material with high thermal conductivity and high tensile strength.

[0057] Based on Figure 1The disclosed design method for semi-solid casting-forging aluminum alloy materials involves constructing a dataset containing various metal element contents, non-metal element contents, and various process parameters, and establishing a first relationship equation between the data in the dataset and the thermal conductivity and tensile strength of the semi-solid casting-forging aluminum alloy. Based on the dataset, multiple candidate semi-solid casting-forging aluminum alloy composition points with high thermal conductivity and high tensile strength are determined. Then, the first relationship equation is used to evaluate the multiple candidate semi-solid casting-forging aluminum alloy composition points to obtain a first evaluation result, and the first relationship equation is corrected according to the first evaluation result and preset experimental data to obtain a second relationship equation. The second relationship equation is used to evaluate the multiple candidate semi-solid casting-forging aluminum alloy composition points again to obtain a second evaluation result, and finally, an aluminum alloy material meeting the preset performance is obtained according to the second evaluation result. Through the above solution, the present invention can be unrestricted by a specific alloy system, improving the design efficiency of aluminum alloys and the accuracy of performance prediction, thereby enabling automatic search for aluminum alloy materials meeting specified performance parameters.

[0058] When applying the design method for semi-solid casting-forging aluminum alloy materials provided by the present invention, it is not necessary to execute according to Figure 1 the order of the steps shown. The specific execution order of each step can be determined according to needs, and the present invention does not limit this.

[0059] In addition, in one or more embodiments of the present invention, after constructing a dataset containing various metal element contents, non-metal element contents, and various process parameters, it further includes:

[0060] Using a machine learning model to perform feature screening on the dataset to obtain feature data.

[0061] Among them, the feature data includes: element composition, heat treatment state, thermal conductivity, phase composition, and density.

[0062] In this embodiment, the feature data includes: element composition, heat treatment state, thermal conductivity, phase composition, and density melting point characteristics, where the phase composition is calculated by commercial software Jmatpro and factstage, and the density melting point characteristics are calculated by the magpie third-party library.

[0063] Optionally, the machine learning model is one or more of linear regression, ridge regression, support vector machine, decision tree, random forest, gradient boosting tree, and artificial neural network.

[0064] In this embodiment, the construction steps of the machine learning model include:

[0065] (1) Obtain data for aluminum alloys through thermodynamic theory calculations. By given process parameters and given alloy compositions, calculate the generated phases and their corresponding volume fractions.

[0066] (2) Use the given temperature and given alloy composition for calculation as input data, and use the generated phases and their volume fractions calculated by thermodynamic theory as output data. Take the given temperature, given alloy composition, and the phases and their volume fractions obtained from the corresponding thermodynamic theory calculation as a data point. By calculating a sufficient number of data points, construct a data set.

[0067] (3) Use the data set to train a machine learning model.

[0068] (4) Save the trained machine learning model.

[0069] Among them, in step (1), when the casting and forging integrated process parameters are the same, for a ternary alloy system, the number of thermodynamic calculation data points obtained is not less than 5000; for a quaternary alloy system, the number of data points obtained reaches or exceeds 100,000; for a quinary alloy system, the number of data points reaches or exceeds 1 million; for a senary alloy system, the number of data points obtained reaches or exceeds 10 million; for a septenary alloy system, the number of data points reaches or exceeds 100 million; and for an octonary alloy system, the number of thermodynamic calculation data points obtained reaches or exceeds 800 million.

[0070] The alloy composition in step (1) includes mass fraction or mole fraction, and the process parameters include die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution treatment time, solution treatment temperature, aging time, and aging temperature, a total of ten variables; for a ternary alloy system, it includes die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution treatment time, solution treatment temperature, aging time, aging temperature, and two metal components, a total of eleven variables; for a quaternary alloy system, it includes die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution treatment time, solution treatment temperature, aging time, aging temperature, and three metal components, a total of twelve variables; for a quinary alloy system, it includes die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution treatment time, solution treatment temperature, aging time, aging temperature, and four metal components, a total of thirteen variables; for a senary alloy system, it includes die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution treatment time, solution treatment temperature, aging time, aging temperature, and five metal components, a total of fourteen variables; for a septenary alloy system, it includes die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution treatment time, solution treatment temperature, aging time, aging temperature, and six metal components, a total of fifteen variables; for an octonary alloy system, it includes die temperature, pouring temperature, filling speed, injection chamber hammer pressure, forging hammer pressure, solution treatment time, solution treatment temperature, aging time, aging temperature, and seven metal components, a total of sixteen variables.

