Prediction and evaluation method for multi-axial fatigue life of additive manufacturing K418B high-temperature alloy

Abstract

The invention relates to a method for predicting and evaluating the multi-axial fatigue life of an additive manufacturing K418B nickel-based superalloy. Comprising the following steps: S1, carrying out intrinsic characteristic characterization on a material, including mechanical property test and defect statistical analysis based on industrial CT and a scanning electron microscope; s2, carrying out single-axis, multi-axis and notch fatigue tests of the system, and analyzing the association between the microstructure and the fatigue life; s3, based on a fracture mechanics theory, correcting the traditional model by taking the equivalent crack propagation driving force delta KDF as a special parameter, and constructing a multi-axial fatigue life prediction model; s4, determining a calculation method of an equivalent stress strength factor Keq optimized by adopting a weight coefficient through a multi-axial fatigue test, and establishing a notched part service life prediction model considering crack propagation driving force; and S5, performing fatigue test verification by designing and manufacturing a simulation piece, and verifying the accuracy and reliability of the multi-axial fatigue life prediction model. According to the method, the prediction precision and the engineering applicability of the fatigue life of the additive manufacturing K418B alloy under the high-temperature multi-axial load are remarkably improved.

Description

Technical Field

[0001] This invention pertains to fatigue life prediction technology for high-temperature alloy materials in the aerospace field, specifically involving a method for predicting the fatigue life of additively manufactured K418B nickel-based high-temperature alloy under multiaxial load conditions. Background Technology

[0002] With the rapid development of the global aerospace industry, high thrust-to-weight ratio and low energy consumption have become the main goals of the design of the next generation of aero engines. Against this backdrop, hot-end components of aero engines, especially turbine blades, face severe challenges from extreme service environments, including high temperature, high stress, and complex conditions such as oxidation and corrosion. As the core load-bearing component of the engine, breakthroughs in the material properties and manufacturing processes of turbine blades are directly related to the overall performance and reliability of the engine. Traditionally, K418B nickel-based superalloy has been widely used in the manufacture of turbine blades due to its excellent high-temperature strength (>900 MPa@800℃) and creep resistance. However, when using traditional casting processes to produce K418B alloy parts, problems such as difficulty in forming complex internal cavity structures (error >200 μm) and low material utilization (<30%) are encountered. These problems seriously limit the development needs of engine lightweighting and functional integration.

[0003] In existing technologies, additive manufacturing (AM), especially selective laser melting (SLM), offers a disruptive solution to the aforementioned problems due to its layer-by-layer deposition principle and the advantages of digital design. AM technology can achieve near-net-shape forming of complex structures in K418B alloy (dimensional accuracy ±50 μm), increasing material utilization to >95%, significantly improving production efficiency and material performance. However, the inherent rapid solidification characteristics of AM technology (cooling rate >10^6 K / s) lead to the formation of heterogeneous microstructures (such as columnar crystals and molten pool boundaries) within the material, while introducing subsurface porosity (porosity 0.5-5%), unfused defects (size 50-200 μm), and residual stress (peak value >500 MPa). These characteristics significantly alter the fatigue damage mechanism of the material, making its failure behavior fundamentally different from that of traditional casting materials.

[0004] In terms of fatigue life prediction, traditional fatigue life models (such as Miner's linear cumulative damage theory and stress-life method) show good applicability in cast / forged materials (error <15%); however, for additive manufacturing of K418B nickel-based superalloy, due to its unique microscopic heterogeneity and complex defect characteristics, the prediction bias of traditional models increases significantly (error >50%).

[0005] Although researchers have proposed fatigue models based on defect statistics, process-performance correlation models, and multi-scale mechanical models in recent years, which have improved the prediction accuracy to some extent, systematic research on nickel-based superalloys (especially K418B) is still insufficient; in particular, under multi-axis load conditions, the complex stress state of turbine blades further increases the difficulty of fatigue life prediction.

[0006] Currently, a fatigue life prediction model for additively manufactured K418B nickel-based superalloy under multiaxial loads has not been established. However, turbine blades often bear multi-directional loads in actual working conditions, which not only increases the complexity of stress but also significantly affects the initiation and propagation path of cracks. Therefore, constructing a fatigue life prediction model for additively manufactured K418B under multiaxial loads is of great significance for accurately assessing the service life and reliability of turbine blades. Summary of the Invention

[0007] To address the aforementioned issues, this invention proposes a multiaxial load fatigue life prediction model for additively manufactured K418B nickel-based superalloy. This model aims to effectively assess the fatigue life of this alloy under high-temperature multiaxial load conditions, providing theoretical support for the serviceability of key hot-end components such as additively manufactured turbine blades. Through in-depth analysis of multiaxial crack driving forces, a new multiaxial fatigue life prediction framework is constructed, providing a theoretical tool for strength and life assessment of additively manufactured components in the aerospace field. This promotes breakthroughs in the application of AM technology in high-performance nickel-based alloys, contributing to performance improvement and reliable application of key components such as turbine blades.

[0008] To achieve the above objectives, the present invention provides a multiaxial load fatigue life prediction model for additive manufacturing of K418B nickel-based superalloy, comprising the following steps: The intrinsic properties of additive manufacturing material K418B were characterized by tensile testing to determine its key mechanical parameters, such as tensile strength, yield strength and elongation. Uniaxial, multiaxial and notched fatigue tests were conducted on K418B alloy. The fatigue test data and fracture morphology were analyzed to reveal the intrinsic relationship between the material's microstructure and fatigue life. Based on fracture mechanics theory, this paper analyzes the key factors affecting fatigue life, using the equivalent crack propagation driving force Δ as an example. K DF Using characteristic parameters, the traditional fatigue life prediction model is modified, and a multiaxial fatigue life prediction model suitable for additive manufacturing of K418B alloy is constructed. Design and implement multiaxial fatigue tests to determine the equivalent stress intensity factor. K eq The calculation method is used to construct a fatigue life prediction model for notched components that considers crack propagation driving force; We designed and manufactured fatigue simulation parts of K418B alloy, conducted fatigue tests, and verified the accuracy and reliability of the multiaxial fatigue life prediction model.

