Abstract: Oxidation damage and cracks are inevitable during the service of turbine blades. The prediction of crack behavior and its influence by oxidation damage are of great significance for service safety and life management. Aiming at the oxidation damage and crack problems of high-pressure turbine blades of aircraft engines in high-temperature environment for a long time, fatigue crack growth tests were carried out on compact tension (CT) specimens at 850℃, and the fatigue crack growth rates of normal specimens and specimens after oxidation damage were obtained; the Paris model was used to consider the crack growth in the presence of oxidation damage, and the influence of oxidation damage on fatigue crack growth rate was compared. Taking a certain high-pressure turbine blade of an aircraft engine as an example, this paper focuses on the influence of oxidation damage on the crack growth stage, and uses ANSYS and Franc3D software to analyze and compare the crack growth life of turbine blades with and without oxidation damage. The results show that under the influence of oxidation damage, the cycle life of the leading edge crack of the blade is reduced to 44.02% of that without oxidation damage on average, and the cycle life of the trailing edge crack of the blade is reduced to 50.22% of that without oxidation damage on average. It provides a basic reference for the life assessment of turbine blades under the actual service environment working conditions, and has important practical significance for correctly evaluating materials and predicting the service life and design strength of engineering hot end parts in other actual projects.
Keywords: turbine blades; nickel-based alloys; oxidation damage; CT specimens; crack propagation
With the continuous improvement of the working efficiency of aircraft engines, as the core components of high-performance aircraft engines, the service environment faced by turbine blades has become more severe. In addition to the centrifugal load caused by high-speed rotation, turbine blades are also subjected to complex aerodynamic and thermal stresses, as well as vibration, corrosion and oxidation. Therefore, for turbine blade components, it is inevitable that cracks will appear in the blades under such conditions for a long time. At the same time, long-term exposure to high temperature will also cause oxidation damage to the blades. Among the various types of damage to turbine blades, oxidation damage is the most common and most likely to occur. Through the investigation of data records such as the borescope of a domestic civil aviation engine blade, it was found that crack-type damage accounted for 11.4%, while oxidation damage accounted for as high as 73.3%. Discoloration of turbine blades is an early typical feature of oxidation damage. However, since the consequences are not serious, it is usually treated differently from other damage during engine maintenance. However, oxidation damage is one of the important factors causing other damage. Since the basic mechanical and fatigue properties of most metal materials are temperature-dependent, and high temperature environment is the key factor affecting the fatigue crack growth of nickel-based high-temperature alloys, and is closely related to the oxidation of materials, the crack growth of turbine blades is bound to be affected by oxidation damage. Kitaguchi used scanning electron microscopy and X-ray energy spectrometer to study the oxide intrusion formed in front of the crack tip of nickel-based high-temperature alloys after exposure to air at 650℃. Li Haiyan et al. proposed a new mechanism that considers the interaction between crack tip oxidation and stress relaxation by conducting load-holding fatigue and benchmark fatigue tests under vacuum conditions. Schulz et al. conducted isothermal fatigue crack growth tests on coarse-grained RR1000 in air at 700℃, and discussed the potential factors of fatigue crack growth resistance from the perspective of load-holding time. Jiang Rong et al. evaluated the role of oxygen in strengthening fatigue cracking in nickel-based superalloys during the initiation and propagation of fatigue cracks. Karabela et al. studied oxygen diffusion and crack propagation in nickel-based superalloys under fatigue-oxidation conditions, and performed microscopic characterization and numerical simulation. The study found that oxygen infiltration into the material and related internal oxidation can lead to material embrittlement and failure. Osinkolu et al. used single-edge notch tensile specimens to study the fatigue crack propagation rate of polycrystalline IN718 superalloy at 650°C with different grain sizes.
Most of the above scholars studied the effect of oxidation damage on crack propagation from a microscopic perspective, starting from the crack initiation time of grain boundary cracks, analyzing the changes in oxygen diffusion capacity caused by different ambient temperatures and loads, and then affecting the crack propagation rate in the initiation stage. They have a clearer understanding of the fatigue crack propagation law of superalloys under the influence of alloy microstructure, temperature, load and other factors, but there are relatively few studies on the effect of oxidation damage on fatigue cracks in the crack propagation stage.
