Titanium alloy weld joint fatigue life prediction method considering plasma welding process key parameters and loading mode

Abstract

The invention discloses a titanium alloy weld joint fatigue life prediction method considering plasma welding process key parameters and a loading mode, and relates to the technical field of welding. Comprising the following three parts: firstly, detecting internal defect total information of a welding component with known process parameters by controlling welding process parameters and using a combined welding seam defect detection method; secondly, establishing a data model between the process parameters and the defect information of the welding component through a correlation analysis and random modeling method; and finally, through a fatigue experiment and high-precision simulation, establishing a prediction method for the fatigue life of the titanium alloy in consideration of the process parameters and the loading form. According to the method, the fatigue life of the weld joint structure under the action of the complex load stress spectrum can be predicted, the fatigue performance of the structure can be completely and accurately evaluated, guidance can be provided for welding engineers, optimization of welding design, selection of a proper welding process and adoption of corresponding post-welding treatment measures are helped, and the welding quality is improved. The fatigue performance of the welding seam structure is improved; and the service life is prolonged.

Description

Technical Field

[0001] The present invention belongs to the field of welding technology, and more specifically, relates to a method for predicting fatigue life of titanium alloy welds taking into account key parameters of plasma welding process and loading methods. Background Art

[0002] With the rapid development of materials science and technology, titanium alloys are widely used in aerospace, weapons and equipment, medical equipment and nuclear power. In the aerospace field, titanium alloys are used to manufacture aircraft fuselages, wings, skins and other parts; in the field of weapons and equipment, they are used for towing artillery, self-propelled artillery and armored vehicles; in the medical field, they are used for the manufacture of biomedical equipment; in the field of nuclear power, they are used to manufacture nuclear power equipment. Since titanium alloy materials are mostly used in complex structural parts in high-tech fields, these components are often complex in shape and have high requirements for connection accuracy, strength and durability. Compared with other connection methods, welding technology can better meet the various requirements of titanium alloy components and achieve efficient and high-quality connections. With the application and research and development of plasma welding technology, more and more studies have shown that during the welding process, due to the influence of various process factors, defects such as cracks, unfusion, pores, slag inclusions, etc. may occur at the weld of the component, making it easier to produce fatigue cracks when the weld of the component is subjected to alternating loads, reducing the strength and toughness of the weld, and then fatigue fracture may occur, affecting the fatigue life. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a titanium alloy weld fatigue life prediction method taking into account the key parameters and loading methods of the plasma welding process, which can predict the fatigue life of the weld structure under the action of a complex load stress spectrum and conduct a complete and accurate evaluation of the fatigue performance of the structure; it can also provide guidance for welding engineers, help optimize welding design, select appropriate welding processes, and take corresponding post-welding treatment measures to improve the fatigue performance of the weld structure and extend its service life.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A method for predicting the fatigue life of titanium alloy welds considering key parameters of plasma welding process and loading methods includes three parts: first, by controlling the welding process parameters and using a combined weld defect detection method to detect the full information of internal defects of welded components with known process parameters; then, through correlation analysis and stochastic modeling methods, a data model between process parameters and welded component defect information is established; finally, through fatigue experiments and high-precision simulations, a method for predicting the fatigue life of titanium alloys considering process parameters and loading methods is established.

[0006] Preferably, by controlling the welding process parameters and using the newly developed combined weld defect detection method, the full information of internal defects of the welded component with known process parameters is detected, including the following:

[0007] M1: Determine and discretize welding process parameters, including welding current, welding speed, and plasma gas flow rate, and select the process parameter combination closest to the actual working conditions for plasma welding test;

[0008] M2: Use phased array ultrasonic testing method to measure the microscopic defects of the main weld structure along its length; obtain and classify the welding defects of plasma welding for several different process conditions; based on the phased array ultrasonic testing results, directly or indirectly obtain the descriptive parameters of the weld microscopic defects, namely, the effective weld width, effective weld depth and effective weld cross-sectional area, etc.;

[0009]

[0010] ED j,i =g[z j,i ,T i ] (2)

[0011]

[0012] Where i represents the type number of the defect, T i Indicates the type of weld defect; j indicates the same component T i The jth defect under defect type, and Indicates T i The starting and ending positions of the jth defect under defect type, z j,i Indicates T i Depth position of the jth defect under defect type, EW j,i represents the effective weld width of the jth defect under weld defect type i, ED j,i represents the effective weld depth of the jth defect under weld defect type i, EA j,i represents the effective weld cross-sectional area of ​​the jth defect under weld defect type i, f represents the effective weld width describing the jth defect under weld defect type i, which is related to the weld defect type and the starting and ending positions of the jth defect under its defect type, g represents the effective weld depth describing the jth defect under weld defect type i, which is related to the weld defect type and the depth position of the jth defect under its defect type, and h represents the effective weld cross-sectional area describing the jth defect under weld defect type i, which is related to the weld defect type and the starting and ending positions of the jth defect under its defect type, and the weld defect type and the depth position of the jth defect under its defect type;

[0013] M3: Based on the quantitative analysis of the first two steps, the process parameters, namely welding current, welding speed and plasma gas flow, as well as the descriptive parameters of weld micro defects can be obtained. The relationship between defect parameters and related process parameters can be expressed by the following formula:

[0014]

[0015] Among them, r represents the Pearson correlation coefficient. The closer |r| is to 1, the higher the correlation is, and the closer |r| is to 0, the lower the correlation is. i Indicates the input process parameter value; Y i Indicates the output of the descriptive parameter value of weld micro defects; Indicates the average value of input process parameters; It indicates the average value of descriptive parameters of weld micro defects;

[0016] Taking the highly correlated process parameters as input and the mean of the descriptive parameters of each type of weld defect as input, a mathematical model between input and output is established through the multivariate regression method.

[0017]

[0018]

[0019]

[0020] Where I represents welding current, v represents welding speed, Q V Indicates the plasma gas flow rate, represents the effective width of weld defect type i, represents the effective depth of weld defect type i, represents the effective weld cross-sectional area under weld defect type i, h i The effective weld width under the weld defect type i is related to the welding current, welding speed and plasma gas flow rate. m i The effective weld depth under the weld defect type i is related to the welding current, welding speed and plasma gas flow rate. i The representative description of the effective weld cross-sectional area under weld defect type i is related to welding current, welding speed and plasma gas flow rate;

[0021] Based on different working conditions, the mean of the descriptive parameters corresponding to each type of defect can be obtained. Noise is added to the mean so that its distribution is consistent with the statistical data in the previous step. Then, a random model of the micro defect parameters can be obtained for subsequent analysis.

[0022] Preferably, a data model between process parameters and welding component defect information is established through correlation analysis and stochastic modeling methods, including the following contents:

[0023] N1: Obtain various material properties of the weld material and conduct material tests and fatigue tests; the tests conducted include tensile tests, triaxial stress tests, crack tip opening displacement tests, and crack extension tests;

[0024] N2: Conduct fatigue mechanical properties test of weld structure under multiple load conditions;

[0025] N3: Based on the material properties and geometric properties obtained from the above tests, a high-precision numerical simulation model is established and finite element fatigue strength analysis is performed;

[0026] N4: Considering that the fatigue failure mode of structural welds containing micro defects is usually manifested as fracture failure, in order to simulate the fracture failure mode and study the influence of weld micro defects on its fatigue strength and life, it is necessary to make an equivalent initial crack assumption;

[0027] N5: In order to analyze the influence of weld defects caused by different process parameters on component fatigue life and failure mode under different load conditions, based on the random model of weld micro-defects of strong discrete process, it is necessary to characterize the mathematical relationship between the micro-defect descriptive parameters, load conditions and equivalent initial crack geometric parameters and establish the corresponding mathematical model, and then load the equivalent initial crack parameters into the finite element model for crack propagation calculation and fatigue life prediction, so as to complete the prediction method of weld fatigue mechanical properties under process parameters and load conditions; the following specific steps are adopted:

[0028] ① Combining fracture mechanics and statistical methods, the micro-defect descriptive parameters obtained after processing the test results, the determined load conditions and the geometric parameters of the equivalent initial crack are characterized, and the mathematical relationship between the three is constructed;