[0009] Furthermore, the intrinsic characterization of additively manufactured K418B material also includes: using industrial CT scanning analysis software and scanning electron microscopy-based fatigue fracture analysis to statistically analyze the size, location, and morphology of internal defects in the material, providing defect data support for fatigue life prediction.

[0010] When constructing a multiaxial fatigue life prediction model, the equivalent stress intensity factor is calculated using the square root formula that includes weighting coefficients. K eq The optimization method for weighting coefficients is to find a set of optimal weighting coefficients that ensures the lifetime data points under all stress states are weighted evenly. K eq On the lifetime plot with parameters, the curves converge to a single, minimally dispersed curve. The optimization results include: At 500℃, the weighting coefficients are respectively ω 1 = 1.00, ω 2 = 6.00, ω 3 = 3.37; At 700℃, the weighting coefficients are respectively ω 1 = 1.00, ω 2 = 5.99, ω 3 = 2.96.

[0011] Furthermore, when revising the traditional fatigue life prediction model, the equivalent crack propagation driving force Δ is used. K DF As the core parameter, the uniaxial fatigue data is converted from the SN system to Δ. K DF -N system.

[0012] Furthermore, when designing and implementing multiaxial fatigue tests, the complex multiaxial stress state of the turbine blade during actual service is considered, and the load conditions under actual working conditions are simulated.

[0013] Furthermore, when constructing a fatigue life prediction model for notched components that considers crack propagation driving forces, the model is based on the Neuber notch fatigue model and incorporates the equivalent stress strength factor. K eq To build.

[0014] Furthermore, when designing and manufacturing fatigue simulation parts for K418B alloy, the simulation part structure includes a turbine blade disk to simulate the service conditions of actual hot-end components.

[0015] Furthermore, when designing and conducting fatigue tests on engineering turbine blade disk simulators to verify the accuracy and reliability of the multiaxial fatigue life prediction model, the difference between the predicted life and the actual test life should be compared to evaluate the engineering applicability of the model. The difference should be within three times the dispersion band.

[0016] Furthermore, when using industrial CT scan analysis software to statistically analyze internal material defects, if the fatigue fracture surface cannot clearly show the actual defect size of the fatigue source, the average value of the maximum values ​​of the six regions in the CT scan results is selected as the default value for near-surface defects used in the simulation.

[0017] Furthermore, the method is applicable to fatigue life prediction of K418B nickel-based superalloys used in additive manufacturing of hot-end components of aero-engines under high-temperature multiaxial load conditions.

[0018] The additive manufacturing K418B nickel-based superalloy multiaxial load fatigue life prediction model of the present invention has many significant advantages, as follows: In real-world operating conditions, turbine blades are subjected to multi-directional loads. Multiaxial loads increase stress complexity and significantly affect crack initiation and propagation paths. This model is specifically designed for multiaxial load conditions and can accurately assess the service life and reliability of additively manufactured K418B turbine blades under complex operating conditions. This provides crucial assurance for the safe operation of aero-engines and effectively reduces the risk of flight accidents caused by blade fatigue failure.

[0019] While traditional fatigue life models perform well in cast / forged materials, their prediction bias increases significantly (error > 50%) in additively manufactured K418B nickel-based superalloys due to the material's unique microscopic heterogeneity and complex defect characteristics. This new model, based on the actual conditions of additively manufactured K418B alloys, uses crack propagation driving force as the core parameter to specifically modify the traditional model, overcoming its limitations and significantly improving prediction accuracy.

[0020] Currently, no fatigue life prediction model has been established for nickel-based superalloys (especially K418B) under multiaxial loading, and existing research mainly focuses on titanium alloys and stainless steel. The proposed model fills this research gap, providing a systematic theory and method for fatigue life prediction of additively manufactured K418B nickel-based superalloys, and promoting the research development in this field.

[0021] This invention provides a powerful theoretical tool for assessing the strength and lifespan of additively manufactured components in the aerospace field. Accurately predicting the fatigue life of additively manufactured K418B alloys helps optimize additive manufacturing process parameters, improve manufacturing quality, promote the widespread application of AM technology in high-performance nickel-based alloys, and facilitate technological upgrading in the aerospace manufacturing industry.

[0022] As a core load-bearing component of aero-engines, the performance of turbine blades directly affects the overall performance of the engine. The application of this model can guide the design and manufacturing of turbine blades, improve the fatigue performance of turbine blades by rationally controlling internal material defects and optimizing the structure, thereby improving the thrust-to-weight ratio and reliability of the engine, reducing energy consumption, and meeting the needs of the aerospace industry to develop towards high thrust-to-weight ratio and low energy consumption.

[0023] This model has undergone rigorous experimental verification, including characterization of intrinsic material properties, fatigue life modeling, and engineering verification on a typical hot-end component of aero-engines—the turbine blade disk structure. Verification results show that the model's predictions all fall within a 3x error band, demonstrating good engineering applicability and reliability, and can be directly applied to fatigue life assessment and design optimization in practical engineering. Attached Figure Description

