A method for evaluating the fatigue life of the weld seam of the hydrogen turbine intake casing of an engine

Abstract

The present invention discloses a method for evaluating the fatigue life of the weld of the hydrogen turbine intake casing of an engine. First, a finite element model of the turbine intake casing weld is constructed by obtaining the three-dimensional characteristics of the weld area and combining the material parameters. On this basis, the transfer function and stress power spectrum density function of the structure are calculated through modal response analysis and random vibration analysis, and a random vibration fatigue life prediction model based on the equivalent SN curve criterion is obtained. The S-N curve containing average stress obtained in the experiment is equivalent to the S-N curve under zero average stress, which solves the problem of interference of average stress on random vibration fatigue life analysis and significantly reduces the experimental cost and time. Combined with a variety of stress range probability empirical formulas, a multi-dimensional and accurate prediction of fatigue life is achieved.

Description

Technical Field

[0001] The present invention relates to structural fatigue life assessment, and in particular to a method for assessing fatigue life of a weld of an engine hydrogen turbine intake casing. Background Art

[0002] With the advancement of aero-engine technology, turbine engines are increasingly used in the aviation field. Their operating environments are extremely harsh, subject to challenges such as high temperatures, high pressures, and complex vibration loads. The weld area of ​​the turbine inlet casing, in particular, is subject to long-term coupled thermal and mechanical stresses, making it a vulnerable link in structural fatigue failure. Therefore, assessing the fatigue life of this weld area is of great engineering significance.

[0003] Unlike the steel or aluminum alloys used in traditional industrial welds, aircraft engine weld areas often utilize specialized alloy materials with unique mechanical and thermal properties designed to meet the demands of extreme operating conditions. For example, nickel-based superalloys and titanium alloys are widely used in engine weld areas to ensure their strength and oxidation resistance in high-temperature environments. However, due to the nonlinear behavior and complex fatigue properties of these materials, traditional fatigue life assessment methods based on standard SN curves cannot be directly applied. Summary of the Invention

[0004] Purpose of the invention: In view of the above shortcomings, the present invention provides a method for evaluating the fatigue life of the weld of the engine hydrogen turbine intake casing made of special materials.

[0005] Technical solution: To solve the above problems, the present invention adopts a method for evaluating the fatigue life of the weld of the engine hydrogen turbine intake casing, which includes the following steps:

[0006] Step 1: Obtain the structural characteristics and material parameters of the weld zone of the standard specimen of the engine hydrogen turbine intake casing weld, establish a finite element model of the standard specimen of the engine hydrogen turbine intake casing weld based on the obtained structural characteristics and material parameters, and perform fatigue testing on the standard specimen of the engine hydrogen turbine intake casing weld to obtain the SN curve;

[0007] Step 2: Based on the finite element model, perform frequency response analysis on the engine hydrogen turbine intake housing structure and construct the transfer function of the engine hydrogen turbine intake housing structure;

[0008] Step 3: Perform zero mean stress correction on the SN curve obtained from the fatigue test;

[0009] Step 4: Calculate the stress power spectrum density of the engine's hydrogen turbine intake casing structure through the transfer function to determine the frequency range of random vibration. Based on the frequency range, determine the empirical formula for probabilistic prediction of the stress range. Then, combine the modified SN curve with the linear cumulative damage theory to establish a random vibration fatigue life prediction model.

[0010] Step 5: Based on the random vibration fatigue life prediction model, the life of the engine hydrogen turbine intake casing weld is predicted.

[0011] Furthermore, the specific steps of obtaining the structural characteristics and material parameters of the weld area of ​​the engine hydrogen turbine intake casing weld standard sample are: grinding and polishing, slicing, CT scanning and mechanical testing the engine hydrogen turbine intake casing weld standard sample to obtain the structural characteristics and material parameters of the weld, including three-dimensional dimensions, mechanical characteristics, etc., and establishing a finite element model of the engine hydrogen turbine intake casing weld in finite element modeling software.

[0012] Furthermore, the frequency response analysis of the engine hydrogen turbine intake housing structure is performed to calculate the displacement, velocity, acceleration or stress responses of the structure under different frequency excitations, determine the natural modes and modal vibration shapes of the structure, and obtain the transfer function of the structure, that is, the input-output relationship. The expression of the transfer function is usually as follows:

[0013] ;

[0014] in, is the transfer function, which represents the input With output The relationship between them. is the frequency domain representation of the input signal, such as applied force, acceleration, or displacement, is the frequency domain response of the output signal, such as the displacement, velocity, or stress response of a structure, where f is the frequency. By comparing the input excitation signal and the corresponding response signal at different frequencies before and after frequency response analysis, a transfer function can be constructed.