[0029]

[0030]

[0031]

[0032]

[0033]

[0034] Among them, x 0 ,y 0 is the coordinate of the equivalent initial crack position, a 0 is the equivalent initial crack depth, b 0is the equivalent initial crack length, θ 0 is the equivalent initial crack direction;

[0035] l represents the coordinate x describing the equivalent initial crack position 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0036] u represents the coordinate y describing the equivalent initial crack position 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0037] p represents the equivalent initial crack depth a 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0038] q represents the equivalent initial crack length b 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0039] s represents the equivalent initial crack direction θ 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0040] ② Using the stereographic fitting technique, a mathematical model was constructed through Matlab, where the x-axis and y-axis represent different load conditions and micro-defects, respectively, and the z-axis represents the geometric parameters of the equivalent initial crack;

[0041] ③ Based on the above functions and the random model of weld microdefects based on strong discrete process, under the conditions of given process parameters and load conditions, the weld microdefect parameters can be determined by the process parameters, and the geometric parameters of the equivalent initial crack can be calculated under given fatigue cycle load conditions, that is, the position of the equivalent initial crack, the equivalent initial crack depth, the equivalent initial crack length and the equivalent initial crack direction. The equivalent initial crack is introduced into the finite element fatigue model for crack propagation calculation and fatigue life prediction, thereby obtaining the fatigue strength, life and failure mode of the component.

[0042] Preferably, in step N3, the steps of establishing a finite element model are as follows:

[0043] a1 Determine the material and geometric properties of the test piece, including the material properties of the weld structure and the geometric information of the component obtained from the test, as well as the position, size and shape of the weld;

[0044] a2 The finite element model was established by ABAQUS software, in which the solid elements of the weld structure were divided into hexahedral grids. To ensure the calculation accuracy, the upper structure was meshed and refined. When arranging the global network seeds, the unit size was set to 10 mm, and when arranging the local network seeds of the upper structure, the unit size was set to 5 mm.

[0045] a3 assigns the corresponding titanium alloy material properties to the geometric model after the mesh division;

[0046] a4 Set the analysis step, select "Static, General", reduce the initial incremental step and the minimum incremental step, set the maximum incremental step to 1, and the number of steps to 10000;

[0047] a5 specifies boundary conditions, completely fixing the base around and at the bottom, and applying loads in two perpendicular directions to the circular structure;

[0048] a6 performs finite element analysis to obtain the stress distribution of the structure and the size and position of the maximum stress point in the weld area.

[0049] Preferably, the position, depth, length and direction of the equivalent initial crack are directly related to the weld microscopic defects and stress state. In step N4, the specific operation of performing the equivalent initial crack assumption is as follows:

[0050] b1 Preliminary estimation of equivalent initial crack

[0051] The position of the equivalent initial crack is assumed to be a semi-elliptical crack. The horizontal plane at the component weld is taken as the reference plane. Without considering the weld height, the centroid of the plane component is defined as the reference origin o to establish the xoy coordinate system. The maximum weld stress calculated by the finite element model is assumed to be the crack centroid position (x 0 ,y 0 ); secondly, the direction of the equivalent initial crack, considering the vertical load condition, assuming that θ is 45° downward from the horizontal plane; finally, the size of the equivalent initial crack, assuming that the equivalent initial crack depth a 0 is 1 mm, and the equivalent initial crack length b 0 1mm;

[0052] b2 Establishment of the fatigue numerical model of weld structure considering equivalent initial crack

[0053] In FRANC3D software, the initial crack is introduced into the finite element model of the component according to the above-mentioned assumed geometric parameters, and the corresponding crack propagation method and corresponding parameters are selected to predict fatigue crack propagation and fatigue life;

[0054] b3 Equivalent initial crack verification and correction

[0055] b4 Finite Element Model Verification

[0056] After the geometric parameters of the equivalent initial crack are corrected, the finite element mesh density, analysis step parameters, and loading method of the boundary conditions are readjusted. The initial crack is reintroduced into the adjusted finite element model in FRANC3D to predict crack propagation and fatigue life. The simulation results are compared with the fatigue test results. The finite element model is verified by fatigue life, fatigue strength, and fatigue failure mode. The finite element model parameters are continuously corrected and the above steps are repeated to ensure the calculation accuracy of the finite element model.

[0057] Preferably, in step b3, the equivalent initial crack is verified and corrected by comparing the fatigue mechanical properties calculated by the simulation model with the actual fatigue mechanical properties obtained by the component fatigue test, and the verification and correction method is as follows:

[0058] ①Verify the location of the equivalent initial crack

[0059] According to the fatigue test results of the component, the crack on the surface of the component is inferred from the interface of fracture failure. Assuming that the crack is centered at the centroid point, its depth and length expand in equal proportion when expanding, and the centroid position (x 1 ,y 1 ), to verify the equivalent initial crack position (x 0 ,y 0 ) and make corrections; geometric expansion formula:

[0060]

[0061] Among them, β is the geometric expansion coefficient, a is the crack depth when the crack expands, b is the crack length when the crack expands, a 0 is the equivalent initial crack depth, b 0 is the equivalent initial crack length;

[0062] ②Verify the size of the equivalent initial crack

[0063]

[0064] Based on the Paris model assumption, a is the crack depth during crack propagation, N is the number of stress cycles, C and m are material-related parameters, C and m are basic parameters describing the fatigue crack propagation performance of materials, where m is specifically the crack propagation index, and the ΔK stress intensity factor range is calculated through the integral formula, where the lower limit in the integral formula is a 1 , the upper limit is a 2 ;

[0065] If a 1 Substitute the equivalent initial crack depth a 0 , a 2 Substitute the final crack depth into the initial crack depth a and reverse the solution. 0 ;

[0066] Since it is assumed that the crack depth and length expand in a constant ratio β during crack extension, the final crack depth a is measured by f and crack depth b f , substitute into formula 8 to calculate the equivalent initial crack length b 0 ;

[0067] The following is the Paris model,

[0068]

[0069] ΔK=K max -K min (15-b)

[0070]

[0071] Where a is the crack depth when the crack propagates, N is the number of stress cycles, ΔK is the change in stress intensity factor, K is the stress intensity factor, σ is the nominal stress, and Y is the shape factor;

[0072] ③Verify the equivalent initial crack direction

[0073] According to the crack propagation direction in the fracture failure mode of the fatigue test results, the direction of the crack at the time of crack initiation is found for verification, and the equivalent initial crack direction angle is corrected by comparing the crack propagation path with the propagation path in the simulation model;

[0074] ④Finally, correct the crack propagation direction and adjust the calculation criteria of the kink angle

[0075] The calculation of crack propagation direction can be determined by the kink angle, and the expansion calculation is implemented through FRANC3D software; the kink angle is determined by the stress intensity factor. After the calculation criterion of the kink angle is selected, the loading cycle mode of the load spectrum is selected, the stress ratio is set, and the crack propagation calculation is performed to obtain the fracture failure mode, fatigue strength and fatigue life; the crack propagation direction, fracture failure mode, fatigue strength and fatigue life in the FRANC3D simulation and fatigue test results are compared to verify whether the equivalent initial crack length, depth and direction are assumed to be accurate. By comparing the crack propagation path, the calculation criterion scheme closest to the test results is selected to complete the correction of the crack propagation direction.