[0024] Figure 1 This is a flowchart of the additive manufacturing method for predicting the multiaxial fatigue life of K418B alloy provided by the present invention. Figure 2 Here are schematic diagrams of the shapes of the fatigue specimens in Example 1: (a) uniaxial smooth specimen; (b) multiaxial fatigue specimen; (c) uniaxial notched specimen. Figure 3 This is a schematic diagram of the design process of the aero-engine turbine blade disk simulator in this invention; Figure 4 This is a flowchart of the additive manufacturing material fatigue model construction based on defect-driven crack propagation in this invention; Figure 5 This is an extended schematic diagram of the fatigue life prediction framework that considers notch stress concentration in this invention. Figure 6 This is a schematic diagram of the stretching curve of AM K418B in this invention; Figure 7 The fatigue test results of AM K418B in this invention are as follows: (a) 500℃; (b) 700℃; Figure 8 This is a schematic diagram of the fatigue fracture morphology of the AM K418B smooth component in this invention at 500°C. Figure 9 This is a schematic diagram of the fatigue fracture morphology of the AM K418B smooth component in this invention at 700°C. Figure 10 This is the present invention. K eq The optimization results of the weighted coefficients are as follows: (a) 500℃; (b) 700℃; Figure 11 The fatigue prediction model results of AM K418B in this invention are: (a) 500℃; (b) 700℃; Figure 12 This invention presents a comparison of fatigue life prediction and experimental results for a simulated K418B turbine blade disk manufactured using additive manufacturing. Detailed Implementation

[0025] The following is a reference to the instruction manual. Figure 1-11 The multiaxial load fatigue life prediction model for additive manufacturing K418B nickel-based superalloy proposed in this invention is described in detail below: S1, Material Preparation Alloy composition design: This embodiment uses K418B nickel-based superalloy, with a nominal chemical composition (wt%) of: Al 5.5 - 6.4, Cr 11.0 - 13.0, Mo 4.0 - 5.0, Nb 1.5 - 2.5, with the balance being Ni. It also contains trace amounts of B and Zr (total amount approximately 0.49 - 1.635 wt%) to optimize high-temperature mechanical properties.

[0026] Additive manufacturing process: Selective laser melting (SLM) technology was used to form the K418B alloy specimens. To reduce the accumulation of residual stress during the printing process, a zigzag scanning strategy with 67° interlayer rotation was employed. All samples were manufactured on a commercial SLM device (EasyAdd 3D EP-M25) under an argon protective atmosphere.

[0027] Raw materials: Gas-atomized spherical powder with an average particle size range of 15-55 μm is selected.

[0028] Sample size and heat treatment: The sample is formed directly onto the substrate. After forming, the sample is sealed in a vacuum quartz tube and placed in a five-sided heated box furnace for heat treatment.

[0029] S2. Characterization of intrinsic material properties Tensile testing: Tensile tests were performed on heat-treated K418B alloy specimens to determine key mechanical parameters such as tensile strength, yield strength, and elongation. The tensile tests were conducted on a universal testing machine, following relevant standards, and the stress-strain curves during the tensile process were recorded to obtain data such as elastic modulus, yield strength, and tensile strength.

[0030] Fatigue test design: Uniaxial fatigue testing: Smooth round bar specimens were prepared and uniaxial tensile-tensile fatigue tests were performed on an MTS hydraulic servo fatigue testing machine. The experiments were conducted at two key service temperatures: 500℃ and 700℃. The applied nominal stress level was set between 450MPa and 1000MPa to fully cover the transition region of the material from high-cycle fatigue to low-cycle fatigue. The loading waveform was a sine wave, with a uniform frequency of 10Hz, and the stress ratio was... R = σ min / σ max = 0.05.

[0031] Multiaxial fatigue testing: Hollow thin-walled specimens were designed, and multiaxial fatigue tests with combined tensile and torsional loading were conducted on an MTS tensile-torsional fatigue testing machine. The multiaxial load level was defined by the von Mises equivalent stress, which was set in the range of 700 MPa to 1000 MPa, consistent with the uniaxial high-stress zone.

[0032] Notched fatigue test: Notched specimens were prepared with the same overall geometry as the smooth specimens, except that a semi-circular circumferential notch with a theoretical stress concentration factor of 2.5 was machined in the gauge length section. The experimental conditions for the notched specimens, including the test temperature (500℃ and 700℃) and the nominal stress range (450MPa to 1000MPa), were the same as those for the smooth specimens.

[0033] Fracture morphology analysis: After the fatigue test, the fracture surfaces of all fractured specimens were analyzed using a field emission scanning electron microscope (SEM, Zeiss GeminiSigma 300). The fracture morphology characteristics were comprehensively and meticulously observed, including the initiation and propagation paths of microcracks and the morphological details of the final fracture region, to analyze the damage evolution mechanism under different fatigue loading conditions.

[0034] S3. Material Internal Defect Detection Based on Industrial CT Industrial CT scan: Industrial CT technology was used to scan two round bar samples of additive manufacturing K418B.

[0035] The scanning parameters were set as follows: X-ray source voltage 150kV, current 60μA, exposure time per image 0.5s, and 5 frames were merged to ensure sufficient transmittance (>30%). The total scanning time was expected to be 1.5h. The effective pixel size of the test results was 5μm.

[0036] Defect Analysis and Statistics: The commercial software Avizo was used to process industrial CT scan results to achieve three-dimensional (3D) visualization of defects. The morphology and size distribution of defects within the K418B material were statistically analyzed. When the fatigue fracture surface could not clearly show the actual defect size of the fatigue source, the average of the maximum values ​​of the six regions in the CT scan results, 49 μm, was selected as the default value for near-surface defects used in the simulation.

[0037] S4. Construction of Fatigue Life Prediction Model for K418B under Multiaxial Loading Equivalent stress intensity factor K eq Determination: Introducing the equivalent stress intensity factor Keq To comprehensively evaluate crack propagation under mixed modes, the weighting coefficients are calibrated by fitting uniaxial and multiaxial fatigue test data to a square root formula containing weighting coefficients through an optimization algorithm.

[0038] Crack propagation threshold Δ K th Estimation: The crack propagation threshold is estimated based on the empirical formula proposed by Murakami. The Vickers hardness of the material and the known defect size are experimentally determined, and combined with the stress ratio, Δ is calculated. K th .

[0039] Normalized crack propagation driving force Δ K DF Definition: Define the normalized crack propagation driving force Δ K DF When Δ K DF When Δ is ≤1, the crack is in a dormant state; when Δ is ≤1, the crack is in a dormant state. K DF When the value is greater than 1, the crack begins to propagate and eventually leads to fatigue failure.

