During the commissioning test of a central transmission, the teeth of the active spiral bevel gear were broken, causing varying degrees of damage to other components. The contact marks of the active spiral bevel gear were inclined towards the small end, close to the tooth bottom, while those of the driven bevel gear were distributed at the top of the small end teeth, with incomplete marks partially beyond the working face. To determine the cause of the fracture failure in the 15Cr14Co12Mo5Ni2 steel spiral bevel gear, various analyses, including macro-inspection, fractographic analysis, metallographic examination, hardness testing, composition analysis, and simulation and verification of contact patterns, were conducted. The results showed that the fracture of the active spiral bevel gear was caused by high-cycle fatigue, originating from the corner between the small end face and the convex root of the tooth, unrelated to material and metallurgical defects. Abnormal occlusion during operation was identified as the primary cause of early fatigue cracking in the gear. It is recommended to optimize the processing parameters of the gear pairs and improve the meshing quality to avoid similar failures.
1. Introduction
With the increase in thrust and thrust-to-weight ratio of aeroengines, the power and speed transmitted by gears have continually increased, necessitating higher temperatures on the gear meshing surfaces. This poses more stringent requirements on the materials used for gear manufacturing. Spiral bevel gears are widely used in engine central transmission systems due to their smooth transmission, strong load-bearing capacity, and compact structure. The central transmission spiral bevel gear serves as a pivotal component for power transmission; its failure can lead to the loss of function of the entire transmission system and the interruption of flight missions.
Failure modes of aviation bevel gears often manifest as pitting, scuffing, wear, plastic deformation, and fracture [1-3]. Scholars at home and abroad have conducted in-depth research on multiple gear fracture failures that have occurred during engine operation. Pierre et al. [4] analyzed the fracture failure of a spur bevel gear in a GE90-11B engine gearbox and attributed it to the combined effect of residual tensile stress from local decarburization and working stress, later resolved by shot peening. Chen Conghui [5] and Song Lemin [6] studied the failure mechanisms of bevel gears in aviation engines and proposed solutions based on their findings.
However, there have been no reports on fatigue failures caused by abnormal loads on spiral bevel gears due to edge contact, and it is difficult to define the differences from traditional resonance failure modes. During a central transmission commissioning test, a fracture failure occurred in the teeth of the active spiral bevel gear. The failed gear was made of the 3rd generation new low-carbon high-alloy steel 15Cr14Co12Mo5Ni2, which is known for its high temperature resistance, corrosion resistance, and excellent strength and toughness. Similar failures have not been reported previously for gears made of this material, making it essential to investigate the failure mechanism to further improve design and manufacturing processes. This paper analyzes the failed spiral bevel gear through macro-inspection, fractographic analysis, metallographic examination, composition analysis, hardness testing, and simulation and verification of contact patterns to determine its failure mode and causes and proposes corresponding improvement measures.
2. Test Procedures and Results
2.1 Macro-Inspection
The macroscopic appearance of the failed spiral bevel gear is shown in Figure 1. The active spiral bevel gear has four entire teeth broken off at the web below the teeth, which can be roughly reassembled into a complete form, while the driven spiral bevel gear remains intact. The concave surface of the active spiral bevel gear and the convex surface of the driven spiral bevel gear serve as the working surfaces of the gear pair, with contact marks clearly visible on both sides. The contact marks are elongated and eyebrow-shaped. The contact marks on the entire circumference of the failed active spiral bevel gear are uneven, biased towards the small end, and close to the tooth bottom. The mark morphology can be roughly divided into three types, as shown in Figure 2. The first type is tooth #5 (cracked tooth), with a mark width of approximately 1.0 mm, a length of approximately 23.0 mm, and a distance of about 5.0 mm from the tooth tip. The second type includes teeth #6 to #9 (broken teeth), with a mark width of about 2.0 mm, a length of about 20.0 mm, and a distance of about 5.3 mm from the tooth tip. The third type includes the remaining teeth (other than teeth #5 to #9), with a mark width of about 1.7 mm, a length of about 18.0 mm, and a distance of about 5.3 mm from the tooth tip. The contact marks on the driven spiral bevel gear are distributed at the top of the small end teeth, incomplete, and partially beyond the working surface.
