Method for predicting fracture toughness of TC18 titanium alloy die forging

Abstract

The invention provides a method for predicting the fracture toughness of a TC18 titanium alloy die forging. The method comprises the following steps: S1, collecting tensile fracture toughness data of the TC18 titanium alloy die forging; s1.1, samples of the tensile sample are extracted at different deformation positions of the die forging; s1.2, sampling in the crack propagation direction of each position of the sample; s1.3, respectively testing twice in each direction to obtain tensile fracture toughness data; s2, testing a microscopic structure; s2.1, carrying out microscopic structure sample preparation on the fractured sample; s2.2, obtaining a microscopic structure picture; s3, training data; s3.1, numbering the tensile fracture toughness data and the microscopic structure picture, and importing the data and the picture into an AlexNet network structure for programming; s3.2, carrying out data set training; s3.3, obtaining a prediction model after data set training; and S4, predicting data. And the fracture toughness of the TC18 titanium alloy die forging is predicted, so that the development period and cost can be greatly shortened, and the production of the forging can be effectively guided.

Description

Technical Field

[0001] The invention relates to the technical field of titanium alloy performance, in particular to a method for predicting the fracture toughness of a TC18 titanium alloy die forging. Background Art

[0002] TC18 titanium alloy is a high-strength and high-toughness near-β titanium alloy widely used in the aviation industry. Due to its excellent damage tolerance performance, it is widely used in the manufacture of aircraft structural parts; fracture toughness is an important indicator of the service of TC18 titanium alloy die forgings, and has an important impact on its service safety. However, as a typical near-β high-strength titanium alloy, the fracture toughness and strength of TC18 are in a contradictory relationship, which makes it difficult to regulate. From the perspective of microstructure, the grain boundary α continuity, the morphology of the original β grains, and the characteristics of the primary α of TC18 titanium alloy have a great influence on the fracture toughness. In addition, the measurement cycle of the material fatigue crack growth curve is long and the cost is high. In particular, the determination of the fatigue crack threshold is a long and laborious process. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method for predicting the fracture toughness of TC18 titanium alloy die forgings. Predicting the fracture toughness of TC18 titanium alloy die forgings can greatly shorten the development cycle and cost and effectively guide the production of forgings.

[0004] The technical solution adopted by the present invention to solve its technical problem is:

[0005] A method for predicting the fracture toughness of a TC18 titanium alloy die forging comprises the following steps:

[0006] S1. Collect the tensile fracture toughness data of TC18 titanium alloy die forgings;

[0007] S1.1. Extract tensile test specimens at different deformation positions of die forgings;

[0008] S1.2. Sampling is performed in the crack propagation direction at each position of the sample, wherein the crack propagation direction includes the vertical direction and the horizontal direction;

[0009] S1.3. Test twice in each direction to obtain tensile fracture toughness data;

[0010] S2, test microstructure;

[0011] S2.1. Prepare the microstructure of the sample after breaking, i.e. take cross-section samples at the fracture surface;

[0012] S2.2. After sampling, optical microscope and scanning electron microscope tests are performed to obtain microstructure images;

[0013] S3, training data;

[0014] S3.1. Number the tensile fracture toughness data and microstructure images, import them into the AlexNet network structure for programming, and train the data;

[0015] S3.2, during the training data programming process, the grain boundary α and original β grain contours identified by optical microscopy, and the length, width and content of primary α identified by scanning electron microscopy are defined as prior knowledge, and data set training is performed;

[0016] S3.3, obtain the prediction model after training the data set;

[0017] S4. Prediction data.

[0018] The TC18 titanium alloy die forging is used for quasi-β forging, and the TC18 titanium alloy die forging is double annealed before forging.

[0019] The range of different deformation amounts of the quasi-β forging die forgings is 15%-85%.

[0020] The double annealing includes a first stage annealing and a second stage annealing, wherein the first stage annealing is performed at a temperature of 830-850° C. and kept at that temperature for 0.5-4 hours; the second stage annealing is performed at a temperature of 740-760° C. and kept at that temperature for 0.5-4 hours, and the aging temperature is 600-620° C. and kept at that temperature for 2-10 hours.

[0021] In step S2.2, the test magnification of the optical microscope is 20-100 times, and the magnification of the scanning electron microscope is 200-4000 times.

[0022] In step S3.1, the imported data of tensile fracture toughness and microstructure images are ≥ 2000 groups, and the training data are ≥ 1500 groups.

[0023] In step S4, the predicted data is ≥ 500 groups.

