TRIP titanium alloy component analysis method based on molybdenum equivalent and iron equivalent
By combining the molybdenum equivalent and iron equivalent diagrams with the Stacking model and SHAP analysis, the problem of low prediction accuracy in the composition design of TRIP titanium alloys was solved, efficient and intuitive alloy composition analysis was achieved, and the research and development of new TRIP titanium alloys was promoted.
Patent Information
- Application Number
- CN202510908272.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-02
- Publication Date
- 2025-10-03
AI Technical Summary
Existing TRIP titanium alloy composition design methods have low accuracy in predicting plastic deformation mechanisms, especially for alloy systems with complex elemental composition, which limits the development speed of new TRIP titanium alloys. In addition, machine learning methods have high barriers to entry and are not intuitive and convenient enough in application.
Using molybdenum equivalent and iron equivalent as the main parameters, a new data set was generated through the Stacking model and the SHAP analysis method was used to draw a molybdenum equivalent-iron equivalent diagram to predict the probability of the TRIP effect, thereby simplifying the alloy composition design process.
It improves the prediction accuracy and convenience of TRIP titanium alloy composition design, realizes fast and accurate alloy composition analysis, and lowers the application threshold of machine learning methods.
Smart Images
Figure CN120748532A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of alloy design, and in particular relates to a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent. Background Art
[0002] TRIP titanium alloys are considered an ideal alternative to traditional metastable β titanium alloys because they can achieve a good balance of strength and toughness by adjusting the plastic deformation mechanism. This type of alloy can better meet the increasingly high requirements for lightweight and mechanical properties of structural materials due to the rapid development of aerospace vehicles. Due to its simplicity and directness, the TRIP titanium alloy composition design method has become the most widely used. This method has been verified on alloys such as Ti-12Mo, Ti-1023, and Ti-10Mo-1Fe. However, The accuracy of the graph in predicting plastic deformation mechanisms is relatively low, especially for alloy systems with relatively complex elemental compositions. This problem has greatly limited the development speed of new TRIP titanium alloys.
[0003] With the development of artificial intelligence, machine learning methods have been tried and applied to alloy design, achieving promising results. Although machine learning methods have demonstrated their potential in alloy design, they rely heavily on accurate datasets and highly accurate machine learning models. The establishment of datasets and the development of machine learning models raise the bar for applying machine learning methods, and their application in TRIP titanium alloy design is also not intuitive and convenient. These two issues limit the widespread application of machine learning methods in TRIP titanium alloy design. Summary of the Invention
[0004] In view of the shortcomings of the existing technology, this application proposes a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent.
[0005] In the first aspect, the present application proposes a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent, comprising:
[0006] Obtain the values of molybdenum equivalent and iron equivalent in titanium alloy;
[0007] In a molybdenum equivalent-iron equivalent diagram, the probability of the occurrence of the TRIP effect is obtained according to the molybdenum equivalent value and the iron equivalent value in the titanium alloy. The molybdenum equivalent-iron equivalent diagram is drawn by: establishing a coordinate system with the molybdenum equivalent and iron equivalent in the new data set as the horizontal and vertical coordinates, and using the probability of the occurrence of the TRIP effect as the value of the coordinate point in the coordinate system. The new data set is generated using the Stacking model.
[0008] When the probability of a TRIP effect occurring is greater than a preset threshold, it is considered that a TRIP effect has occurred.
[0009] The new dataset is generated using the Stacking model and includes:
[0010] Establishing a metastable β titanium alloy original data set, dividing the metastable β titanium alloy original data set into a training set and a test set, wherein the metastable β titanium alloy original data set includes: multiple factors affecting the TRIP effect;
[0011] Using multiple basic machine learning models, training them on the training set and testing them on the test set, to obtain machine learning models of multiple factors affecting the TRIP effect;
[0012] Integrate machine learning models of various factors affecting the TRIP effect to establish a Stacking model;
[0013] According to the prediction results of the test set obtained in the stacking model, the most important influencing factors of the TRIP effect are screened. The most important influencing factors of the TRIP effect are molybdenum equivalent and iron equivalent;
[0014] Based on molybdenum equivalent and iron equivalent, a new data set was generated using the Stacking model.
[0015] The probability of the TRIP effect occurring is obtained by:
[0016] Input the new data set into the Stacking model to obtain new prediction results;
[0017] The SHAP analysis method is used to analyze the new prediction results and obtain the probability of the TRIP effect.
