A method for predicting the relationship between defect welding repair and fatigue life of TC17 titanium alloy blade

Abstract

The application discloses a prediction method for the relationship between welding repair defects and fatigue life of TC17 titanium alloy blades, and comprises the following steps: step one, establishing a fatigue life prediction model related to repair defects; step two, preparing a rotary bending fatigue sample; step three, heat treating a surfacing sample; step four, processing the rotary bending fatigue sample; step five, testing the rotary bending fatigue; step six, analyzing a fracture and counting defects; and step seven, establishing a rotary bending fatigue life prediction model. The application selects an electric arc welding wire repair test block, tests the rotary bending fatigue performance of the test block, observes and counts the defect morphology, size and position of the fracture, analyzes the corresponding relationship between the defects and the fatigue life, establishes a fatigue life prediction model related to the repair defects, and determines the influence of the pore defects on the fatigue life of the repaired blade through the model.

Description

Technical Field

[0001] This invention relates to the field of manufacturing and repair technology of titanium alloy blades for aero-engines, and in particular to a method for predicting the relationship between welding repair defects and fatigue life of TC17 titanium alloy blades. Background Technology

[0002] The titanium alloy compressor blades and integral bladed disks for aero-engines are made of TC17. During use, these blades are susceptible to wear, foreign object damage (FOD), cracks, corrosion, and fatigue damage. To reduce manufacturing costs and extend blade life, welding or laser deposition methods are commonly used for repair. Because these components must withstand high rotational speeds and centrifugal loads during operation, the fatigue life requirements for both materials and structures are very high.

[0003] Titanium alloys are chemically highly reactive, and their ability to absorb oxygen, nitrogen, and hydrogen increases significantly with rising temperatures. Therefore, after welding or laser deposition repair, titanium alloy blades and integral bladed disks inevitably contain varying degrees of porosity defects within the repaired material. These defects significantly impact the fatigue performance of the blades. Summary of the Invention

[0004] To address the aforementioned technical issues, a method for predicting the relationship between welding repair defects and fatigue life in TC17 titanium alloy blades is proposed. The specific technical solution is as follows:

[0005] A method for predicting the relationship between welding repair defects and fatigue life of TC17 titanium alloy blades, characterized by the following steps:

[0006] Step 1: Establish a fatigue life prediction model related to defect repair

[0007] Test blocks repaired by electric arc fuse were selected, and their bending fatigue performance was tested. The morphology, size, and location of defects at the fracture were observed and statistically analyzed. The correspondence between defects and fatigue life was analyzed, and a fatigue life prediction model related to defect repair was established.

[0008] Step 2: Preparation of bending fatigue specimens

[0009] TC17 forgings were selected as the test plate base, and arc deposition was carried out on the test plate using an automatic argon arc welding machine. The welding wire was TC17 with a diameter of 1.6 mm.

[0010] Step 3: Heat treatment of the weld overlay specimen

[0011] The test plate after arc deposition was subjected to vacuum heat treatment. The heat treatment regime was: 550℃, holding for 4 hours, and then cooling with the furnace.

[0012] Step 4: Take bending fatigue specimens by wire cutting and process them by turning and grinding to meet the standard requirements, ensuring that the interface between the matrix and the accumulation zone is in the center of the specimen.

[0013] Step 5: Bending Fatigue Test

[0014] (1) The bending fatigue performance test was carried out by gradient loading. The samples were divided into 3 groups, with initial loading stresses of 350MPa, 450MPa and 500MPa respectively.

[0015] (2) Set the number of samples in each group and conduct slewing fatigue performance tests at stress levels of 350MPa, 470MPa and 500MPa.

[0016] Step Six: Fracture Analysis and Defect Statistics

[0017] The fracture surface of the bending fatigue test specimen was observed, and the shape, size and location of the defects were statistically analyzed.

[0018] Step 7: Establish a bending fatigue life prediction model

[0019] Based on the dependence of material yield strength on parameters such as composition, the following general functional form of fatigue life is selected: σN o =C

[0020] Where σ is the fatigue stress, and a and C are constants;

[0021] Considering that fatigue stress is closely related to defect repair, and that the morphology, location, size, and distribution of defects all affect service life, fatigue stress is considered only as a function of defect size and location: σ = σ(D, S).

[0022] Where D and S are the location and size of the defect, respectively;

[0023] Based on experimental data, under the constraints of nominal fatigue stress and sample macroscopic dimensions, the expression for fatigue life obtained using supervised machine learning is as follows:

[0024] Where R is the radius of the bent sample, D is the distance between the defect and the sample surface, and S is the diameter of the defect;

[0025] The model predicts the lifespan and the actual lifespan, with the upper and lower 10% confidence intervals marked by dashed lines. Based on the numerical model, the general model for bending fatigue life can be written in the following form:

[0026] Where N and σ oHere, denoted as fatigue life and nominal fatigue stress, respectively; D is the vertical distance from the defect to the surface; R is the sample radius; S is the defect size; and the rest are constants.

