A method and apparatus for grain refinement and mechanical property optimization of aluminum alloys

Abstract

The application provides an aluminum alloy coarse grain inhibition and mechanical property optimization method and device, relates to the technical field of metal material design, and the method comprises the following steps: obtaining component sample data of prepared aluminum alloy and observation data of the aluminum alloy sample to construct a training set, and training a machine learning model through the training set; calling the trained machine learning model to predict model input characteristics of component element parameters when the aluminum alloy is prepared, so as to obtain coarse grain probability prediction values and mechanical property indexes when the aluminum alloy is prepared; based on the coarse grain probability prediction values and the mechanical property indexes, at least one target element in the component element parameters is subjected to feature analysis, so as to obtain the contribution degree of the target element to coarse grain inhibition and mechanical property optimization when the aluminum alloy is prepared, and the component element parameters are updated based on the contribution degree. The application is used to solve the technical problems that coarse grain inhibition and mechanical property optimization cannot be realized when the aluminum alloy is developed in the prior art, and the development cycle is long.

Description

Technical Field

[0001] This invention relates to the field of metal material design technology, specifically to a method and apparatus for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys. Background Technology

[0002] The current research and development of high-strength aluminum alloys faces some challenges. Specifically, the coarse grain problem that occurs during the research and development process can lead to a decrease in the mechanical properties of aluminum alloys. Traditional methods can improve this by adding rare earth elements such as yttrium, lanthanum, and cerium. However, there is a lack of quantitative models, making it impossible to accurately control the synergistic effect of yttrium, lanthanum, and cerium, and thus difficult to achieve coarse grain suppression and optimization of mechanical properties.

[0003] Existing technologies offer several methods for controlling the content ratios of various constituent elements and rare earth elements during aluminum alloy preparation to suppress coarse grains and optimize mechanical properties, but all have drawbacks. For example, while regression modeling combines coarse grains and mechanical properties for correlation analysis during preparation, the predicted mechanical property indicators have high errors and cannot effectively guide aluminum alloy development. Characterization techniques can analyze the microstructure of aluminum alloys but cannot guide mechanical property optimization. Experimental methods, while capable of determining the content ratios of constituent elements and rare earth elements under the simultaneous consideration of coarse grains and mechanical properties, require 6-8 months, resulting in an excessively long development cycle. Summary of the Invention

[0004] In view of this, it is necessary to provide a method and apparatus for suppressing coarse grains and optimizing mechanical properties of aluminum alloys, so as to solve the technical problems of difficulty in suppressing coarse grains and optimizing mechanical properties and long development cycle in the research and development of aluminum alloys in the prior art.

[0005] To address the above problems, this invention provides a method for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys, comprising: The compositional sample data and observational data of the aluminum alloy sample are obtained. The compositional sample data includes the content of rare earth elements used in the preparation of the aluminum alloy. The aluminum alloy sample is prepared based on the content of rare earth elements. The observational data includes the microcrystalline data and mechanical property data of the aluminum alloy sample. A training set is constructed based on the component sample data and the observation data, and a machine learning model is trained using the training set. The model input features are constructed based on the elemental parameters of the aluminum alloy during preparation, and the trained machine learning model is called to predict the model input features to obtain the coarse grain probability prediction value and mechanical property index of the aluminum alloy during preparation. Based on the predicted coarse grain probability and the mechanical property index, feature analysis is performed on at least one target element in the composition element parameters to obtain the contribution of the target element to coarse grain suppression and mechanical property optimization during aluminum alloy preparation, and the composition element parameters are updated based on the contribution.

[0006] In one possible implementation, the composition sample data also includes a set iron content, wherein the rare earth element content is determined according to a preset mass ratio of multiple rare earth elements.

[0007] In one possible implementation, the step of constructing model input features based on the elemental composition parameters during aluminum alloy preparation includes: Determine the main chemical composition parameters of the aluminum alloy and the process parameters for preparing the aluminum alloy; Extracting crystal parameters from the microcrystalline data of aluminum alloys; The component element parameters, main chemical composition parameters, crystal parameters, and process parameters used in the preparation of aluminum alloys are vectorized and concatenated to obtain the model input features.

[0008] In one possible implementation, the feature analysis of at least one target element in the compositional element parameters based on the coarse grain probability prediction value and the mechanical property index is performed to obtain the contribution of the target element to the optimization of coarse grain suppression and mechanical property optimization during the preparation of aluminum alloy, including: Multiple combinations of first elements are determined based on any two target parameters among the component element parameters; Based on the predicted coarse grain probability and the mechanical performance index, the SHAP algorithm is called to calculate the single contribution value of each target element in the first element combination to coarse grain suppression and mechanical performance optimization. Multiple combinations of second elements are determined based on at least three target parameters among the component element parameters; Based on the predicted coarse grain probability and the mechanical performance index, the SHAP algorithm is called to calculate the synergistic contribution of the second element combination to coarse grain suppression and mechanical performance optimization. The contribution synergy index of the second element combination is calculated based on the single contribution value and the synergistic contribution value corresponding to the target element in the second element combination.