[0018] The metastable β titanium alloy original data set includes: alloy elements, solution treatment temperature, solution treatment time, value, The value, molybdenum equivalent and iron equivalent are used as the input of the metastable β titanium alloy original data set, and the input of the metastable β titanium alloy original data set is the influencing factors of various TRIP effects; whether the TRIP effect occurs is used as the output of the metastable β titanium alloy original data set.
[0019] A variety of basic machine learning models were used, trained on the training set and tested on the test set, to obtain machine learning models of various factors affecting the TRIP effect, including:
[0020] The Random Forest model, the XGBoost model, the LightGBM model, and the ExtraRandomized Tree model are respectively trained on the training set to obtain machine learning models of various factors affecting the TRIP effect;
[0021] The machine learning models based on the influencing factors of the various TRIP effects are tested on the test set to obtain test results of the various machine learning models;
[0022] According to the test results of various machine learning models, the area under the ROC curve was used to evaluate the classification prediction accuracy of the test results of various machine learning models, so as to perform hyperparameter tuning and obtain optimized machine learning models of various factors affecting the TRIP effect.
[0023] The prediction results obtained from the stacking model based on the test set are used to screen the most important factors affecting the TRIP effect, including:
[0024] Input the test set into the Stacking model to obtain the prediction results;
[0025] The prediction results were analyzed using the SHAP analysis method to obtain the SHAP analysis value;
[0026] The SHAP analysis values were sorted from high to low, and the two factors with the highest analysis values were taken as the most important influencing factors of the TRIP effect.
[0027] When the probability of a TRIP effect occurring is greater than a preset threshold, a TRIP effect is considered to have occurred, including: an area with a probability of a TRIP effect occurring greater than 0.85 is considered to be extremely likely to occur TRIP; an area with a probability of a TRIP effect occurring within the range of 0.75 to 0.85 is considered to be prone to TRIP; an area with a probability of a TRIP effect occurring within the range of 0.5 to 0.75 is considered to have the possibility of TRIP occurring; an area with a probability of a TRIP effect occurring within the range of 0.35 to 0.5 is considered to be unlikely to occur TRIP; and an area with a probability of a TRIP effect occurring less than 0.35 is considered to be unlikely to occur TRIP.
[0028] Beneficial effects:
[0029] This application proposes a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent, which can achieve rapid and accurate design of TRIP titanium alloy. The molybdenum equivalent-iron equivalent plot, based on machine learning, identifies two parameters that have the greatest impact on the TRIP effect: molybdenum equivalent and iron equivalent, improving prediction accuracy. The machine learning approach also expanded the TRIP titanium alloy dataset, further enhancing the new method's predictive capabilities for TRIP titanium alloys. Furthermore, compared to machine learning methods, the molybdenum equivalent-iron equivalent plot can be separated from datasets and machine learning models, offering a more intuitive and simplified approach, making it more readily applicable to TRIP titanium alloy composition design. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 Flowchart of a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent according to an embodiment of the present application;
[0031] Figure 2 Schematic diagram of molybdenum equivalent-iron equivalent diagram of the embodiment of the present application;
[0032] Figure 3 BC image of the Ti-Mo-V-Cr-Al-Nb alloy prepared in Example 1 of the present application after solution treatment;
[0033] Figure 4 Transmission electron microscope image of the deformation zone of the Ti-Mo-V-Cr-Al-Nb alloy prepared in Example 1 of the present application. DETAILED DESCRIPTION
[0034] The specific implementation of the present application is further described in detail below with reference to the accompanying drawings and examples.
[0035] based on The accuracy of the graph in predicting plastic deformation mechanisms is relatively low, especially for alloy systems with relatively complex elemental compositions. This problem has greatly limited the development speed of new TRIP titanium alloys.
[0036] This application proposes a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent. The method calculates the alloy [Mo] eq (Molybdenum equivalent) and [Fe] eq (iron equivalent), and then in [Mo] eq -[Fe] eq Quickly locate the corresponding alloy in the figure to determine whether the alloy has TRIP effect. [Mo] eq -[Fe] eq The diagram not only meets the high-precision requirements of TRIP titanium alloy design, but also has the advantages of being intuitive and simple.
