Method and equipment for predicting mechanical property of extruded magnesium alloy

Abstract

The invention provides a method and equipment for predicting mechanical properties of extruded magnesium alloy, which comprises the following steps of: setting descriptors for representing components-process-properties before and after magnesium alloy extrusion forming, screening descriptors beneficial to machine learning in all descriptors by utilizing a feature selection method, eliminating useless or harmful descriptors, and predicting mechanical properties of the extruded magnesium alloy. And finally, a machine learning model capable of accurately predicting the mechanical properties of the extruded magnesium alloy is trained. Through the trained prediction model, the mechanical property of the magnesium alloy after extrusion forming can be accurately predicted under the condition that any magnesium alloy component and any extrusion process are known.

Description

TECHNICAL FIELD

[0001] The present application relates to a method and device for predicting the mechanical properties of extruded magnesium alloys. BACKGROUND

[0002] The mechanical properties of magnesium alloys are influenced by various factors such as alloy composition, processing techniques, heat treatment, etc. Changes in these factors can result in significant fluctuations in the mechanical properties of magnesium alloys.

[0003] Currently, the main methods for predicting the mechanical properties of extruded magnesium alloys are based on experimental data and empirical formulas. However, due to insufficient experimental conditions and data, as well as the limitations of empirical formulas, these prediction methods often have inaccuracies and are unreliable. SUMMARY

[0004] The present application aims to provide a method and device for predicting the mechanical properties of extruded magnesium alloys.

[0005] To solve the above problems, the present application provides a method for predicting the mechanical properties of extruded magnesium alloys, comprising:

[0006] Selecting descriptors that describe the composition information of the extruded magnesium alloy; selecting descriptors that describe the extrusion process of the extruded magnesium alloy; selecting descriptors that describe the mechanical properties of the extruded magnesium alloy;

[0007] Based on the descriptors that describe the composition information of the extruded magnesium alloy, the descriptors that describe the extrusion process, and the descriptors that describe the mechanical properties, a standard Xgboost model is used to establish an initial prediction model based on the descriptors that describe the composition information of the extruded magnesium alloy and the descriptors that describe the extrusion process of the extruded magnesium alloy.

[0008] Based on the initial prediction model and using the sequential backward selection method, the descriptors that describe the composition information of the extruded magnesium alloy and the descriptors that describe the extrusion process of the extruded magnesium alloy corresponding to the descriptors that describe the mechanical properties of the extruded magnesium alloy are screened to obtain the screened descriptors that describe the composition information and the descriptors that describe the extrusion process.

[0009] Based on the screened descriptors that describe the composition information and the descriptors that describe the extrusion process, and the data distribution corresponding to each descriptor, the initial prediction model is adjusted, and the best hyperparameter combination is obtained using grid search to obtain a trained prediction model.

[0010] The values of the descriptors that describe the composition information and the values of the descriptors that describe the extrusion process of the magnesium alloy to be predicted are input into the trained prediction model to obtain the output values of the descriptors that describe the mechanical properties of the magnesium alloy to be predicted.

[0011] Further, in the above method, the descriptors that describe the composition information of the extruded magnesium alloy include:

[0012] (1) the mass percentage of magnesium element in the alloy;

[0013] (2) the mass percentage of the element with the highest content in the alloy except magnesium;

[0014] (3) the atomic number of the element with the highest content in the alloy except magnesium;

[0015] (4) the mass percentage of the second most abundant element in the alloy except magnesium;

[0016] (5) the atomic number of the second most abundant element in the alloy except magnesium;

[0017] (6) the mass percentage of the third most abundant element in the alloy except magnesium;

[0018] (7) the atomic number of the third most abundant element in the alloy except magnesium;

[0019] (8) the average value of atomic radius:

[0020]

[0021] wherein w is the mass percentage of magnesium and other elements with the highest content in the alloy, r is the atomic radius of magnesium and other elements with the highest content in the alloy, Mg represents the content of magnesium, 1 represents the content of the most abundant element except magnesium, 2 represents the content of the second most abundant element except magnesium, and 3 represents the content of the third most abundant element except magnesium;

[0022] (9) the standard deviation of atomic radius:

[0023]

[0024] wherein w is the mass percentage of magnesium and other elements with the highest content in the alloy, r is the atomic radius of magnesium and other elements with the highest content in the alloy, is the average value of atomic radius;

[0025] (10) the average value of electronegativity:

[0026]

[0027] wherein w is the mass percentage of magnesium and other elements with the highest content in the alloy, χ is the electronegativity of magnesium and other elements with the highest content in the alloy;

[0028] (11) the standard deviation of electronegativity:

[0029]

[0030] wherein w is the mass percentage of magnesium and other elements with the highest content in the alloy, χ is the electronegativity of magnesium and other elements with the highest content in the alloy, is the average value of the electronegativity of the alloy.