[0040] Establishment of a multiaxial fatigue life prediction model: Based on the widely used semi-logarithmic linear relationship Wohler model in engineering, the model replaces the stress description with Δ... K DF A quantitative predictive relationship from driving force to lifespan was established, and a multiaxial fatigue life prediction model suitable for additive manufacturing of K418B alloy was obtained through data fitting.

[0041] Life prediction model extension considering notch effect: Model parameter calibration, introducing fatigue notch sensitivity coefficient to correlate the theoretical stress concentration factor of the notch with the fatigue notch coefficient. According to Neuber's theory, notch sensitivity is related to the geometry and material properties of the notch root, and the characteristic length related to the material is estimated. The experimental fatigue notch coefficient at the life point is calculated using the formula, and then the characteristic length of the material is calibrated by combining the theoretical stress concentration factor of the notched specimen and the notch radius.

[0042] Fatigue life prediction: For a new notch geometry, using the known stress concentration factor, notch radius, and calibrated material parameters, the fatigue notch coefficient of the new component is calculated using a formula, establishing a mapping relationship between the fatigue behavior of the new component and the fatigue behavior of the smooth specimen (reference). By traversing all data points on the S-N curve of the smooth specimen and using the above conversion relationship, a series of data points corresponding to the new component are calculated. By fitting these data points, the fatigue life prediction curve of the new notched component is obtained, realizing the life prediction of any notched component.

[0043] S5. Design and verification of fatigue simulation components for structural parts. Simulation design: Following the principles of material consistency, temperature consistency, and stress gradient consistency, a three-dimensional model and finite element analysis (FEA) of the actual turbine disk structure were performed to determine the key stress concentration areas and stress gradient distribution. Based on the stress distribution at key points, an appropriate simulation component type was selected. The geometric dimensions of the simulation component (such as notch radius R and diameter D) and the nominal load σ applied to the far end of the specimen were adjusted through iterative optimization to ensure that the stress distribution at critical points was consistent with the stress distribution in the key areas of the structural component.

[0044] Simulation component manufacturing and testing: Fatigue simulation components are manufactured according to the design results and fatigue tests are conducted on the simulation components using an MTS tensile-torsional fatigue testing machine. Except for the experimental load, all other settings are consistent with those of the uniaxial test to ensure comparability of experimental conditions.

[0045] Model Validation: The fatigue life of the simulated components was compared with the results of the prediction model to verify the accuracy and reliability of the proposed life prediction framework. In this embodiment, all test data points of the simulated components, whether at 500℃ or 700℃, were within the three-fold dispersion band, demonstrating the model's good engineering applicability.

[0046] Example: like Figure 1-11 As shown, this embodiment will provide theoretical support for the service life of key hot-end components through a multiaxial load fatigue life prediction model for additive manufacturing of K418B nickel-based superalloy. This embodiment consists of three modules: material intrinsic property characterization, measuring key mechanical parameters, conducting fatigue tests, and revealing the correlation between microstructure and fatigue life; fatigue life modeling, developing a multiaxial fatigue life prediction model based on fracture mechanics to modify traditional models, and constructing a fatigue life prediction model for notched components that considers crack propagation driving forces; and model engineering verification, designing and manufacturing fatigue simulation components and conducting tests to verify the accuracy and reliability of the prediction method.

[0047] S1, Material Preparation The K418B alloy in this embodiment is a nickel-based superalloy with added elements such as Al, Cr, Mo, Ti, Nb, Fe, B, and Zr. This alloy has low carbon and titanium content, which significantly improves its plasticity and impact toughness at room temperature and enhances its microstructure stability. Its nominal chemical composition is shown in Table 1 below, with the main components (wt%) including: Al (5.5–6.4), Cr (11.0–13.0), Mo (4.0–5.0), Nb (1.5–2.5), and the balance being Ni; it also contains trace amounts of B and Zr (total approximately 0.49–1.635 wt%) to further optimize its high-temperature mechanical properties.

[0048] Table 1 Chemical composition (%) of K418B powder The K418B alloy sample used in this embodiment was formed using Selective Laser Melting (SLM) technology. Specific process parameters were as follows: laser power 250 W, scanning speed 1500 mm / s, scanning spacing h = 0.09 mm, and layer thickness t = 0.03 mm. To effectively reduce the accumulation of residual stress during the printing process, a zigzag scanning strategy with 67° interlayer rotation was employed. All samples were manufactured using a commercial SLM device (EasyPlus EP-M25) under an argon protective atmosphere.

[0049] The raw material used in this embodiment is gas-atomized spherical powder with an average particle size ranging from 15 to 55 μm. The sample size is 50 mm × 16 mm × 16 mm, directly formed on the substrate. After forming, the sample is sealed in a vacuum quartz tube and placed in a five-sided heated box furnace for heat treatment: first, it is held at 1180°C for 2 hours, then cooled to room temperature by purging with argon gas; then, it undergoes aging treatment, held at 930°C for 8 hours, to regulate the microstructure and improve the overall mechanical properties.

[0050] S2. Characterization of intrinsic material properties In this embodiment, to systematically obtain the experimental data required for calibrating the fatigue life prediction model, a series of fatigue tests covering static tensile and uniaxial, multiaxial, and notched stress states were designed and conducted for additively manufactured K418B nickel-based superalloy. Figure 2 As shown. All experiments were conducted on an MTS hydraulic servo fatigue testing machine using constant amplitude load control. The loading waveform was a sine wave, and the frequency was uniformly set to 10Hz. The stress ratio was... R = σ min / σ max All are 0.05. Details are as follows: First, uniaxial tensile fatigue tests were performed on smooth round bar specimens: the experiments were conducted at two key service temperatures (500℃ and 700℃). The applied nominal stress levels were set between 450MPa and 1000MPa to comprehensively cover the transition region of the material from high-cycle fatigue to low-cycle fatigue, thereby providing broad and reliable baseline data for model building.