[0037] Example 1:
[0038] This embodiment proposes a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent. Figure 1 Shown, including:
[0039] Step S1: obtaining the molybdenum equivalent value and the iron equivalent value in the titanium alloy;
[0040] Step S2: In a molybdenum equivalent-iron equivalent diagram, the probability of the occurrence of the TRIP effect is obtained according to the molybdenum equivalent value and the iron equivalent value in the titanium alloy. The molybdenum equivalent-iron equivalent diagram is drawn by establishing a coordinate system with the molybdenum equivalent and iron equivalent in the new data set as the horizontal and vertical coordinates, and using the probability of the occurrence of the TRIP effect as the value of the coordinate point in the coordinate system. The new data set is generated using the stacking model.
[0041] Step S3: When the probability of the TRIP effect occurring is greater than a preset threshold, it is considered that the TRIP effect has occurred.
[0042] In this embodiment, the traditional Instead of using a graph, two parameters that have a greater impact on the TRIP effect, molybdenum equivalent and iron equivalent, are used to draw a molybdenum equivalent-iron equivalent graph. The probability of the TRIP effect occurring is obtained from the molybdenum equivalent-iron equivalent graph. If the probability of the TRIP effect occurring is greater than a preset threshold, for example, greater than 0.75, it is considered that the TRIP effect has occurred. The method of this embodiment improves the accuracy of predicting the TRIP effect.
[0043] The new dataset is generated using the Stacking Integrated Model (Stacking Integrated Model), including:
[0044] Step S101: establishing a metastable β titanium alloy original data set, dividing the metastable β titanium alloy original data set into a training set and a test set, wherein the metastable β titanium alloy original data set includes: multiple factors affecting the TRIP effect;
[0045] The metastable β titanium alloy original data set includes: alloy elements, solution treatment temperature, solution treatment time, value (the number of d electron binding, reflecting the d orbital electron binding strength), The input of the metastable β titanium alloy original data set is the influencing factors of various TRIP effects. The output of the metastable β titanium alloy original data set is whether the TRIP (TRansformation Induced Plasticity) effect occurs.
[0046] In this example, relevant literature was searched to establish a metastable β titanium alloy original data set, including 216 sets of data. All data were collected from relevant literature on metastable β titanium alloys. The input includes: alloying elements (titanium Ti, aluminum Al, molybdenum Mo, vanadium V, chromium Cr, iron Fe, zirconium Zr, tin Sn, niobium Nb), solution treatment temperature, solution treatment time, value, Value, [Mo] eq Value and [Fe] eq The output is whether there is a TRIP effect; the original dataset is randomly divided into a training set and a test set in a ratio of 7:3.
[0047] Step S102: Using multiple basic machine learning models, training on the training set and testing on the test set, to obtain machine learning models of multiple factors affecting the TRIP effect, including:
[0048] The Random Forest model, the XGBoost model, the LightGBM model, and the Extra Randomized Tree model are used to perform training on the training set to obtain machine learning models of various factors affecting the TRIP effect.
[0049] The machine learning models based on the influencing factors of the various TRIP effects are tested on the test set to obtain test results of the various machine learning models;
[0050] Based on the test results of various machine learning models, the area under the ROC curve (AUC, Area Under Curve, the area under the ROC curve and the coordinate axis) was used to evaluate the classification prediction accuracy of the test results of various machine learning models, so as to perform hyperparameter tuning and obtain optimized machine learning models of various factors affecting the TRIP effect.
[0051] In this example, long-term research has found that each model is sensitive to certain factors influencing the TRIP effect but not others. Using only one model to process data can lead to a tendency to accurately predict only certain factors. Therefore, this example employs multiple basic machine learning models to avoid this tendency. All basic models are trained separately on the training set; the area under the ROC curve (AUC) is used to evaluate the model's classification prediction accuracy for hyperparameter tuning.
[0052] Step S103: Integrate the machine learning models of various factors affecting the TRIP effect to establish a stacking model;
[0053] In this embodiment, a stacking algorithm is used to integrate machine learning models of various factors affecting the TRIP effect to obtain a stacking model, and the accuracy of the model is verified on a test set.
[0054] Step S104: Screening the most important factors affecting the TRIP effect based on the prediction results obtained from the test set in the Stacking model. The most important factors affecting the TRIP effect are molybdenum equivalent and iron equivalent.
[0055] In this example, SHAP analysis (SHapley Additive exPlanations, an interpretable method based on game theory) was used to analyze the most important factors affecting the TRIP effect in the Stacking model. The results showed that the most important factors affecting the TRIP effect are [Mo] eq and [Fe] eq ;
[0056] Specifically, the most important factors affecting the TRIP effect are screened based on the prediction results obtained from the test set in the Stacking model, including:
[0057] Step S104.1: Input the test set into the Stacking model to obtain the prediction result;
[0058] Step S104.2: Analyze the prediction results using the SHAP analysis method to obtain a SHAP analysis value;
[0059] Step S104.3: Sort the SHAP analysis values from high to low, and take the two factors with the highest analysis values as the most important influencing factors of the TRIP effect.