[0009] In one possible implementation, calculating the contribution synergy index of the second element combination based on the single contribution value corresponding to the target element in the second element combination and the synergistic contribution value includes: The difference between the collaborative contribution value and the single contribution value is determined, and the ratio of the difference to the single contribution value is used as the contribution collaboration index of the second element combination.

[0010] In one possible implementation, updating the component element parameters based on the contribution includes: Determine the content ranges and boundaries of rare earth elements and iron elements for preparing aluminum alloys. Determine the process conditions boundary for preparing aluminum alloys; The contribution, process condition boundary, and content range boundary are input into the target optimization model for optimization to obtain the optimal content of each component element when preparing aluminum alloy; The component element parameters are adjusted according to the optimal content.

[0011] In one possible implementation, the constraints of the objective optimization model include: The content of each main chemical component of the aluminum alloy does not exceed the preset content range; The iron content of aluminum alloys shall not exceed the specified range; The two elements with the largest contribution synergistic index among the constituent elements in the preparation of aluminum alloys must meet the set mass ratio. The process parameters for preparing aluminum alloys shall not exceed the set parameter range; The optimization objective of the target optimization model is to minimize the objective function. The value of the objective function is obtained by weighted summation of the coarse grain probability prediction value and the iron content deviation penalty value. The iron content deviation penalty value is the median value of the iron content exceeding the specified content range.

[0012] The present invention also provides a device for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys, comprising: The acquisition module is used to acquire compositional sample data for preparing aluminum alloys and observational data of aluminum alloy samples. The compositional sample data includes the content of rare earth elements used in preparing the aluminum alloy, and the aluminum alloy sample is prepared based on the rare earth element content. The observational data includes the microcrystalline data and mechanical property data of the aluminum alloy sample. The training module is used to construct a training set based on the component sample data and the observation data, and to train a machine learning model using the training set. The prediction module is used to construct model input features based on the elemental parameters of the aluminum alloy during preparation, and call the trained machine learning model to predict the model input features to obtain the coarse grain probability prediction value and mechanical property index during the preparation of the aluminum alloy. The analysis module is used to perform feature analysis on at least one target element in the composition element parameters based on the coarse grain probability prediction value and the mechanical performance index, respectively, to obtain the contribution of the target element to coarse grain suppression and optimization of mechanical properties when preparing aluminum alloy, and to update the composition element parameters based on the contribution.

[0013] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a program; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps of the above-described method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys.

[0015] The beneficial effects of the above-described implementation method are as follows: The method and apparatus for suppressing coarse grains and optimizing mechanical properties of aluminum alloys provided by this invention train a machine learning model using the compositional sample data and observational data of the aluminum alloy samples. This allows the machine learning model to quickly predict coarse grains and mechanical properties based on the compositional samples during aluminum alloy preparation. Furthermore, during aluminum alloy preparation, the prediction results are used for quantitative analysis of target elements to determine their contribution to suppressing coarse grains and optimizing mechanical properties. This allows for targeted design of compositional element parameters during aluminum alloy development, with the goal of suppressing coarse grains and optimizing mechanical properties. In addition, the entire development process utilizes machine learning model prediction and quantitative analysis, eliminating the need for element-by-element content proportioning and experimental verification, effectively shortening the aluminum alloy development cycle. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0017] Figure 1 A schematic flowchart illustrating the method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys provided by this invention; Figure 2 A schematic diagram illustrating the principle of the method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys provided by the present invention; Figure 3 A schematic diagram of the device for suppressing coarse grains and optimizing mechanical properties of aluminum alloys provided by the present invention; Figure 4 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] In the description of the embodiments of this application, unless otherwise stated, "a plurality of" means two or more.

[0020] In this embodiment of the invention, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, apparatus, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product or device.

[0021] The naming or numbering of steps in the embodiments of the present invention does not mean that the steps in the method flow must be executed in the time / logical order indicated by the naming or numbering. The execution order of the named or numbered process steps can be changed according to the technical purpose to be achieved, as long as the same or similar technical effect can be achieved.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] This invention provides a method and apparatus for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys. It can be applied to the research and development of a rare earth aluminum alloy. The execution entity can be a server, terminal, or remote cloud device. By designing experimental materials in the research and development scenario, obtaining the compositional sample data and observational data of the aluminum alloy samples, a training set is created and uploaded. Then, the machine learning model is trained by calling the method for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys provided by this invention. Furthermore, during the actual preparation of the aluminum alloy, the compositional element parameters during the preparation of the aluminum alloy are input for the machine learning model to make predictions. Finally, the updated compositional element parameters are output as the basis for the content ratio of the prepared aluminum alloy.

[0024] The following section details the method for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys provided by this invention.

[0025] Figure 1 This is a schematic flowchart illustrating the method for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys provided by the present invention. Figure 1 As shown, the method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys can be achieved through the following steps 101 to 104, which are explained in detail below.

[0026] Step 101: Obtain the compositional sample data and observation data of the aluminum alloy sample.

