In view of the crack failure of the compressor rotor blade during the test run of the turbofan engine, this paper analyzes the strength and vibration characteristics of the blade and concludes that the blade has the risk of resonance during operation. The vibration of the blade was measured by non-contact measurement method, and the vibration stress of the rotor blade without corrosion pits was obtained. The working reliability of the blade in the non-corrosion state was verified by carrying out high-cycle fatigue test of the blade. The crack extension threshold of the blade in the corrosion state was reversed by Pairs formula, and the cause of blade cracking was analyzed. The main reason for the fracture is that the blade first forms corrosion pits, and then fails due to corrosion fatigue under the action of high-cycle alternating loads. This paper focuses on the improvement from the process perspective, controls the material tempering temperature range, adds the aluminum infiltration process on the blade surface, effectively prevents the blade from fracture failure, and improves the working reliability of the blade.

The blade is one of the important parts of the engine, and undertakes the important task of converting thermal energy into mechanical energy. Due to its high speed, large load, and complex working conditions, it is easy to fail during operation. Aviation stainless steel has high strength, good plasticity, toughness and fatigue resistance, and is low in price. It is widely used in the aviation industry to manufacture engine blades. Steam turbines used in marine aircraft and ships, due to meteorological factors in the marine atmosphere, such as high temperature, high humidity, high salinity, and many foggy areas, will directly affect the corrosion behavior of steel alloy materials. Engine blades are very prone to stress corrosion and fatigue corrosion, which not only reduces the efficiency of the engine, but also increases maintenance time and costs.

Stress corrosion cracking is often a brittle fracture that occurs without any obvious macroscopic deformation. Once formed, stress corrosion cracks expand faster than other types of local corrosion, and are the most destructive type of corrosion known to date. Statistical results show that stress corrosion cracking of stainless steel ranks first among fracture failures, accounting for more than 50%. For decades, scholars in related fields around the world have been committed to the research of corrosion fatigue of high-strength alloy steel structures, laying a solid foundation for the experimental development and mechanism exploration of corrosion fatigue of such structures. For example, Liu et al. studied the corrosion fatigue characteristics of 38CrMoAl high-strength steel and found that corrosion damage would first appear in the local plastic zone of the specimen, accelerating the initiation of fatigue cracks. Guo Hongchao studied the fatigue performance of Q690 high-strength steel in a corrosive environment and found that the fatigue limit decreased by 30.15% and 38.89% when the corrosion cycle was 60 d and 100 d, respectively. Jing Yongzhi summarized the relevant research on engine blade protective coatings serving in marine environments and summarized the design concept of blade protective coatings.

Aiming at the stress corrosion fracture phenomenon of the first-stage rotor blade of the compressor during the test run of a certain type of engine, this paper analyzed the steady-state stress and vibration characteristics within the working envelope of the blade and concluded that the blade had a resonance risk below the slow speed; carried out a blade vibration monitoring test based on non-contact strain measurement, and obtained the vibration stress of the rotor blade without corrosion pits; combined with the measurement results of the blade high-cycle fatigue test, the working reliability of the blade in the non-corrosion state was verified; the Pairs formula was used to reverse the crack extension threshold of the blade in the corrosive state, and the cause of the blade cracking was analyzed. The analysis results were consistent with the conclusions of the fracture analysis, verifying the effectiveness of the analysis. Corresponding protective measures were taken, and the feasibility of the measures was verified through experiments.

1 Fault Overview
The first-stage blade disk and front journal of a turbofan engine compressor are integrated, using 1Cr12Ni2WMoVNb heat-resistant steel die forging and integral CNC machining. After about 177 hours of test run, it was found that all blades had unevenly distributed pits of various sizes from the root to the tip of the blade, and one blade had a crack. The crack length is about 8.3 mm, located near the inlet edge, about 4.8 mm from the edge plate, and the appearance of the cracked blade is shown in Figure 1.
The macroscopic morphology of the fracture source area is shown in Figure 2, where typical fatigue arcs and radial ridges can be seen. The source area is black within about 0.2 mm, indicating that there are corrosion products in the source area. The extended area is gray-black and light yellow, and a large number of fatigue arcs can be seen.

2 Cause Analysis
In order to further clarify the cause and mechanism of the failure, static strength analysis, vibration analysis, crack extension analysis and fracture analysis were carried out on the first-stage rotor blade of the compressor.