[0060] Step S105: Generate a new data set using the Stacking model.
[0061] The probability of the TRIP effect occurring is obtained by:
[0062] Step S106: Input the new data set into the Stacking model to obtain a new prediction result;
[0063] Step S107: Analyze the new prediction result using the SHAP analysis method to obtain the probability of the TRIP effect occurring.
[0064] In this embodiment, the new data set contains 40,000 sets of data. The generated data are all within the range of the original data set. The Stacking model is used to predict the new prediction results, and the probability of the TRIP effect occurring, P(TRIP), is obtained based on the SHAP analysis.
[0065] Step S108: A coordinate system is established with the molybdenum equivalent and iron equivalent in the new data set as the horizontal and vertical coordinates, and a molybdenum equivalent-iron equivalent graph is drawn with the probability of the TRIP effect occurring as the value of the coordinate point in the coordinate system, as shown in FIG. Figure 2 As shown;
[0066] In this embodiment, the molybdenum equivalent-iron equivalent diagram is drawn, as shown in FIG. Figure 2 As shown, the probability P(TRIP) of the TRIP effect increasing from the dark area to the light area, the white area has the highest probability of the TRIP effect, that is, greater than 0.85, and the darkest gray area has the lowest probability of the TRIP effect, that is, below 0.35.
[0067] Based on the obtained molybdenum equivalent and iron equivalent values in the titanium alloy, the probability of the TRIP effect occurring is obtained from the plotted molybdenum equivalent-iron equivalent graph. When the probability of the TRIP effect occurring is greater than a preset threshold, the TRIP effect is considered to have occurred. The following are the regions where the probability of the TRIP effect occurring is greater than 0.85 and is considered to be highly prone to TRIP; regions where the probability of the TRIP effect occurring is between 0.75 and 0.85 and is considered to be prone to TRIP; regions where the probability of the TRIP effect occurring is between 0.5 and 0.75 and is considered to have the possibility of TRIP occurring; regions where the probability of the TRIP effect occurring is between 0.35 and 0.5 and is considered to be unlikely to occur; and regions where the probability of the TRIP effect is less than 0.35 and is considered unlikely to occur. When designing the TRIP titanium alloy composition, the ideal effect should be achieved by selecting regions where the probability of the TRIP effect is above 0.75.
[0068] In order to prove that based on [Mo eq -[Fe] eq To illustrate the effectiveness of predicting the TRIP effect, three examples are listed below:
[0069] Example 1:
[0070] The prediction of the TRIP effect in titanium alloys using a molybdenum equivalent-iron equivalent diagram includes the following steps:
[0071] (1)According to [Mo] eq -[Fe] eqFigure 1 shows the alloy composition design within the range of P(TRIP) greater than 0.75. The nominal composition is Ti-5Mo-5V-3Cr-3Al-2Nb (wt.%), where [Mo] eq is 10.71, [Fe] eq is 2.87;
[0072] (2) Pure metals (Ti, Cr) and intermediate alloys (Mo-Al, V-Al, Nb-Al) are melted three times in a suspension furnace to obtain alloy ingots with uniform composition, and then forged in the β phase region;
[0073] (3) subjecting the forged alloy plate to a solution treatment at 860°C / 1h;
[0074] (4) Microstructure observation and room temperature tensile test of the alloy after solution treatment;
[0075] (5) Observe the deformation zone of the alloy after tensile fracture.
[0076] like Figure 3 、 Figure 4 As shown in Figure 2, the Ti-5Mo-5V-3Cr-3Al-2Nb alloy in this example exhibits a single β phase with a grain size of 163 μm after solution treatment at 860°C for 1 h. α″ martensite is found in the alloy after deformation, demonstrating the TRIP phenomenon. [Mo] eq -[Fe] eq The figure correctly predicts the occurrence of TRIP effect in this alloy.