[0031] is the average value of the valence electron number of the alloy.

[0032]

[0033] wherein w is the mass percentage of magnesium and other most content elements in the alloy, and e is the valence electron number of magnesium and other most content elements in the alloy.

[0034] is the average value of the mixing enthalpy of the alloy.

[0035]

[0036] wherein, is the binary mixing enthalpy of magnesium and other elements, and w is the mass percentage of magnesium and other most content elements in the alloy.

[0037] is the total content of the second phase.

[0038]

[0039] wherein A is the solubility limit of other elements in magnesium, S is the mass percentage of magnesium and other most content elements in the alloy,

[0040] is the average value of the twin boundary segregation energy of the alloy.

[0041]

[0042] wherein w is the mass percentage of magnesium and other most content elements in the alloy, and E is the twin boundary segregation energy of magnesium and other most content elements in the alloy.

[0043] is the average value of the diffusion coefficient of the alloy.

[0044]

[0045] wherein w is the mass percentage of magnesium and other most content elements in the alloy, and D is the diffusion coefficient of other most content elements in magnesium in the alloy.

[0046] Further, in the above method, the descriptors describing the extrusion process of the extruded magnesium alloy are selected, including:

[0047] solid solution temperature, solid solution time, extrusion temperature, extrusion ratio and extrusion speed.

[0048] Further, in the above method, the descriptors describing the mechanical properties of the extruded magnesium alloy are selected, including:

[0049] ultimate tensile strength, yield strength and elongation.

[0050] Furthermore, in the above method, the descriptors for the composition information of the extruded magnesium alloy and the descriptors for the extrusion process corresponding to the descriptors for predicting the mechanical properties of the extruded magnesium alloy are screened to obtain screened descriptors for the composition information and the extrusion process, including:

[0051] For ultimate tensile strength, the descriptors of the screened composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the atomic number of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number of the second most abundant element other than magnesium in the alloy, the mass percentage of the third most abundant element other than magnesium in the alloy, the atomic number of the third most abundant element other than magnesium in the alloy, the average atomic radius, the standard deviation of atomic radius, the average electronegativity, the average valence electron number, the total content of the second phase, and the average twin boundary segregation energy.

[0052] For ultimate tensile strength, the descriptors for the selected extrusion process include: solution temperature, solution time, extrusion ratio, and extrusion speed.

[0053] Furthermore, in the above method, the descriptors for the composition information of the extruded magnesium alloy and the descriptors for the extrusion process corresponding to the descriptors for predicting the mechanical properties of the extruded magnesium alloy are screened to obtain screened descriptors for the composition information and the extrusion process, including:

[0054] For yield strength, the descriptors of the screened composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the atomic number of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number of the second most abundant element other than magnesium in the alloy, the mass percentage of the third most abundant element other than magnesium in the alloy, the atomic number of the third most abundant element other than magnesium in the alloy, the average atomic radius, the standard deviation of atomic radius, the standard deviation of electronegativity, the enthalpy of mixing of the alloy, the total content of the second phase, and the average diffusion coefficient.

[0055] For yield strength, the descriptors for the selected extrusion process include: solution treatment time, extrusion temperature, and extrusion speed.

[0056] Furthermore, in the above method, the descriptors for the composition information of the extruded magnesium alloy and the descriptors for the extrusion process corresponding to the descriptors for predicting the mechanical properties of the extruded magnesium alloy are screened to obtain screened descriptors for the composition information and the extrusion process.

[0057] For elongation, the descriptors of the filtered composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number, average atomic radius, standard deviation of atomic radius, average electronegativity, average valence electron number, alloy mixing enthalpy, total content of the second phase, average twin boundary segregation energy, and average diffusion coefficient.

[0058] For elongation, the descriptors for the selected extrusion process include: solution temperature, solution time, extrusion temperature, extrusion ratio, and extrusion speed.

[0059] Furthermore, in the above method, the trained prediction model includes:

[0060] The first model used to predict ultimate tensile strength is obtained by adjusting the traditional Xgboost model. It is a medium-complexity regression model with a maximum depth of 7 layers, a pruning parameter of 1, a regularization parameter of 1, and 100 weak learners.

[0061] The second model used to predict yield strength is obtained by adjusting the traditional Xgboost model. It is a medium-complexity regression model with a maximum depth of 8 layers, a pruning parameter of 1, a regularization parameter of 1, and 100 weak learners.

[0062] The third model used to predict the elongation rate is an adjusted version of the traditional Xgboost model. It is a medium-complexity regression model with a maximum depth of 7 layers, a pruning parameter of 1, a regularization parameter of 1, and 150 weak learners.

[0063] Furthermore, in the above method, the values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the values ​​of the descriptors for the extrusion process are input into the trained prediction model to obtain the output values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted, including:

[0064] Based on the composition and extrusion process parameters of the magnesium alloy to be predicted, determine whether the magnesium alloy is extrudable.