2.1 Static Strength Analysis
According to the cyclic symmetric structural characteristics of the first-stage blade of the compressor, a 1/31 disk body and a complete blade were taken as the calculation model, and the static strength analysis was performed using the ANSYS software platform. The axial and circumferential degrees of freedom of the bolt hole nodes of the web were constrained, and the load took into account the temperature, speed and aerodynamic force. The cyclic symmetric boundary conditions were applied on the cyclic symmetry surface. The finite element model is shown in Figure 3, and the stress distribution of the blade body under the maximum working state is shown in Figure 4. The calculation results show that the stress in the middle area of ​​the back root of the blade is the largest, and the stress at the crack initiation of the blade is relatively low, which meets the strength design requirements.

2.2 Vibration Analysis
The modal analysis of the first-stage rotor blade of the compressor was carried out. The first-order vibration mode and relative vibration stress distribution of the blade are shown in Figure 5. It can be seen from Figure 5 that the position of the first-order maximum vibration stress coincides with the position of the blade crack. The resonance speed diagram of the blade is shown in Figure 6.

Among them, the excitation orders that need to be analyzed are: K = 1, 2, 3, 4, corresponding to the inlet airflow distortion and low-order excitation of the engine; the number of the front-stage guide blades is 38, and the number of the rear-stage guide blades is 52. As shown in Figure 6, within the engine operating speed range, there is a resonance point between the K = 3 times excitation line and the first-order natural frequency line of the blade. The corresponding engine operating speed is the slow speed, the resonance point is below the slow speed, and the resonance margin is 5.4%.

In order to verify the first-order resonance risk of the blade under K=3 times excitation, the vibration of the first-stage rotor blade of the compressor was measured using a non-contact blade vibration measurement system. The first-stage rotor blades of the compressor were inspected before the test, and no corrosion pits were found.
In order to measure the maximum vibration stress that may occur in the blade within the envelope range, the test run considered the combination of different guide vane opening angles and inlet temperature conditions, and a total of 6 combination state tests were carried out. The speed test spectrum is shown in Figure 7.

The basic principle of non-contact strain testing is divided into two steps: the first step is to test the tip amplitude value of the blade under the resonance condition under the actual working condition of the blade; the second step is to calculate the strain result of the required strain measurement point at resonance based on the conversion relationship between the blade strain and the tip amplitude. The vibration displacement, resonant speed and frequency results of the blade in the first cycle speed-up process under state 1 are shown in Figure 8. The horizontal axis in the figure is the blade number, and the vertical axis is displacement, resonant speed, and resonant frequency from top to bottom. The first-order vibration stress of the blade obtained after conversion is shown in Table 1.

Referring to HB 5277-84, the high-cycle vibration fatigue limit of the blade was measured by the lifting method, and 15 valid data were obtained. The 107-cycle fatigue limit-3σ value of the blade with a 5% error limit (i.e., 95% confidence level, 99.73% survival rate) was 485MPa. The high-cycle fatigue reserve analysis using the fatigue limit-3σ value of the blade is shown in Figure 9, where the ordinate is the vibration stress and the abscissa is the steady-state stress. As can be seen from Figure 9, the vibration stress at the crack of the blade is distributed below the Goodman curve with a fatigue reserve of 1.7, and the high-cycle fatigue reserve calculated using the maximum vibration stress is 5.2, so the blade will not suffer high-cycle fatigue damage.