[0076] Preferably, through fatigue experiments and high-precision simulations, the steps for predicting the fatigue life of TC4 titanium alloy under process parameters and loading forms are as follows:

[0077] C1: Obtain the main SN curve of titanium alloy weld structure

[0078] ① Based on the above method of obtaining microscopic defect parameters and load conditions to obtain equivalent initial cracks, load them into a large number of fatigue models to calculate structural stress and life; by simulating the stress cycle under actual working conditions, stress analysis is performed on the weld components, and the structural stress at any point on the structure can be calculated under given conditions;

[0079] σ s =σ me +σ b (16)

[0080] Among them, σ s represents the structural stress, σ me represents the membrane stress, σ b represents bending stress;

[0081]

[0082] Among them, ΔS s represents the range of equivalent structural stress variation, m is the crack extension index, t is the plate thickness, and I(r) is a dimensionless function of the load-bending ratio r

[0083]

[0084] Where a is the crack depth during crack expansion, M kn is the stress intensity magnification factor, are dimensionless functions for determining the range of stress intensity factors when membrane stress and bending stress act alone, is the differential of a / t, M kn is the stress intensity amplification factor;

[0085] ② The calculated structural stress is converted into equivalent structural stress by formula 16, and the fatigue life N corresponding to different equivalent structural stresses is calculated by finite element method. Since the dimensionless function I(r) of load-bending ratio is introduced into the equivalent structural stress formula, the main SN curve obtained by the equivalent structural stress method takes into account the influence of the initial crack depth and crack extension;

[0086] ③ Plot all points in the coordinate system to draw a main SN curve of a titanium alloy weld structure with a narrow discrete band. This curve can take into account the influence of initial cracks on fatigue life. In engineering, the curve can be directly referenced to determine the corresponding fatigue life by calculating the equivalent structural stress, thus completing the prediction of the structural fatigue life;

[0087] C2: Based on the main SN curve of the titanium alloy weld structure obtained above, under the specific environmental conditions and stress ratio of the actual project, analyze the stress type and range that the structure is subjected to under actual working conditions, including the maximum stress and the minimum stress; by calculating the stress amplitude (that is, half of the difference between the maximum stress and the minimum stress), under the known stress amplitude, find the corresponding number of cycles from the SN curve, the horizontal axis value corresponds to the number of cycles before fatigue failure, the vertical axis value corresponds to the stress range, and the number of cycles is the maximum number of cycles that the material can withstand before fatigue failure at this stress level; in order to compensate for possible calculation errors, material performance fluctuations, unconsidered load conditions and other uncertain factors, it is necessary to introduce a safety factor in the design; use the standard safety factor to adjust the life prediction value previously read from the SN curve to ensure that the fatigue life of the structure can still be guaranteed under the most unfavorable conditions, thereby completing the prediction and evaluation of the fatigue life;

[0088] C3: In order to accurately evaluate the life of the weld structure, the Rainflow Counting Method is used to deal with the influence of multi-segment cyclic loads; for multi-axis load conditions, the load projections of the x, y, and z axes are firstly calculated by linear superposition to obtain the equivalent structural stress component σ x , σ y , σ z , and then decompose the complex variable frequency load into multiple horizontal cyclic loads through the rain flow counting method, calculate the fatigue life N within the stress range of the x-axis, y-axis, and z-axis directions respectively, and then calculate the fatigue damage accumulation in the x, y, and z directions respectively through the Miner criterion, and finally sum up and calculate the total accumulated damage of the structure, and then perform fatigue assessment;

[0089] After applying the rain flow counting method to stress analyze the weld structure, the Palmgren-Miner criterion is used to consider the accumulation of fatigue damage for each different constant amplitude stress range; the damage fraction d is calculated by summing the fatigue damage contribution of each stress range; the fatigue damage contribution of each stress range is calculated by dividing the number of cycles by the number of failure cycles of the stress range;

[0090]

[0091] Where S represents the damage score, n i Indicates that under constant amplitude cyclic stress σ i The number of cycles under the current or scheduled loading, N i Indicates that under constant amplitude cyclic stress σ i The fatigue life under the condition of constant amplitude cyclic stress σ is calculated by FRANC3D i The fatigue life under the condition can be obtained;

[0092] Add up the damage contributions of all stress ranges. If the total damage fraction D>1, it is considered that fatigue failure has occurred. Otherwise, it is considered that fatigue failure has occurred. This is used to evaluate whether the structure will suffer from fatigue failure.

[0093] The beneficial effects of adopting the above technical solution are:

[0094] 1. The present invention obtains micro-defect parameters by inputting strong discrete process parameters into the random model with micro-defects, and calculates fatigue life by loading the initial crack into the fatigue model under given working conditions. The fatigue life of the weld structure under the action of complex load stress spectrum can be predicted by combining with the equivalent structural stress method, and the fatigue performance of the structure can be fully and accurately evaluated. It can also provide guidance for welding engineers, help optimize welding design, select appropriate welding processes, and take corresponding post-welding treatment measures to improve the fatigue performance of the weld structure and extend its service life.

[0095] 2. The present invention is highly customized and accurate. By inputting strong discrete process parameters into the micro defect stochastic model, the present invention can accurately capture various micro defect characteristics generated during the welding process, such as cracks, inclusions, pores, etc. These parameters are often simplified or ignored in traditional methods. This highly customized method makes the prediction results closer to reality and improves the accuracy of fatigue life assessment.

[0096] 3. The present invention embodies the adaptability of multiple working conditions. The existing technology usually predicts fatigue life based on a single or simplified working condition, while the present invention can treat micro defects as initial cracks and load them into the fatigue model under given complex working conditions, thereby more comprehensively considering the impact of multiple stress states on weld fatigue life in the actual working environment, and improving the comprehensiveness and practicality of the prediction.

[0097] 4. The present invention highlights the optimization of efficient structure and design. By combining the prediction results with the equivalent structural stress method, the present invention can not only predict fatigue life, but also quickly identify weak areas in the structure, providing welding engineers with a powerful design optimization tool. Engineers can adjust welding parameters, optimize weld layout or select more suitable materials based on the prediction results, thereby significantly improving the fatigue performance of the structure during the design stage. Promote the development of intelligent welding technology: The implementation of the present invention relies on advanced computing models and data analysis technologies, which makes it possible for the intelligentization and automation of welding processes. By continuously accumulating data and optimizing models, an intelligent welding process recommendation system based on big data and machine learning can be realized in the future to further improve production efficiency and product quality.

[0098] 5. The present invention has the advantages of efficient utilization of resources and enhanced safety and reliability. Accurate fatigue life prediction helps to avoid over-design and reduce unnecessary material use and processing costs. At the same time, by guiding welding engineers to take targeted post-processing measures (such as heat treatment, surface treatment, etc.), the service life of the weld structure can be effectively extended, maintenance costs and replacement frequency can be reduced, and more efficient utilization of resources can be achieved. In fields with high safety requirements such as aerospace, weapons and equipment, medical equipment, and nuclear power, the application of the present invention can significantly improve the accuracy of fatigue strength assessment of the structure, reduce the risk of safety accidents caused by fatigue failure, and enhance the safety and reliability of the overall system.

[0099] 6. The present invention provides unprecedented accuracy and flexibility for fatigue life prediction of weld structures by comprehensively considering multiple factors such as microscopic defects, complex working conditions, structural stress, etc., which not only promotes the development of welding science and technology, but also provides strong support for structural safety design and optimization in engineering practice. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Figure 1 is a flow chart of the present invention;

[0101] Figure 2 This is a prediction chart for the fatigue life of TC4 titanium alloy welds. DETAILED DESCRIPTION

[0102] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0103] The present invention considers the use of plasma welding process key parameters and loading methods to predict the fatigue life of titanium alloy welds. The technical method includes three parts: by controlling the welding process parameters and using the newly developed combined weld defect detection method, the full information of internal defects of the welded components with known process parameters is detected; through correlation analysis and random modeling methods, a data model between process parameters and welded component defect information is established; finally, through fatigue experiments and high-precision simulations, a prediction method for the fatigue life of (specific) TC4 titanium alloys under process parameters and loading forms is established, such as Figure 1 shown.

[0104] The present invention firstly controls the welding process parameters and uses the newly developed combined weld defect detection method to detect the full information of internal defects of welding components with known process parameters. According to different welding process conditions, non-destructive testing technology, namely digital ray and phased array ultrasonic testing, is applied to detect the micro defects of the components, parameterize and describe the micro defects, and establish a probability model of micro defect parameters. Through Pearson correlation analysis, the correlation between micro defect parameters and process parameters is studied, and a random model of defect parameters based on a strong discrete plasma welding process is established.

[0105] During the welding process, due to different choices of welding process parameters, the microstructure characteristics of the welded joint will change significantly, resulting in various welding defects. In order to analyze the influence of different welding process parameters on welding defects, this project intends to discretely quantify the plasma welding (TIG) process and study the influence of different process parameter combinations on weld micro defects.