[0027] Here, the compositional sample data includes the content of rare earth elements used in the preparation of the aluminum alloy. For example... Figure 2 As shown, the composition was first designed according to the GB / T3190-2020 standard required for aluminum alloy research and development. A composition gradient sample was designed. The iron content was first determined to be ≤0.25wt%. For rare earth elements, the yttrium content range was set to 0.2-1.7wt%, the lanthanum content range to 0.02-0.15wt%, and the cerium content range to 0.02-0.15wt%.

[0028] The aluminum alloy samples were prepared based on the rare earth element content. Next, the rare earth element content from these compositional sample data can be configured for experimental verification. Specifically, deformation and solution treatment temperatures are set to perform hot compression tests or solution treatments to prepare aluminum alloy samples. The process flow is designed with specific conditions, such as controlling the deformation for hot compression tests to 20% to 70%, setting the solution treatment temperature for 7-series aluminum alloys to 460℃-475℃, and for 6-series aluminum alloys to 525℃-565℃. Of course, the coarse grain structure and mechanical properties of the samples are not considered during sample preparation; the corresponding samples are prepared directly based on the provided compositional sample data.

[0029] Furthermore, observational data were collected by sampling the prepared aluminum alloy samples. This data included microcrystalline data and mechanical property data. Microcrystalline data, such as the proportion of coarse grains and grain size, could be determined using electron backscatter diffraction (EBSD) grain boundary diagrams and coarse grain statistics. Mechanical property data, including various mechanical properties such as tensile strength and elongation, were estimated based on the correlation between coarse grains and mechanical properties obtained through microcrystalline analysis of the aluminum alloy samples.

[0030] In one possible implementation, the composition sample data also includes a set iron content, and the rare earth element content is determined according to a preset mass ratio of multiple rare earth elements. For example, according to the GB / T3190-2020 standard, the iron content is strictly controlled within the range of 0.15-0.25 wt%, and the rare earth elements yttrium, lanthanum, and cerium need to be added in a preset mass ratio of (8-12):(0.8-1.2):1.

[0031] By controlling the iron and rare earth element content of the sample, the machine learning model can be trained on data samples that meet the GB / T3190-2020 standard, ensuring the rationality and practicality of the model's prediction of coarse grains and mechanical properties.

[0032] Step 102: Construct a training set based on the component sample data and observation data, and train the machine learning model using the training set.

[0033] Here, the training data for the machine learning model consists of 14-dimensional compositional and process parameters. From the compositional sample data, nine parameters can be extracted: the main chemical components Si, Fe, Cu, Mn, Mg, Zn, and the contents of rare earth elements Y, La, and Ce. Two parameters, coarse grain ratio and average grain size, can be extracted from the observational data. Additionally, three process parameters involved in the process flow are included: deformation time, solution temperature, and holding time. Thus, a total of 14 parameters are extracted to construct the 14-dimensional compositional and process parameter features as training samples. Furthermore, for the crystal parameters in the aluminum alloy sample observational data, the number of coarse grains can be statistically analyzed, and the coarse grain ratio can be calculated as a label to construct the training set.

[0034] Machine learning models can be implemented using deep neural networks (DNNs) or XGBoost models. Taking a DNN network as an example, its hierarchical topology consists of an input layer, hidden layers, and an output layer. In the input layer, the core of machine learning uses 14-dimensional compositional process parameters as feature inputs, which are fed into the hidden layers. The hidden layers contain three layers with 128, 64, and 32 nodes respectively. Finally, the output layer maps a one-dimensional probability value, representing the coarse-grained probability prediction value.

[0035] During training, a machine learning model is trained using a constructed training set. The training samples are input into the machine learning model to predict a coarse grain probability. This probability is then compared with the coarse grain ratio label to construct a loss function. The parameters of the machine learning model are optimized using the loss function, enabling the machine learning model to predict coarse grains based on 14-dimensional features.

[0036] Step 103: Construct model input features based on the elemental parameters of the aluminum alloy during preparation, and call the trained machine learning model to predict the model input features to obtain the coarse grain probability prediction value and mechanical property index of the aluminum alloy during preparation.

[0037] When preparing aluminum alloys, the first step is to determine the elemental parameters of the alloy composition, including the content of various chemical elements and the corresponding process parameters. In addition, the observation data of the prepared aluminum alloy, including the coarse grain ratio and average grain size, are also used to construct the model input features, which serve as the input to the machine learning model.

[0038] In one possible implementation, the model input features are constructed based on the elemental parameters of the aluminum alloy during its preparation. This can be achieved in the following ways, which will be explained in detail below.

[0039] First, determine the main chemical composition parameters of the aluminum alloy and the process parameters for preparing the aluminum alloy. The main chemical composition parameters are the main chemical components that make up the aluminum alloy sample: Si, Fe, Cu, Mn, Mg, and Zn. The process parameters are the three process parameters involved in the aluminum alloy preparation process, including deformation time, solution temperature, and holding time.