2.3 Crack propagation analysis
In order to determine whether the blade may undergo fatigue propagation under the action of high-cycle alternating loads, a crack propagation analysis of the blade is now performed.
The fatigue crack growth law is shown in Figure 10. It can be seen from Figure 10 that there are three regions between the fatigue crack growth rate da/dN and the stress intensity factor ΔK.
a) The first region is the slow fatigue crack growth stage. There is a fatigue crack growth threshold value ΔKth. When ΔK is lower than ΔKth, the fatigue crack does not grow or grows extremely slowly;
b) The fatigue crack growth in the second region follows the power function law. The fatigue crack growth rate da/dN can be expressed by the power function of the stress intensity factor amplitude ΔK. The Paris formula is widely used to express it;
c) The third region is the rapid growth stage. When the crack slowly grows close to or reaches KIC (1 – R), the crack grows rapidly. As can be seen from Figure 1, the blade crack starts from the corrosion pit, and the fatigue crack growth is generated in the local area near the tip of the corrosion pit. Vibration analysis shows that the first-order vibration stress at the blade crack is the tensile stress along the intake edge, and the initial crack belongs to the I-type crack. The stress field and displacement field near the I-type crack tip can be simplified as: See formula (1) and (2) in the figure.
Where: KI is the stress intensity factor of the I-type crack tip; r is the polar radius of the crack tip in polar coordinates; fij(I) (θ) and g(ijI) (θ) are the stress function and displacement function respectively.
According to linear elastic fracture mechanics, the expression of the stress intensity factor is as shown in formula (3), where: Δσ is the stress amplitude; a is the crack size; Y is the shape coefficient. Since the shape of the corrosion pit is approximately an elliptical surface crack, the shape coefficient Y is taken as 1.12. Transform formula (3) to obtain (4).
Where: a0 is the critical crack size for fatigue cracking. If the crack size is less than a0, fatigue cracking will not occur in the blade.
For martensitic steel, Barsom obtained the following empirical relationship (5). Where: R is the stress ratio. That is, as the stress ratio increases, the threshold value of the stress intensity factor of martensitic steel will decrease.

The sample data of the measured vibration stress are statistically analyzed, and the frequency distribution of the blade vibration stress is analyzed. The histogram of the vibration stress frequency distribution is shown in Figure 11. As shown in Figure 11, the vibration stress distribution conforms to the normal distribution, and the fitting curve obeys the X~N (36.86, 323.336) distribution. The vibration stress +3σ value (i.e. 95% confidence level, 0.13% survival rate) is calculated to be 88 MPa.

Based on the vibration stress +3σ value and the steady-state stress at the resonant speed of the blade, the stress ratio R at the crack initiation of the blade is calculated to be 0.2. From formula (5), it can be calculated that the threshold value of the stress intensity factor ΔKth corresponding to the stress ratio R of 0.2 is 5.31 MPa·m1/2. From formula (4), it can be calculated that the critical crack size a0 of fatigue cracking is 0.23 mm. The depth of the corrosion pit measured comprehensively is 0.25 mm. From the above calculation, it can be seen that when the vibration stress takes the value of +3σ, the depth of the corrosion pit can reach the critical crack size and the crack will expand. Since the vibration stress distribution follows the normal distribution, the vibration stress part less than the value of +3σ cannot meet the conditions for crack expansion. Analysis shows that this is related to the reduction of material properties after the blade is corroded.

Since the corrosive environment will reduce the stress intensity factor of the metal material, making the blade more prone to cracking, this stress intensity factor amplitude is called the corrosion fatigue stress intensity factor amplitude threshold value, represented by ΔKthCF. Now the stress intensity factor threshold value of the blade under the corrosive environment is reversed. Assuming that the critical crack size of the blade is 0.25 mm, the average value of the vibration stress is 36.86 MPa and formula (3) are used to calculate the stress intensity factor threshold value of the blade under the corrosive environment to be 2.31MPa·m1/2. The analysis shows that the corrosive environment reduces the threshold value of the stress intensity factor of the blade. When the stress intensity factor at the crack initiation point of the blade reaches the threshold value of crack extension in the corrosive environment, the corrosion fatigue crack initiates, and then fatigue extension occurs.

2.4 Fracture analysis
The fracture analysis of the cracked blade shows the microscopic morphology of the fracture source area in Figure 12. Typical intergranular features can be seen in the source area, and fine corrosion pit morphology can be seen on the grain surface. The fracture micromorphology is shown in Figure 13. The crack extends toward the exhaust edge, and typical fatigue band features can be seen before, during, and after the extension.

A metallographic sample was cut from the cracked blade parallel to the crack direction. The sample was ground and polished to observe the microstructure. The morphology is shown in Figure 14. As can be seen from Figure 14, a large number of intergranular cracks can be seen on the intake edge of the cracked blade. The crack depth is relatively shallow, about 0.25mm, and fine intergranular cracking characteristics can be seen near the grain boundary, indicating that the pits on the intake edge of the blade are caused by corrosion.