In this paper, crack growth rate curves of specimens with and without oxidation damage are obtained by designing control tests, joint simulation is carried out using Franc3D and ANSYS software, crack growth life is calculated using the Paris model, and the crack growth law of high-pressure turbine blades in the stable growth stage with and without oxidation damage is compared.
1 Fatigue crack growth test
1.1 Materials and specimens
The material of the engine turbine blade is DZ125 high-performance nickel-based alloy, which is often used in engineering applications under high temperature environments. It can maintain good mechanical properties, creep resistance and corrosion resistance in high temperature environments, and has a long service life and reliability. Its chemical composition is shown in Table 1.
In this paper, CT specimens are selected for crack growth test research. The specimen design is based on the standard of metal materials-fatigue test-fatigue crack growth method, and the crack growth direction is set parallel to the directional solidification direction. The specific dimensions are shown in Figure 1.
1.2 Test plan
The purpose of the test is to obtain the crack growth rate curve of the specimens with and without oxidation damage, and to fit the Paris parameters. Considering that it is too complicated to use real blades and change the test environment conditions, this paper uses heat-treated CT specimens to simulate oxidation damage for simplified analysis.
In this test, two specimens were tested for crack growth. Specimen No. 1 was not subjected to high-temperature oxidation treatment, and specimen No. 2 was subjected to heat treatment at 850℃ for 1000h in an SXL-1400C box-type test electric furnace to simulate the material properties of the nickel-based alloy after oxidation. The specimen is shown in Figure 2.
After studying the temperature distribution of the turbine blade, it was calculated that the average temperature of the turbine blade was around 850℃, so the high-temperature environment of the test process in this section was set under this temperature condition. All specimens were pre-cracked at room temperature, and the test system is shown in Figure 3. During the test, the specimen was subjected to fatigue crack growth test in a high-temperature 850℃ air environment, heated by a resistance furnace, and the error between it and the test setting was controlled to be within ±5℃. Sine wave loading was used, with an initial maximum cyclic load of 3.5 kN, a frequency of 5 Hz, and a stress ratio of R = 0.1, until the test specimen broke and the test was terminated.
The compliance method was used to measure the crack length during the test. The compliance method is an indirect method for measuring the crack extension length of metal materials. The crack length is determined by measuring the displacement difference on both sides of the crack [16]. The relationship between compliance and normalized crack length is usually expressed by dimension one (formerly dimensionless) compliance BEVx /P and normalized crack length a/W. The specific formulas are as follows (1) and (2). In the formula, B and W are the thickness and width of the CT specimen; E is the elastic modulus of the material; Vx is the displacement of the measuring point; P is the load applied during the test; a is the crack length; C0~C5 are related parameters.
1.3 Test results
The a-N curves of the two specimens during the test were recorded. The fatigue crack growth rate formula can describe the relationship between the crack growth rate and the stress intensity factor range under cyclic loading, and is used to study the fatigue fracture mechanism of materials and predict material life. The calculation formula of stress intensity factor is based on the following standards (3), (4). In the formula, ΔP is the load change; W and B are the width and thickness of the specimen respectively; g(α) is the shape factor; α = a/W.
When selecting the method for processing actual data of crack growth rate, GB/T 6398-2017 recommends the use of secant method and seven-point increasing polynomial method, and the seven-point increasing polynomial method has a higher fitting degree. Therefore, this paper adopts the seven-point increasing polynomial method for processing. In order to fit the parameters in the Paris formula da/dN = C(ΔK)n, the calculated results (da/dN)i and ΔKi are sorted into a set of data, and linearly fitted under double logarithmic coordinates, as shown in Figure 4. The Paris parameters obtained by fitting are shown in Table 2.
1.4 Numerical simulation verification
This paper conducts numerical simulation analysis on the crack growth test of normal CT specimens to verify the feasibility and accuracy of the numerical simulation method. The material parameters of the CT specimen are the same as those of the blade. The loading conditions of the numerical simulation are consistent with those of the test. The initial maximum cyclic load is 3.5 kN, the loading stress ratio R=0.1, and the cycle life is calculated using the parameters fitted in Table 2. Figure 5 is a schematic diagram of the numerical simulation.