[0040] Then, the crystal parameters in the microcrystalline data of the aluminum alloy are extracted, that is, the observation data are extracted from the prepared aluminum alloy, including the crystal parameters in the microcrystalline data, namely the coarse grain ratio and the average grain size.

[0041] Finally, the elemental parameters, main chemical composition parameters, crystal parameters, and process parameters used in the preparation of aluminum alloys are vectorized and concatenated to obtain the model input features.

[0042] See also Figure 2 The composition element parameters are the contents of rare earth elements such as Y, La, and Ce used in the preparation of aluminum alloys. Since the core of machine learning is to input 14-dimensional composition and process parameters, the composition element parameters, chemical composition parameters, crystal parameters, and process parameters are vectorized together to obtain 14-dimensional model input features.

[0043] In this embodiment of the invention, the model input features are constructed based on the various compositional process parameters for preparing aluminum alloys, enabling the machine learning model to map the probability of coarse grains to aluminum alloys and achieve quantitative correlation analysis from rare earth elements to microstructure.

[0044] Once the model input features are constructed, the trained machine learning model is invoked to predict the model input features, thereby obtaining the predicted coarse grain probability value and mechanical property index when preparing aluminum alloy.

[0045] Here, based on the input features of the model, the trained machine learning model will make predictions and ultimately determine the predicted coarse grain probability value for the aluminum alloy preparation. The value ranges from 0 to 100%, where 0 indicates no coarse grains and 100% indicates all coarse grains. Furthermore, based on the predicted coarse grain probability value, the number of coarse grains can be determined. Then, combined with grain size analysis, the microstructure of the aluminum alloy sample is analyzed. Finally, based on the correlation between coarse grains and mechanical properties, the mechanical property indicators of the aluminum alloy sample, such as tensile strength and elongation, are estimated.

[0046] Therefore, based on the number of aluminum alloys prepared, multiple sets of predictive data can be obtained. For example, preparing 100 kinds of aluminum alloys can yield 100 sets of predictive data for subsequent analysis.

[0047] Step 104: Based on the coarse grain probability prediction value and mechanical property index, perform feature analysis on at least one target element in the composition element parameters to obtain the contribution of the target element to coarse grain suppression and mechanical property optimization when preparing aluminum alloy, and update the composition element parameters based on the contribution.

[0048] like Figure 2 As shown, after the machine learning model (DNN network) makes predictions, the feature contributions are then analyzed using the SHAP algorithm to determine the composition range and predicted performance indicators of the aluminum alloy.

[0049] Based on the output prediction data, quantitative analysis can be performed on the target elements in the compositional element parameters. The contribution of these target elements' content to coarse grain suppression and mechanical property optimization can be analyzed using coarse grain probability prediction values ​​and mechanical property indices. This contribution includes the individual contribution value of a single target element, as well as the synergistic contribution value of a single target element within a combination of target elements and its corresponding synergistic contribution index. Quantitative analysis can be achieved using the SHAP algorithm, which will be explained in detail below.

[0050] In one possible implementation, at least one target element in the composition element parameters is analyzed based on the coarse grain probability prediction value and mechanical property index to obtain the contribution of the target element to coarse grain suppression and mechanical property optimization during the preparation of aluminum alloy.

[0051] First, multiple combinations of first elements are determined based on any two target parameters from the composition element parameters. For example, the composition element parameters include the content of 13 chemical elements that constitute aluminum alloys. Two chemical elements are randomly selected as target parameters, thus forming a total of 2... 13 (i.e., 8192) combinations of the first element. For example, if iron (Fe) and yttrium (Y) are chosen as the first element combination, denoted as S=[Fe, Y], it means that in this quantitative analysis, only the contributions of iron (Fe) and yttrium (Y) to the suppression of coarse grains and the optimization of mechanical properties of aluminum alloys are considered.

[0052] Next, based on the predicted coarse grain probability and mechanical performance indicators, the SHAP algorithm is called to calculate the single contribution of each target element in the first element combination to coarse grain suppression and mechanical performance optimization.

[0053] The SHAP algorithm specifically calculates the Shapley value, used for feature attribution tasks, interpreting the predictions of machine learning models, and measuring the contribution of each feature to the model's output. In the first element combination S, the SHAP algorithm calculates the marginal contribution. That is, for all possible feature subset combinations, it calculates the marginal contribution of feature i (the target element) added to the first element combination S and averages them according to weights. The weights are determined by factorial terms to ensure that different first element combination sizes and different addition orders are fairly weighted, ultimately yielding a fair contribution of feature i to the machine learning model's output. A rough formula can be expressed as follows: (1) Where i represents the feature, i.e., the target element. This represents the Shapley value of feature i on the prediction result of the machine learning model on the first element combination S, which is also the attribution score of feature i, and serves as the contribution value. The prediction result is the coarse grain probability prediction value and mechanical performance index. Let S represent the subset of features that does not contain feature i, and F be the set of all features (target elements) in the component element parameters. Denotes the set after removing feature i from F. , , This represents the factorial term related to the number of combinations, used to calculate the weight of the operation of adding the first element S to feature i. This represents the change in the prediction result of the machine learning model after feature i is added to subset S.