[0051] To investigate the fatigue behavior of materials under complex stress states and to provide a calibration basis for the equivalent stress strength factor formula in the model, this embodiment designed hollow thin-walled specimens and conducted multiaxial fatigue tests under combined tensile and torsional loading. By applying simultaneous axial and torsional loads on an MTS tensile-torsional fatigue testing machine, the complex stresses experienced by the material under actual working conditions were simulated. In this embodiment, the ratio of tensile normal stress to shear stress was set to 1.73. To effectively compare with uniaxial experiments, the multiaxial load level was defined using von Mises equivalent stress. (1) in, σ It is tensile stress. τ This refers to shear stress. The equivalent stress level is set within the range of 700MPa to 1000MPa, consistent with the uniaxial high-stress region. By comparing the fatigue life of multiaxial and uniaxial stresses under the same equivalent stress, the influence of multiaxial stress on the fatigue performance of materials can be analyzed in depth.

[0052] In addition, to account for the stress concentration effect caused by the macroscopic notch, uniaxial fatigue tests were also conducted on the notched specimens in this embodiment. The overall geometry of the notched specimens was consistent with that of the smooth specimens, except that a semi-circular circumferential notch with a theoretical stress concentration factor of 2.5 was machined only in the gauge length section. The experimental conditions for the notched specimens, including the test temperatures (500℃ and 700℃) and the nominal stress range (450MPa to 1000MPa), were exactly the same as those for the smooth specimens to ensure the comparability of the two sets of experimental results.

[0053] After the fatigue test, the fracture surfaces of all fractured specimens were microscopically analyzed using a field emission scanning electron microscope (SEM, Zeiss Gemini Sigma 300). This allowed for a comprehensive and detailed observation of the fracture morphology, including the initiation and propagation paths of microcracks, as well as the morphological details of the final fracture region. By analyzing the microscopic morphology of the fracture surfaces, the damage evolution mechanism of the material under different fatigue loading conditions was explored in depth, achieving an effective correlation between macroscopic fatigue performance and microscopic damage behavior. This provides key microscopic information and theoretical basis for constructing an accurate fatigue life prediction model.

[0054] To verify the applicability of the proposed additive manufacturing model for multiaxial load fatigue life prediction of K418B nickel-based superalloy in actual engineering structural components, this embodiment fabricated multi-sample simulation test specimens (simulation specimens). The design of the simulation specimens followed the principles of material consistency, temperature consistency, and stress gradient consistency. The specific design process is as follows: Figure 3 As shown: First, a three-dimensional model and finite element analysis (FEA) of the actual turbine disk structure are performed to determine the key stress concentration areas and stress gradient distribution. Subsequently, based on the stress distribution at key points, a suitable simulation component type is selected, entering a crucial iterative optimization loop: adjusting the geometry of the simulation component (such as the notch radius). R ,diameter D and the nominal load applied to the distal end of the specimen. σ This ensures that the stress distribution at the critical points is consistent with the stress distribution in the key areas of the structural components.

[0055] Finally, after the above iterative optimization design, the geometric dimensions of the simulation part were determined as follows: Figure 3 As shown in step 6, the nominal tensile load applied to the clamping end of the simulation part is 100 MPa. Under this load condition, the nominal stress at the central notch of the simulation part will be 711 MPa.

[0056] Fatigue simulation parts were fabricated based on the design results and subjected to fatigue tests on an MTS tensile-torsional fatigue testing machine. Except for the experimental load, all other settings were kept consistent with uniaxial experiments to ensure comparability of experimental conditions. The accuracy and reliability of the proposed life prediction framework were verified by comparing the fatigue life of the simulation parts with the results of the prediction model.

[0057] S3. Material Internal Defect Detection Based on Industrial CT In this embodiment, accurate characterization of internal defects (such as porosity, unfused zones, and microcrack initiation sites) is a prerequisite for establishing a high-precision fatigue model during fatigue life prediction of additively manufactured K418B alloy components. Traditional metallographic observation has certain limitations in terms of defect resolution, three-dimensional reconstruction, and quantitative analysis. Industrial CT (Computed Tomography) technology, as an advanced non-destructive testing method, can provide high-precision three-dimensional structural information for defect detection and quantitative evaluation, thereby providing more reliable initial defect parameter input for the fatigue life prediction model.

[0058] In this embodiment, two round bar samples of additively manufactured K418B were scanned using industrial CT. The scanned area of ​​the test piece was 10 mm long and 5 mm in diameter. The scanning parameters were set as follows: X-ray source voltage 150 kV, current 60 μA, exposure time per image 0.5 s, and 5 frames were merged to ensure sufficient transmittance (>30%). The total scanning time was estimated to be 1.5 h. The effective pixel size of the test results was 5 μm. After obtaining the test results, the defects were visualized in three dimensions using the commercial software Avizo, and the defect morphology and size distribution inside the K418B material were statistically analyzed. These detailed defect feature data will serve as the initial input parameters for the subsequent fatigue life prediction model, helping to improve the accuracy and reliability of the model prediction.