[0065] If extrusion is possible, the values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the descriptor for the extrusion process are input into the trained prediction model to obtain the output values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted.

[0066] Furthermore, in the above method, based on the composition of the magnesium alloy to be predicted and the extrusion process parameters, determining whether the magnesium alloy to be predicted is extrudable includes:

[0067] Establish a dataset of solidus temperatures for magnesium alloys and a dataset of actual extrusion temperatures;

[0068] Based on the composition of the magnesium alloy to be predicted, the corresponding solidus temperature of the magnesium alloy is obtained from the magnesium alloy solidus temperature dataset.

[0069] Based on the extrusion process parameters of the magnesium alloy to be predicted, the corresponding actual extrusion temperature is obtained from the actual extrusion temperature dataset.

[0070] Determine whether the actual extrusion temperature is greater than the solidus temperature of the magnesium alloy. If so, it is determined to be extrudable; otherwise, it is determined to be non-extrudable.

[0071] Furthermore, in the above method, a dataset of solidus temperatures for magnesium alloys is established, including:

[0072] By combining the solidus temperatures of ten commonly used ternary magnesium alloys (Mg-Al-Ca, Mg-Al-Mn, Mg-Al-Sn, Mg-Al-Zn, Mg-Gd-Y, Mg-Nd-Y, Mg-Y-Al, Mg-Y-Zn, Mg-Ca-Zn, and Mg-Zn-Zr) under different compositions according to the alloy phase diagram, a dataset of magnesium alloy solidus temperatures was established.

[0073] Furthermore, in the above method, a dataset of actual extrusion temperatures is established, including:

[0074] Using the preprocessor of DEFORM simulation software, the process parameters for simulating the extrusion process of magnesium alloys are set, including: extrusion die movement speed, die temperature and magnesium alloy material temperature (extrusion temperature). The actual temperature during the magnesium alloy extrusion process is numerically simulated by finite element calculation. The highest temperature of the material during the forming process of magnesium alloy under different extrusion process parameters is statistically analyzed to establish an actual extrusion temperature dataset.

[0075] According to another aspect of the present invention, a computer-readable storage medium is also provided, on which computer-executable instructions are stored, wherein when executed by a processor, the computer-executable instructions cause the processor to perform the method described in any of the above embodiments.

[0076] According to another aspect of the present invention, a calculator device is also provided, comprising:

[0077] Processor; and

[0078] A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method described in any of the above embodiments.

[0079] Compared with existing technologies, this application sets descriptors to represent the "composition-process-property" of magnesium alloys before and after extrusion molding. It uses feature selection methods to filter descriptors that are beneficial to machine learning, eliminating useless or harmful descriptors, and ultimately training a machine learning model that can accurately predict the mechanical properties of extruded magnesium alloys. With the trained prediction model, the mechanical properties of any magnesium alloy after extrusion molding can be accurately predicted given the known composition and extrusion process. Attached Figure Description

[0080] Figure 1 This is a flowchart illustrating the difference in effects of using the "total content of the second phase" in the standard Xgboost model according to an embodiment of the present invention.

[0081] Figure 2a This is a flowchart of a feature selection process according to an embodiment of the present invention;

[0082] Figure 2b This is a diagram illustrating the effect of feature filtering according to an embodiment of the present invention;

[0083] Figure 3 This is a schematic diagram comparing model predictions with actual experimental data according to an embodiment of the present invention. Detailed Implementation

[0084] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0085] This invention provides a method for predicting the mechanical properties of extruded magnesium alloys, comprising:

[0086] Step S1: Select a descriptor that describes the composition information of the extruded magnesium alloy; select a descriptor that describes the extrusion process of the extruded magnesium alloy; select a descriptor that describes the mechanical properties of the extruded magnesium alloy.

[0087] Step S2: Based on the descriptors describing the composition information, extrusion process, and mechanical properties of the extruded magnesium alloy, an initial prediction model for the mechanical properties of the extruded magnesium alloy based on the descriptors describing the composition information and extrusion process of the extruded magnesium alloy is established using the standard Xgboost model.

[0088] Step S3: Based on the initial prediction model, and using the sequential backward selection method, the descriptors of the composition information of the extruded magnesium alloy and the descriptors of the extrusion process corresponding to the descriptors of the mechanical properties of the extruded magnesium alloy are screened to obtain the screened descriptors of the composition information and the descriptors of the extrusion process.

[0089] Specifically, such as Figure 2aAs shown, firstly, descriptors of the composition information and extrusion process of all extruded magnesium alloys are used as features for model learning. Then, a certain feature is iteratively deleted, and the remaining features are used for model learning. Poor features are deleted, and good features are retained. Finally, the set of retained features is the final feature selection result. The effect of feature selection is as follows: Figure 2b As shown.