[0027] The beneficial effects of this invention are:

[0028] This invention utilizes test blocks repaired with arc-fused wire to test their bending fatigue performance. It observes and statistically analyzes the morphology, size, and location of defects at the fracture surface, analyzes the correlation between defects and fatigue life, and establishes a fatigue life prediction model related to repaired defects. This model can determine the impact of porosity defects on the fatigue life of the repaired blade. This invention can be directly applied to the welding and laser deposition repair of TC17 material blades and integral bladed disk components, guiding the establishment of welding and laser deposition repair quality acceptance standards to ensure that the performance of the repaired blades meets fatigue performance requirements. This extends the service life of parts and reduces replacement repair costs. This invention can be extended to the prediction of fatigue life after welding repair of rotor components made of other materials, possessing broad application potential and economic benefits. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of an arc deposition sample.

[0030] Figure 2 This is a schematic diagram of a bending specimen;

[0031] Figure 3 This is a schematic diagram of the bending fatigue life prediction model and experimental values. Detailed Implementation

[0032] The following is in conjunction with the appendix Figure 1-3 The embodiments further illustrate the present invention in detail.

[0033] Example 1

[0034] A method for predicting the relationship between welding repair defects and fatigue life of TC17 titanium alloy blades, characterized by the following steps:

[0035] Step 1: Establish a fatigue life prediction model related to defect repair

[0036] Test blocks repaired by electric arc fuse were selected, and their bending fatigue performance was tested. The morphology, size, and location of defects at the fracture were observed and statistically analyzed. The correspondence between defects and fatigue life was analyzed, and a fatigue life prediction model related to defect repair was established.

[0037] Step 2: Preparation of bending fatigue specimens

[0038] TC17 forgings were selected as the test plate base, and arc deposition was carried out on the test plate using an automatic argon arc welding machine. The welding wire grade was TC17, with a diameter of 1.2mm-1.6mm.

[0039] Step 3: Heat treatment of the weld overlay specimen

[0040] The test plate after arc deposition was subjected to vacuum heat treatment. The heat treatment regime was: 550℃±10℃±, holding for 4 hours to 4 hours and 10 minutes, and then cooled with the furnace.

[0041] Step 4: Take bending fatigue specimens by wire cutting and process them by turning and grinding to meet the standard requirements, ensuring that the interface between the matrix and the accumulation zone is in the center of the specimen.

[0042] Step 5: Bending Fatigue Test

[0043] (1) The slewing fatigue performance test was carried out by gradient loading. The specimens were divided into 3 groups, with 6 specimens in each group. The initial loading stresses were 350MPa, 450MPa and 500MPa respectively. If the specimens did not break after 1 million cycles, the stress level was increased to 500MPa for a second loading.

[0044] (2) The bending fatigue performance test was carried out by single stress loading. The specimens were divided into two groups with loading stresses of 470MPa and 500MPa, respectively. There were 6 specimens in each group of 470MPa and 16 specimens in each group of 500MPa.

[0045] Step Six: Bending Fatigue Test Results

[0046] (1) The test results of gradient loading are shown in Table 1. Under the stress of 350MPa, the life of all 6 specimens reached 1 million cycles. Under the stress of 450MPa, the life of 2 specimens was less than 1 million cycles. Under the stress of 500MPa, the life of 4 specimens was less than 1 million cycles. It can be seen that with the increase of stress level, the possibility of fatigue failure and fracture increases.

[0047] Table 1 Results of Gradient Loading Room Temperature Rotational Bending Fatigue Test

[0048]

[0049]

[0050] (2) The results of single stress loading are shown in Table 2. Under the stress condition of 470MPa, the life of two specimens was less than 1 million cycles, which is similar to the case when the initial stress was 450MPa in the gradient loading test. Three specimens did not break after cycling up to 10 million cycles. Under the stress condition of 500MPa, only one specimen did not break. The fatigue life of the fractured specimens ranged from about 20,000 cycles to 4 million cycles, with obvious fluctuations.

[0051] Table 2 Results of single-stressed room temperature rotating bending fatigue test

[0052]

[0053]

[0054] Step 7: Fracture Analysis and Defect Statistics

[0055] The results of bending fatigue tests and fracture analysis were summarized, and the shape, size, and distribution of defects were statistically analyzed, as shown in Tables 3 and 4. When defects were located on the surface, the fatigue life was significantly lower, generally not exceeding 100,000 cycles; when defects were located inside, the fatigue life was affected by the size and distribution of the defects.

[0056] Table 3. Statistics of fracture surface defects during gradient loading and bending fatigue.

[0057]

[0058]

[0059] Table 4. Statistics of fracture surface defects under 500MPa single stress loading bending fatigue

[0060]

[0061] Fracture surfaces of gradient-loaded bending fatigue specimens with an initial stress of 350 MPa were observed and sorted by fatigue life from low to high at 550 MPa. Circular pores were visible on the fracture surfaces, serving as crack initiation points. This indicates that when defects are present in the specimens, a certain stress level is required for crack initiation. When the stress is below a certain level, i.e. below the fatigue limit, even with certain defects, the specimens can be considered safe. Statistical analysis of pore size and distribution revealed that pore distribution significantly affects fatigue life. The fatigue life of the three specimens with pores located on the surface was less than 100,000 cycles. The farther the pores were from the surface, even larger ones could achieve a fatigue life exceeding 1 million cycles.