Energy spectrum analysis at the grain boundary shows that the fracture source area mainly contains corrosive elements such as O, S, and C, and there is also a certain amount of O element in the expansion area. There are also corrosive elements such as S and O in other pit areas and surfaces of the blade, see Table 2.

The fracture analysis results show that the pits on the blade inlet edge and the fracture source area along the grain are caused by corrosion. From the perspective of the degree of corrosion damage and the cracking position, the crack source area is basically close to the blade root, indicating that the fatigue extension of the blade is not only related to the degree of surface corrosion damage, but also to the relatively large vibration stress borne by this position during operation. The blade may first undergo corrosion cracking along the grain, and then fatigue extension occurs under the action of working stress.

3 Comprehensive Cause Analysis

The reasons for blade failure and fracture are summarized as follows: rotor blades often work in coastal and inland humid and hot areas. The atmosphere contains high levels of corrosive media such as sulfur and chlorine, and the pH value is low. Under the influence of the environment, the blades first corrode, and uneven pits and holes are formed on the air inlet edge. The formation of corrosion pits produces local stress concentration, so the corrosion fatigue cracks of the blades originate from the corrosion pits.

Corrosion greatly weakens the bonding force between material grains and reduces the threshold value of the stress intensity factor of the material. Under the action of high-cycle vibration stress, the corrosion pits begin to transform into cracks. When the stress intensity factor value of the equivalent crack at the corrosion pit of the blade reaches the threshold value of the stress intensity factor for the extension of corrosion fatigue cracks, the corrosion fatigue cracks initiate. Afterwards, under the combined action of the corrosive environment and high-cycle alternating loads, the corrosion fatigue cracks are promoted to expand, and finally the corrosion fatigue failure of the blades is caused.

4 Improvement measures and verification

4.1 Improvement measures
Since the rotor blades meet the structural and aerodynamic performance requirements in terms of structural design, the following two improvements are considered from the process perspective:
a) During the forging process of the forging, the tempering temperature is controlled to improve the corrosion resistance of the material;
b) Low-temperature aluminizing process is added to the blade surface to improve the corrosion resistance of the blade.

4.2 Verification of measures
In order to verify the effectiveness of the measures, salt spray corrosion tests on the same material specimens were carried out. According to the requirements of GJB150.11A-2009[19], the specimens were designed, and the dimensions are shown in Figure 15. Three specimens tempered at 590℃ without aluminizing, three specimens tempered at 580℃ without aluminizing, and three specimens tempered at 580℃ with aluminizing were taken for salt spray corrosion tests, and the influence of aluminizing process and tempering temperature on the salt spray corrosion resistance of 1Cr12Ni2WMoVNb material was explored. The test parameters of the test process are shown in Table 3, and the appearance of the test piece after 96 hours of salt spray corrosion is shown in Figure 16.

The test results show that the corrosion resistance of the 580 ℃ tempered sample is significantly better than that of the 590 ℃ tempered sample; the aluminized layer significantly delays the corrosion of the substrate and plays a role in resisting salt spray corrosion.
After implementing the above improvement measures, the rotor blades of the engine that has reached the end of its service life were disassembled and inspected, and no corrosion or fracture occurred, indicating that the measures were verified to be effective.

Conclusion

Related research was carried out on the corrosion and fracture of blades during the test of a certain type of engine, and the following conclusions can be drawn:

According to the simulation analysis, it can be seen that the blades have resonance below the slow speed; according to the non-contact strain measurement test of the whole machine and the high-cycle fatigue test measurement results of the blades, it can be proved that the blades work reliably in a non-corrosive state.

The main reason for the fracture is that the blade is corroded first, and the corrosion reduces the fatigue crack extension threshold of the material. When the stress intensity factor value of the equivalent crack at the corrosion pit of the blade reaches the stress intensity factor threshold value of the corrosion fatigue crack extension, the corrosion fatigue crack is initiated, and then fatigue failure occurs under the action of high cycle alternating load. The influence of corrosion on the fatigue crack extension threshold is related to the corrosive medium, the organization and properties of the material, temperature, stress ratio and load form, which is relatively complex and requires further in-depth research.

It is necessary to pay attention to the design concept of the blade protective coating. For example, the low-temperature aluminizing process can effectively improve the corrosion resistance and service life of the blade. However, the low-temperature aluminizing process may affect parameters such as the crack extension threshold, and the extent of its influence requires in-depth research through relevant experiments.