[0015] Furthermore, the SN curve obtained from the fatigue test is corrected for zero mean stress, and the fatigue test SN curve with non-zero mean stress is equivalent to the SN curve with zero mean stress. The zero mean stress correction methods include the Goodman method and the Soderberg method, which are as follows:

[0016] ;

[0017] ;

[0018] in, represents the equivalent zero mean stress; represents the stress amplitude of the i-th point of the fatigue test SN curve; is the stress mean value of the i-th point of the fatigue test SN curve, represents the yield strength, Indicates tensile strength.

[0019] The Goodman method appears as a straight line on a stress-strain diagram, with the ultimate strength and fatigue limit representing the two extreme points on the coordinate axis. As the mean stress increases, the amplitude of the alternating stresses that the material can withstand gradually decreases until the material yields or fails. The Soderberg criterion appears as a straight line on a stress-strain diagram, but its endpoint is at the yield strength rather than the ultimate strength.

[0020] Furthermore, the resulting stress power spectrum density is solved. The input of random vibration is usually expressed as the acceleration power spectrum density , defines the energy distribution of the excitation acceleration in the frequency domain, f is the frequency, It represents the power of the acceleration at a certain frequency. Then the transfer function The power spectrum density of stress at frequency f can be solved as ;

[0021] ;

[0022] This formula shows that the power spectrum density of stress is proportional to the power spectrum density of acceleration and the square of the transfer function. At each frequency point, the transfer function The input energy of acceleration excitation is converted into output energy of stress response.

[0023] Furthermore, the Dirlik method, Steinberg method and Bendat method are used to predict the probability of stress range.

[0024] The Dirlik stress range probability formula is used to estimate fatigue damage caused by random vibration from the stress power spectrum density (PSD). This formula is based on power spectrum density data and uses frequency domain methods to calculate the probability distribution of the stress range. The specific formula is as follows:

[0025] ;

[0026] Among them, the probability density function It represents the probability of stress range S appearing in a certain range. Parameters M, Z, Q, R, D1, D2 and D3 are coefficients related to the power spectrum characteristics and can be calculated from the stress power spectrum density. The details are as follows:

[0027] ; ; ; ; ; .

[0028] Gauge stress power spectrum moment: ,but:

[0029] ; .

[0030] For the Steinberg method, it is usually assumed that the peak distribution of stress follows a normal distribution model, so this method directly gives the distribution probability of the stress range. Specifically, the probability of stress occurrence at the 1σ level is 68.3%, the probability of stress occurrence at the 2σ level is 27.1%, and the probability of stress occurrence at the 3σ level is 4.43%. Where σ is the root mean square stress, which can be combined with the stress power spectrum density function Solution:

[0031] ;

[0032] The Bendat method assumes that the stress range S follows a Rayleigh distribution, and its probability density function is:

[0033] ;

[0034] Where σ is the root mean square stress, and the solution is the same as above. All three methods are suitable for predicting stress ranges in random vibration environments, especially for engineering applications with complex random loads, such as aviation components and mechanical equipment.

[0035] Furthermore, it is assumed that the damage to fatigue life of each load cycle is linearly accumulated. When the accumulated damage reaches 1, the structure fails due to fatigue. The expression is:

[0036] ;

[0037] in, For cumulative damage, is the number of cycles at a certain stress level, is the maximum number of cycles obtained from the SN curve at this stress level. When the structure reaches its fatigue life, failure occurs.

[0038] The SN curve can usually be described by the following power relationship:

[0039] ;

[0040] Here, S is the stress amplitude, or stress range; N is the number of cycles; m is the slope of the material's SN curve; and C is the fatigue strength coefficient. The SN curve (stress-life curve) is a fundamental representation of a material's fatigue performance. It describes the number of fatigue cycles (N) a material can withstand at different stress amplitudes (S).

[0041] Furthermore, the Miner linear cumulative damage criterion and the SN curve of equivalent zero mean stress are combined to solve the weld fatigue life under the three stress range prediction probability methods of Dirlik, Steinberg and Bendat respectively.