[0106] M1: Determine and discretize welding process parameters, including welding current I, welding speed v, plasma gas flow Q v These parameters directly affect the heat input, stability and protection effect of the weld, and have a great impact on the generation of weld defects. Finally, due to the difficulty in controlling the plasma gas flow and the mismatch of parameters, plasma welding could not be formed, and the process parameter combination closest to the actual working conditions was selected for welding tests.

[0107] M2: Nondestructive testing, namely phased array ultrasonic testing, is used to measure microscopic defects along the length of the main weld structure. Microscopic defects are determined by the process and do not depend on external load conditions, so external load effects are not considered. For several different process conditions, plasma welding defects are obtained and classified, including side wall lack of fusion (T 1 ), root not fully penetrated (T 2 ), blunt edge not fully penetrated (T 3 ) etc. Considering the high process precision, the order of magnitude difference between defects and component geometric dimensions is large, so the following assumptions are made: ① The thickness of the defect is not considered, only the effective weld width (EW) and effective weld depth (ED) are considered. ② The shape of the defect is not considered, and its area is calculated as the product of width and depth, that is, the effective weld cross-sectional area (EA). ③ The influence of the defect direction on subsequent calculations is ignored.

[0108] Based on the phased array ultrasonic testing results, the descriptive parameters of the weld micro defects are directly or indirectly obtained, namely the effective weld width (EW), effective weld depth (ED) and effective weld cross-sectional area (EA). Under specific working conditions, each type of defect T i The descriptive parameters, namely EW, ED and EA, can be obtained by calculating the test results, namely formulas (1)-(3).

[0109]

[0110] ED j,i =g[z j,i ,T i ] (2)

[0111]

[0112] Where i represents the type number of the defect, T i Indicates the type of weld defect; j indicates the same component T i The jth defect under defect type, and Indicates T i The starting and ending positions of the jth defect under defect type, z j Indicates T i The depth position of the jth defect under defect type. EW j,i represents the effective weld width of the jth defect under weld defect type i, ED j,i represents the effective weld depth of the jth defect under weld defect type i, EA j,i It represents the effective weld cross-sectional area of ​​the jth defect under weld defect type i, f represents the effective weld width describing the jth defect under weld defect type i, which is related to the weld defect type and the starting and ending positions of the jth defect under its defect type, g represents the effective weld depth describing the jth defect under weld defect type i, which is related to the weld defect type and the depth position of the jth defect under its defect type, and h represents the effective weld cross-sectional area describing the jth defect under weld defect type i, which is related to the weld defect type and the starting and ending positions of the jth defect under its defect type and the weld defect type and the depth position of the jth defect under its defect type.

[0113] Finally, the descriptive parameter distribution diagram of weld micro defects corresponding to each working condition was statistically analyzed, but the number of samples needs to be further expanded to determine the distribution characteristics of the descriptive parameters of weld micro defects. and rules, thereby establishing its probability model.

[0114] M3: Based on the quantitative analysis of the first two steps, the process parameters can be obtained, namely welding current I, welding speed v, and plasma gas flow Q v , as well as the descriptive parameters of micro defects, namely the effective weld width (EW), effective weld depth (ED) and effective weld cross-sectional area (EA). Therefore, this section intends to select the relationship between defect parameters and related process parameters under specific defect types through further Pearson correlation analysis, namely formula (4).

[0115]

[0116] Where r represents the Pearson correlation coefficient. The closer |r| is to 1, the higher the correlation is, and the closer |r| is to 0, the lower the correlation is. i Indicates the input process parameter value; Y i Indicates the output of the descriptive parameter value of weld micro defects; Indicates the average value of input process parameters; Represents the average value of descriptive parameters of weld micro defects.

[0117] With highly correlated process parameters as input and the mean of descriptive parameters of each type of weld defect as input, a mathematical model between input and output can be established through the multivariate regression method (Formulas (5)-(7)).

[0118]

[0119]

[0120]

[0121] Where I represents welding current, v represents welding speed, Q V Indicates the plasma gas flow rate, represents the effective width of weld defect type i, represents the effective depth of weld defect type i, represents the effective weld cross-sectional area under weld defect type i, h i The effective weld width under the weld defect type i is related to the welding current, welding speed and plasma gas flow rate. m i The effective weld depth under the weld defect type i is related to the welding current, welding speed and plasma gas flow rate. i The effective weld cross-sectional area representing weld defect type i is related to welding current, welding speed and plasma gas flow rate.

[0122] Based on different working conditions, the mean of the descriptive parameters corresponding to each type of defect can be obtained. Noise is added to the mean so that its distribution is consistent with the statistical data in the previous step. Then, a random model of the micro defect parameters can be obtained for subsequent analysis.

[0123] Secondly, a data model between process parameters and weld component defect information was invented through correlation analysis and random modeling methods. That is, the material property data of the weld was first collected, and then the fatigue performance of the weld under multiple load conditions was analyzed. On this basis, a finite element fatigue model including initial cracks was constructed, and the correctness of the model was verified by comparing numerical simulation with test results. Then, a mathematical model of the relationship between micro-defect descriptive parameters and initial cracks was established, and the mathematical model was combined with the fatigue model, thus providing effective theoretical support for predicting weld fatigue life.

[0124] N1: Obtain the material properties of the weld material and conduct a series of material tests and fatigue tests.

[0125] ① Tensile test: The yield strength (ReH) and elastic modulus (E) of the material can be obtained through the tensile test. In the test, the material sample is subjected to gradually increasing tension until plastic deformation occurs, and the stress (σ) and strain (ε) at this time are recorded. The stress-strain curve can be drawn through the data. The yield strength is usually determined at the proportional limit or 0.2% offset method on the curve, and the elastic modulus is the ratio of stress to strain (E = σ / ε). 1 Triaxial stress test: The Poisson's ratio (ν) state of the material under uniaxial tension or compression can be measured through the triaxial stress test, and the Poisson's ratio can be calculated by measuring the strain in different directions. ③ Crack tip opening displacement (COD) test: The fracture toughness (KIC) of the material can be determined through the crack tip opening displacement test. A crack is prefabricated on the sample, and then a load is applied to measure the relationship between the crack tip opening displacement and the stress intensity factor, thereby calculating the fracture toughness. ④ Crack extension test: Usually a slow loading crack extension test (such as compact tensile test CT test) is used to apply slowly increasing stress to the sample with pre-cracks, measure the relationship between the crack extension rate and the stress intensity factor, and determine the crack extension material constant C and exponent m through data analysis.

[0126] N2: Fatigue mechanical properties test of weld structure is carried out under multiple load conditions. Fatigue mechanical properties refer to the mechanical behavior and characteristics of materials under repeated or cyclic loads. It mainly includes fatigue strength, fatigue life, fatigue crack growth rate, and SN curve. In order to obtain the fatigue mechanical properties of weld structure, fatigue mechanical properties test is required.

[0127] Fatigue tests under various load conditions need to consider the following load conditions: Vertical concentrated load: refers to the load perpendicular to the welded component, such as deadweight, snow load, etc. Horizontal load: refers to the load parallel to the welded component, such as wind load, lateral load, etc. Non-uniform surface pressure: refers to the unevenly distributed load on the surface of the welded component, such as vehicle load, ice pressure, etc. In actual engineering applications, welded components may be subjected to multiple loads at the same time, and these loads may appear in different combinations. Based on the partner's requirements for the load spectrum and the specific load range, it needs to be determined according to the actual working conditions.

[0128] N3: Based on the material properties and geometric properties obtained from the above material tests, a high-precision numerical simulation model is established and a finite element fatigue strength analysis is performed to lay the foundation for subsequent verification of the fatigue strength, life and failure mode of components considering initial cracks. The steps to establish the finite element model are as follows:

[0129] a1 Determine the material and geometric properties of the test piece, including the structural material properties of the weld obtained from the test, the geometric information of the component, and the position, size, and shape of the weld.