[0071] In step (2), the dataset can be constructed through the following two approaches:

[0072] (1’) Construct the dataset by collecting data on semi-solid casting and forging of aluminum alloys in the literature.

[0073] (2’) Conduct thermodynamic calculations within the temperature and composition ranges at a fixed step size to construct the dataset.

[0074] Among them, the output options of the dataset in step (2) include:

[0075] Only generate the type of phase, simultaneously generate the type of phase and the volume fraction corresponding to the phase type. Among them, the volume fraction index corresponds to the hash value of the phase type. Based on this, the task of predicting the volume fraction and type of the phase is transformed into predicting the volume fraction of the phase and ensuring that the volume fraction matches the correct index column.

[0076] In the solution shown in this embodiment, by collecting data and screening feature data through machine learning, a relationship equation between features and the thermal conductivity of the alloy is constructed, and data dimensionality reduction and candidate composition point recommendation are performed. After multiple rounds of calculation and experimental iteration, a high-strength and high-thermal-conductivity aluminum alloy is finally prepared. The thermal conductivity of the aluminum alloy is 184.193 W / (m·K) and the tensile strength is 326 Mpa, which reflects that the method of the present invention has wide applicability to cast and forged aluminum alloys, is not restricted by a specific alloy system, significantly improves the alloy design efficiency and performance prediction accuracy, and can provide a relationship equation with good interpretability and clear physical meaning, having broad application prospects.

[0077] In addition, in one or more embodiments of the present invention, after using a machine learning model to screen the features of the dataset to obtain feature data, it further includes:

[0078] Adjust the hyperparameters of the machine learning model according to different alloy systems included in the dataset to obtain an optimized machine learning model.

[0079] Schematically, referring to Figure 2 , it is the result after grid search through a multi-objective jellyfish optimization algorithm to achieve automated hyperparameter adjustment. The abscissa is the mean value Mean, and the ordinate is the standard deviation (Standard Deviation, STD). Among them, Figure 2 in (a) shows the comparison between the predicted value and the actual value of the tensile strength, Figure 2 in (b) shows the comparison between the predicted value and the actual value of the elongation rate. The diagonal line represents the predicted value, and the points near the diagonal line represent the actual values. As can be seen from Figure 2 , the predicted values are near the diagonal line, indicating that the generated data has good quality.

[0080] The feature data is screened again using an optimized machine learning model to obtain optimized feature data.

[0081] In this embodiment, according to different alloy systems included in the data set, the hyperparameters of the machine learning model are adjusted, which can continuously correct the hyperparameters of the machine learning model during the process of obtaining feature data, continuously improve the ability of the machine learning model to screen features, and improve the correlation between the screened features and the semi-solid cast-forged aluminum alloy material, thereby further improving the accuracy of predicting the properties of the semi-solid cast-forged aluminum alloy material.

[0082] In addition, in one or more embodiments of the present invention, according to the data set, a plurality of candidate semi-solid cast-forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength are determined, specifically including:

[0083] The data set is dimensionally reduced using a tabular conditional adversarial neural network, and a first performance region where the thermal conductivity and the tensile strength are synergistically improved is identified in the reduced latent space.

[0084] In this embodiment, each output value of the layer before the output layer of the tabular conditional adversarial neural network needs to be guaranteed to be greater than 0. When the output only selects alloy element components, the softmax function is used as the activation function for the output layer of the adversarial neural network model; when the output is the thermal conductivity and the tensile strength, the specifically constructed Normalize normalization function is used as the activation function for the output layer of the adversarial neural network model. Among them, the softmax function converts a set of input values into probability values between 0 and 1, so that the sum of all output values is 1, and each output value respectively represents the relative possibility of each category.

[0085] The first performance region is reconstructed through a decoder to obtain a plurality of candidate semi-solid cast-forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength.