[0059] S4. Construction of Fatigue Life Prediction Model for K418B under Multiaxial Loading In this embodiment, a theoretical model is established that can accurately predict the fatigue life of additively manufactured (AM) K418B nickel-based superalloy under complex working conditions, such as... Figure 4 As shown. Compared with conventionally manufactured metallic materials, additively manufactured materials exhibit unique fatigue properties, with their high-cycle / ultra-high-cycle fatigue life typically controlled by microscopic defects unavoidably generated during the manufacturing process. These defects, such as lack of fusion (LOF) or spheroidized pores, act as inherent stress concentration sources and are the primary locations for fatigue crack initiation. To quantify the impact of defects on fatigue strength, Murakami et al. proposed a widely used theoretical framework. The core idea of ​​this theory is to correlate the fatigue limit with the size of the largest defect within the material. The defect size is defined as the square root of the area projected onto the defect perpendicular to the direction of the maximum principal stress, i.e. This parameter, acting as an equivalent crack size, cleverly simplifies the complex defect geometry into a single scalar. However, As a purely geometric parameter, its inherent limitations are also quite obvious: it cannot distinguish between defects with the same projected area but vastly different geometric shapes. For example, a sharp, LOF defect and a smooth, spherical hole, if their... The models would assign the same level of harm to both. However, in reality, the stress concentration effect of the former is far greater than that of the latter, and the damage to fatigue performance is also more severe.

[0060] To more accurately describe the true severity of defects, this embodiment must shift from geometric description to physical and mechanical analysis. Linear Elastic Fracture Mechanics (LEFM) provides a powerful tool for this purpose. Within the LEFM framework, the stress intensity factor (SIF), denoted as , is not the macroscopic size of the defect that determines the stress field intensity at the defect tip and drives crack propagation. K This can be obtained through finite element method (LEFM) simulation. This LEFM-based method can distinguish the severity of defects of different shapes from their physical nature, thus overcoming the limitations of simply relying on... The model has its shortcomings.

[0061] Aerospace structural components often endure complex multiaxial loads during service, leading to internal cracks operating under a composite mode of Type I (opening), Type II (slip), and Type III (tear) loading. Studies have shown a significant interaction between stress intensity factors of different loading modes, collectively determining fatigue crack propagation behavior. Therefore, the concept of crack propagation driving force must be extended from uniaxial to multiaxial stress states. To comprehensively assess crack propagation under this mixed mode, academia and engineering typically introduce the equivalent stress intensity factor. K eq The concept of , a commonly used equivalent stress intensity factor is the square root formula: (2) in, These are undetermined weighting coefficients. These coefficients reflect the differences in the material's sensitivity to different types of loading modes, and their values ​​need to be calibrated by fitting uniaxial and multiaxial fatigue test data using an optimization algorithm. The core idea of ​​calibration is to find an optimal set of weighting coefficients such that the life data points under all stress states are within the range of... K eq On a lifetime plot with parameters, the curves can converge to a single curve with minimal dispersion.

[0062] According to LFEM theory, when K eq Reaching or exceeding the material's critical value K IC When the crack propagates rapidly and unstablely, it leads to structural failure. For fatigue problems, cyclic loading alone is insufficient to cause crack propagation. Crack propagation must overcome the inherent resistance of the material, which is determined by the crack propagation threshold Δ. K th To define. Only when the applied Δ K eq Exceeding Δ K th Only when the crack reaches a stable subcritical propagation will it proceed. Based on extensive experimental data, Murakami proposed a rapid estimation method for Δ. K th The famous empirical formula: (3) in, R The stress ratio of the fatigue load. HV The Vickers hardness of the material; αHere, represents a material parameter. This formula links the macroscopic hardness of a material to the size of its microscopic defects, providing a basis for assessing the material's resistance to microcrack propagation. Based on the above analysis, to establish a fatigue life model that better reflects the physical mechanism, this embodiment defines a normalized crack driving force Δ. K DF Its expression is: (4) The dimensionless parameter Δ K DF The physical meaning is quite clear: it quantifies the ratio between the equivalent driving force generated by external multiaxial cyclic loading and the material's intrinsic crack propagation resistance (i.e., the threshold value). When Δ K DF When Δ is ≤1, the crack is in a dormant state; when Δ is ≤1, the crack is in a dormant state. K DF When the value is greater than 1, the crack begins to propagate and eventually leads to fatigue failure. Therefore, Δ K DF It can serve as a unified damage parameter to replace the stress description in traditional fatigue models. σ To correlate fatigue life under different working conditions N Taking the widely used semi-log-linear Wohler model as an example, a quantitative predictive relationship from driving force to life is established by replacing the stress description: (5) in m , n These are the material fatigue performance constants obtained by fitting experimental data. This equation forms the core of this prediction model; once the material constants are calibrated, Δ under any working condition can be calculated. K DF Values ​​are used to predict the corresponding fatigue life. N .

[0063] Furthermore, real-world engineering components often exhibit geometric discontinuities such as holes and fillets, i.e., notches, leading to local stresses far exceeding nominal stresses. To apply the above model to components with notches, this embodiment extends the life prediction framework to arbitrary notch coefficients based on Neuberger's rule. The entire process is divided into two stages: model parameter calibration and fatigue life prediction, as detailed below. Figure 5 As shown.

[0064] In this embodiment, a fatigue notch sensitivity coefficient is introduced during the model parameter calibration stage. q The theoretical stress concentration factor associated with the notch K tand fatigue notch coefficient K f : (6) (7) in, It is the driving force for crack propagation in smooth parts. It is the driving force for crack propagation in notched components. K f This reflects the actual reduction in fatigue strength caused by the notch. According to Neuber's theory, notch sensitivity is related to the geometry and material properties at the notch root, and can be estimated using the following formula: (8) In the formula r It is the radius of the notch root. p It is a material-related feature length. p It is considered an inherent property of materials, reflecting the response of the material's internal microstructure to macroscopic stress gradients. Its calibration process is as follows: First, the experimental fatigue curves of smooth specimens and specimens with known notches are used as input. Within the life range where both curves are valid, several specific fatigue life values ​​are selected. N i For each selected lifespan N i The corresponding fatigue driving force parameters can be obtained from the two curves, i.e., the fatigue driving force parameters of the smooth specimen. and notched specimens .

[0065] Subsequently, the experimental fatigue notch coefficient at this life point was calculated using equation (6). K f Finally, combining this K f Values ​​and theoretical stress concentration factors of notched specimens K t and the radius of the gap r Substituting these equations into equations (7) and (8) will allow us to determine the characteristic length of the material. a .