The normal test piece crack is selected to expand from 18 mm to 24 mm for verification. The stress intensity factor corresponding to the 24 mm crack in the test is set as the termination condition of the numerical simulation, and the number of cycles and crack extension length of the numerical simulation and the test are compared [18]. The test results show that when the crack length is 24 mm, the corresponding number of cycles is 14400, and the stress intensity factor K=28.14 MPa √m, so this value is set as the termination condition of the numerical simulation. The number of cycles and crack extension length calculated by the test and numerical simulation are shown in Figure 6. The corresponding calculation errors are 4.5% and 1.6%, respectively, proving the feasibility of the numerical simulation method.
2 Determine the dangerous parts of the turbine blade
2.1 Model establishment and mesh division
For the high-pressure turbine blade structure, a reverse modeling method based on HandySCAN3D three-dimensional scanner is used to obtain an accurate three-dimensional geometric model, thereby improving the accuracy of finite element numerical simulation. The point cloud of the measured blade is obtained by scanning the physical blade, as shown in Figure 7 (a). Then, SolidWorks software is used to establish a high-precision overall model with the point cloud data as a reference.
The blade finite element model is established using ANSYS software, with a grid size of 0.2 mm. The grid of the pore area is appropriately encrypted, as shown in Figure 7 (b), and the total number of grid units is 100×105. The blade material is DZ125 nickel-based alloy, which is an anisotropic material. The material parameters at 850℃ are shown in Table 3, and the density is ρ = 8.48g/cm3.
The external flow channel grid is divided using ANSYS ICEM software, and the grid division result is shown in Figure 8 (a), and the number of flow channel grids is 75×105. In order to make the meshing more accurate and the simulation results more accurate, local encryption is performed for the area with larger curvature, as shown in Figure 8 (b).
2.2 Boundary conditions
Figure 9 shows the general mission speed spectrum of the aircraft at the location of the turbine blade. The temperature field in the turbine blade changes with the flight state. From the analysis of the flight spectrum, the temperature field and the speed change are basically consistent. Therefore, the worst state for the analysis of the influence of cracks and oxidation on crack propagation can be simply based on the speed spectrum.
Consider the load cycle form of the aircraft as 0-max-0. The maximum cruise state is selected as the input condition for finite element analysis and crack propagation. According to the engine load spectrum data, the engine exhaust temperature (EGT) and high-pressure compressor speed (N2) are extracted. Based on the extracted data, the boundary conditions of the high-pressure turbine blade flow field under the maximum working condition are calculated, as shown in Table 4.
Set the fluid domain and solid domain in Fluent. As shown in Figure 10, the fluid domain contains both the flow area of the gas outside the blade and the flow area of the cold air inside the blade, and the solid domain is composed of the blade itself. The total pressure inlet is set at the inlet of the outer flow channel, the static pressure outlet is set at the outlet, and the mass flow inlet is set at the inlet of the inner flow channel, and the direction is radially outward.
2.3 Simulation results of turbine blades
The flow field analysis is carried out through the above boundary conditions, and the calculation results are imported into the Steady-State Thermal module to study the temperature distribution of the turbine blade. The temperature distribution of the blade basin and the back of the turbine blade is shown in Figure 11.
As can be seen from Figure 11, the temperature distribution of the coupling surface. The turbine blade as a whole presents the characteristics of high temperature and large temperature difference. The temperature at the inlet end of the flow field is significantly higher than that at the outlet end, and there is a highest temperature point on the inlet side of the blade, reaching 1086.2℃. Affected by the impact of the airflow, the leading edge temperature of the blade is higher than the trailing edge temperature, and the temperature of the blade basin area is higher than that of the blade back area. The leading edge of the blade has the highest temperature, which is the main cause of oxidation or ablation. Compared with the actual engine turbine blades in service, as shown in Figure 12, the figure was taken by a DVM6 digital microscope at 43.77 times magnification. The leading edge of the blade is the most severely ablated area, proving that the temperature in this area is the highest, which is consistent with the simulation results.
The temperature and aerodynamic stress distribution are imported into the Static Structure module, and the stress distribution of the turbine blade is studied. The stress results show that the maximum stress of the blade body is 524.8MPa near the blade root on the back of the blade. The stress in the blade basin area is higher than that in the blade back area, and the stress and strain values in the blade root area are larger. This is because the centrifugal force generated by the blade body part acts completely on the blade root area during the high-speed rotation of the blade. The maximum stress point is at the fixed constraint. The stress cloud diagram is shown in Figure 13.