[0077] Example 2:
[0078] The prediction of the TRIP effect in titanium alloys using a molybdenum equivalent-iron equivalent diagram includes the following steps:
[0079] (1)According to [Mo] eq -[Fe] eq Figure 1 shows the alloy composition design within the range of P(TRIP) greater than 0.75. The nominal composition is Ti-5Mo-5V-3Cr-3Al-2Nb (wt.%), where [Mo] eq is 10.71, [Fe] eq is 2.87;
[0080] (2) Pure metals (Ti, Cr) and intermediate alloys (Mo-Al, V-Al, Nb-Al) are melted three times in a suspension furnace to obtain alloy ingots with uniform composition, and then forged in the β phase region;
[0081] (3) subjecting the forged alloy plate to a solution treatment at 800°C / 1h;
[0082] (4) Microstructure observation and room temperature tensile test of the alloy after solution treatment;
[0083] (5) Observe the deformation zone of the alloy after tensile fracture.
[0084] The Ti-5Mo-5V-3Cr-3Al-2Nb alloy in this example exhibits a dual-phase structure with both β and α phases after solution treatment at 800°C for 1 hour, with a grain size of 137 μm. α″ martensite was found in the alloy after deformation, demonstrating the TRIP phenomenon. [Mo] eq -[Fe] eq The figure correctly predicts the occurrence of TRIP effect in this alloy.
[0085] Example 3:
[0086] The prediction of the TRIP effect in titanium alloys using a molybdenum equivalent-iron equivalent diagram includes the following steps:
[0087] (1)According to [Mo] eq -[Fe] eq Figure 1 shows the alloy composition design within the range of P(TRIP) greater than 0.75. The nominal composition is Ti-4Mo-8V-3Cr-4Al-2Zr (wt.%), where [Mo] eq is 10.16, [Fe] eq is 2.86;
[0088] (2) Pure metals (Ti, Cr, Zr) and intermediate alloys (Mo-Al, V-Al) are melted three times in a suspension furnace to obtain alloy ingots with uniform composition, and then forged in the β phase region;
[0089] (3) subjecting the forged alloy plate to a solution treatment at 860°C / 1h;
[0090] (4) Microstructure observation and room temperature tensile test of the alloy after solution treatment;
[0091] (5) Observe the deformation zone of the alloy after tensile fracture.
[0092] The Ti-4Mo-8V-3Cr-4Al-2Zr alloy in this example exhibited a single β phase with a grain size of 142 μm after solution treatment at 860°C for 1 h. After deformation, α″ martensite was found in the alloy, demonstrating TRIP. [Mo] eq -[Fe] eq The figure correctly predicts the occurrence of TRIP effect in this alloy.
[0093] Comparative Example 1:
[0094] The composition and mass percentage of Ti-Mo-V-Cr-Al-Nb alloy, smelting, forging and solution treatment are the same as those in Example 1. The difference is that the traditional The graphical method is used to predict whether the alloy will undergo TRIP effect. The figure shows that the Ti-Mo-V-Cr-Al-Nb alloy is outside the TRIP region in the figure, i.e. The prediction result of the graph method is that the alloy does not have TRIP effect. Figure,[Mo] eq -[Fe] eq The graph shows a higher prediction accuracy.
[0095] Comparative Example 2:
[0096] The composition and mass percentage of Ti-Mo-V-Cr-Al-Zr alloy, smelting, forging and solution treatment are the same as those in Example 3. The difference is that the traditional The graphical method is used to predict whether the alloy will undergo TRIP effect. The figure shows that the Ti-Mo-V-Cr-Al-Zr alloy is outside the TRIP region in the figure, i.e. The prediction result of the graph method is that the alloy does not have TRIP effect. Figure,[Mo] eq -[Fe] eq The graph shows a higher prediction accuracy.
[0097] In summary, this embodiment proposes a TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent, establishes the original data set of titanium alloy, and integrates the Random Forest model, XGBoost model, LightGBM model and Extra Randomized Tree model based on the Stacking model to achieve classification prediction of plastic deformation mechanism. The Stacking model is based on solution treatment temperature, solution treatment time, alloying elements, [Mo] eq , [Fe] eq And empirical parameters Value and The value is used as input and whether the TRIP effect occurs is used as output. Based on machine learning, [Mo] eq and [Fe] eq The most important factor affecting the TRIP effect. 40,000 sets of data were fitted within the corresponding data range using a machine learning model to expand the data set and determine whether the TRIP effect occurred. [Mo] was drawn based on the machine learning results. eq -[Fe] eq The graph can then be used to implement TRIP titanium alloy composition design and analysis without machine learning and data sets.
[0098] The various embodiments in this application are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.