[0054] Therefore, the individual contribution value of each target element in the first element combination to coarse grain suppression and mechanical property optimization can be calculated using the above formula (1).

[0055] Furthermore, multiple combinations of second elements are determined based on at least three target parameters in the compositional element parameters. For example, the compositional element parameters include the content of 13 chemical elements that constitute the aluminum alloy. By arbitrarily selecting at least three chemical elements as target parameters, multiple combinations of second elements can be formed. For example, selecting yttrium (Y), lanthanum (La), and cerium (Ce) as the second element combination, denoted as S=[Ce, La, Ce], indicates that in this quantitative analysis, the synergistic contribution of yttrium (Y), lanthanum (La), and cerium (Ce) to the suppression of coarse grains and the optimization of mechanical properties is considered simultaneously.

[0056] Based on the predicted coarse grain probability and mechanical performance indicators, the SHAP algorithm is used to calculate the synergistic contribution of the second element combination to coarse grain suppression and mechanical performance optimization.

[0057] Here, the SHAP algorithm calculates the contribution of multiple features to the output of the machine learning model. Using the same method, feature i is denoted as yttrium (Y), lanthanum (La), and cerium (Ce). Formula (1) is called to calculate the fair contribution of yttrium (Y), lanthanum (La), and cerium (Ce) to the coarse-grain probability prediction and mechanical performance indicators of the machine learning model output. This fair contribution is recorded as the synergistic contribution value of the three rare earth elements. .

[0058] Finally, the contribution synergy index of the second element combination is calculated based on the single contribution value and synergistic contribution value corresponding to the target element in the second element combination.

[0059] Here, the individual contribution value corresponding to the target element in the second element combination can be calculated one by one from the above formula (1), for example, the individual contribution value corresponding to yttrium Y. Then, the contribution synergy index of the second element combination is calculated using the synergy contribution value.

[0060] In one possible implementation, the contribution synergy index of the second element combination is calculated based on the single contribution value and synergistic contribution value corresponding to the target element in the second element combination. This can be achieved in the following way, which is explained in detail below.

[0061] First, the difference between the collaborative contribution value and the single contribution value is determined, and then the ratio of the difference to the single contribution value is used as the contribution collaboration index of the second element combination.

[0062] Taking yttrium (Y) as the target element and yttrium (Y), lanthanum (La), and cerium (Ce) as the second element combination, the formula for calculating the contribution synergy index is as follows: (2) in, This indicates a feature subset containing lanthanum (La) and cerium (Ce) in addition to yttrium (Y), i.e., a second element combination. This indicates a combination of elements containing only yttrium (Y). This represents the single contribution of the second element combination containing only yttrium (Y) to the coarse-grained probability prediction and mechanical performance indicators output by the machine learning model. This represents the synergistic contribution of the second element combination, including yttrium (Y), lanthanum (La), and cerium (Ce), to the coarse-grain probability prediction and mechanical performance indicators output by the machine learning model. The contribution synergy index represents the combination of the second element of yttrium (Y), lanthanum (La), and cerium (Ce).

[0063] Based on the characteristic contribution of SHAP analysis, the influence of the content of each target element on coarse grain suppression and mechanical property optimization can be determined. For example, when yttrium > 0.6 wt%, the probability of coarse grains in the prepared aluminum alloy is reduced by 40%, and the tensile strength is increased by 18%. In the synergy of lanthanum and cerium, the grain boundary pinning efficiency is increased by 30%.

[0064] In this embodiment of the invention, the SHAP algorithm is used to perform quantitative analysis on each target element. This allows for the analysis of the contribution of a single target element and combinations of multiple target elements to coarse grain suppression and mechanical property optimization. This provides a targeted basis for subsequent updates to the content of constituent elements, ensuring the effective implementation of coarse grain suppression and mechanical property optimization.

[0065] Finally, the component element parameters are updated based on their contribution.

[0066] Here, updating the elemental composition parameters means updating the content of each chemical element used in the preparation of the aluminum alloy, including the main chemical composition and added rare earth elements. However, it is necessary to ensure that the content ranges of each rare earth element and iron element do not exceed the limits of GB / T 3190-2020 standard, and that the process parameters therein do not exceed the feasible process limits. Therefore, various constraints need to be met when determining the content. To address this, this invention provides a multi-objective optimization algorithm to determine each element composition.

[0067] In one possible implementation, updating the component element parameters based on contribution can be achieved in the following way, which is explained in detail below.

[0068] First, determine the content range boundaries of each rare earth element and iron element in the preparation of the aluminum alloy. These content range boundaries must meet the GB / T 3190-2020 standard, including an iron content range of 0.15-0.25 wt%, a yttrium content range of 0.2-1.7 wt%, a lanthanum content range of 0.02-0.15 wt%, and a cerium content range of 0.02-0.15 wt%.

[0069] Secondly, the process conditions for preparing aluminum alloys must be defined. For example, the solubility of rare earth elements must not exceed the limit value, and the solution temperature must not exceed the temperature threshold to prevent calcination.