[0066] In the fatigue life prediction stage, firstly, for the new notch geometry, the known stress concentration factor is used... K t and the radius of the gap r and the calibrated material parameters p The fatigue notch coefficient of the new component can be calculated using equations (8) and (7). K fNext, a mapping relationship is established between the fatigue behavior of the new component and the fatigue behavior of the smooth specimen (reference). According to the physical meaning of equation (6), for any given life... N i The fatigue driving force required for the new component and the smooth specimen satisfies the following relationship: (9) This formula constitutes the "transformation function" from the baseline curve to the target prediction curve. Finally, by traversing the smooth sample... SN Using the transformation relationship described above, a series of data points corresponding to the new component can be calculated from all the data points on the curve. By fitting these data points, the fatigue life prediction curve of the new notched component can be obtained, thus realizing the life prediction of any notched component.

[0067] S5. Design and verification of fatigue simulation components for structural parts. In this embodiment, the tensile curve of the additively manufactured K418B alloy was obtained through material tensile testing, as shown below. Figure 6 As shown, its elastic modulus is 168 GPa, yield strength is 780 MPa, and tensile strength is 875 MPa. High-temperature fatigue performance tests were conducted on K418B material at 500℃ and 700℃. Fatigue lives were obtained for smooth specimens, notched specimens, multiaxial specimens (compared using equivalent stress), and turbine blade disk fatigue simulation specimens under different loading stresses. The relevant results are as follows. Figure 7 As shown.

[0068] The fatigue test results in this embodiment show that, under the same stress level, the fatigue life of the smooth fatigue specimen at 700℃ is significantly lower than that at 500℃. This is because at higher temperatures, the yield strength of the material decreases, and the material softens, leading to a reduction in fatigue resistance. It is worth noting that the fatigue life of the smooth fatigue specimen at 700℃ exhibits a certain degree of dispersion under high-cycle fatigue. This may be due to the stress state being close to the material's fatigue limit. Fatigue lives near the material's fatigue limit typically exhibit significant dispersion because the fatigue life at this point is highly sensitive to stress magnitude; even small stress changes can lead to substantial differences in fatigue life. To further clarify the failure mechanism of K418B additive manufacturing material under uniaxial fatigue conditions, microscopic analysis was performed on the fatigue failure fracture surface of the smooth specimen. Figure 8 and Figure 9The fracture morphology of smooth fatigue specimens subjected to 800 MPa at 500℃ and 700℃ is shown. Red circles indicate fatigue crack initiation sites for the corresponding specimens. High-magnification SEM observation of the crack initiation sites reveals that at 800 MPa, almost all fatigue crack initiation sites are located on the subsurface of the material, and most cracks initiate from defects. Defect interfaces easily become stress concentration points, directly inducing crack initiation. Furthermore, testing of smooth fatigue specimens at other stress levels at 500℃ and 700℃ revealed similar results on their fracture surfaces, indicating that AM K418B alloy is highly sensitive to defects under high stress levels and easily initiates cracks from defects. The fatigue failure mode of AM K418B alloy in this embodiment is primarily defect-driven failure.

[0069] In this embodiment, the defect statistics function of industrial CT scan analysis software is used to statistically analyze the effects of each defect on the vertical stress loading direction. The results are shown in Table 2. As can be seen from the table, the internal defect size distribution of K418B is relatively uniform, and most are below 50 μm. The defect size used in finite element simulation of fatigue specimens is usually derived from the microscopic fracture surface image. However, for fatigue safety and design cost considerations, when the fatigue fracture surface cannot clearly show the actual defect size of the fatigue source, six regions from the CT scan results are selected. The average value of the maximum value is 49 μm, which is the default value for near-surface defects used in the simulation.

[0070] Table 2. Statistical results of defects in various regions of AM K418B (unit: μm) To extend the lifetime prediction framework proposed in this embodiment from uniaxial to multiaxial stress states, an equivalent stress intensity factor that can reasonably describe the crack driving force of hybrid modes is constructed. K eq This is very important. This embodiment uses a square root formula that includes weighting coefficients to achieve this purpose. The weighting coefficients of the equivalent stress in the optimized model are respectively at 500℃. ω 1 = 1.00, ω 2 = 6.00, ω 3 = 3.37; 700℃ are respectively ω 1 = 1.00, ω 2 = 5.99, ω 3 = 2.96 Figure 10 This demonstrates the effect of weight coefficient optimization. In the original... SN In the life diagram, even after converting the stress from multiaxial experiments to equivalent stress, the results from uniaxial and multiaxial experiments still show significant discrepancies, making it impossible to reach a unified conclusion. Figure 7As shown in the figure, under the optimized driving force system, multi-axis and single-axis data successfully converged into a more compact region. This indicates that the optimized equivalent stress intensity factor can effectively integrate the influence of multi-axis stress states, laying a solid foundation for subsequent normalization processing.

[0071] Subsequently, based on the calculation formula for crack propagation driving force, all uniaxial fatigue data were... SN The system is converted to Δ K DF - N The system is established, and the fatigue life prediction model is solved (see flowchart). Figure 4 The final results are shown in Table 3.

[0072] Table 3 AM K418B Fatigue Life Prediction Model Figure 11 The final fatigue life prediction model is shown, demonstrating the life prediction results for the model component. The green dashed line in the figure represents the model's prediction of the life of the simulated component. K t The fatigue life prediction curve is calculated as (1.09). Figure 11 As can be clearly observed in (a) and (b), under both temperature conditions, the experimental data points (purple stars) of the simulated component are closely distributed around their respective predicted curves, demonstrating the good accuracy of the proposed model considering the notch effect. To more intuitively and quantitatively evaluate the model's engineering applicability, Figure 12 The predicted lifespan of all simulated components was compared with their actual test lifespan. As can be seen from the figure, the test data points of all simulated components, whether at 500℃ or 700℃, are all within the three-fold dispersion band.