[0042] For the Dirlik method, the number of cycles per unit time for a certain stress range S is n(S):

[0043] ;

[0044] Where p(S) is the probability density of the Dirlik stress range, and E(peak) is the peak factor, which can be calculated by the following formula:

[0045] ;

[0046] Combining Miner's linear cumulative damage criterion with the SN curve, the fatigue damage per unit time can be obtained as D Dirlik :

[0047] ;

[0048] Let D Dirlik = 1, and the fatigue life T Dirlik for:

[0049] ;

[0050] The Steinberg method directly gives the distribution probability of the stress range, and the number of cycles per unit time is:

[0051] ; ; ;

[0052] Where v0 is the zero crossing rate, which is expressed as follows:

[0053] ;

[0054] Combining Miner's linear cumulative damage criterion with the SN curve, the fatigue damage per unit time D Steinberg for:

[0055] ;

[0056] Let DSteinberg = 1, and the fatigue life T Steinberg for:

[0057] ;

[0058] For the Bendat method, both the Dirlik method and the Bendat method give the probability density of the stress range. However, due to the different frequency bands analyzed, the final peak number E(peak) is replaced by the zero crossing rate v0, so the frequency life expression is the same as T Dirlik Consistent with:

[0059] ;

[0060] The present invention also adopts a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0061] The present invention also adopts a computer-readable storage medium having a computer program stored thereon, and the computer program implements the steps of the above method when executed by a processor.

[0062] Beneficial effects: Compared with the existing technology, the present invention has significant advantages, especially for special fatigue test weld materials. By obtaining structural characteristics, it can achieve accurate evaluation of weld fatigue life under random vibration conditions. The model fully considers the influence of factors such as weld size, morphology, and position on random vibration fatigue behavior. Based on the three-dimensional finite element model of the engine hydrogen turbine inlet shell weld, the frequency domain response characteristics of the structure under random vibration are calculated. At the same time, the present invention uses multiple methods to perform zero-mean stress equivalent processing on the SN curve. In fatigue analysis, the actual working conditions (including alternating stress and average stress) are converted into equivalent full alternating stress working conditions, which simplifies the life prediction process under complex loads, reduces experimental costs, and provides conservative life assessment for engineering design. The present invention also combines multiple stress range probability distribution prediction methods to make fatigue life prediction more accurate and comprehensive. The method is highly innovative and has good robustness. It is particularly suitable for fatigue life assessment of special weld materials such as aircraft engines and has broad prospects for promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic diagram of the evaluation method of the present invention.

[0064] Figure 2 This is a schematic diagram of the finite element model of the weld area of ​​the engine hydrogen turbine intake casing in the present invention.

[0065] Figure 3This is a schematic diagram of the SN experimental curve of the engine hydrogen turbine intake casing weld sample made of the GH4169 nickel-based high-temperature alloy material in the present invention.

[0066] Figure 4 This is a schematic diagram of the SN equivalent zero-average stress curve of the engine hydrogen turbine intake casing weld sample made of the GH4169 nickel-based high-temperature alloy material in the present invention.

[0067] Figure 5 This is a schematic diagram of the fatigue life of random vibration welds predicted in the present invention. DETAILED DESCRIPTION

[0068] like Figure 1 As shown, in this embodiment, a method for evaluating the fatigue life of a weld of an engine hydrogen turbine intake casing includes the following steps:

[0069] Step 1: Grind and polish the standard sample of the engine hydrogen turbine intake housing weld, slice and sample, CT scan and mechanical test to obtain the structural characteristics and material parameters of the weld, including three-dimensional dimensions, mechanical characteristics, etc., and establish a finite element model of the engine hydrogen turbine intake housing weld in the finite element modeling software. Figure 2 As shown. Fatigue tests were performed on the weld specimens obtained by section sampling, and their SN curves were obtained and shown in Table 1. Since the stress ratio of the weld specimens was 0.1 and the mean stress was not zero, the SN curve correction method with equivalent zero mean stress was subsequently used to eliminate the influence of mean stress on the fatigue behavior of the structure.

[0070] Table 1 SN curve data of the engine hydrogen turbine intake housing weld made of GH4169 nickel-based superalloy

[0071]

[0072] Step 2: Perform frequency response analysis on the engine hydrogen turbine intake casing structure and calculate the transfer function of the overall structure based on the input excitation and output response.

[0073] Step 3: Use the Goodman method and Soderberg method, two equivalent zero-mean stress methods, to correct the experimental SN curve. The corrected data are shown in Table 2.

