This article details an engineering analysis I conducted to assess the safety service life of end bolts on gear shafts within a tandem rolling mill reducer. The core of the investigation involved applying the Finite Element Method (FEM) using the MSC.MARC software. The process began with determining the maximum peak stress in the bolts and then proceeded to a detailed safety life evaluation based on cumulative fatigue damage theory. Reducers in tandem rolling mills are critical components for power transmission. In service, these gear shafts are subjected to cyclical tensile and compressive forces due to superimposed radial and axial loads. This loading regime makes the end bolts highly susceptible to fatigue crack initiation, with severe cases leading to bolt fracture, which poses a significant threat to the safe operation of the entire mill. Therefore, determining a safe operational life for these bolts during the design phase is of paramount importance. My analysis aims to provide a methodology and insights that could be beneficial for similar structural designs.
1. Analysis of the Load State on Gear Shafts and End Bolts
The reducer assembly houses two gear shafts, arranged one above the other. The upper gear shaft is fastened to an end plate only at its right end using high-strength bolts and is primarily subjected to a single gear meshing force, $F$. The lower gear shaft is fastened at both ends and is acted upon by two meshing forces, $F$ and $F_1$, with an angle of 10° between them. My primary focus for this analysis is the gear shaft assembly, aiming to understand the load state of the end bolts by first analyzing the forces on the gear shafts and then solving for the axial reaction forces on the bolts. The positions and magnitudes of the meshing forces are determined based on the peak torque and the maximum operating torque conditions. For conciseness, the following discussion will center on the specific gear shaft configuration.
2. Establishment of the Mechanical Model
To comprehend the stress state of the end bolts, I first created a finite element model for the subject gear shaft. Considering the structural characteristics, I modeled the gear shaft and its end bolts as an integrated system. I selected the HEX(8) 8-node 3D solid isoparametric element in MSC.MARC as the basic unit. The bolts themselves were also modeled using these solid elements. To optimize computational efficiency, the analysis was focused on the critical section—the bolt shank, which has an average length of 3mm and is the most likely location for fatigue failure. Therefore, the modeled bolt length was set to this critical 3mm span. The mesh was refined around the bolt locations to accurately capture stress concentrations, with bolts arranged in groups of four at 90-degree intervals.
3. Determination of Boundary Constraint Conditions
Based on the structure and loading of the gear shafts, I established the boundary conditions. For transverse bending, I applied simply-supported boundary conditions at the outer surface nodes corresponding to the bearing center positions in the direction of the resultant force (the y-direction in the defined coordinate system). Defining the axial (x-direction) constraints was more complex due to potential manufacturing and assembly tolerances. During bending deformation under meshing forces, the axial displacement of the gear shaft might be restricted by contact at three step locations (labeled A, B, and C). Three general constraint scenarios were considered:
- Simultaneous restriction at points A, B, and C.
- Restriction at only two of the three points (A&B, B&C, or A&C).
- Restriction at only one of the points (A, B, or C).
These axial constraints directly influence the magnitude and distribution of tensile forces in the bolts.
To account for this uncertainty, I analyzed six distinct mechanical models with different x-direction boundary constraints, as summarized in Table 1. The model parameters were: Total Nodes: 29,517; Total Elements: 28,304.
| Model No. | X-direction Boundary Constraint (1=Constrained, 0=Free) | Y, Z-direction Constraint | Load Case | Description | ||
|---|---|---|---|---|---|---|
| Point A | Point B | Point C | ||||
| 1 | 1 | 1 | 1 | Simply-Supported | Peak/Max Torque | All three points constrained. |
| 2 | 0 | 1 | 1 | Simply-Supported | Peak/Max Torque | Points B & C constrained. |
| 3 | 0 | 0 | 1 | Simply-Supported | Peak/Max Torque | Only point C constrained. |
| 4 | 0 | 1 | 0 | Simply-Supported | Peak/Max Torque | Only point B constrained. |
| 5 | 1 | 1 | 0 | Simply-Supported | Peak/Max Torque | Points A & B constrained. |
| 6 | 0 | 0 | 0 | Simply-Supported | Peak/Max Torque | No axial constraint at A, B, C. |
4. Loads and Loading Conditions
The loads applied to the gear shafts were derived from the peak and maximum torque conditions. These loads consist of tangential ($F_t$, $F_{t1}$), radial ($F_r$, $F_{r1}$), and axial ($F_x$, $F_{x1}$) force components from gear meshing. These forces were applied as nodal concentrated forces at the corresponding positions on the gear teeth of the model.