[0062] All samples that fractured at 450 MPa had defects on their surfaces. The three samples that fractured at 550 MPa had similar lifespans, but the defect distributions differed. The defect in sample 450-3# was very close to the surface but only about 13 μm in size. No defects were found at the crack initiation site of sample 450-4#. The pores in sample 450-5# were larger but more than 500 μm from the surface. This shows that smaller defect sizes and greater distances from the surface result in higher bending fatigue life.

[0063] Most of the gradient-loaded fatigue specimens with an initial stress of 500 MPa fractured at 500 MPa. Therefore, their fracture surfaces were classified and analyzed in the same category as those of fatigue specimens subjected to single stress loading at 500 MPa. When pores were present on the specimen surface, the fatigue life was significantly shorter. As the size of the pores decreased and they moved further away from the specimen surface, the fatigue life gradually increased.

[0064] Step 8: Establish a bending fatigue life prediction model

[0065] Based on the dependence of material yield strength on parameters such as composition, the following general functional form of fatigue life is selected: σN o =C

[0066] Where σ is the fatigue stress, and a and C are constants;

[0067] Considering that fatigue stress is closely related to defect repair, and that the morphology, location, size, and distribution of defects all affect service life, fatigue stress is considered only as a function of defect size and location: σ = σ(D, S).

[0068] Where D and S are the location and size of the defect, respectively;

[0069] Based on experimental data, under the constraints of nominal fatigue stress and sample macroscopic dimensions, the expression for fatigue life obtained using supervised machine learning is as follows:

[0070] Where R is the radius of the bent sample, D is the distance between the defect and the sample surface, and S is the diameter of the defect;

[0071] The model predicts the lifespan and the actual lifespan, with the upper and lower 10% confidence intervals marked by dashed lines. Based on the numerical model, the general model for bending fatigue life can be written in the following form:

[0072] Where N and σ o Here, denoted as fatigue life and nominal fatigue stress, respectively; D is the vertical distance from the defect to the surface; R is the sample radius; S is the defect size; and the rest are constants.

Claims

1. A method for predicting the relationship between welding repair defects and fatigue life of TC17 titanium alloy blades, characterized in that: The steps are as follows: Step 1: Establish a fatigue life prediction model related to defect repair Test blocks repaired by electric arc fuse were selected, and their bending fatigue performance was tested. The morphology, size and location of defects at the fracture were observed and statistically analyzed. The correspondence between defects and fatigue life was analyzed, and a fatigue life prediction model related to the repaired defects was established. Step 2: Preparation of bending fatigue specimens TC17 forgings were selected as the test plate base, and arc deposition was carried out on the test plate using an automatic argon arc welding machine. The welding wire was TC17 with a diameter of 1.6 mm. Step 3: Heat treatment of the weld overlay specimen The test plate after arc deposition was subjected to vacuum heat treatment. The heat treatment regime was: 550℃, holding for 4 hours, and then cooling with the furnace. Step 4: Take bending fatigue specimens by wire cutting and process them by turning and grinding to meet the standard requirements, ensuring that the interface between the matrix and the accumulation zone is in the center of the specimen. Step 5: Bending Fatigue Test (1) The bending fatigue performance test was carried out by gradient loading. The specimens were divided into 3 groups, and the initial loading stresses were 350MPa, 450MPa and 500MPa, respectively. (2) Set the number of samples and conduct slewing fatigue performance tests at stress levels of 350MPa, 470MPa and 500MPa. Step Six: Fracture Analysis and Defect Statistics The fracture surface of the bending fatigue test specimen was observed, and the shape, size and location of the defects were statistically analyzed. Step 7: Establish a bending fatigue life prediction model Based on experimental data, under nominal fatigue stress and sample macroscopic dimensions, a supervised machine learning expression for fatigue life was obtained, and a general model for bending fatigue life was established based on the numerical model.

2. The method for predicting the relationship between welding repair defects and fatigue life of TC17 titanium alloy blades according to claim 1, characterized in that: In step seven, based on the dependence of material yield strength on composition, the following general functional form of fatigue life is selected: in, The stress is fatigue stress, where a and C are constants; Considering that fatigue stress is closely related to defect repair, and that the morphology, location, size, and distribution of defects all affect service life, fatigue stress can be viewed solely as a function of defect size and location. Where D and S are the location and size of the defect, respectively; Based on experimental data, under the constraints of nominal fatigue stress and sample macroscopic dimensions, the expression for fatigue life obtained using supervised machine learning is as follows: Where R is the radius of the bent sample, D is the distance between the defect and the sample surface, and S is the diameter of the defect; The model predicts the lifespan and the actual lifespan, with the upper and lower 10% confidence intervals marked by dashed lines. Based on the numerical model, the general model for bending fatigue life can be written in the following form: Where N and Here, denoted as fatigue life and nominal fatigue stress, respectively; D is the vertical distance from the defect to the surface; R is the sample radius; S is the defect size; and the rest are constants.

Images