(a) Gear Pair
(b) Active Spiral Bevel Gear
(c) Driven Spiral Bevel Gear
Figure 1: Local Macroscopic Morphology of the Failed Gear
(a) Cracked Tooth
(b) Broken Tooth
(c) Other Teeth
Figure 2: Macroscopic Morphology of the Mating Marks on the Concave Surface of the Active Spiral Bevel Gear
2.2 Fractographic Analysis
The broken tooth fracture surfaces of the failed active spiral bevel gear can be divided into radial 1, circumferential, and radial 2 fractures. The radial 1 and circumferential fracture surfaces are delicate, with visible fatigue stripes and radiating ridges, indicating that the broken tooth fracture surfaces are of fatigue nature, as shown in Figure 3(a). The direction of the fatigue stripes and radiating ridges indicates that the fatigue originated from the corner between the small end face and the convex root of the tooth, starting from a point source, as shown in Figure 3(b). The circumferential fracture surface is formed by the fatigue propagation of the radial 1 crack. The radial 2 fracture surface is rougher, accounting for less than 20% of the entire broken tooth fracture surface area, and is considered an instantaneous fracture surface formed in the later stage of fatigue propagation. As visible in the figure, the crack first propagates from the small end face to the large end face and then simultaneously along the axial and radial directions. It then propagates circumferentially along the rotation direction after reaching the web transition area, crossing four teeth before finally causing overload fracture and tooth loss along the radial direction.
(a) Overall View
(b) Enlarged Source Area
Figure 3: Macroscopic Morphology of the Broken Tooth Fracture Surface
A radial 1 fracture surface was cut out using wire electrical discharge machining and observed under a scanning electron microscope. The fracture surface shows clear fatigue stripes and radiating ridge characteristics, further indicating that the fatigue originated from the corner between the small end face and the convex root of the tooth, starting from a point source. The source area exhibits obvious damage traces but no significant metallurgical defects, as shown in Figure 4(a) and (b). The area near the source of the broken tooth fracture surface mainly exhibits granular carbide morphology at high magnification, with local visible fatigue striations, as shown in Figure 4(c). The propagation zone of the fracture surface shows a dense fatigue striation morphology, as shown in Figure 4(d). These features collectively indicate that the fracture of the failed gear is of high-cycle fatigue nature.
(a) Low Magnification of Source Area
(b) Enlarged Source Area
(c) High Magnification Near Source Area
(d) High Magnification of Propagation Zone
Figure 4: Microscopic Morphology of the Radial 1 Fracture Surface of the Broken Tooth
2.3 Microstructural Examination
A complete double tooth was cut from the area near the fracture surface of the failed active spiral bevel gear for microstructural observation. The carbide grade of the tooth profile carburized layer is 3, meeting the standard requirement (grades 1-4), as shown in Figure 5(a). Deformation features can be observed on the surface of the concave meshing area, as shown in Figure 5(b). The core microstructure consists of tempered martensite and a small amount of retained austenite, with no abnormalities observed, as shown in Figure 5(c). White blocky structures exist at the interface between the carburized layer and the matrix, as shown in Figure 5(d). Microhardness tests conducted at this location yielded hardness values of 450, 507, and 445, significantly lower than the core hardness of 552.6, indicating that these white blocky structures are retained austenite.
(a) Tooth Profile
(b) Meshing Area of Concave Surface
(c) Core
(d) Transition Zone of Carburized Layer
Figure 5: Microstructural Morphology of the Failed Spiral Bevel Gear
2.4 Hardness Testing
Hardness tests were conducted on the concave surface of the failed active spiral bevel gear, yielding results of HRC = 63.1, 63.4, and 63.2 (specification requirement: HRC = 60-67). Hardness tests on the core yielded results of HRC = 50.1, 51.5, and 51.4 (specification requirement: HRC = 48-52). Single-point microhardness tests were also performed on the carburized layer region of the tooth fillet radius of the failed active spiral bevel gear, with the results shown in Figure 6. The figure reveals the presence of a “soft spot” zone at the transition between the carburized layer and the core microstructure, corresponding to the location of the white blocky structures (as shown in Figure 5(a) and (b)).