[0130] a2 The finite element model was established using ABAQUS software, in which the solid elements of the weld structure were divided into hexahedral grids. To ensure the calculation accuracy, the mesh of the superstructure was refined. When arranging the global network seeds (global seeds), the unit size was set to 10 mm, and when arranging the local network seeds (local seeds) of the superstructure, the unit size was set to 5 mm.

[0131] a3 assigns the corresponding titanium alloy material properties to the geometric model completed by the above mesh division.

[0132] a4 sets the analysis step, select "Static, General", reduce the initial incremental step and minimum incremental step, set the maximum incremental step to 1, and the number of steps to 10000.

[0133] a5 specifies the boundary conditions, completely fixing the base around and at the bottom, and applying loads in two perpendicular directions to the circular structure.

[0134] a6 performs finite element analysis to obtain the stress distribution of the structure and the size and position of the maximum stress point in the weld area (x 0 ,y 0 ).

[0135] N4: Considering that the fatigue failure mode of structural welds containing microscopic defects is usually manifested as fracture failure, in order to simulate the fracture failure mode and study the influence of weld microscopic defects on its fatigue strength and life, it is necessary to make an equivalent initial crack assumption, that is, the position of the equivalent initial crack (x 0 ,y 0), equivalent initial crack depth (a 0 ), equivalent initial crack length (b 0 ) and the equivalent initial crack direction (θ 0 ) is directly related to the microscopic defects and stress state of the weld. The specific operations are as follows:

[0136] b1 Preliminary estimation of equivalent initial crack

[0137] The position of the equivalent initial crack is assumed to be a semi-elliptical crack. The horizontal plane at the component weld is taken as the reference plane. Without considering the weld height, the centroid of the plane component is defined as the reference origin o to establish the xoy coordinate system. Assume that the maximum weld stress calculated by the finite element model is the crack centroid position (x 0 ,y 0 ). Secondly, the direction of the equivalent initial crack, considering the vertical load condition, assumes that θ is 45° downward from the horizontal plane. Finally, the size of the equivalent initial crack is empirically assumed to be its initial depth a 0 is 1 mm, initial length b 0 is 1mm.

[0138] b2 Establishment of the fatigue numerical model of weld structure considering equivalent initial crack

[0139] In FRANC3D software, the initial crack is introduced into the finite element model of the component according to the above-mentioned assumed geometric parameters, and the corresponding crack propagation method and corresponding parameters are selected to predict fatigue crack propagation and fatigue life.

[0140] b3 Equivalent initial crack verification and correction

[0141] The equivalent initial crack is verified by comparing the fatigue mechanical properties calculated by the simulation model with the actual fatigue mechanical properties obtained from the component fatigue test.

[0142] ① Verify the position of the equivalent initial crack. According to the fatigue test results of the component, the surface crack of the component is inferred from the interface of the fracture failure. Assuming that the crack is centered at the centroid point, its depth and length expand in a geometric ratio of β when expanding (Formula 8). Determine the centroid position of the surface crack of the test component (x 1 ,y 1 ), to verify the equivalent initial crack position (x 0 ,y 0 ) for accuracy and make revisions.

[0143]

[0144] Where β is the geometric expansion coefficient, a is the crack depth when the crack expands, and b is the crack length when the crack expands. 0 is the equivalent initial crack depth, b 0is the equivalent initial crack length.

[0145] ② Verify the size of the equivalent initial crack, based on the Paris model assumption, a is the crack depth when the crack propagates, N is the number of stress cycles, ΔK is the range of stress intensity factor variation, through the integral formula, where the lower limit in the integral formula is a 1 , the upper limit is a 2 ; If a 1 Substitute the equivalent initial crack depth a 0 , a 2 Substitute the final crack depth into the initial crack depth a and reverse the solution. 0 ;

[0146]

[0147] Since it is assumed that the crack depth and length expand in a constant ratio β during crack propagation (Formula 8), the final crack depth a is measured by f and crack depth b f , substitute into formula 8 to calculate the equivalent initial crack length b 0 In the Paris model, that is, in Formula 9, C and m are basic parameters describing the fatigue crack growth performance of the material, where m is the crack growth index, a is the crack depth, N is the number of cycles, and ΔK is the stress intensity factor change, which can be obtained by stress measurement in the test.

[0148]

[0149] ΔK=K max -K min (10-b)

[0150]

[0151] The stress intensity factor K is a physical quantity that describes the strength of the stress field at the crack tip. It is related to the crack geometry, load size, and material elastic properties, and is used to quantitatively evaluate the trend or driving force of crack extension.

[0152] According to a large number of simulation experiments, fatigue crack propagation is mainly dominated by type I (opening) cracks in actual engineering. This paper mainly studies type I and type II cracks, and the variation law of their composite stress intensity factors. In the test, the stress intensity is measured by sticking strain gauges on the components, and the stress intensity factor can be calculated by formula (10-a. In formula 10-a: σ is the nominal stress, which refers to the average stress calculated by a simplified mechanical model without considering stress concentration and uneven actual stress distribution. The calculation formula is simply the force divided by the force area, and the stress calculated at the crack position is calculated as if there is no crack; a is the crack geometry; Y is the shape coefficient (related to the crack geometry, position, etc.). For a given welded joint, the Y value is a constant corresponding to it.

[0153] ③Verify the direction of the equivalent initial crack θ 0 According to the direction of crack propagation in the fracture failure mode of the fatigue test results, the direction of the crack at the time of crack initiation is found for verification, and the equivalent initial crack direction angle is corrected by comparing the crack propagation path with the propagation path in the simulation model.

[0154] ④Finally, the crack propagation direction is corrected and the calculation criteria for the kink angle are adjusted. The calculation of the crack propagation direction can be determined by the kink angle, and the expansion calculation is implemented through FRANC3D software. The kink angle is determined by the stress intensity factor. When calculating the kink angle, K Ⅰ K Ⅱ The crack propagation direction can be obtained by calculating according to the selected calculation criteria. Two calculation criteria are given priority: maximum tensile stress and strain energy release rate (SERR) criteria. After selecting the calculation criteria for the kink angle, select the loading cycle mode of the load spectrum, set the stress ratio, and perform crack propagation calculations to obtain the fracture failure mode, fatigue strength, and fatigue life. Compare the crack propagation direction, fracture failure mode, fatigue strength, and fatigue life in the FRANC3D simulation and fatigue test results to verify whether the equivalent initial crack length, depth, and direction are assumed to be accurate. By comparing the crack propagation path, select the calculation criterion scheme that is closest to the test results to complete the correction of the crack propagation direction.

[0155] b4 Finite Element Model Verification

[0156] After the geometric parameters of the equivalent initial crack are corrected, the finite element mesh density, analysis step parameters, and loading method of the boundary conditions are readjusted. The initial crack is reintroduced into the adjusted finite element model in FRANC3D to predict crack propagation and fatigue life. The simulation results are compared with the fatigue test results. The finite element model is verified by fatigue life, fatigue strength, and fatigue failure mode. The finite element model parameters are continuously corrected and the above steps are repeated to ensure the calculation accuracy of the finite element model.

[0157] N5: In order to analyze the influence of weld defects caused by different process parameters on component fatigue life and failure mode under different load conditions, based on the random model of weld micro-defects of strong discrete process, it is necessary to characterize the mathematical relationship between micro-defect descriptive parameters, load conditions and equivalent initial crack geometric parameters and establish a corresponding mathematical model, and then load the equivalent initial crack parameters into the finite element model for crack propagation calculation and fatigue life prediction, so as to complete the prediction method of weld fatigue mechanical properties under process parameters and load conditions. This project intends to adopt the following specific steps:

[0158] ① Previously, all processes and defects of all structural types provided have been collected and statistically processed, the stress state of cracks under different load conditions has been analyzed, and the descriptive parameters of microscopic defects of different test components and the geometric parameters of equivalent initial cracks under different load conditions have been obtained. Combining fracture mechanics and statistical methods, the descriptive parameters of microscopic defects, that is, the effective weld width obtained after processing the test results, have been obtained. Effective weld depth Effective weld area With the determined load case P x ,P y ,P z ,e x ,e y ,e z and the geometric parameters of the equivalent initial crack, i.e., the equivalent initial crack position (x 0 ,y 0 ), equivalent initial crack depth a 0 , equivalent crack initial length b 0 and the equivalent initial crack direction θ 0 The mathematical relationship between the three components is characterized. The micro-defect descriptive parameters and load conditions are used as input parameters to construct the output equivalent initial crack position (x 0 ,y 0 ), equivalent initial crack depth a 0 , equivalent initial crack length b 0 and the equivalent initial crack direction θ 0 The function (public

[0159] Formula 11-15).