[0073] The above results demonstrate the effectiveness of the fatigue life prediction framework constructed in this embodiment for complex stress states of AM K418B. This model successfully unifies the intrinsic defects, external multiaxial stress states, and geometric notch effects of AM K418B components under a fracture mechanics-based theoretical framework, providing a reliable theoretical basis for the life assessment of additively manufactured K418B high-temperature alloy components with defects under complex working conditions.

[0074] In summary, this embodiment focuses on the additive manufacturing of K418B nickel-based superalloy for key hot-end components of aero-engines. It systematically conducts fatigue life prediction research under high-temperature multiaxial service conditions from multiple dimensions, including material performance testing, defect characterization, life modeling, and engineering verification. Through uniaxial and multiaxial fatigue tests at different temperatures (500℃ and 700℃), combined with fracture surface analysis and industrial CT inspection results, the key influence mechanism of internal defects (including size, location, and morphology) on fatigue life was clarified. By drawing on Murakami's fatigue analysis framework based on maximum defects and introducing a multiaxial stress correction factor, a multiaxial fatigue life prediction model suitable for additively manufactured K418B alloy was proposed. This model uses crack propagation driving force as the core parameter, achieving a normalized expression of fatigue life under different loading conditions and defect scales. Finally, a typical turbine blade disk structure fatigue simulation component was designed and manufactured, and fatigue experiments were conducted for verification. The results show that the prediction results of the proposed life prediction model all fall within a 3x error band, demonstrating good engineering applicability and reliability.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting and assessing the multi-axial fatigue life of additively manufactured K418B superalloy, characterized by: The method comprises the following steps: The intrinsic characteristics of the additive manufacturing K418B material are characterized, and key mechanical parameters such as tensile strength, yield strength and elongation are determined by a tensile test; Uniaxial, multiaxial and notched fatigue tests of the K418B alloy are carried out, fatigue test data and fracture morphology are analyzed, and the internal correlation between the microstructure characteristics of the material and the fatigue life is revealed; Based on the theory of fracture mechanics, the key factors affecting fatigue life are analyzed, and the equivalent crack propagation driving force ΔK is introduced K DF The characteristic parameters are used to modify the traditional fatigue life prediction model, and a multi-axial fatigue life prediction model suitable for K418B alloy manufactured by additive manufacturing is constructed; Design and implement multi-axial fatigue tests to determine the equivalent stress intensity factor K eq A notched fatigue life prediction model is constructed by considering the driving force of crack propagation. Fatigue simulation parts of the K418B alloy are designed and manufactured, fatigue tests are carried out, and the accuracy and reliability of the multiaxial fatigue life prediction model are verified.

2. The method of claim 1, wherein the method is for predicting and evaluating the multi-axial fatigue life of an additively manufactured K418B superalloy. The intrinsic characteristics of the additive manufacturing K418B material are characterized, and key mechanical parameters such as tensile strength, yield strength and elongation are determined by a tensile test; The size, position and morphology of internal defects of the material are analyzed by using an industrial CT scanning analysis software and a scanning electron microscope based fatigue fracture analysis, and defect data support is provided for fatigue life prediction.

3. The method of claim 1, wherein the method is for predicting and evaluating the multi-axial fatigue life of an additively manufactured K418B superalloy. The square root formula containing the weight coefficient is used to calculate the equivalent stress intensity factor when the multi-axial fatigue life prediction model is constructed K eq The optimization method of the weight coefficient is to find a set of optimal weight coefficients, so that all the life data points under all stress states can be collected near a single curve with the minimum dispersion on the life diagram with K eq the parameter. The optimization results include: At 500°C, the weight coefficients are respectively ω 1 = 1.00, ω 2 = 6.00, ω 3 = 3.37; At 700°C, the weight coefficients are respectively ω 1 = 1.00, ω 2 = 5.99, ω 3 = 2.

96.

4. The method of claim 1, wherein the method is for predicting and evaluating the multi-axial fatigue life of an additively manufactured K418B superalloy. The conventional fatigue life prediction model is modified to use equivalent crack growth driving force Δ K DF The uniaxial fatigue data is converted from S-N system to Δ K DF -N system with Δ 5. The method of claim 1, wherein: When the multiaxial fatigue test is designed and implemented, the complex multiaxial stress state in the actual service process of the turbine blade is considered, and the load conditions under the actual working conditions are simulated.

6. The method of claim 1, wherein: The construction considers the crack propagation driving force of the notch fatigue life prediction model, based on Neuber notch fatigue model, combined with equivalent stress intensity factor K eq was constructed.

7. The method of claim 1, wherein the method is for predicting and evaluating the multi-axial fatigue life of an additively manufactured K418B superalloy. When the fatigue simulation parts of the K418B alloy are designed and manufactured, the simulation part structure includes a turbine blade disc, which is used to simulate the service conditions of the actual hot end components.

8. The method of claim 1, wherein: When the fatigue test of the engineering turbine blade disc simulation part is designed and carried out to verify the accuracy and reliability of the multiaxial fatigue life prediction model, the difference between the predicted life and the actual test life is compared, the engineering applicability of the model is evaluated, and the difference should be within three times of the dispersion band.

9. The method of claim 2, wherein the method is for predicting the multi-axial fatigue life of an additively manufactured K418B superalloy. When the internal defects of the material are analyzed by using the industrial CT scanning analysis software, when the fatigue fracture cannot clearly show the actual defect size of the fatigue source, the average value of the maximum value of the six regions in the CT scanning result is selected as the default value of the near-surface defect used in simulation.

10. A method of predicting and assessing the multi-axial fatigue life of an additively manufactured K418B superalloy according to any one of claims 1 to 9, characterized in that: The method is suitable for fatigue life prediction of the additive manufacturing K418B nickel-based high-temperature alloy for hot end components of an aero-engine under high-temperature multiaxial load working conditions.

Images