[0090] Step S4: Based on the descriptors of the screened component information and the descriptors of the extrusion process and the data distribution corresponding to each descriptor, adjust the initial prediction model and use grid search to obtain the optimal hyperparameter combination to obtain the trained prediction model.

[0091] Here, "data distribution" refers to the magnitude of the value range of each descriptor, such as a component range of 0-100 (hundreds), a yield strength range of 0-1000 (thousands), and so on.

[0092] Step S5: Input the values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the values ​​of the descriptors for the extrusion process into the trained prediction model to obtain the values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted.

[0093] This application uses descriptors to represent the "composition-process-property" of magnesium alloys before and after extrusion molding. Feature selection methods are used to filter descriptors that are beneficial to machine learning, eliminating useless or harmful descriptors. Ultimately, a machine learning model capable of accurately predicting the mechanical properties of extruded magnesium alloys is trained. With the trained prediction model, the mechanical properties of any magnesium alloy after extrusion molding can be accurately predicted given the known composition and extrusion process.

[0094] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S1, selecting descriptors describing the composition information of the extruded magnesium alloy, includes: extracting 16 relevant descriptors for describing the composition information of the extruded magnesium alloy, specifically including:

[0095] (1) Mass percentage of magnesium in the alloy;

[0096] (2) The mass percentage of the most abundant element in the alloy, excluding magnesium;

[0097] (3) The atomic number of the most abundant element in the alloy besides magnesium;

[0098] (4) The mass percentage of the second most abundant element in the alloy, excluding magnesium;

[0099] (5) The atomic number of the second most abundant element in the alloy besides magnesium;

[0100] (6) The mass percentage of the third most abundant element in the alloy, excluding magnesium;

[0101] (7) The atomic number of the third most abundant element in the alloy besides magnesium;

[0102] (8) Average atomic radius:

[0103]

[0104] Where w is the mass percentage of magnesium and other most abundant elements in the alloy, r is the atomic radius of magnesium and other most abundant elements in the alloy; Mg represents the magnesium content; 1 represents the content of the most abundant element other than magnesium; 2 represents the content of the second most abundant element other than magnesium; 3 represents the content of the third most abundant element other than magnesium. (9) Standard deviation of atomic radius:

[0105]

[0106] Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and r is the atomic radius of magnesium and other most abundant elements in the alloy. This represents the average atomic radius.

[0107] (10) Average electronegativity:

[0108]

[0109] Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and χ is the electronegativity of magnesium and other most abundant elements in the alloy.

[0110] (11) Electronegativity standard deviation:

[0111]

[0112] Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and χ is the electronegativity of magnesium and other most abundant elements in the alloy. The average electronegativity is (10).

[0113] (12) Average number of valence electrons:

[0114]

[0115] Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and e is the number of valence electrons of magnesium and other most abundant elements in the alloy;

[0116] (13) Mixed enthalpy of alloy:

[0117]

[0118] in, ν is the binary enthalpy of mixing of magnesium with other elements, and w is the mass percentage of magnesium and other elements that are most abundant in the alloy.

[0119] (14) Total content of the second phase:

[0120]

[0121] Where A represents the solubility limit of other elements in magnesium, and S represents the mass percentage of magnesium and other most abundant elements in the alloy.

[0122] (15) Average twin boundary segregation energy:

[0123]

[0124] Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and E is the twin boundary segregation energy of magnesium and other most abundant elements in the alloy.

[0125] (16) Average diffusion coefficient

[0126]

[0127] Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and D is the diffusion coefficient of other most abundant elements in magnesium.

[0128] Here, as Figure 1 As shown, the addition of a descriptor for the total content of the second phase in this embodiment can give the same model better predictive ability.

[0129] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S1, selecting a descriptor describing the extrusion process of the extruded magnesium alloy, includes:

[0130] Five relevant descriptors were extracted to describe the extrusion process, specifically:

[0131] (1) Solution temperature (°C): In the extrusion process, the solution temperature refers to the temperature to which the metal material is heated before extrusion. This temperature is usually high enough to give the material sufficient plasticity so that it can pass smoothly through the extrusion die, while also ensuring that the material does not lose its structural stability during the extrusion process.

[0132] (2) Solution treatment time (h): Solution treatment time refers to the time that the metal material is heated in the furnace. The main purpose of this time is to ensure that the material is heated sufficiently so that it can reach the required plastic state during the extrusion process.

[0133] (3) Extrusion temperature (°C): Extrusion temperature refers to the temperature of the material during the extrusion process. This temperature usually needs to be controlled within a certain range to maintain the plasticity and stability of the material, and also affects the thermal effects that may be generated during the extrusion process;

[0134] (4) Extrusion Ratio (%): The extrusion ratio is the ratio of the diameter of the extrusion cylinder to the diameter of the extrusion die. This ratio determines the compression ratio of the material through the die during the extrusion process. Generally speaking, a larger extrusion ratio means that the material has a higher density and a finer grain structure after extrusion.