[0086] In this embodiment, the original data set is dimensionally reduced by a tabular conditional adversarial neural network (CTGAN), and divided according to the size of the original data set in the reduced latent space. Refer to Figure 3 , which is a generation result graph of a low-performance region and a high-performance region provided by the present invention, where Figure 3 (a) in Figure 3 is the curve distribution graph of the original data and the generated data of the magnesium element content, Figure 3 (b) in Figure 3(d) in it is the curve distribution diagram of the original data and the generated data of time, Figure 3 (e) in it is the curve distribution diagram of the original data and the generated data of the ultimate tensile strength, Figure 3 (f) in it is the curve distribution diagram of the original data and the generated data of the elongation. It can be seen from Figure 3 that 90% of them are low-performance regions and 10% are potential high-performance regions.

[0087] In order to simplify the multi-objective optimization process, a comprehensive index is constructed by multiplying the tensile strength and the thermal conductivity. The decoder is used to reconstruct 10% to 20% of the number of the dataset size in the potential high-performance region, and the number of features in the dataset is screened to 10% to 15%.

[0088] In this embodiment, by using the tabular condition against the neural network to perform dimensionality reduction processing on the dataset, and identifying the first performance region where the thermal conductivity and the tensile strength are synergistically improved in the latent space after dimensionality reduction, and then reconstructing the first performance region through the decoder, a plurality of candidate semi-solid cast-forged aluminum alloy composition points with high thermal conductivity and high tensile strength are obtained, which can improve the accuracy of determining the candidate semi-solid cast-forged aluminum alloy composition points.

[0089] In addition, in one or more embodiments of the present invention, the first relationship equation is corrected according to the first evaluation result and the preset experimental data to obtain the second relationship equation, specifically including:

[0090] According to the first evaluation result and the preset experimental data, the collective decision-making of the decision tree and the tabular condition against the neural network are improved to obtain the improved collective decision-making of the decision tree and the improved tabular condition against the neural network.

[0091] A regression method with the improved collective decision-making of the decision tree is used to construct the second relationship equation between the characteristic data and the thermal conductivity and the tensile strength of the semi-solid cast-forged aluminum alloy.

[0092] In this embodiment, compared with the first relationship equation, the second relationship equation can more accurately measure the relationship between the characteristic data and the thermal conductivity and the tensile strength of the semi-solid cast-forged aluminum alloy.

[0093] In this embodiment, the collective decision-making of the decision tree and the tabular condition against the neural network are improved through the evaluation result and the experimental data to realize the improvement of the collective decision-making of the decision tree and the tabular condition against the neural network. The improved collective decision-making of the decision tree and the tabular condition against the neural network are used to construct the second relationship equation between the characteristic data and the thermal conductivity and the tensile strength of the semi-solid cast-forged aluminum alloy. The iteration from constructing the relationship equation to evaluating the candidate alloy composition points is realized, the accuracy of evaluating the candidate alloy composition points is improved, and finally the obtained aluminum alloy has high thermal conductivity and high tensile strength.

[0094] Optionally, after evaluating multiple candidate semi-solid cast and forged aluminum alloy composition points, it further includes:

[0095] Select the semi-solid cast and forged aluminum alloy composition points with the top 10% of the predicted thermal conductivity results for experimental verification.

[0096] Schematically, Figure 4 in (a) and Figure 4 in (b) show the comparison diagrams of the original data and the data after calibration augmentation, where the abscissa is different performance indicators of the alloy material and the ordinate is the indicator value. Figure 5 is the experimental result diagram of a decoder provided by the present invention for reconstructing candidate semi-solid cast and forged aluminum alloy composition points. In Figure 4 , by uniformly observing the index data of both, it is found that the error values before and after rejection are both less than 0.1, and the systematic error is relatively small. Especially in the error index of Mean Squared Error (MBE), it is lower than 10 -3 orders of magnitude, and it can achieve a better prediction effect. By separately observing, it is found that the error indicators (Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), MBE) after rejecting impurity data are all lower than those without rejection; in terms of the prediction effect, the prediction effect index (R2) of the training set and test set after rejection is increased by 1.7% compared with the prediction effect of the data without rejection, and the prediction effect index (R2) after rejection is increased by 7.2% compared with the prediction effect of the data without rejection. The comparison of the performance indicators before and after rejection is as Figure 4 shown. In Figure 5 , based on the optimal process parameters, the semi-solid squeeze casting Al–Si–Mg–Fe alloy casting is subjected to T6 heat treatment. Figure 5 is the stress-strain curve of the semi-solid squeeze casting Al–Si–Mg–Fe alloy casting after T6 heat treatment, and its tensile strength, yield strength and elongation are 352.5 MPa, 280.4 MPa and 7.3% respectively.