Figure 6: Microhardness Curve of the Carburized Layer in the Tooth Fillet Radius
2.5 Composition Analysis
The energy dispersive spectroscopy (EDS) analysis results of the matrix of the failed bevel gear are shown in Table 1. The mass fractions of the main alloy elements do not show significant abnormalities.
Table 1: EDS Analysis Results of the Matrix of the Failed Driven Bevel Gear (wt %)
| Element | V | Cr | Co | Fe | Ni | Mo |
|---|---|---|---|---|---|---|
| Mass Fraction | 0.61 | 13.1 | 13.0 | 66.7 | 2.0 | 4.5 |
2.6 Simulation and Verification of Meshing Marks
In the technical specifications for spiral bevel gears, the contact patterns of gear pairs are divided into two types: static contact patterns and dynamic contact patterns. Static contact patterns are obtained on the working surfaces of the gears after the installed gear pair is rotated under slight braking force on a rolling inspection machine. Dynamic contact patterns expand upon the original static contact patterns within the elastic range of the gear material after loading. The design requirements for the color meshing marks of the failed spiral bevel gear specify that the marks should form an uninterrupted ellipse always within the tooth surface and without edge contact. The inspection positions include the theoretical meshing installation position and eight inspection points, totaling nine positions. For the theoretical meshing installation position inspection, the gear pair is adjusted to the theoretical meshing installation position (where the cone apexes of the large and small gears coincide). The eight inspection points are set primarily considering the changes in the working position of the gears due to deformation and bearing clearance variations under the axial force of the engine rotor.
Based on the grinding and inspection requirements specified in the color application instructions, commercial software and finite element software were used to simulate and analyze the tooth surface contact patterns, which were then verified using a dedicated gear rolling inspection machine. The analysis and verification results are shown in Figures 7-9. The results indicate that the static contact patterns analyzed by the commercial software are generally consistent with the rolling inspection results of the actual parts, with the tooth surface contact patterns at individual inspection positions close to the tooth edge, making them prone to edge contact under working loads. The dynamic contact patterns analyzed by both software packages are generally consistent. The contact patterns on the tooth surface of the active spiral bevel gear at the theoretical position and position 1 are far from the tooth edge, with a contact stress of approximately 1100 MPa. However, at position 8, the contact patterns are close to the tooth edge, resulting in edge contact with a significantly increased contact stress of 1600-2000 MPa.
(a) Static Simulation by Commercial Software
(b) Actual Static
(c) Dynamic Simulation by Commercial Software
(d) Finite Element Dynamic Simulation
Figure 7: Simulation and Verification Results of Meshing Marks at the Theoretical Position (Failure Inspection)
(a) Static Simulation by Commercial Software
(b) Actual Static
(c) Dynamic Simulation by Commercial Software
(d) Finite Element Dynamic Simulation
Figure 8: Simulation and Verification Results of Meshing Marks at Position 1
(a) Static Simulation by Commercial Software
(b) Actual Static
(c) Dynamic Simulation by Commercial Software
(d) Finite Element Dynamic Simulation
Figure 9: Simulation and Verification Results of Meshing Marks at Position 8
3. Analysis and Discussion
3.1 Fracture Analysis
The fracture inspection of the failed active spiral bevel gear reveals that the crack originated from the corner between the small end face and the convex root of the tooth, starting from a point source. It first propagated from the small end face to the large end face and then simultaneously along the axial and radial directions. After reaching the web transition area, it propagated circumferentially along the rotation direction, crossing four teeth before finally causing overload fracture and tooth loss along the radial direction. The propagation zone accounts for approximately 80% of the entire fracture surface area, indicating sufficient fatigue propagation. Microscopically, the propagation zone exhibits extensive tearing ridge texture features and dense fatigue striation morphology, combined with the criteria for high-cycle fatigue fractures, indicating that the failure nature of the crack in the failed active spiral bevel gear is high-cycle fatigue.
3.2 Macroscopic Examination Analysis
The macroscopic examination of the failed active spiral bevel gear shows that the contact marks on its entire circumference are uneven, biased towards the small end, and close to the tooth bottom. The contact marks on the driven spiral bevel gear are distributed at the top of the small end teeth, incomplete, and partially beyond the working surface, suggesting that edge contact occurred during the operation of the failed spiral bevel gear pair.