[0160]

[0161]

[0162]

[0163]

[0164]

[0165] l represents the coordinate x describing the equivalent initial crack position 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0166] u represents the coordinate y describing the equivalent initial crack position 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0167] p represents the equivalent initial crack depth a 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0168] q represents the equivalent initial crack length b 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0169] s represents the equivalent initial crack direction θ 0 Effective width of weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related;

[0170] ② Using the stereographic fitting technique, a mathematical model was constructed through Matlab, where the x-axis and y-axis represent different load conditions and micro-defects, respectively, and the z-axis represents the geometric parameters of the equivalent initial crack.

[0171] ③ Based on the above functions and the random model of weld micro-defects based on strong discrete process, under the conditions of given process parameters and load conditions, the weld micro-defect parameters can be determined by the process parameters, and the geometric parameters of the equivalent initial crack can be calculated under given fatigue cycle load conditions. The location of the equivalent initial crack (x 0 ,y 0 ), equivalent initial crack depth a 0 , equivalent initial crack length b 0 and the equivalent initial crack direction θ 0 That is, the equivalent initial crack is introduced into the finite element fatigue model to calculate the crack propagation and predict the fatigue life, so as to obtain the fatigue strength, life and failure mode of the component.

[0172] Finally, a method for predicting the fatigue life of (specific) TC4 titanium alloy under process parameters and loading forms was invented through fatigue experiments and high-precision simulations. The finite element fatigue model obtained through the above steps was simulated in large quantities, and its stress analysis was the equivalent structural stress method, from which the equivalent structural stress and life of a large number of loaded components can be obtained. Through a large number of statistical analyses, the main SN curve is drawn with the equivalent structural stress variation range and fatigue life, so as to achieve the fatigue life of the welded structure being evaluated with a main SN curve.

[0173] C1: Obtain the main SN curve of the titanium alloy weld structure.

[0174] ① Based on the above method of obtaining microscopic defect parameters and load conditions to obtain equivalent initial cracks, load them into a large number of fatigue models to calculate structural stress and life. By simulating the stress cycle under actual working conditions, stress analysis is performed on the weld components, and the structural stress at any point on the structure can be calculated under given conditions (Formula 16).

[0175] σ s =σ me +σ b (16)

[0176] Among them, σ s represents the structural stress, σ me represents the membrane stress, σ b Represents bending stress.

[0177]

[0178] Among them, ΔS s represents the equivalent structural stress variation range, m is the crack growth index, t is the plate thickness, and I(r) is the dimensionless function of the load-bending ratio r (18), which can be obtained from the literature.

[0179]

[0180] Where a is the crack depth during crack extension, which is the same as the definition in equation (8); M kn is the stress intensity magnification factor, They are dimensionless functions for determining the range of stress intensity factors when membrane stress and bending stress act alone.

[0181] ② The calculated structural stress is converted into equivalent structural stress by formula 17, and the corresponding fatigue life N under different equivalent structural stresses is calculated by finite element method. Among them, since the dimensionless function I(r) of the load bending ratio is introduced into the equivalent structural stress formula, that is, formula 18, the main SN curve obtained by the equivalent structural stress method takes into account the influence of the initial crack depth and crack extension.

[0182] ③ Plot all points in the coordinate system and draw a main SN curve of the titanium alloy weld structure with a narrow discrete band. This curve can take into account the influence of the initial crack on the fatigue life. The curve can be directly referenced in engineering, and the corresponding fatigue life can be determined by calculating the equivalent structural stress to complete the prediction of the structural fatigue life.

[0183] C2: Based on the main SN curve of the titanium alloy weld structure obtained above, under the specific environmental conditions and stress ratios of the actual project, the stress type and range, including the maximum stress and the minimum stress, that the structure is subjected to under actual working conditions are analyzed. By calculating the stress amplitude (i.e. half of the difference between the maximum stress and the minimum stress), under the known stress amplitude, the corresponding number of cycles is found from the SN curve. The horizontal axis value corresponds to the number of cycles before fatigue failure, and the vertical axis value corresponds to the stress range. The number of cycles is the maximum number of cycles that the material can withstand before fatigue failure at this stress level. In order to compensate for possible calculation errors, fluctuations in material properties, unconsidered load conditions and other uncertainties, a safety factor needs to be introduced into the design. The life prediction value previously read from the SN curve is adjusted using the standard safety factor to ensure that the fatigue life of the structure can still be guaranteed under the most unfavorable conditions, thereby completing the prediction and evaluation of the fatigue life.

[0184] C3: In order to accurately evaluate the life of the weld structure, the Rainflow Counting Method is used to deal with the impact of multi-segment cyclic loads. The Rainflow Counting Method is an effective cyclic load data analysis method that simplifies the complex cyclic load history into multiple simple banner cycles and extracts key statistical parameters such as maximum stress, minimum stress, and stress range. For multi-axis load conditions, the load projections of the x, y, and z axes are firstly calculated by linear superposition to obtain the equivalent structural stress component σ x , σ y , σ z , and then the complex variable frequency load is decomposed into multiple horizontal cyclic loads through the rain flow counting method, and the fatigue life N within the stress range of the x-axis, y-axis, and z-axis directions is calculated respectively. Then, the fatigue damage accumulation in the x, y, and z directions is calculated respectively through the miner criterion, and finally the total accumulated damage of the structure is calculated by summing up, and then fatigue assessment is carried out.

[0185] After applying the rain flow counting method to stress analyze the weld structure, the Palmgren-Miner criterion is used to consider the accumulation of fatigue damage for each different constant amplitude stress range. The damage fraction d is calculated by summing the fatigue damage contribution of each stress range. The fatigue damage contribution of each stress range is calculated by dividing the number of cycles by the number of cycles to failure of the stress range. (Formula 18).

[0186]

[0187] Where D represents the damage fraction, n i Indicates that under constant amplitude cyclic stress σ i The number of cycles under the current or scheduled loading, N i Indicates that under constant amplitude cyclic stress σ iThe fatigue life under the condition of constant amplitude cyclic stress σ is calculated by FRANC3D i The fatigue life can be obtained.

[0188] Add up the damage contributions of all stress ranges. If the total damage fraction D>1, it is considered that fatigue failure has occurred. Otherwise, it is considered that fatigue failure has occurred. This is used to evaluate whether the structure will suffer from fatigue failure.

[0189] In summary, by inputting the strongly discrete process parameters into the random model with micro-defects, the micro-defect parameters are obtained, and the fatigue life is calculated by loading the initial crack into the fatigue model under given working conditions. The fatigue life of the weld structure under the action of the complex load stress spectrum is predicted by combining the equivalent structural stress method mentioned above, and the fatigue performance of the structure is fully and accurately evaluated. It can also provide guidance for welding engineers, help optimize welding design, select appropriate welding processes, and take corresponding post-welding treatment measures to improve the fatigue performance of the weld structure and extend its service life.

[0190] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A titanium alloy weld fatigue life prediction method considering key parameters and loading methods of plasma welding process, characterized in that: It includes three parts: first, by controlling the welding process parameters and using a combined weld defect detection method to detect the full information of internal defects of welded components with known process parameters; then, through correlation analysis and stochastic modeling methods, a data model between process parameters and welding component defect information is established; finally, through fatigue experiments and high-precision simulations, a prediction method for the fatigue life of titanium alloys considering process parameters and loading forms is established.