[0099] The scope of protection of this application is not limited to the above-described embodiments. Obviously, those skilled in the art may make various modifications and variations to this disclosure without departing from the scope and spirit of this disclosure. If such modifications and variations fall within the scope of equivalent technologies of this disclosure, then the disclosure is intended to include such modifications and variations.
Claims
1. A TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent, characterized in that: include: Obtain the values of molybdenum equivalent and iron equivalent in titanium alloy; In the molybdenum equivalent-iron equivalent diagram, the probability of the TRIP effect occurring is obtained according to the molybdenum equivalent value and the iron equivalent value in the titanium alloy; The molybdenum equivalent-iron equivalent diagram is drawn by: establishing a coordinate system with the molybdenum equivalent and iron equivalent in the new data set as the horizontal and vertical coordinates, and using the probability of the TRIP effect as the value of the coordinate point in the coordinate system, wherein the new data set is generated using the Stacking model; When the probability of a TRIP effect occurring is greater than a preset threshold, it is considered that a TRIP effect has occurred.
2. The TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent according to claim 1, characterized in that: The new dataset is generated using the Stacking model and includes: Establishing a metastable β titanium alloy original data set, dividing the metastable β titanium alloy original data set into a training set and a test set, wherein the metastable β titanium alloy original data set includes: multiple factors affecting the TRIP effect; Using multiple basic machine learning models, training them on the training set and testing them on the test set, to obtain machine learning models of multiple factors affecting the TRIP effect; Integrate machine learning models of various factors affecting the TRIP effect to establish a Stacking model; According to the prediction results of the test set obtained in the stacking model, the most important influencing factors of the TRIP effect are screened. The most important influencing factors of the TRIP effect are molybdenum equivalent and iron equivalent; Based on molybdenum equivalent and iron equivalent, a new data set was generated using the Stacking model.
3. The TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent according to claim 1, characterized in that: In the molybdenum equivalent-iron equivalent diagram, the probability of the TRIP effect occurring is obtained by: Input the new data set into the Stacking model to obtain new prediction results; The SHAP analysis method is used to analyze the new prediction results and obtain the probability of the TRIP effect.
4. The TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent according to claim 2, characterized in that: The metastable β titanium alloy original data set includes: alloy elements, solution treatment temperature, solution treatment time, value, The value, molybdenum equivalent and iron equivalent are used as the input of the metastable β titanium alloy original data set, and the input of the metastable β titanium alloy original data set is the influencing factors of various TRIP effects; whether the TRIP effect occurs is used as the output of the metastable β titanium alloy original data set.
5. The TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent according to claim 2, characterized in that: A variety of basic machine learning models were used, trained on the training set and tested on the test set, to obtain machine learning models of various factors affecting the TRIP effect, including: The Random Forest model, the XGBoost model, the LightGBM model, and the ExtraRandomized Tree model are respectively trained on the training set to obtain machine learning models of various factors affecting the TRIP effect; The machine learning models based on the influencing factors of the various TRIP effects are tested on the test set to obtain test results of the various machine learning models; According to the test results of various machine learning models, the area under the ROC curve was used to evaluate the classification prediction accuracy of the test results of various machine learning models, so as to perform hyperparameter tuning and obtain optimized machine learning models of various factors affecting the TRIP effect.
6. The TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent according to claim 2, characterized in that: Based on the prediction results of the test set obtained in the stacking model, the most important factors affecting the TRIP effect are screened, including: Input the test set into the Stacking model to obtain the prediction results; The prediction results were analyzed using the SHAP analysis method to obtain the SHAP analysis value; The SHAP analysis values were sorted from high to low, and the two factors with the highest analysis values were taken as the most important influencing factors of the TRIP effect.
7. The TRIP titanium alloy composition analysis method based on molybdenum equivalent and iron equivalent according to claim 1, characterized in that: When the probability of a TRIP effect occurring is greater than a preset threshold, a TRIP effect is considered to have occurred, including: an area with a probability of a TRIP effect occurring greater than 0.85 is considered to be extremely likely to occur TRIP; an area with a probability of a TRIP effect occurring within the range of 0.75 to 0.85 is considered to be prone to TRIP; an area with a probability of a TRIP effect occurring within the range of 0.5 to 0.75 is considered to have the possibility of TRIP occurring; an area with a probability of a TRIP effect occurring within the range of 0.35 to 0.5 is considered to be unlikely to occur TRIP; and an area with a probability of a TRIP effect occurring less than 0.35 is considered to be unlikely to occur TRIP.