[0070] Then, the contribution, process condition boundary, and content range boundary are input into the target optimization model for optimization to obtain the optimal content of each component element when preparing aluminum alloy.

[0071] In one possible implementation, the constraints of the objective optimization model include the following four constraints, which are explained one by one below.

[0072] First, the content of each main chemical component of the aluminum alloy shall not exceed the preset content range, that is, the main chemical components Si, Fe, Cu, Mn, Mg, Zn, etc., and the content of these elements shall also meet the relevant standards of GB / T 3190-2020. For example, the content range of silicon Si is 0.4-0.8wt%, and the content range of copper Cu is 1.2-2.0wt%, etc.

[0073] Second, the iron content of the aluminum alloy must not exceed the specified range. Limiting the iron content is crucial in the preparation of aluminum alloys. According to the relevant standard GB / T 3190-2020, the specified iron content range is 0.15-0.25 wt%.

[0074] Third, the two elements with the largest contribution synergistic index among the constituent elements of the aluminum alloy must meet the set mass ratio.

[0075] For example, the two rare earth elements, La and Ce, have the largest contribution synergistic index in the SHAP algorithm calculation. Therefore, in order to make the prepared aluminum alloy have lower coarse grains and better mechanical properties, it is necessary to ensure the configuration of these two elements, for example, setting their mass ratio to meet 0.8:1.2.

[0076] Fourth, the process parameters for preparing aluminum alloys must not exceed the set parameter range. For example, the solution temperature must not exceed 475℃ to ensure that the corresponding process flow is met.

[0077] The objective of the target optimization model is to minimize the objective function. The value of the objective function is obtained by weighted summing of the predicted coarse grain probability value and the penalty value for deviation of iron content. The penalty value for deviation of iron content is the median value of the iron content exceeding the specified range. The objective function is denoted as J and expressed by the following formula: (3) in, This represents the predicted probability value for coarse grains. Indicates the iron content, This indicates that the iron content deviates from the penalty value, that is, the iron content exceeds the median value of the specified content range. The specified content range for iron is 0.15-0.25wt%, and the median can be 0.20wt. and These are preset weight parameters, such as 0.7 and 0.3.

[0078] The purpose of minimizing the objective function J is to ensure that the objective optimization model, in multi-objective optimization, maximizes the coarse-grained probability prediction value. Smaller size achieves the effect of suppressing coarse grains, while also ensuring that the iron content is within the specified range of 0.15-0.25wt.

[0079] In this embodiment of the invention, by constructing a target optimization model, the constraints and objective function are determined so that the optimal content of each component element during the preparation of aluminum alloy can meet the various standards for preparation, and on this basis, the effect of coarse grain suppression is optimized, thereby improving the mechanical properties of aluminum alloy.

[0080] Here, by designing four constraints and an objective function to construct a multi-objective optimization model, the contribution, process condition boundary, and content range boundary can be input into the objective optimization model for optimization. Under the constraints, the objective optimization model can optimize the content of each component element in the preparation of aluminum alloy with minimizing the objective function as the optimization objective, and finally generate the optimal content of each component element, including the content of each chemical element in the preparation of aluminum alloy.

[0081] Finally, the component element parameters are adjusted according to the optimal content. These optimal contents output by the multi-objective optimization model are the optimal formulation for suppressing coarse grains and optimizing mechanical properties of aluminum alloys. Based on this configuration, the component element parameters during the preparation of aluminum alloys can be updated and adjusted. For example, a formulation of 0.82 wt% yttrium, 0.06 wt% lanthanum, 0.05 wt% cerium, and 0.20 wt% iron can be designed. After verification of the composition range and predicted performance indicators, the aluminum alloy prepared according to this optimal formulation has a tensile strength of 560 MPa, an elongation of 10.2%, and a coarse grain ratio of only 4.3%, thus achieving suppression of coarse grains and optimization of mechanical properties in aluminum alloys.

[0082] In this embodiment of the invention, the contribution calculated by the SHAP algorithm is used as the optimization basis. Under the constraints of various component elements and process parameters, the optimal content of each component element is directly determined by a multi-objective optimization algorithm. Without the need for content ratio and experimental verification for each element, the coarse grains of aluminum alloys can be effectively suppressed and the mechanical properties optimized, thus effectively shortening the research and development cycle of aluminum alloys.

[0083] In summary, this invention trains a machine learning model using compositional sample data and observational data of aluminum alloy samples. This allows the machine learning model to rapidly predict coarse grains and mechanical properties based on the compositional samples during aluminum alloy preparation. Furthermore, the prediction results are used for quantitative analysis of target elements to determine their contribution to coarse grain suppression and mechanical property optimization. This enables targeted design of compositional element parameters during aluminum alloy development, achieving both coarse grain suppression and mechanical property optimization. Moreover, the entire development process utilizes machine learning model prediction and quantitative analysis, eliminating the need for element-by-element content proportioning and experimental verification, effectively shortening the aluminum alloy development cycle.

[0084] The following is a detailed description of the aluminum alloy coarse grain suppression and mechanical property optimization device provided by the present invention.