The shape, size, and position of the contact area on the tooth surface of a spiral bevel gear directly affect its smooth operation, load distribution, service life, and noise. The meshing mark is an important indicator for evaluating the meshing quality of spiral bevel gears and is a crucial issue in their development. A company pioneered the technology for forming spiral bevel gear processing parameters based on the principle of local conjugation and adopted tooth contact analysis (TCA) technology to simulate the meshing process of spiral bevel gears to obtain tooth surface contact patterns and transmission error curves, thereby predicting the rationality of the obtained spiral bevel gear processing parameters. The geometric and processing parameters of the active spiral bevel gear are determined, and a reference point on the large gear tooth surface is selected. Using the principle of local conjugation, a point conjugate to the reference point of the driving gear is found on the driven spiral bevel gear, and the normal curvature and normal vector of this point are calculated to derive the processing parameters of the driven spiral bevel gear. In actual production and processing, based on the rolling inspection test results, it is often necessary to adjust the parameters repeatedly to obtain a suitable contact pattern and good meshing performance. This shows that the color meshing marks of spiral bevel gears are closely related to the adjustment of processing parameters. The simulation and verification of the typical inspection position meshing marks on the working surface of the failed active spiral bevel gear reveal that the static contact patterns of the gears processed according to the current detailed design parameters of the cone gear tooth surface are at the tooth profile edge at individual typical inspection positions. During operation, under the action of load, the dynamic contact patterns on the concave surface of the active spiral bevel gear move towards the small end and the tooth root, resulting in edge contact and increased stress at the tooth root. This periodic and reciprocating abnormal occlusion promotes the initiation of fatigue cracks at the corner between the small end face and the convex root of the tooth. Insufficient consideration of the gear tooth surface processing parameters is the primary cause of the failure. Therefore, optimizing the gear tooth surface processing parameters, simulating the meshing marks, and conducting experimental verification to ensure that the gears fully meet the design requirements for color meshing marks after trial production and improve the meshing quality during gear operation are the main measures to avoid such failures.
3.3 Material Analysis
The metallographic structure and microhardness test results of the failed active spiral bevel gear indicate the presence of a “soft spot” zone at the transition between the carburized layer and the core microstructure (approximately 0.8 mm from the surface), corresponding to the location of retained austenite structures. This phenomenon is likely due to the difficulty in carburizing the 15Cr14Co12Mo5Ni2 material due to its high alloy element content. During prolonged carburization, if the carbon potential is too high and the diffusion time is insufficient, it can lead to the accumulation of carbon and alloy elements in the carburization transition zone, resulting in retained austenite bands (white blocky structures) and the formation of hardness soft spots. Therefore, controlling and adjusting the carburization process and parameters to reduce the stability of austenite during cooling is the most effective way to eliminate and reduce the retained austenite in the carburized layer of 15Cr14Co12Mo5Ni2. The crack in the failed active spiral bevel gear originated from the surface of the corner between the small end face and the convex root of the tooth, with no significant retained austenite or other structural defects observed in the source area. The base material composition, structure, and hardness meet the standard requirements, indicating that they are not significantly related to the initiation of fatigue cracks in the gear.
4. Conclusion
- The fracture of the failed active spiral bevel gear is of high-cycle fatigue nature and is unrelated to material and metallurgical defects.
- Insufficient consideration of the gear tooth surface processing parameters results in abnormal occlusion during gear operation and increased stress at the tooth root.
- The fractographic examination of the failed active spiral bevel gear revealed that the crack initiation was located at the corner between the small end face and the convex root of the tooth, displaying a point-source characteristic. The crack initially propagated from the small end face towards the large end face, then simultaneously along the axial and radial directions. It subsequently extended circumferentially along the rotation direction after reaching the web transition area, traversing four teeth before finally fracturing along the radial direction due to overload. The fatigue propagation zone accounted for approximately 80% of the entire fracture surface, indicating sufficient fatigue propagation. Microscopically, the propagation zone exhibited extensive tearing ridges and dense fatigue striations, which, combined with the criteria for high-cycle fatigue fractures, confirm that the failure mode of the active spiral bevel gear was high-cycle fatigue.