[0135] (5) Extrusion speed (mm / s): Extrusion speed refers to the speed at which the material passes through the extrusion die. This speed needs to be controlled within a certain range to ensure the stability of the extrusion process and the quality of the product. Too slow or too fast may lead to product quality problems.

[0136] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S1, selecting a descriptor describing the mechanical properties of the extruded magnesium alloy, includes:

[0137] Three relevant descriptors were extracted to describe the mechanical properties of the extruded magnesium alloy, specifically:

[0138] (1) Ultimate tensile strength (MPa): Tensile strength is the maximum load-bearing capacity of a magnesium alloy when subjected to tensile load. It reflects the ability of a magnesium alloy to resist tensile deformation.

[0139] (2) Yield strength (MPa): Yield strength is the minimum stress value of a magnesium alloy when subjected to shear or bending loads. When the stress of a magnesium alloy reaches its yield strength, the material will begin to undergo plastic deformation, that is, its shape and size will undergo irreversible changes.

[0140] (3) Elongation (%): Elongation is the percentage of plastic deformation that a magnesium alloy can undergo when subjected to stress. It reflects the plasticity of the magnesium alloy.

[0141] The above 24 descriptors can comprehensively represent the state of extruded magnesium alloys before and after forming.

[0142] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S3 involves filtering the descriptors of the composition information of the extruded magnesium alloy and the descriptors of the extrusion process corresponding to the descriptors for predicting the mechanical properties of the extruded magnesium alloy, to obtain the filtered descriptors of the composition information and the descriptors of the extrusion process, including:

[0143] For ultimate tensile strength, the descriptors of the screened composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the atomic number of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number of the second most abundant element other than magnesium in the alloy, the mass percentage of the third most abundant element other than magnesium in the alloy, the atomic number of the third most abundant element other than magnesium in the alloy, the average atomic radius, the standard deviation of atomic radius, the average electronegativity, the average valence electron number, the total content of the second phase, and the average twin boundary segregation energy.

[0144] For ultimate tensile strength, the descriptors for the selected extrusion process include: solution temperature, solution time, extrusion ratio, and extrusion speed.

[0145] Here, for ultimate tensile strength, the deleted features are the average diffusion coefficient, alloy mixing enthalpy, extrusion temperature, and electronegativity standard deviation.

[0146] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S3 involves filtering the descriptors for the composition information of the extruded magnesium alloy and the descriptors for the extrusion process corresponding to the descriptors for predicting the mechanical properties of the extruded magnesium alloy, to obtain the filtered descriptors for the composition information and the extrusion process, including:

[0147] For yield strength, the descriptors of the screened composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the atomic number of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number of the second most abundant element other than magnesium in the alloy, the mass percentage of the third most abundant element other than magnesium in the alloy, the atomic number of the third most abundant element other than magnesium in the alloy, the average atomic radius, the standard deviation of atomic radius, the standard deviation of electronegativity, the enthalpy of mixing of the alloy, the total content of the second phase, and the average diffusion coefficient.

[0148] For yield strength, the descriptors for the selected extrusion process include: solution treatment time, extrusion temperature, and extrusion speed.

[0149] Here, for yield strength, the deleted features are: solution temperature, average twin boundary segregation energy, extrusion ratio, average valence electron number, and average electronegativity.

[0150] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys according to the present invention, step S3 involves filtering the descriptors for the composition information of the extruded magnesium alloy and the descriptors for the extrusion process corresponding to the descriptors for predicting the mechanical properties of the extruded magnesium alloy, to obtain the filtered descriptors for the composition information and the extrusion process.

[0151] For elongation, the descriptors of the filtered composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number, average atomic radius, standard deviation of atomic radius, average electronegativity, average valence electron number, alloy mixing enthalpy, total content of the second phase, average twin boundary segregation energy, and average diffusion coefficient.

[0152] Specifically, Figure 1 In this study, the predictive ability of the model was improved after introducing "second phase content," i.e., domain knowledge; and the predictive ability of the model was further improved after feature selection.

[0153] For elongation, the descriptors for the selected extrusion process include: solution temperature, solution time, extrusion temperature, extrusion ratio, and extrusion speed.

[0154] Here, for elongation, the deleted features are: the atomic number of the most abundant element in the alloy other than magnesium, the standard deviation of electronegativity, the mass percentage of the third most abundant element in the alloy other than magnesium, and the atomic number of the second most abundant element in the alloy other than magnesium.

[0155] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S4, obtaining the trained prediction model includes:

[0156] The first model used to predict ultimate tensile strength is obtained by adjusting the traditional Xgboost model. It is a regression model of medium complexity with a maximum depth of 7 layers, a pruning parameter of 1, a regularization parameter of 1, and 100 weak learners (trees).