[0085] Figure 3 This is a schematic diagram of the structure of the aluminum alloy coarse grain suppression and mechanical property optimization device provided by the present invention, as shown below. Figure 3 As shown, the device for suppressing coarse grains and optimizing mechanical properties of aluminum alloys specifically includes: an acquisition module 301, a training module 302, a prediction module 303, and an analysis module 304.

[0086] The acquisition module 301 is used to acquire the composition sample data and observation data of the aluminum alloy sample for preparing the aluminum alloy. The composition sample data includes the content of rare earth elements used in preparing the aluminum alloy, and the aluminum alloy sample is prepared based on the content of rare earth elements. The observation data includes the microcrystalline data and mechanical property data of the aluminum alloy sample. Training module 302 is used to construct a training set based on the component sample data and the observation data, and to train a machine learning model using the training set; The prediction module 303 is used to construct model input features based on the composition element parameters during the preparation of aluminum alloy, and call the trained machine learning model to predict the model input features to obtain the coarse grain probability prediction value and mechanical property index during the preparation of aluminum alloy. The analysis module 304 is used to perform feature analysis on at least one target element in the composition element parameters based on the coarse grain probability prediction value and the mechanical performance index, respectively, to obtain the contribution of the target element to coarse grain suppression and optimization of mechanical properties when preparing aluminum alloy, and to update the composition element parameters based on the contribution.

[0087] The aluminum alloy coarse grain suppression and mechanical property optimization device provided in the above embodiments can realize the technical solutions described in the above aluminum alloy coarse grain suppression and mechanical property optimization method embodiments. The specific implementation principle of each module or unit can be referred to the corresponding content in the above aluminum alloy coarse grain suppression and mechanical property optimization method embodiments, and their technical effects can also be referred to each other, which will not be repeated here.

[0088] like Figure 4 As shown, the present invention also provides an electronic device 400. The electronic device 400 includes a processor 401, a memory 402, and a display 403. Figure 4 Only some components of the electronic device 400 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0089] In some embodiments, memory 402 may be an internal storage unit of electronic device 400, such as a hard disk or memory of electronic device 400. In other embodiments, memory 402 may also be an external storage device of electronic device 400, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 400.

[0090] Furthermore, the memory 402 may include both internal storage units of the electronic device 400 and external storage devices. The memory 402 is used to store application software and various types of data installed on the electronic device 400.

[0091] In some embodiments, processor 401 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 402 or process data, such as the method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys in this invention.

[0092] In some embodiments, display 403 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 403 is used to display information from electronic device 400 and to display a visual user interface. Components 401-403 of electronic device 400 communicate with each other via a system bus.

[0093] In some embodiments of the present invention, when the processor 401 executes the aluminum alloy preparation program in the memory 402, the following steps can be implemented: acquiring compositional sample data and observational data of the aluminum alloy sample, wherein the compositional sample data includes the content of rare earth elements used in preparing the aluminum alloy, the aluminum alloy sample is prepared based on the rare earth element content, and the observational data includes the microcrystalline data and mechanical property data of the aluminum alloy sample; constructing a training set based on the compositional sample data and the observational data, and training a machine learning model through the training set; constructing model input features based on the compositional element parameters in preparing the aluminum alloy, and calling the trained machine learning model to predict the model input features to obtain the coarse grain probability prediction value and mechanical property index in preparing the aluminum alloy; performing feature analysis on at least one target element in the compositional element parameters based on the coarse grain probability prediction value and the mechanical property index to obtain the contribution of the target element to coarse grain suppression and mechanical property optimization in preparing the aluminum alloy, and updating the compositional element parameters based on the contribution value.

[0094] It should be understood that when the processor 401 executes the aluminum alloy preparation program in the memory 402, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0095] Furthermore, the embodiments of the present invention do not specifically limit the type of electronic device 400 mentioned. Electronic device 400 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, electronic device 400 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0096] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program performs the coarse grain suppression and mechanical property optimization method for aluminum alloys provided by the methods described above. This method includes: acquiring compositional sample data for preparing the aluminum alloy and observational data of the aluminum alloy sample, wherein the compositional sample data includes the content of rare earth elements used in preparing the aluminum alloy, the aluminum alloy sample is prepared based on the rare earth element content, and the observational data includes microcrystalline data and mechanical property data of the aluminum alloy sample; constructing a training set based on the compositional sample data and the observational data, and training a machine learning model using the training set; constructing model input features based on the compositional element parameters used in preparing the aluminum alloy, and calling the trained machine learning model to predict the model input features to obtain a coarse grain probability prediction value and mechanical property index for preparing the aluminum alloy; performing feature analysis on at least one target element in the compositional element parameters based on the coarse grain probability prediction value and the mechanical property index, respectively, to obtain the contribution of the target element to coarse grain suppression and mechanical property optimization during aluminum alloy preparation, and updating the compositional element parameters based on the contribution.