[0097] The present invention screens the feature data by the recursive elimination method and constructs a machine learning model in a recursive manner. After multiple rounds of calculation and experimental iteration, a high thermal conductivity and high electrical conductivity aluminum alloy is finally prepared. The feature with the smallest weight is removed until the number of features in the original data set is 5 to 15, which can improve the accuracy of screening the feature data.

[0098] In addition, in one or more embodiments of the present invention, when the machine learning model is a random forest model, the hyperparameters specifically include:

[0099] The number of decision trees, the maximum depth of decision trees, the maximum number of leaf nodes, the minimum number of samples in leaf nodes, and the number of features to be considered when finding the best split.

[0100] Optionally, the machine learning model integrates a grid search parameter tuning algorithm and a Bayesian optimization algorithm.

[0101] In this embodiment, the grid search ensures a comprehensive exploration of the parameter space, while the Bayesian optimization efficiently locates the optimal parameters based on performance feedback. The combination of the two significantly improves the speed and accuracy of hyperparameter tuning.

[0102] The above is the design method of the semi-solid cast-forged aluminum alloy material provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding design system for the semi-solid cast-forged aluminum alloy material, as Figure 6 shown, the system includes:

[0103] A construction module for constructing a data set containing the contents of various metal elements, non-metal elements, and various process parameters.

[0104] A determination module for constructing a first relationship equation between the data in the data set and the thermal conductivity and tensile strength of the semi-solid cast-forged aluminum alloy through a regression method with collective decision-making of decision trees; determining multiple candidate semi-solid cast-forged aluminum alloy composition points with a preset thermal conductivity and a preset tensile strength according to the data set.

[0105] An evaluation module for evaluating multiple candidate semi-solid cast-forged aluminum alloy composition points using the first relationship equation to obtain a first evaluation result, and correcting the first relationship equation according to the first evaluation result and preset experimental data to obtain a second relationship equation, and evaluating the multiple candidate semi-solid cast-forged aluminum alloy composition points again using the second relationship equation to obtain a second evaluation result.

[0106] A manufacturing module for obtaining an aluminum alloy material that meets the preset performance according to the second evaluation result.

[0107] For the specific limitations of the design system of the semi-solid cast-forged aluminum alloy material, reference can be made to the limitations of the design method of the semi-solid cast-forged aluminum alloy material in the above text, which will not be elaborated here. Each module in the above design system of the semi-solid cast-forged aluminum alloy material can be implemented in whole or in part through software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0108] The present invention also provides a computer-readable storage medium, which stores a computer program that can be used to execute the above Figure 1The provided design method for semi-solid cast-forged aluminum alloy materials.

[0109] The present invention also provides Figure 7 The structural schematic diagram of the computer device shown, as Figure 7 shown, at the hardware level, this computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the above Figure 1 The provided design method for semi-solid cast-forged aluminum alloy materials.

[0110] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the described embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the various methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided by the present invention can include at least one of non-volatile and volatile memories. The non-volatile memory can include a read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical memory, etc. The volatile memory can include a random access memory (RAM) or an external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as a static random access memory (SRAM) or a dynamic random access memory (DRAM), etc.

[0111] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope recorded by the present invention.