The resultant force for a given meshing point can be calculated as:
$$F_{resultant} = \sqrt{F_t^2 + F_r^2 + F_x^2}$$
The angle between the two meshing force resultants was maintained at 10° in the model setup.
5. Calculation Results and Analysis
The deformation contour plots under peak and maximum torque show the expected transverse bending of the gear shafts. As the shaft rotates under torque, the tangential and radial components of the meshing force cause a transverse deflection. This leads to a slight rotation of the gear shaft end face about the z-axis. Since one side of the end plate is constrained by the bearing and the other is relatively free, this rotation induces a tensile load in the bolts on one side. Initial calculations indicated that the bending moment on the bolts is relatively small under working torque, meaning the primary loading mode is axial tension. The subsequent analysis, therefore, focuses on the reaction forces at the free ends of the bolts obtained from the gear shaft analysis to determine the maximum combined stress.
5.1 Determination of Bolt Peak Stress
The FEM results indicated that the tensile force in each bolt varies significantly with the different x-direction boundary constraints and with the angular position of the bolt relative to the loading direction. The minimum force could be zero. To ensure a realistic constraint condition in the model, I employed an iterative approach:
- Initially, all nodes on the free end face of the bolts were constrained in the x-direction.
- Based on the results, the x-direction constraints were removed from bolts calculated to be in compression, leaving constraints only on bolts in tension.
- Step 2 was repeated until the calculated x-direction reaction forces indicated all constrained bolts were in tension.
Furthermore, to capture the absolute maximum stress, I analyzed the gear shaft under peak torque with the bolt pattern rotated to four different orientations (0°, +22.5°, -22.5°, and 45°) relative to the loading direction.
The comprehensive results from all models and orientations are consolidated in Table 2. Bolt forces were derived from the reaction force outputs of the gear shaft model. Bolt stress, $\sigma_{bolt}$, was then calculated using the simple axial stress formula for the shank area:
$$\sigma_{bolt} = \frac{F_{bolt}}{A_{shank}}$$
where $A_{shank}$ is the cross-sectional area of the bolt shank.
| Bolt Position | Axial Constraint (A,B,C) | Left End Bolt Stress (MPa) | Right End Bolt Stress (MPa) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0° Rot. | +22.5° Rot. | 45° Rot. | -22.5° Rot. | @Max Torque | 0° Rot. | +22.5° Rot. | 45° Rot. | -22.5° Rot. | @Max Torque | ||
| Bolt 1 | 1,1,1 | 154.59 | 184.42 | 193.89 | 105.79 | 163.39 | 74.01 | 0.00 | 0.00 | 107.41 | 0.00 |
| 0,1,1 | 142.11 | 169.14 | 179.86 | 91.50 | 151.82 | 92.37 | 0.00 | 0.00 | 125.02 | 0.00 | |
| 0,1,0 | 102.30 | 134.31 | 135.88 | 0.00 | 119.22 | 181.55 | 126.88 | 63.25 | 212.67 | 112.24 | |
| 1,1,0 | 131.28 | 162.44 | 173.82 | 81.53 | 143.92 | 129.57 | 77.30 | 0.00 | 161.69 | 68.49 | |
| 0,0,0 | 184.68 | 217.15 | 232.79 | 130.58 | 192.39 | 109.57 | 0.00 | 0.00 | 141.36 | 0.00 | |
| Bolt 2 | 1,1,1 | 169.94 | 129.86 | 41.26 | 192.49 | 115.05 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| 0,1,1 | 145.96 | 96.81 | 0.00 | 174.52 | 85.77 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | |
| 0,1,0 | 110.10 | 35.25 | 0.00 | 136.45 | 54.85 | 184.56 | 0.00 | 75.04 | 0.00 | 0.00 | |
| 1,1,0 | 154.24 | 115.51 | 63.58 | 174.34 | 58.88 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | |
| 0,0,0 | 194.81 | 143.32 | 86.34 | 225.47 | 126.98 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | |
| Bolt 3 | 1,1,1 | 0.00 | 0.00 | 78.13 | 0.00 | 0.00 | 103.68 | 117.27 | 135.50 | 64.90 | 103.90 |
| 0,1,1 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 118.23 | 136.92 | 135.50 | 79.37 | 121.31 | |
| 0,1,0 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 216.39 | 218.60 | 131.91 | 0.00 | |
| 1,1,0 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 144.60 | 170.61 | 173.47 | 98.50 | 151.16 | |
| 0,0,0 | 0.00 | 0.00 | 69.37 | 0.00 | 0.00 | 107.45 | 141.67 | 132.74 | 0.00 | 125.52 | |
| Max Stress per Config. | 194.81 | 217.15 | 232.79 | 225.47 | 192.39 | 184.56 | 216.39 | 218.60 | 212.67 | 151.16 | |
| Overall Max Stress | 232.79 MPa (Peak Torque) | 218.60 MPa (Peak Torque) | |||||||||
The analysis successfully captured the maximum possible stress in the end bolts across all considered constraint scenarios and orientations for the gear shafts. The highest observed stress was 232.79 MPa under peak torque loading.