3 Oxidation damage and crack extension analysis
3.1 Oxidation damage analysis
Oxidation damage of turbine blades refers to the phenomenon that the blade surface is exposed to high temperature, high pressure and high speed working environment for a long time, and the surface reacts with oxygen molecules in the air to cause oxidation and corrosion, resulting in the degradation of the performance of the blade surface and internal materials. This oxidation damage can lead to problems such as blade thinning, deformation, and cracks, which in severe cases can threaten the safety and reliability of the engine.
High-temperature oxidation of the turbine blade surface is mainly caused by direct contact between the blade surface and oxygen at high temperatures. The alloy elements on the blade surface will be oxidized and gradually detach from the alloy structure, causing the blade surface to crack and peel off. In order to slow down the oxidation damage of turbine blades, surface covering, metal spraying, composite materials and other measures are usually used to protect the turbine blades, and advanced manufacturing processes and materials are used to improve the oxidation resistance of blade materials. Since crack initiation and propagation occur after the thermal barrier coating falls off, this paper conducts research without thermal barrier coating.
Discoloration of turbine blades is a typical feature of oxidation damage, as shown in Figure 14, which is taken by a DVM6 digital microscope at 101 times magnification. The left side is the area that has experienced oxidation damage, and the right side is the area that has not undergone oxidation damage. By comparison, it is found that the left area has obviously turned yellow. Therefore, it is very necessary to accurately identify and evaluate the oxidation damage of turbine blades, understand the impact of oxidation damage on it after cracks appear, and take appropriate measures for repair and maintenance.
3.2 Franc3D crack numerical simulation
Franc3D uses M-integral to calculate stress intensity factor, and the M-integral energy expression is (5). Wherein, Γ is the integral loop around the crack tip; W(1,2) = σ(1) ij ε(2) ij = σ(2) ij ε(1) ij. The relationship between M-integral and stress intensity factor is (6). Wherein, KI, KII, KIII are the stress intensity factors corresponding to the three basic crack forms. For the case of composite crack extension, considering the three basic crack modes, the equivalent stress intensity factor is used to characterize the stress field at the crack tip, and its expression is (7), (8). Wherein, ΔKe is the equivalent stress intensity factor; βII, βIII are weight factors.
A conservation integral calculation is performed on the element ring surrounding the crack tip. The integration domain includes an inner ring of a 15-node singular wedge element and an outer ring of a 20-node hexahedral element, as shown in Figure 15.
The crack propagation direction can significantly affect the simulation results. This article adopts the maximum circumferential stress criterion. The maximum circumferential stress criterion defines that cracks will propagate along the direction perpendicular to the maximum circumferential stress σθ. The condition for determining the crack propagation direction is (9), and the expressions for calculating the crack propagation cracking angle are (10) and (11). The step length of crack propagation is also an important factor affecting the accuracy of crack propagation.
white. In this paper, the crack expansion step is set by specifying the expansion distance at the midpoint of the crack front, and the crack expansion step is obtained based on the proportional relationship between the growth rate of the stress intensity factor at each point of the crack front and the midpoint (12).
Next, we study the effects of different initial crack locations and oxidation damage on crack growth. The cross-sectional position and stress at this position are shown in Figure 16. Because cracks are not allowed to occur in the blade basin and blade back of this type of turbine blade, combined with the analysis of stress distribution, this article will study the crack propagation problem on the dangerous sections (sections 3, 4 and 5) of the leading edge and trailing edge. Sections 3, 4 and 5 are named A1, A2 and A3 at the leading edge nodes and B1, B2 and B3 at the trailing edge nodes.
3.3 Analysis of cracks appearing on the leading edge of blades
When inspecting turbine blades for cracks, inspectors can detect cracks with a length of 0.05mm and above, so this article inserts an initial crack with an initial crack size of r=0.1mm on the leading edge. The initial crack setup is shown in Figure 17.
The leading edge is shown using the A3 area (the area with the maximum stress) as an example. The calculation methods for the A1 and A2 areas are the same and will not be described again. Stress intensity factors are calculated from the resulting stresses using M-integration. The crack front is normalized, as shown in Figure 18, and the calculated initial crack stress intensity factor is shown in Figure 19.
It can be seen from Figure 19 that the initial stress intensity factor of type I is not much larger than that of type II and type III. The crack is a mixed type crack in the early stage. The key to crack expansion lies in the calculation of the expansion direction and step length. The crack expansion direction is determined by Equation (11), and the expansion step length is determined by Equation (12). After calculating the crack expansion angle and expansion step length, a new crack front is obtained. After re-gridding, the above calculation is repeated until stopped.