[0074] Table 2 SN curve data of equivalent zero mean stress

[0075]

[0076] Step 4: Calculate the stress power spectrum density of the structure through the transfer function to determine the vibration frequency range of random vibration. Determine the empirical formula for stress range probability prediction based on the vibration frequency range. Based on the determined empirical formula, combine the modified SN curve parameters with the linear cumulative damage theory to establish a random vibration fatigue life prediction model.

[0077] Step 5: Based on the random vibration fatigue life prediction model, a high-precision evaluation of the random vibration fatigue life of the engine hydrogen turbine intake casing weld is achieved.

[0078] The specific implementation steps are as follows: First, the geometric structural characteristics of the weld area of ​​the engine's hydrogen turbine inlet casing are acquired, including information such as weld size, morphology, and location. These structural characteristics directly determine the stress distribution and fatigue life of the weld under fatigue loading. In addition, material parameters (such as elastic modulus, Poisson's ratio, and yield strength) are crucial to the fatigue performance of the weld. Based on these structural and material parameters, a three-dimensional finite element model of the turbine inlet casing weld is constructed to simulate its mechanical response under actual operating conditions. Figure 2 A finite element mesh model of the weld seam of an engine's hydrogen turbine inlet casing is presented. A hexahedral mesh is used for the weld seam, while a tetrahedral mesh is used for the remaining areas. This finite element model not only accurately reflects the local stress concentration in the weld seam, but also provides stress and strain data for subsequent analysis.

[0079] Next, a frequency response analysis is performed on the constructed 3D integral finite element model. The purpose of frequency response analysis is to determine the dynamic characteristics of the weld structure under random vibration, specifically the transfer function of the overall structure. The transfer function describes the relationship between the system's input excitation (such as random vibration load) and its output response (such as stress or displacement) and is a key step in random vibration fatigue analysis. This step allows the determination of the weld's vibration response amplitude at different frequencies, providing the necessary foundation for subsequent stress power spectral density calculations.

[0080] Subsequently, due to the particularity of the weld material, the standard SN curve cannot be directly applied. The SN curve of the weld material must be obtained through fatigue testing, such as Figure 3 As shown in , if the stress ratio of the experimental conditions is not -1, the equivalent zero mean stress method must be used to correct the experimental SN curve. By using methods such as the Goodman method or the Soderberg method, the fatigue data containing mean stress is equivalent to the data under zero mean stress. Figure 4 The following figure shows the modified SN curves using the Goodman and Soderberg methods. This step not only considers the stress changes in the weld under complex loading conditions, but also takes into account the fatigue properties of the actual material, ensuring the accuracy of the fatigue life prediction model.

[0081] Next, the transfer function is used to calculate the stress power spectral density (PSD) of the weld structure. The stress power spectral density describes the stress energy distribution within different frequency ranges under random vibration excitation. Combining various prediction formulas for stress range probability distribution (such as the Dirlik method, the Steinberg method, and the Bendat method) allows for further assessment of weld fatigue damage. These methods are each applicable to different vibration frequency bandwidths. The Dirlik method is suitable for describing broadband random vibration processes, while the Steinberg and Bendat methods are effective in describing the stress range distribution of welds under narrowband random vibration. Considering all three methods simultaneously allows for a more comprehensive assessment of fatigue life. By combining these stress probability distribution formulas with the modified SN curve and Miner's linear cumulative damage theory, a random vibration fatigue life prediction model for welds is established. This model accurately estimates the cumulative fatigue damage of welds under different loading frequencies and stress amplitudes.

[0082] Finally, based on the established fatigue life prediction model, the fatigue life of the engine hydrogen turbine intake housing weld is predicted. Figure 5 The figure shows the predicted weld fatigue life. During this process, the model comprehensively considers the complexity of random vibration, the structural characteristics and material properties of the weld, and various fatigue damage accumulation mechanisms. This process provides an estimate of the fatigue life of the weld under random vibration conditions, providing a reliable basis for the design, maintenance, and life assessment of key engine components. This model is not only applicable to the fatigue life assessment of the hydrogen turbine inlet casing weld in the current study, but can also be extended to fatigue life analysis of other similar structures and special weld materials.