[0157] The second model used to predict yield strength is an adjusted version of the traditional Xgboost model. It is a moderately complex regression model with a maximum depth of 8 layers, a pruning parameter of 1, a regularization parameter of 1, and 100 weak learners (trees).

[0158] The third model used to predict the elongation rate is an adjusted version of the traditional Xgboost model. It is a medium-complexity regression model with a maximum depth of 7 layers, a pruning parameter of 1, a regularization parameter of 1, and 150 weak learners (trees).

[0159] Specifically, the prediction results of the prediction model trained in this invention are compared with the actual experimental data, such as... Figure 3 As shown,

[0160] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys according to the present invention, step S5 involves inputting the values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the values ​​of the descriptors for the extrusion process into the trained prediction model to obtain the output values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted, including:

[0161] Step S51: Based on the composition of the magnesium alloy to be predicted and the extrusion process parameters, determine whether the magnesium alloy to be predicted is extrudable.

[0162] Step S52: If extrusion is possible, input the values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the descriptor for the extrusion process into the trained prediction model to obtain the output values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted.

[0163] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S51, determining whether the magnesium alloy to be predicted is extrudable based on the composition of the magnesium alloy to be predicted and the extrusion process parameters, includes:

[0164] Step S511: Establish a dataset of solidus temperature of magnesium alloy and a dataset of actual extrusion temperature;

[0165] Step S512: Based on the composition of the magnesium alloy to be predicted, obtain the corresponding magnesium alloy solidus temperature from the magnesium alloy solidus temperature dataset.

[0166] Step S513: Based on the extrusion process parameters of the magnesium alloy to be predicted, obtain the corresponding actual extrusion temperature from the actual extrusion temperature dataset.

[0167] Step S514: Determine whether the actual extrusion temperature is greater than the solidus temperature of the magnesium alloy. If yes, it is determined that it can be extruded; otherwise, it is determined that it cannot be extruded.

[0168] Specifically, in actual production, magnesium alloys often become unusable due to poor surface quality after extrusion. This phenomenon occurs because the actual temperature of the material during hot extrusion exceeds the solidus temperature of the alloy, resulting in the appearance of a liquid phase. This embodiment establishes separate datasets for magnesium alloy solidus temperatures and actual extrusion temperatures, allowing for comparison of whether the actual extrusion temperature exceeds the solidus temperature of the magnesium alloy, thereby enabling the determination of the extrudability of the magnesium alloy.

[0169] The prediction of extrudability mainly relies on the construction of two datasets. The first is to establish a thermodynamic database (magnesium alloy solidus temperature dataset) using magnesium alloy data to predict the solidus temperature of magnesium alloys with specified compositions. The second is to construct a dataset of actual extrusion temperatures of magnesium alloys under different extrusion process parameters through finite element simulation.

[0170] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys of the present invention, step S511, establishing a solidus temperature dataset for magnesium alloys, includes:

[0171] By combining the solidus temperatures of ten commonly used ternary magnesium alloys (Mg-Al-Ca, Mg-Al-Mn, Mg-Al-Sn, Mg-Al-Zn, Mg-Gd-Y, Mg-Nd-Y, Mg-Y-Al, Mg-Y-Zn, Mg-Ca-Zn, and Mg-Zn-Zr) under different compositions according to the alloy phase diagram, a magnesium alloy solidus temperature dataset (thermodynamic dataset) was established.

[0172] In one embodiment of the method for predicting the mechanical properties of extruded magnesium alloys according to the present invention, step S511, establishing an actual extrusion temperature dataset, includes:

[0173] Using the preprocessor of DEFORM simulation software, the process parameters for simulating the extrusion process of magnesium alloys are set, including: extrusion die movement speed, die temperature and magnesium alloy material temperature (extrusion temperature of the extrusion process descriptor). The actual temperature during the magnesium alloy extrusion process is numerically simulated by finite element calculation. The highest temperature of the material during the forming process of magnesium alloy under different extrusion process parameters is statistically analyzed to establish an actual extrusion temperature dataset.

[0174] Finally, by constructing the two datasets above, the solidus temperature and actual extrusion temperature of a magnesium alloy with a specific composition under certain extrusion parameters can be obtained. If the solidus temperature is lower than the actual extrusion temperature, the alloy is determined to have good extrudability under the current process. This allows for the next step of predicting extrusion performance.

[0175] According to another aspect of the present invention, a computer-readable storage medium is also provided, on which computer-executable instructions are stored, wherein when executed by a processor, the computer-executable instructions cause the processor to perform the method described in any of the above embodiments.

[0176] According to another aspect of the present invention, a calculator device is also provided, comprising:

[0177] Processor; and

[0178] A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method described in any of the above embodiments.