Claims

1. A design method for semi-solid cast and forged aluminum alloy material, characterized in that: include: Construct a data set containing various metal element contents, non-metal element contents and various process parameters; Constructing a first relationship equation between the data in the data set and the thermal conductivity and tensile strength of the semi-solid cast and forged aluminum alloy by a regression method with a decision tree collective decision; Determining, based on the data set, a plurality of candidate semi-solid cast and forged aluminum alloy composition points having a preset thermal conductivity and a preset tensile strength; Using the first relationship equation to evaluate the plurality of candidate semi-solid cast and forged aluminum alloy composition points to obtain a first evaluation result, and modifying the first relationship equation according to the first evaluation result and preset experimental data to obtain a second relationship equation, and using the second relationship equation to evaluate the plurality of candidate semi-solid cast and forged aluminum alloy composition points again to obtain a second evaluation result; An aluminum alloy material that meets preset performance is obtained according to the second evaluation result.

2. The method for semi-solid casting and forging of aluminum alloy material according to claim 1, characterized in that: After constructing the data set including the contents of multiple metal elements, the contents of non-metal elements and multiple process parameters, the following is further included: Using a machine learning model to perform feature screening on the data set to obtain feature data; The characteristic data include: element composition, heat treatment state, thermal conductivity, phase composition and density.

3. The method for semi-solid casting and forging of aluminum alloy material according to claim 2, characterized in that: After the feature screening of the data set by using the machine learning model to obtain feature data, the method further includes: According to different alloy systems included in the data set, the hyperparameters of the machine learning model are adjusted to obtain an optimized machine learning model; The feature data is screened again using the optimized machine learning model to obtain optimized feature data.

4. The method for semi-solid casting and forging of aluminum alloy material according to claim 1, characterized in that: Determining, based on the data set, a plurality of candidate semi-solid cast and forged aluminum alloy composition points having a preset thermal conductivity and a preset tensile strength, specifically comprising: Using a tabular conditional adversarial neural network to perform dimensionality reduction processing on the data set, and identifying a first performance region where thermal conductivity and tensile strength are synergistically improved in the latent space after dimensionality reduction; The first performance region is reconstructed by a decoder to obtain a plurality of candidate semi-solid cast and forged aluminum alloy composition points having a preset thermal conductivity and a preset tensile strength.

5. The method for semi-solid casting and forging of aluminum alloy material according to claim 4, characterized in that: The first relationship equation is corrected according to the first evaluation result and the preset experimental data to obtain the second relationship equation, which specifically includes: According to the first evaluation result and the preset experimental data, the decision tree collective decision and the tabular conditional adversarial neural network are improved to obtain an improved decision tree collective decision and an improved tabular conditional adversarial neural network; A second relationship equation between the characteristic data and the thermal conductivity and the tensile strength of the semi-solid cast and forged aluminum alloy is constructed using a regression method with the improved decision tree collective decision.

6. The method for semi-solid casting and forging of aluminum alloy material according to claim 3, characterized in that: When the machine learning model is a random forest model, the hyperparameters specifically include: The number of decision trees, the maximum depth of a decision tree, the maximum number of leaf nodes, the minimum number of samples per leaf node, and the number of features to consider when finding the best split.

7. The method for semi-solid casting and forging of aluminum alloy material according to claim 2, characterized in that: The machine learning model integrates a grid search parameter adjustment algorithm and a Bayesian optimization algorithm.

8. A design system for semi-solid cast and forged aluminum alloy materials, characterized in that: include: A construction module is used to construct a data set containing various metal element contents, non-metal element contents and various process parameters; A determination module, for constructing a first relationship equation between the data in the data set and the thermal conductivity and tensile strength of the semi-solid cast and forged aluminum alloy by a regression method with a decision tree collective decision; Determining, based on the data set, a plurality of candidate semi-solid cast and forged aluminum alloy composition points having a preset thermal conductivity and a preset tensile strength; an evaluation module, configured to evaluate the plurality of candidate semi-solid cast and forged aluminum alloy composition points using the first relationship equation to obtain a first evaluation result, and to modify the first relationship equation according to the first evaluation result and preset experimental data to obtain a second relationship equation, and to evaluate the plurality of candidate semi-solid cast and forged aluminum alloy composition points again using the second relationship equation to obtain a second evaluation result; A manufacturing module is used to obtain an aluminum alloy material that meets preset performance according to the second evaluation result.

9. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method for semi-solid casting and forging of aluminum alloy material according to any one of claims 1 to 7 is implemented.

10. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for semi-solid casting and forging of aluminum alloy material as claimed in any one of claims 1 to 7 is implemented.

Images