[0024] The beneficial effects of the present invention are:

[0025] 1. After establishing the training model, the microstructure of TC18 titanium alloy can be input to predict the corresponding fracture toughness. It has the advantages of simple operation and high stability. It fully extracts the morphological characteristics such as grain boundary α continuity, original β grain morphology, primary α content and size in the microstructure of TC18 titanium alloy forgings, and has the characteristics of high prediction accuracy.

[0026] 2. By considering the influence of fracture resistance on the fracture toughness of titanium alloy, a fracture toughness prediction model based on the tensile properties and crack propagation path of titanium alloy was established. The fracture toughness of TC18 titanium alloy was successfully predicted, solving the problems of long cycle and high cost in traditional crack propagation rate test determination.

[0027] 3. After double annealing, TC18 titanium alloy can obtain a basketweave structure, which contains primary α sheets of 5-20 microns in length and secondary α of tens of nanometers. This structure has both strength, plasticity and fracture toughness and has excellent mechanical properties. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is the metallographic diagram of position No. 1;

[0029] Figure 2 is the scanning electron microscope image of position 1;

[0030] Figure 3 This is the metallographic diagram of position 2;

[0031] Figure 4 This is the SEM image of position 2;

[0032] Figure 5 The metallographic diagram of the predicted position No. 1 of the forging in Example 1;

[0033] Figure 6 This is a scanning electron microscope image of the predicted position No. 1 of the forging in Example 1;

[0034] Figure 7 The metallographic diagram of the predicted second position of the forging in Example 1;

[0035] Figure 8 This is a scanning electron microscope image of the predicted No. 2 position of the forging in Example 1;

[0036] Fig. 9 is the model accuracy curve;

[0037] Fig.10 Schematic diagram of double annealing. DETAILED DESCRIPTION

[0038] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0039] Example 1

[0040] like Figures 1 to 4 As shown, a method for predicting the fracture toughness of a TC18 titanium alloy die forging comprises the following steps:

[0041] S1. Collect the tensile fracture toughness data of TC18 titanium alloy die forgings;

[0042] S1.1. Extract tensile test specimens at different deformation positions of die forgings;

[0043] S1.2. Sampling is performed in the crack propagation direction at each position of the sample, and the crack propagation direction includes the vertical direction (“vertical direction” means “vertical direction”, which is called “horizontal direction” in the industry) and the horizontal direction;

[0044] S1.3, each direction was tested twice to obtain the tensile fracture toughness data. The sample data of the tensile specimens at different positions are shown in Table 1;

[0045] Serial number <![CDATA[Fracture toughness / MPa.m 1 / 2 > 1 78 2 82 3 90 4 60 …… ……

[0046] Table 1. List of tensile properties of TC18 titanium alloy at different positions

[0047] S2, test microstructure;

[0048] S2.1. Prepare the microstructure of the sample after breaking, i.e., take cross-section samples near the fracture;

[0049] S2.2. After sampling, optical microscope and scanning electron microscope tests are performed to obtain microstructure images;

[0050] Microstructure pictures of different positions (such as position 1 and position 2) Figures 1 to 4 As shown, Figure 1 Used for original β grain identification; Figure 2 Used for identification of primary α and grain boundary α; Figure 3 Used for original β grain identification; Figure 4 Used for identification of primary α and grain boundary α;

[0051] S3, training data;

[0052] S3.1. Number the tensile fracture toughness data and the optical microscope and scanning electron microscope microstructure images of the corresponding samples one by one, import them into the AlexNet network structure for programming, and train the data;

[0053] S3.2, during the training data programming process, the grain boundary α and original β grain contours identified by optical microscopy, and the length, width and content of primary α identified by scanning electron microscopy are defined as prior knowledge, and data set training is performed;

[0054] S3.3, obtain the prediction model after data set training (the prediction model is a program defined after big data training and recognition);

[0055] Use the trained prediction model in practice;

[0056] The microstructure of a forging at different positions is imported into the prediction model for prediction. The microstructure of the sample is as follows: Figures 5 to 8 The predicted and measured mechanical properties are shown in Table 2.

[0057]

[0058] Table 2 Comparison of predicted and measured mechanical properties

[0059] S4, prediction data;

[0060] The accuracy curves for predicting fracture toughness are as follows: Fig. 9 As shown, the accuracy of the model in fracture toughness is 99%.