[0097] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0098] The above provides a detailed description of the method and apparatus for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys, characterized in that, include: The compositional sample data and observational data of the aluminum alloy sample are obtained. The compositional sample data includes the content of rare earth elements used in the preparation of the aluminum alloy. The aluminum alloy sample is prepared based on the content of rare earth elements. The observational data includes the microcrystalline data and mechanical property data of the aluminum alloy sample. A training set is constructed based on the component sample data and the observation data, and a machine learning model is trained using the training set. The model input features are constructed based on the elemental parameters of the aluminum alloy during preparation, and the trained machine learning model is called to predict the model input features to obtain the coarse grain probability prediction value and mechanical property index of the aluminum alloy during preparation. Based on the predicted coarse grain probability and the mechanical property index, feature analysis is performed on at least one target element in the composition element parameters to obtain the contribution of the target element to coarse grain suppression and mechanical property optimization during aluminum alloy preparation, and the composition element parameters are updated based on the contribution.

2. The method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys according to claim 1, characterized in that, The component sample data also includes a set iron content, and the rare earth element content is determined according to a preset mass ratio of multiple rare earth elements.

3. The method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys according to claim 1, characterized in that, The step of constructing model input features based on the elemental composition parameters during aluminum alloy preparation includes: Determine the main chemical composition parameters of the aluminum alloy and the process parameters for preparing the aluminum alloy; Extracting crystal parameters from the microcrystalline data of aluminum alloys; The component element parameters, main chemical composition parameters, crystal parameters, and process parameters used in the preparation of aluminum alloys are vectorized and concatenated to obtain the model input features.

4. The method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys according to claim 1, characterized in that, The method involves performing feature analysis on at least one target element in the compositional element parameters based on the predicted coarse grain probability value and the mechanical property index, respectively, to obtain the contribution of the target element to the optimization of coarse grain suppression and mechanical property optimization during aluminum alloy preparation, including: Multiple combinations of first elements are determined based on any two target parameters among the component element parameters; Based on the predicted coarse grain probability and the mechanical performance index, the SHAP algorithm is called to calculate the single contribution value of each target element in the first element combination to coarse grain suppression and mechanical performance optimization. Multiple combinations of second elements are determined based on at least three target parameters among the component element parameters; Based on the predicted coarse grain probability and the mechanical performance index, the SHAP algorithm is called to calculate the synergistic contribution of the second element combination to coarse grain suppression and mechanical performance optimization. The contribution synergy index of the second element combination is calculated based on the single contribution value and the synergistic contribution value corresponding to the target element in the second element combination.

5. The method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys according to claim 4, characterized in that, The calculation of the contribution synergy index of the second element combination based on the single contribution value and the synergistic contribution value corresponding to the target element in the second element combination includes: The difference between the collaborative contribution value and the single contribution value is determined, and the ratio of the difference to the single contribution value is used as the contribution collaboration index of the second element combination.

6. The method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys according to claim 1, characterized in that, The step of updating the component element parameters based on the contribution includes: Determine the content ranges and boundaries of rare earth elements and iron elements for preparing aluminum alloys. Determine the process conditions boundary for preparing aluminum alloys; The contribution, process condition boundary, and content range boundary are input into the target optimization model for optimization to obtain the optimal content of each component element when preparing aluminum alloy; The component element parameters are adjusted according to the optimal content.

7. The method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys according to claim 1, characterized in that, The constraints of the objective optimization model include: The content of each main chemical component of the aluminum alloy does not exceed the preset content range; The iron content of aluminum alloys shall not exceed the specified range; The two elements with the largest contribution synergistic index among the constituent elements in the preparation of aluminum alloys must meet the set mass ratio. The process parameters for preparing aluminum alloys shall not exceed the set parameter range; The optimization objective of the target optimization model is to minimize the objective function. The value of the objective function is obtained by weighted summation of the coarse grain probability prediction value and the iron content deviation penalty value. The iron content deviation penalty value is the median value of the iron content exceeding the specified content range.

8. A device for suppressing coarse grains and optimizing the mechanical properties of aluminum alloys, characterized in that, include: The acquisition module is used to acquire compositional sample data for preparing aluminum alloys and observational data of aluminum alloy samples. The compositional sample data includes the content of rare earth elements used in preparing the aluminum alloy, and the aluminum alloy sample is prepared based on the rare earth element content. The observational data includes the microcrystalline data and mechanical property data of the aluminum alloy sample. The training module is used to construct a training set based on the component sample data and the observation data, and to train a machine learning model using the training set. The prediction module is used to construct model input features based on the elemental parameters of the aluminum alloy during preparation, and call the trained machine learning model to predict the model input features to obtain the coarse grain probability prediction value and mechanical property index during the preparation of the aluminum alloy. The analysis module is used to perform feature analysis on at least one target element in the composition element parameters based on the coarse grain probability prediction value and the mechanical performance index, respectively, to obtain the contribution of the target element to coarse grain suppression and optimization of mechanical properties when preparing aluminum alloy, and to update the composition element parameters based on the contribution.

9. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps of the method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for suppressing coarse grains and optimizing mechanical properties of aluminum alloys as described in any one of claims 1 to 7.

Images