2. The titanium alloy weld fatigue life prediction method considering key parameters and loading methods of plasma welding process according to claim 1 is characterized in that: By controlling the welding process parameters and using the newly developed combined weld defect detection method, the full information of internal defects of welded components with known process parameters can be detected, including the following: M1: Determine and discretize welding process parameters, including welding current, welding speed, and plasma gas flow rate, and select the process parameter combination closest to the actual working conditions for plasma welding test; M2: Use phased array ultrasonic testing method to measure the microscopic defects of the main weld structure along its length; obtain and classify the welding defects of plasma welding for several different process conditions; based on the phased array ultrasonic testing results, directly or indirectly obtain the descriptive parameters of the weld microscopic defects, namely, the effective weld width, effective weld depth and effective weld cross-sectional area, etc.; ED j,i =g[z j,i ,T i ] (2) Where i represents the type number of the defect, T i Indicates the type of weld defect; j indicates the same component T i The jth defect under defect type, and Indicates T i The starting and ending positions of the jth defect under defect type, z j,i Indicates T i Depth position of the jth defect under defect type, EW j,i represents the effective weld width of the jth defect under weld defect type i, ED j,i represents the effective weld depth of the jth defect under weld defect type i, EA j,i represents the effective weld cross-sectional area of ​​the jth defect under weld defect type i, f represents the effective weld width describing the jth defect under weld defect type i, which is related to the weld defect type and the starting and ending positions of the jth defect under its defect type, g represents the effective weld depth describing the jth defect under weld defect type i, which is related to the weld defect type and the depth position of the jth defect under its defect type, and h represents the effective weld cross-sectional area describing the jth defect under weld defect type i, which is related to the weld defect type and the starting and ending positions of the jth defect under its defect type, and the weld defect type and the depth position of the jth defect under its defect type; M3: Based on the quantitative analysis of the first two steps, the process parameters, namely welding current, welding speed and plasma gas flow, as well as the descriptive parameters of weld micro defects can be obtained. The relationship between defect parameters and related process parameters can be expressed by the following formula: Among them, r represents the Pearson correlation coefficient. The closer |r| is to 1, the higher the correlation is, and the closer |r| is to 0, the lower the correlation is. i Indicates the input process parameter value; Y i Indicates the output of the descriptive parameter value of weld micro defects; Indicates the average value of input process parameters; It indicates the average value of descriptive parameters of weld micro defects; Taking the highly correlated process parameters as input and the mean of the descriptive parameters of each type of weld defect as input, a mathematical model between input and output is established through the multivariate regression method. Where I represents welding current, v represents welding speed, Q V Indicates the plasma gas flow rate, represents the effective width of weld defect type i, represents the effective depth of weld defect type i, represents the effective weld cross-sectional area under weld defect type i, h i The effective weld width under the weld defect type i is related to the welding current, welding speed and plasma gas flow rate. m i The effective weld depth under the weld defect type i is related to the welding current, welding speed and plasma gas flow rate. i The representative description of the effective weld cross-sectional area under weld defect type i is related to welding current, welding speed and plasma gas flow rate; Based on different working conditions, the mean of the descriptive parameters corresponding to each type of defect can be obtained. Noise is added to the mean so that its distribution is consistent with the statistical data in the previous step. Then, a random model of the micro defect parameters can be obtained for subsequent analysis.

3. The titanium alloy weld fatigue life prediction method considering key parameters and loading methods of plasma welding process according to claim 2 is characterized in that: Through correlation analysis and stochastic modeling methods, a data model between process parameters and welding component defect information is established, including the following contents: N1: Obtain various material properties of the weld material and conduct material tests and fatigue tests; the tests conducted include tensile tests, triaxial stress tests, crack tip opening displacement tests, and crack extension tests; N2: Conduct fatigue mechanical properties test of weld structure under multiple load conditions; N3: Based on the material properties and geometric properties obtained from the above tests, a high-precision numerical simulation model is established and finite element fatigue strength analysis is performed; N4: Considering that the fatigue failure mode of structural welds containing micro defects is usually manifested as fracture failure, in order to simulate the fracture failure mode and study the influence of weld micro defects on its fatigue strength and life, it is necessary to make an equivalent initial crack assumption; N5: In order to analyze the influence of weld defects caused by different process parameters on component fatigue life and failure mode under different load conditions, it is necessary to characterize the mathematical relationship between micro-defect descriptive parameters, load conditions and equivalent initial crack geometric parameters and establish a corresponding mathematical model based on the random model of weld micro-defects of strong discrete process, and then load the equivalent initial crack parameters into the finite element model for crack propagation calculation and fatigue life prediction, so as to complete the prediction method of weld fatigue mechanical properties under process parameters and load conditions; Use the following specific steps: ① Combining fracture mechanics and statistical methods, the micro-defect descriptive parameters obtained after processing the test results, the determined load conditions and the geometric parameters of the equivalent initial crack are characterized, and the mathematical relationship between the three is constructed; Where x0, y0 are the coordinates of the equivalent initial crack position, a0 is the equivalent initial crack depth, b0 is the equivalent initial crack length, and θ0 is the equivalent initial crack direction; l represents the coordinate x0 describing the equivalent initial crack position and the effective width of the weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related; u represents the coordinate y0 describing the equivalent initial crack position and the effective width of the weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related; p represents the effective width of the weld defect type i and the equivalent initial crack depth a0. Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related; q represents the effective width of the weld defect type i and the equivalent initial crack length b0. Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related; s represents the effective width describing the equivalent initial crack direction θ0 and weld defect type i Effective Depth Effective weld cross-sectional area and the horizontal load P x , longitudinal load P y , vertical load P z , horizontal eccentricity e x , longitudinal eccentricity e y , vertical eccentricity e z Related; ② Using the stereographic fitting technique, a mathematical model was constructed through Matlab, where the x-axis and y-axis represent different load conditions and micro-defects, respectively, and the z-axis represents the geometric parameters of the equivalent initial crack; ③ Based on the above functions and the random model of weld microdefects based on strong discrete process, under the conditions of given process parameters and load conditions, the weld microdefect parameters can be determined by the process parameters, and the geometric parameters of the equivalent initial crack can be calculated under given fatigue cycle load conditions, that is, the position of the equivalent initial crack, the equivalent initial crack depth, the equivalent initial crack length and the equivalent initial crack direction. The equivalent initial crack is introduced into the finite element fatigue model for crack propagation calculation and fatigue life prediction, thereby obtaining the fatigue strength, life and failure mode of the component.

4. The titanium alloy weld fatigue life prediction method considering key parameters and loading methods of plasma welding process according to claim 3 is characterized in that: In step N3, the steps for establishing the finite element model are as follows: a1 Determine the material and geometric properties of the test piece, including the material properties of the weld structure and the geometric information of the component obtained from the test, as well as the position, size and shape of the weld; a2 The finite element model was established by ABAQUS software, in which the solid elements of the weld structure were divided into hexahedral grids. To ensure the calculation accuracy, the upper structure was meshed and refined. When arranging the global network seeds, the unit size was set to 10 mm, and when arranging the local network seeds of the upper structure, the unit size was set to 5 mm. a3 assigns the corresponding titanium alloy material properties to the geometric model after the mesh division; a4 Set the analysis step, select "Static, General", reduce the initial incremental step and the minimum incremental step, set the maximum incremental step to 1, and the number of steps to 10000; a5 specifies boundary conditions, completely fixing the base around and at the bottom, and applying loads in two perpendicular directions to the circular structure; a6 performs finite element analysis to obtain the stress distribution of the structure and the size and position of the maximum stress point in the weld area.

5. The titanium alloy weld fatigue life prediction method considering key parameters and loading methods of plasma welding process according to claim 3 is characterized in that: The position, depth, length and direction of the equivalent initial crack are directly related to the weld microscopic defects and stress state. In step N4, the specific operation of the equivalent initial crack assumption is as follows: b1 Preliminary estimation of equivalent initial crack The position of the equivalent initial crack is assumed to be a semi-ellipse. The horizontal plane at the weld of the component is taken as the reference plane. Without considering the height of the weld, the centroid of the plane component is defined as the reference origin o to establish the xoy coordinate system. The maximum stress of the weld calculated by the finite element model is assumed to be the crack centroid position (x0, y0). Secondly, the direction of the equivalent initial crack is considered to be a vertical load condition, and θ is assumed to be 45° downward with the horizontal plane. Finally, the size of the equivalent initial crack is assumed to be 1 mm in depth a0 and 1 mm in length b0. b2 Establishment of the fatigue numerical model of weld structure considering equivalent initial crack In FRANC3D software, the initial crack is introduced into the finite element model of the component according to the above-mentioned assumed geometric parameters, and the corresponding crack propagation method and corresponding parameters are selected to predict fatigue crack propagation and fatigue life; b3 Equivalent initial crack verification and correction b4 Finite Element Model Verification After the geometric parameters of the equivalent initial crack are corrected, the finite element mesh density, analysis step parameters, and loading method of the boundary conditions are readjusted. The initial crack is reintroduced into the adjusted finite element model in FRANC3D to predict crack propagation and fatigue life. The simulation results are compared with the fatigue test results. The finite element model is verified by fatigue life, fatigue strength, and fatigue failure mode. The finite element model parameters are continuously corrected and the above steps are repeated to ensure the calculation accuracy of the finite element model.