[0061] The TC18 titanium alloy die forging is used for quasi-β forging, and the TC18 titanium alloy die forging is double annealed before forging. The quasi-β forging die forging has different deformation ranges of 15%-85%, and the range of 15%-85% can cover all deformation ranges of TC18 titanium alloy die forgings in the engineering manufacturing process.

[0062] like Fig.10 As shown, the double annealing includes a first stage annealing and a second stage annealing, the first stage annealing is at a temperature of 830-850°C and is kept warm for 0.5-4h; the second stage annealing is at a temperature of 740-760°C and is kept warm for 0.5-4h, and the aging temperature is 600-620°C and is kept warm for 2-10h.

[0063] After double annealing, TC18 titanium alloy can obtain a basketweave structure, which contains primary α sheets with a length of 5-20 microns and secondary α with a length of tens of nanometers. This structure has excellent mechanical properties, combining strength, plasticity and fracture toughness.

[0064] In step S2.2, the test magnification of the optical microscope is 20-100 times, and the magnification of the scanning electron microscope is 200-4000 times.

[0065] The imported data of the tensile fracture toughness and microstructure images are ≥ 2000 groups, and the training data are ≥ 1500 groups.

[0066] The predicted data is ≥ 500 groups.

[0067] After establishing the training model, the microstructure of TC18 titanium alloy can be input to predict the corresponding fracture toughness. It has the advantages of simple operation and high stability. It fully extracts the morphological characteristics such as grain boundary α continuity, original β grain morphology, primary α content and size in the microstructure of TC18 titanium alloy forgings, and has the characteristics of high prediction accuracy.

[0068] By considering the influence of fracture resistance on the fracture toughness of titanium alloy, a fracture toughness prediction model based on the tensile properties and crack propagation path of titanium alloy was established. The fracture toughness of TC18 titanium alloy was successfully predicted, solving the problems of long cycle and high cost in traditional crack propagation rate test determination.

[0069] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting the fracture toughness of TC18 titanium alloy die forgings, characterized in that: The following steps are involved: S1. Collect the tensile fracture toughness data of TC18 titanium alloy die forgings; S1.

1. Extract tensile test specimens at different deformation positions of die forgings; S1.

2. Sampling is performed in the crack propagation direction at each position of the sample, wherein the crack propagation direction includes the vertical direction and the horizontal direction; S1.

3. Test twice in each direction to obtain tensile fracture toughness data; S2, test microstructure; S2.

1. Prepare the microstructure of the sample after breaking, i.e. take cross-section samples at the fracture surface; S2.

2. After sampling, optical microscope and scanning electron microscope tests are performed to obtain microstructure images; S3, training data; S3.1, number the tensile fracture toughness data and microstructure images, import them into the AlexNet network structure for programming, and perform training data; S3.2, during the training data programming process, the grain boundary α and original β grain contours identified by optical microscopy, and the length, width and content of primary α identified by scanning electron microscopy are defined as prior knowledge, and data set training is performed; S3.3, obtain the prediction model after the data set is trained; S4. Prediction data.

2. The method for predicting the fracture toughness of a TC18 titanium alloy die forging according to claim 1, characterized in that: The TC18 titanium alloy die forging is used for quasi-β forging, and the TC18 titanium alloy die forging is double annealed before forging.

3. The method for predicting the fracture toughness of a TC18 titanium alloy die forging according to claim 2, characterized in that: The range of different deformation amounts of the quasi-β forging die forgings is 15%-85%.

4. The method for predicting the fracture toughness of a TC18 titanium alloy die forging according to claim 2, characterized in that: The double annealing includes a first stage annealing and a second stage annealing, wherein the first stage annealing is performed at a temperature of 830-850° C. and kept at that temperature for 0.5-4 hours; the second stage annealing is performed at a temperature of 740-760° C. and kept at that temperature for 0.5-4 hours, and the aging temperature is 600-620° C. and kept at that temperature for 2-10 hours.

5. The method for predicting the fracture toughness of a TC18 titanium alloy die forging according to claim 1, characterized in that: In step S2.2, the test magnification of the optical microscope is 20-100 times, and the magnification of the scanning electron microscope is 200-1500 times.

6. The method for predicting fracture toughness of a TC18 titanium alloy die forging according to claim 1, characterized in that: In step S3.1, the imported data of tensile fracture toughness and microstructure images are ≥ 2000 groups, and the training data are ≥ 1500 groups.

7. The method for predicting fracture toughness of a TC18 titanium alloy die forging according to claim 1, characterized in that: In step S4, the predicted data is ≥ 500 groups.

Images