[0179] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0180] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0181] Obviously, those skilled in the art can make various modifications and variations to the invention without departing from the spirit and scope of the invention. Therefore, if these modifications and variations fall within the scope of the claims of the invention and their equivalents, the invention is also intended to include these modifications and variations.

Claims

1. A method for predicting the mechanical properties of extruded magnesium alloys, characterized in that, include: Select a descriptor that describes the composition information of the extruded magnesium alloy; Select a descriptor that describes the extrusion process of the extruded magnesium alloy; Select a descriptor that describes the mechanical properties of the extruded magnesium alloy; Based on descriptors describing the composition information, extrusion process, and mechanical properties of extruded magnesium alloys, an initial prediction model for the mechanical properties of extruded magnesium alloys based on the descriptors describing the composition information and extrusion process of extruded magnesium alloys is established using the standard Xgboost model. Based on the initial prediction model, and using the sequential backward selection method, the descriptors of the composition information of the extruded magnesium alloy and the descriptors of the extrusion process corresponding to the descriptors of the mechanical properties of the extruded magnesium alloy are screened to obtain the screened descriptors of the composition information and the descriptors of the extrusion process. Based on the descriptors of the screened component information and the descriptors of the extrusion process, as well as the data distribution corresponding to each descriptor, the initial prediction model is adjusted, and grid search is used to obtain the optimal combination of hyperparameters to obtain a trained prediction model. The values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the values ​​of the descriptors for the extrusion process are input into the trained prediction model to obtain the values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted.

2. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 1, characterized in that, Select descriptors that describe the compositional information of extruded magnesium alloys, including: (1) Mass percentage of magnesium in the alloy; (2) The mass percentage of the most abundant element in the alloy, excluding magnesium; (3) The atomic number of the most abundant element in the alloy besides magnesium; (4) The mass percentage of the second most abundant element in the alloy, excluding magnesium; (5) The atomic number of the second most abundant element in the alloy besides magnesium; (6) The mass percentage of the third most abundant element in the alloy, excluding magnesium; (7) The atomic number of the third most abundant element in the alloy besides magnesium; (8) Average atomic radius: Where w is the mass percentage of magnesium and other most abundant elements in the alloy, r is the atomic radius of magnesium and other most abundant elements in the alloy; Mg represents the magnesium content; 1 represents the content of the most abundant element other than magnesium; 2 represents the content of the second most abundant element other than magnesium; 3 represents the content of the third most abundant element other than magnesium. (9) Standard deviation of atomic radius: Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and r is the atomic radius of magnesium and other most abundant elements in the alloy. This represents the average atomic radius. (10) Average electronegativity: Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and χ is the electronegativity of magnesium and other most abundant elements in the alloy. (11) Electronegativity standard deviation: Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and χ is the electronegativity of magnesium and other most abundant elements in the alloy. The average electronegativity is (10). (12) Average number of valence electrons: Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and e is the number of valence electrons of magnesium and other most abundant elements in the alloy; (13) Mixed enthalpy of alloy: in, ν is the binary enthalpy of mixing of magnesium with other elements, and w is the mass percentage of magnesium and other elements that are most abundant in the alloy. (14) Total content of the second phase: Where A represents the solubility limit of other elements in magnesium, and S represents the mass percentage of magnesium and other most abundant elements in the alloy. (15) Average twin boundary segregation energy: Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and E is the twin boundary segregation energy of magnesium and other most abundant elements in the alloy. (16) Average diffusion coefficient Where w is the mass percentage of magnesium and other most abundant elements in the alloy, and D is the diffusion coefficient of other most abundant elements in magnesium.

3. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 2, characterized in that, Take the descriptor describing the extrusion process of the extruded magnesium alloy, including: Solution temperature, solution time, extrusion temperature, extrusion ratio, and extrusion speed.

4. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 3, characterized in that, Select descriptors to describe the mechanical properties of extruded magnesium alloys, including: Ultimate tensile strength, yield strength, and elongation.

5. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 4, characterized in that, The descriptors for the composition information of extruded magnesium alloys and the descriptors for the extrusion process corresponding to the descriptors predicting the mechanical properties of extruded magnesium alloys are filtered to obtain the filtered descriptors for the composition information and the extrusion process, including: For ultimate tensile strength, the descriptors of the screened composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the atomic number of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number of the second most abundant element other than magnesium in the alloy, the mass percentage of the third most abundant element other than magnesium in the alloy, the atomic number of the third most abundant element other than magnesium in the alloy, the average atomic radius, the standard deviation of atomic radius, the average electronegativity, the average valence electron number, the total content of the second phase, and the average twin boundary segregation energy. For ultimate tensile strength, the descriptors for the selected extrusion process include: solution temperature, solution time, extrusion ratio, and extrusion speed.

6. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 4, characterized in that, The descriptors for the composition information of extruded magnesium alloys and the descriptors for the extrusion process corresponding to the descriptors predicting the mechanical properties of extruded magnesium alloys are filtered to obtain the filtered descriptors for the composition information and the extrusion process, including: For yield strength, the descriptors of the screened composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the atomic number of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number of the second most abundant element other than magnesium in the alloy, the mass percentage of the third most abundant element other than magnesium in the alloy, the atomic number of the third most abundant element other than magnesium in the alloy, the average atomic radius, the standard deviation of atomic radius, the standard deviation of electronegativity, the enthalpy of mixing of the alloy, the total content of the second phase, and the average diffusion coefficient. For yield strength, the descriptors for the selected extrusion process include: solution treatment time, extrusion temperature, and extrusion speed.

7. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 4, characterized in that, The descriptors for predicting the mechanical properties of extruded magnesium alloys, along with the descriptors for the extrusion process, are filtered to obtain the filtered descriptors for the composition information and the extrusion process. For elongation, the descriptors of the filtered composition information include: the mass percentage of magnesium in the alloy, the mass percentage of the most abundant element other than magnesium in the alloy, the mass percentage of the second most abundant element other than magnesium in the alloy, the atomic number, average atomic radius, standard deviation of atomic radius, average electronegativity, average valence electron number, alloy mixing enthalpy, total content of the second phase, average twin boundary segregation energy, and average diffusion coefficient. For elongation, the descriptors for the selected extrusion process include: solution temperature, solution time, extrusion temperature, extrusion ratio, and extrusion speed.

8. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 4, characterized in that, The trained prediction models include: The first model used to predict ultimate tensile strength is obtained by adjusting the traditional Xgboost model. It is a medium-complexity regression model with a maximum depth of 7 layers, a pruning parameter of 1, a regularization parameter of 1, and 100 weak learners. The second model used to predict yield strength is obtained by adjusting the traditional Xgboost model. It is a medium-complexity regression model with a maximum depth of 8 layers, a pruning parameter of 1, a regularization parameter of 1, and 100 weak learners. The third model used to predict the elongation rate is an adjusted version of the traditional Xgboost model. It is a medium-complexity regression model with a maximum depth of 7 layers, a pruning parameter of 1, a regularization parameter of 1, and 150 weak learners.

9. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 1, characterized in that, The values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the values ​​of the descriptors for the extrusion process are input into the trained prediction model to obtain the output values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted, including: Based on the composition and extrusion process parameters of the magnesium alloy to be predicted, determine whether the magnesium alloy is extrudable. If extrusion is possible, the values ​​of the descriptors for the composition information of the magnesium alloy to be predicted and the descriptor for the extrusion process are input into the trained prediction model to obtain the output values ​​of the descriptors for the mechanical properties of the magnesium alloy to be predicted.

10. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 9, characterized in that, Based on the composition and extrusion process parameters of the magnesium alloy to be predicted, determine whether the magnesium alloy is extrudable, including: Establish a dataset of solidus temperatures for magnesium alloys and a dataset of actual extrusion temperatures; Based on the composition of the magnesium alloy to be predicted, the corresponding solidus temperature of the magnesium alloy is obtained from the magnesium alloy solidus temperature dataset. Based on the extrusion process parameters of the magnesium alloy to be predicted, the corresponding actual extrusion temperature is obtained from the actual extrusion temperature dataset. Determine whether the actual extrusion temperature is greater than the solidus temperature of the magnesium alloy. If so, it is determined to be extrudable; otherwise, it is determined to be non-extrudable.

11. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 10, characterized in that, Establish a dataset of solidus temperatures for magnesium alloys, including: By combining the solidus temperatures of ten commonly used ternary magnesium alloys (Mg-Al-Ca, Mg-Al-Mn, Mg-Al-Sn, Mg-Al-Zn, Mg-Gd-Y, Mg-Nd-Y, Mg-Y-Al, Mg-Y-Zn, Mg-Ca-Zn, and Mg-Zn-Zr) under different compositions according to the alloy phase diagram, a dataset of magnesium alloy solidus temperatures was established.

12. The method for predicting the mechanical properties of extruded magnesium alloys as described in claim 10, characterized in that, Establish a dataset of actual extrusion temperatures, including: Using the preprocessor of DEFORM simulation software, the process parameters for simulating the extrusion process of magnesium alloys were set, including: extrusion die movement speed, die temperature and magnesium alloy material temperature. The temperature of the actual magnesium alloy extrusion process was numerically simulated by finite element calculation. The highest temperature of the material during the forming process of magnesium alloy under different extrusion process parameters was statistically analyzed to establish an actual extrusion temperature dataset.

13. A computer-readable storage medium having stored thereon computer-executable instructions, wherein, When the computer-executable instructions are executed by a processor, the processor causes the processor to perform the method as described in any one of claims 1 to 12.

14. A calculator device, wherein, include: processor; as well as A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method as described in any one of claims 1 to 12.

Images