Claims

1. A method for evaluating the fatigue life of the weld of the hydrogen turbine intake casing of an engine, characterized in that: The following steps are involved: Step 1: obtaining structural characteristics and material parameters of the weld zone of the standard sample of the engine hydrogen turbine air intake casing weld, establishing a finite element model of the standard sample of the engine hydrogen turbine air intake casing weld based on the obtained structural characteristics and material parameters, and performing a fatigue test on the standard sample of the engine hydrogen turbine air intake casing weld to obtain an SN curve; the specific steps of obtaining the structural characteristics and material parameters of the weld zone of the standard sample of the engine hydrogen turbine air intake casing weld are: grinding and polishing, slicing, CT scanning and mechanical testing the standard sample of the engine hydrogen turbine air intake casing weld, wherein the structural characteristics and material parameters include three-dimensional dimensions and mechanical characteristics; Step 2: Based on the finite element model, a frequency response analysis is performed on the engine hydrogen turbine intake casing structure to construct a transfer function of the engine hydrogen turbine intake casing structure; the frequency response analysis includes calculating the displacement response, velocity response, acceleration response or stress response of the engine hydrogen turbine intake casing structure under different frequency excitations, and obtaining the transfer function of the engine hydrogen turbine intake casing structure by comparing the input excitation signals and corresponding response signals at different frequencies before and after the frequency response analysis; Step 3: Perform zero mean stress correction on the SN curve obtained from the fatigue test; Step 4: Calculate the stress power spectrum density of the engine's hydrogen turbine intake casing structure through the transfer function to determine the frequency range of random vibration. Based on the frequency range, determine the empirical formula for probabilistic prediction of the stress range. Then, combine the modified SN curve with the linear cumulative damage theory to establish a random vibration fatigue life prediction model. Step 5: Based on the random vibration fatigue life prediction model, the life of the engine hydrogen turbine intake casing weld is predicted.

2. The method for evaluating the fatigue life of the weld of the hydrogen turbine intake casing of the engine according to claim 1, characterized in that: The SN curve obtained from the fatigue test is subjected to zero mean stress correction, and the SN curve with non-zero mean stress is equivalent to the SN curve with zero mean stress. The zero mean stress correction method includes the Goodman method and the Soderberg method; The expression of Goodman's method is: ； The expression of Soderberg method is: ； in, represents the equivalent zero mean stress; represents the stress amplitude of the i-th point on the SN curve; represents the stress mean of the i-th point on the SN curve, represents the yield strength, Indicates tensile strength.

3. The method for evaluating the fatigue life of the weld of the hydrogen turbine intake casing of the engine according to claim 2, characterized in that: Calculating the stress power spectrum density of the engine hydrogen turbine air intake casing structure by the transfer function in step 4 is specifically as follows: calculating the stress power spectrum density function of the engine hydrogen turbine air intake casing structure by the transfer function and random vibration acceleration power spectrum density excitation of the engine hydrogen turbine air intake casing structure under different frequency excitations, and the stress power spectrum density function G(f) at frequency f is: ； Where H(f) represents the transfer function at frequency f, and g(f) represents the power spectrum density of random vibration acceleration.

4. The method for evaluating the fatigue life of the weld of the hydrogen turbine intake casing of the engine according to claim 3, characterized in that: The empirical formulas for stress range probability prediction include the empirical formula of the Dirlik method, the empirical formula of the Steinberg method and the empirical formula of the Bendat method; The empirical formula of the Dirlik method is: ； in, represents the probability of occurrence of stress range S, M, Z, Q, R, D1, D2 and D3 represent coefficients related to power spectrum characteristics, which are calculated by stress power spectrum density; The peak distribution of stress in the Steinberg method follows a normal distribution model. The empirical formula of the Steinberg method is: the probability of occurrence of stress at the 1σ level is 68.3%, the probability of occurrence of stress at the 2σ level is 27.1%, and the probability of occurrence of stress at the 3σ level is 4.43%. Where σ is the root mean square stress, expressed as: ； The empirical formula of the Bendat method is: ； Among them, the stress range S obeys the Rayleigh distribution.

5. The method for evaluating the fatigue life of the weld of the hydrogen turbine intake casing of the engine according to claim 4, characterized in that: The expression of linear cumulative damage theory is: ； Where D represents the cumulative damage, Represents stress The number of cycles under the level, Represents stress The maximum number of cycles obtained from the SN curve at the horizontal level.

6. The method for evaluating the fatigue life of the weld of the hydrogen turbine intake casing of the engine according to claim 5, characterized in that: The SN curve expression is: ； Among them, S represents the stress range, N represents the number of cycles, m represents the slope of the SN curve of the material, and C represents the fatigue strength coefficient.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images