- The macroscopic examination of the failed gear showed that the contact patterns around the entire circumference of the active spiral bevel gear were uneven, biased towards the small end and close to the tooth root. The contact patterns on the driven spiral bevel gear were distributed at the top of the small end teeth, with incomplete marks and partial extension beyond the working surface. This indicated that edge contact occurred during the operation of the gear pair.
- The shape, size, and location of the contact area on the tooth surface of a spiral bevel gear have a direct impact on its smooth operation, load distribution, service life, and noise levels. The contact pattern is a crucial indicator of the meshing quality of spiral bevel gears and a pivotal issue in their development. A company pioneered the technology for forming spiral bevel gear processing parameters based on the principle of local conjugacy and employed Tooth Contact Analysis (TCA) to simulate the meshing process of spiral bevel gears, thereby obtaining tooth surface contact patterns and transmission error curves to predict the rationality of the processed gear parameters. By determining the geometric and processing parameters of the active spiral bevel gear and selecting a reference point on the large gear tooth surface, the conjugate point on the driven spiral bevel gear corresponding to the reference point on the driving gear was found using the principle of local conjugacy. The principal curvature and normal vector at this point were then determined, leading to the processing parameters of the driven spiral bevel gear. In actual production, multiple iterations of parameter adjustments are often required based on rolling inspection test results to achieve a suitable contact pattern and good meshing performance. This underscores the intimate relationship between the colored contact pattern of spiral bevel gears and the adjustment of processing parameters.
- The simulation and verification of the typical inspection position contact patterns on the working surface of the failed active spiral bevel gear revealed that the static contact patterns at certain typical inspection positions of the gears processed according to the current detailed design parameters for the tooth surface were located at the tooth profile edges. During operation, under load, the dynamic contact patterns on the concave surface of the active spiral bevel gear shifted towards the small end and tooth root, resulting in edge contact and increased stress at the tooth root. This periodic and recurring abnormal meshing promoted the initiation of fatigue cracks at the corner between the small end face and the convex root of the tooth. Inadequate consideration of the gear tooth surface processing parameters was the primary cause of the failure. Therefore, optimizing the gear tooth surface processing parameters, conducting contact pattern simulations, and experimental validations to ensure that the gears fully meet the design requirements for colored contact patterns after trial production and to improve the meshing quality during gear operation are essential measures to prevent such failures.
- Combined with the metallographic structure and microhardness testing of the failed active spiral bevel gear, it was observed that a “soft spot” zone existed at the transition between the carburized layer and the core material (approximately 0.8 mm from the surface), corresponding to the presence of retained austenite. This phenomenon was attributed to the difficulty in carburizing the 15Cr14Co12Mo5Ni2 material due to its high alloy element content. During the prolonged carburizing process, if the carbon potential was too high and the diffusion time was insufficient, it could lead to the aggregation of carbon and alloy elements in the carburized transition zone, forming retained austenite bands (white blocky structures), which resulted in soft spots. Thus, controlling and adjusting the carburizing process and parameters to reduce the stability of austenite during cooling is the most effective way to eliminate and reduce retained austenite in the carburized layer of 15Cr14Co12Mo5Ni2. The crack in the current failed active spiral bevel gear initiated at the surface of the corner between the small end face and the convex root of the tooth, and no obvious organizational defects such as retained austenite were observed in the source area. The composition, structure, and hardness of the base material met the standard requirements, indicating that they were not significantly related to the initiation of gear fatigue cracks.
- 3. Conclusion
- (1) The fracture of the failed active spiral bevel gear was high-cycle fatigue in nature and unrelated to material or metallurgical defects.
- (2) Inadequate consideration of the gear tooth surface processing parameters led to abnormal meshing during gear operation and increased stress at the tooth root, which was the primary cause of gear fatigue fracture.
- It is recommended to optimize the gear tooth surface processing parameters, conduct contact pattern simulations, and experimental validations to ensure that the gears fully meet the design requirements for colored contact patterns after trial production, thereby avoiding abnormal meshing during gear operation.