The crack developed into a penetrating crack after 18 steps of propagation. After that, two crack fronts appeared when the crack continued to develop. Similarly, the respective crack fronts were normalized. It stops after 32 steps of expansion. At this time, the length of the crack on the left side reaches 2.095mm, the length of the crack on the right side reaches 1.954mm, and the total crack is 4.05mm. The front edges of each crack are shown in Figure 20.
After the subsequent crack propagates and develops into a penetration crack, the stress intensity factor of type I is much larger than that of type II and type III, so only the stress intensity factor of type I is displayed. Take the KI value where the normalized coordinate of the crack front is 0.5. Because the crack becomes a penetration crack at step 18 and two crack fronts appear, two KI values appear after that, and each K value is responsible for the crack expansion of the respective crack front. The relationship between KI and crack length is shown in Figure 21.
After obtaining the relationship between crack length and stress intensity factor, the simplified load cycle was compiled into Franc3D, and the Paris formula (da/dN = C(ΔK)n) was used to conduct crack growth life analysis. The parameters are listed in Table 2 given. The Paris parameters before oxidation are used to calculate the crack growth life of the turbine blade without oxidation damage, and the parameters after oxidation are used to calculate the crack growth life with oxidation damage. The calculation results are shown in Figure 22.
The results of crack propagation at other positions on the front edge are shown in Table 5. The cycle life of the turbine blade without oxidation damage and the cycle life of the turbine blade under the influence of oxidation damage were calculated. Through comparison, it was found that under the influence of oxidation damage, the crack growth life of the dangerous section of the leading edge was reduced on average to that without oxidation damage. 44.02%.
3.4 Analysis of cracks appearing on the trailing edge of blades
Due to the presence of high temperature gradients, cracks tend to develop around the air slots (part of the cooling channels). Unlike the cracks at the leading edge, the cracks at the trailing edge are usually through cracks, so the initial crack setup is shown in Figure 23 with the initial crack size r= 0.5mm.
The trailing edge is demonstrated using the B3 area (maximum stress area) as an example. Similarly, the crack front is normalized, as shown in Figure 24, and the calculated initial crack stress intensity factor is shown in Figure 25.
It can be seen from Figure 25 that the initial stress intensity factor of type I is much larger than that of type II and type III, and this crack belongs to a type I opening crack. It stops after 28 steps of expansion. At this time, the crack length is 7.35mm, and the crack front is shown in Figure 26.
During subsequent crack expansion, the stress intensity factor of type I is also much larger than that of type II and type III, so only the stress intensity factor of type I is displayed. Taking the KI value where the normalized coordinate of the crack front is 0.5, the relationship between KI and crack length is shown in Figure 27.
After obtaining the relationship between crack length and stress intensity factor, the crack growth life is analyzed. Each parameter is given in Table 2, and the calculation method is consistent with the previous section. The crack growth life curve is obtained as shown in Figure 28.
The results of cracks at different positions on the trailing edge are shown in Table 6. The cycle life of turbine blades without oxidation damage and the cycle life of turbine blades under the influence of oxidation damage were calculated. Through comparison, it was found that under the influence of oxidation damage, the crack growth life of dangerous sections at the trailing edge was reduced on average to that without oxidation damage. 50.22%. Compared with the leading edge, oxidation damage has less impact on crack growth at the trailing edge.
4 Conclusions
This paper obtains the material parameters of turbine blade crack propagation through experiments, and uses Franc3D and ANSYS software for joint simulation to compare the effect of oxidation damage on turbine blade crack propagation:
(1) The crack propagation rate of directionally solidified nickel-based superalloy DZ125 at 850℃ was studied, and the crack propagation rate of the material without oxidation damage and with oxidation damage was obtained.
(2) When cracks appear in the blade, the closer the crack is to the root area, the shorter its cycle life.
(3) Using Franc3D to calculate the crack propagation of the leading edge of the blade is more affected by oxidation damage than the trailing edge. The life of the leading edge under the influence of oxidation damage is reduced to 44.02% of that without oxidation damage; the life of the trailing edge under the influence of oxidation damage is reduced to 50.22% of that without oxidation damage.