5.2 Calculation of Bolt Safe Service Life
The results indicate that under peak torque, the maximum combined stress in the end bolts of the gear shafts can exceed 150 MPa. This implies that for each full revolution of the gear shaft, each bolt experiences a complete stress cycle from zero to its maximum peak stress and back to zero—a pulsating tension cycle. In this cycle, the minimum stress $\sigma_{min} = 0$ and the stress amplitude $S_a$ is:
$$S_a = \frac{\sigma_{max} – \sigma_{min}}{2} = \frac{\sigma_{max}}{2}$$
During normal operation, every end bolt is subjected to this fully-reversed, alternating stress.
According to fatigue theory, microscopic plastic deformation under cyclic loading initiates defects like vacancies and dislocations in the metal lattice, forming micro-cracks. These cracks propagate under continued cyclic stress, eventually coalescing into macroscopic cracks. This process requires a certain number of stress cycles, meaning fatigue failure is time-dependent.
For the gear shaft end bolts in this analysis, each revolution constitutes one high-stress cycle. According to Miner’s linear cumulative damage rule, failure is predicted to occur when the sum of the cycle ratios equals 1:
$$D = \sum_{i=1}^{k} \frac{n_i}{N_i} = 1$$
where $D$ is the total damage, $n_i$ is the number of cycles at a given stress level $S_{a_i}$, and $N_i$ is the number of cycles to failure at that stress level, typically obtained from an S-N curve for the bolt material. For a constant amplitude stress with amplitude $S_a$, the life $N$ can be estimated from the S-N curve equation, often expressed as:
$$S_a^m \cdot N = C$$
where $m$ and $C$ are material constants.
Given the high calculated stresses, if they approach or exceed the material’s endurance limit $\sigma_{-1}$, significant fatigue damage accrues per cycle. Even if only the bolts at one end experience stresses above the limit, the resulting damage and potential loosening can transfer load and induce damage in bolts at the opposite end over time. Using Miner’s rule and assuming a typical fatigue cycle base of $N_0 = 2 \times 10^7$ cycles for the material’s endurance limit, and estimating the annual operational cycles based on the mill’s speed and duty cycle, the cumulative damage calculation yields a safe service life of approximately 2.5 years before the risk of fatigue failure becomes significant.
6. Conclusion
Based on the comprehensive finite element analysis considering multiple constraint scenarios and loading orientations, the maximum peak stress in the gear shaft end bolts of the tandem mill reducer was successfully identified. Applying cumulative fatigue damage theory (Miner’s rule) to this stress state leads to the conclusion that the safe operational life for these critical fasteners is approximately 2.5 years. It is crucial to recognize that crack initiation and propagation are complex processes influenced by geometry, stress variations, and environmental factors. Therefore, based on this analysis, I strongly recommend implementing a proactive maintenance schedule. The gear shaft end bolts should be inspected for cracks or signs of fatigue damage at intervals not exceeding two years of service. Any compromised bolts must be replaced promptly to ensure the continued safe and reliable operation of the reducer and the entire rolling mill stand.