6. The titanium alloy weld fatigue life prediction method considering key parameters and loading methods of plasma welding process according to claim 5 is characterized in that: In step b3, the fatigue mechanical properties calculated by the simulation model are compared with the actual fatigue mechanical properties obtained by the component fatigue test to verify and correct the equivalent initial crack. The verification and correction methods are as follows: ①Verify the location of the equivalent initial crack According to the fatigue test results of the component, the crack on the surface of the component is inferred from the interface of fracture failure. Assuming that the crack is centered at the centroid point, its depth and length expand geometrically during expansion, and the centroid position (x1, y1) of the surface crack of the test component is determined to test the accuracy of the equivalent initial crack position (x0, y0) and make corrections; the geometric expansion formula is: Where, β is the geometric expansion coefficient, a is the crack depth when the crack expands, b is the crack length when the crack expands, a0 is the equivalent initial crack depth, and b0 is the equivalent initial crack length; ②Verify the size of the equivalent initial crack Based on the Paris model assumption, a is the crack depth during crack propagation, N is the number of stress cycles, C and m are material-related parameters, C and m are basic parameters describing the fatigue crack propagation performance of materials, where m is specifically the crack propagation index, and the ΔK stress intensity factor variation range is calculated through the integral formula, where the lower limit in the integral formula is a1 and the upper limit is a2; If a1 is substituted into the formula as the equivalent initial crack depth a0, and a2 is substituted into the formula as the final crack depth, the initial crack depth a0 can be reversed; Since it is assumed that the crack depth and length expand in a constant ratio β during crack extension, the final crack depth a is measured by f and crack depth b f , substitute into formula 8 to calculate the equivalent initial crack length b0; The following is the Paris model, ΔK=K max -K min (15-b) Where a is the crack depth when the crack propagates, N is the number of stress cycles, ΔK is the change in stress intensity factor, K is the stress intensity factor, σ is the nominal stress, and Y is the shape factor; ③Verify the equivalent initial crack direction According to the crack propagation direction in the fracture failure mode of the fatigue test results, the direction of the crack at the time of crack initiation is found for verification, and the equivalent initial crack direction angle is corrected by comparing the crack propagation path with the propagation path in the simulation model; ④Finally, correct the crack propagation direction and adjust the calculation criteria of the kink angle The calculation of crack propagation direction can be determined by the kink angle, and the expansion calculation is implemented through FRANC3D software; the kink angle is determined by the stress intensity factor. After the calculation criterion of the kink angle is selected, the loading cycle mode of the load spectrum is selected, the stress ratio is set, and the crack propagation calculation is performed to obtain the fracture failure mode, fatigue strength and fatigue life; the crack propagation direction, fracture failure mode, fatigue strength and fatigue life in the FRANC3D simulation and fatigue test results are compared to verify whether the equivalent initial crack length, depth and direction are assumed to be accurate. By comparing the crack propagation path, the calculation criterion scheme closest to the test results is selected to complete the correction of the crack propagation direction.

7. The titanium alloy weld fatigue life prediction method considering key parameters and loading methods of plasma welding process according to claim 6 is characterized in that: Through fatigue experiments and high-precision simulations, the steps for predicting the fatigue life of TC4 titanium alloy considering process parameters and loading forms are as follows: C1: Obtain the main SN curve of titanium alloy weld structure ① Based on the above method of obtaining microscopic defect parameters and load conditions to obtain equivalent initial cracks, load them into a large number of fatigue models to calculate structural stress and life; by simulating the stress cycle under actual working conditions, stress analysis is performed on the weld components, and the structural stress at any point on the structure can be calculated under given conditions; s s =s me +s b (16) Among them, σ s represents the structural stress, σ me represents the membrane stress, σ b represents bending stress; Among them, ΔS s represents the range of equivalent structural stress variation, m is the crack extension index, t is the plate thickness, and I(r) is a dimensionless function of the load-bending ratio r Where a is the crack depth during crack expansion, M kn is the stress intensity magnification factor, are dimensionless functions for determining the range of stress intensity factors when membrane stress and bending stress act alone, is the differential of a / t, M kn is the stress intensity amplification factor; ② The calculated structural stress is converted into equivalent structural stress by formula 16, and the fatigue life N corresponding to different equivalent structural stresses is calculated by finite element method. Since the dimensionless function I(r) of load-bending ratio is introduced into the equivalent structural stress formula, the main SN curve obtained by the equivalent structural stress method takes into account the influence of the initial crack depth and crack extension; ③ Plot all points in the coordinate system to draw a main SN curve of a titanium alloy weld structure with a narrow discrete band. This curve can take into account the influence of initial cracks on fatigue life. In engineering, the curve can be directly referenced to determine the corresponding fatigue life by calculating the equivalent structural stress, thus completing the prediction of the structural fatigue life; C2: Based on the main SN curve of the titanium alloy weld structure obtained above, under the specific environmental conditions and stress ratio of the actual project, analyze the stress type and range that the structure is subjected to under actual working conditions, including the maximum stress and the minimum stress; by calculating the stress amplitude (that is, half of the difference between the maximum stress and the minimum stress), under the known stress amplitude, find the corresponding number of cycles from the SN curve, the horizontal axis value corresponds to the number of cycles before fatigue failure, the vertical axis value corresponds to the stress range, and the number of cycles is the maximum number of cycles that the material can withstand before fatigue failure at this stress level; in order to compensate for possible calculation errors, material performance fluctuations, unconsidered load conditions and other uncertain factors, it is necessary to introduce a safety factor in the design; use the standard safety factor to adjust the life prediction value previously read from the SN curve to ensure that the fatigue life of the structure can still be guaranteed under the most unfavorable conditions, thereby completing the prediction and evaluation of the fatigue life; C3: In order to accurately evaluate the life of the weld structure, the Rainflow Counting Method is used to deal with the influence of multi-segment cyclic loads; for multi-axis load conditions, the load projections of the x, y, and z axes are firstly calculated by linear superposition to obtain the equivalent structural stress component σ x , σ y , σ z , and then decompose the complex variable frequency load into multiple horizontal cyclic loads through the rain flow counting method, calculate the fatigue life N within the stress range of the x-axis, y-axis, and z-axis directions respectively, and then calculate the fatigue damage accumulation in the x, y, and z directions respectively through the Miner criterion, and finally sum up and calculate the total accumulated damage of the structure, and then perform fatigue assessment; After applying the rain flow counting method to stress analyze the weld structure, the Palmgren-Miner criterion is used to consider the accumulation of fatigue damage for each different constant amplitude stress range; the damage fraction d is calculated by summing the fatigue damage contribution of each stress range; the fatigue damage contribution of each stress range is calculated by dividing the number of cycles by the number of failure cycles of the stress range; Where D represents the damage fraction, n i Indicates that under constant amplitude cyclic stress σ i The number of cycles under the current or scheduled loading, N i Indicates that under constant amplitude cyclic stress σ i The fatigue life under the condition of constant amplitude cyclic stress σ is calculated by FRANC3D i The fatigue life under the condition can be obtained; Add up the damage contributions of all stress ranges. If the total damage fraction D>1, it is considered that fatigue failure has occurred. Otherwise, it is considered that fatigue failure has occurred. This is used to evaluate whether the structure will suffer from fatigue failure.

Images