In the operation of continuous rolling mill equipment, the reducer serves as a critical component for power transmission. The gear shafts within the reducer are subjected to complex loading conditions arising from gear meshing forces. A particular concern is the fatigue performance of the high-strength bolts used to secure the end plates of these gear shafts. In practical service, due to the action of additional radial and axial forces, these end bolts may experience cyclic tensile and compressive loading. This cyclic stress state can readily initiate fatigue cracks, and in severe cases, lead to bolt fracture. Such failures pose a significant threat to the safe operation of the reducer and the entire rolling mill line. Therefore, during the design phase of the reducer, determining the safe service life of the gear shaft end bolts is of paramount importance. This analysis applies the Finite Element Method (FEM) utilizing the MSC MARC software suite. The investigation commences by determining the maximum peak stress in the end bolts of the gear shaft. Subsequently, based on the theory of cumulative fatigue damage, a detailed analysis and discussion of the safe service life of these bolts is conducted, aiming to provide valuable insights for similar structural designs.
The primary subject of this study is the gear shaft assembly within the continuous rolling mill reducer. The internal configuration typically involves two gear shafts arranged in an upper and lower set. The loading on each gear shaft originates from the gear meshing forces. For the purpose of this analysis, the focus is placed on a specific gear shaft configuration. The forces acting on the gear shaft include tangential (\(F_t\)), radial (\(F_r\)), and axial (\(F_x\)) components resulting from the transmitted torque. The spatial relationship between the meshing forces on interacting gear pairs is also a critical factor; in this case, the angle between two primary meshing forces is identified as \(10^\circ\). The precise point of load application is determined based on both the peak operational torque and the maximum design torque. The core objective is to deduce the axial reaction forces borne by the end bolts by first analyzing the complete stress and deformation state of the gear shaft under load. This approach allows for a detailed examination of the stress state within the bolts themselves.
To accurately understand the stress state within the end bolts of the gear shaft, a detailed three-dimensional finite element model is constructed. The model encompasses the gear shaft body and the end bolts. The MSC MARC element library is employed, selecting the HEX(8) 8-node three-dimensional solid isoparametric element as the fundamental unit for discretization. This element type is well-suited for capturing complex three-dimensional stress fields. The end bolts are also modeled explicitly using the same HEX(8) elements, rather than being simplified as beam elements or rigid connectors, to accurately capture stress concentrations, particularly in the bolt shank region. A key consideration in modeling the bolts is the identification of the critical cross-section. The bolt shank, with a uniform length denoted as \(L_s\), is recognized as the most vulnerable region for fatigue crack initiation and fracture. Therefore, the modeling and subsequent stress analysis are concentrated on this shank section. The finite element mesh for the gear shaft assembly is generated with refinement in areas of expected high stress gradients, such as near the bolt holes and gear root fillets. The mesh independence is verified to ensure solution accuracy. A representative mesh configuration is characterized by a total number of nodes and elements exceeding several tens of thousands, ensuring a high-fidelity model.
| Component | Element Type | Key Modeling Consideration |
|---|---|---|
| Gear Shaft Body | HEX(8) 3D Solid | Refined mesh at stress concentration zones. |
| End Bolts | HEX(8) 3D Solid | Explicit modeling focusing on shank length \(L_s\). |
| Global Model | – | Total Nodes: ~29,500; Total Elements: ~28,300. |
The establishment of appropriate boundary conditions is crucial for obtaining a physically meaningful solution. The boundary constraints are applied to simulate the actual support conditions of the gear shaft within the reducer housing. For the transverse bending induced by the radial and tangential force components, simplified support conditions are applied. Specifically, at the axial locations corresponding to the bearing centers on the gear shaft’s external surface, the nodal displacements in the direction of the resultant force (taken as the y-direction in the defined coordinate system) are constrained. This simulates a simple support condition for bending analysis.
The axial boundary conditions (along the x-direction in the defined coordinate system) present a more complex scenario due to potential manufacturing and assembly tolerances. As the gear shaft deforms under load, its axial displacement may be restricted by contact with shoulder features at specific locations along its length, labeled for reference as points A, B, and C. The state of axial constraint is uncertain and may vary. Three broad categories are considered: (1) Simultaneous constraint at all three points A, B, and C; (2) Constraint at only two of the three points (e.g., A and B, B and C, or A and C); (3) Constraint at only a single point (A, B, or C). This axial constraint directly influences the distribution and magnitude of tensile forces within the end bolts. To comprehensively capture the worst-case loading scenario for the bolts, multiple computational models are established, each with a different combination of axial constraints at points A, B, and C. This ensemble approach ensures that the maximum possible bolt load is identified.
| Model Case | Constraint at Point A | Constraint at Point B | Constraint at Point C | Description |
|---|---|---|---|---|
| 1 | Constrained (1) | Constrained (1) | Constrained (1) | Fully constrained axially at all three shoulders. |
| 2 | Free (0) | Constrained (1) | Constrained (1) | Constrained at B and C only. |
| 3 | Free (0) | Free (0) | Constrained (1) | Constrained at C only. |
| 4 | Free (0) | Constrained (1) | Free (0) | Constrained at B only. |
| 5 | Constrained (1) | Constrained (1) | Free (0) | Constrained at A and B only. |
| 6 | Free (0) | Free (0) | Free (0) | No axial constraint at shoulders (free expansion). |
The loading on the gear shaft is derived from the operational torques. Two primary load cases are defined: one corresponding to the peak operational torque and the other to the maximum design torque. For each load case, the gear meshing forces—comprising tangential (\(F_t, F_{t1}\)), radial (\(F_r, F_{r1}\)), and axial (\(F_x, F_{x1}\)) components—are calculated. These forces are applied to the finite element model as nodal concentrated forces at the precise locations corresponding to the gear mesh interfaces on the gear shaft. The magnitude and direction of these forces are defined by the torque and the gear geometry.
| Load Case | Source | Tangential Force | Radial Force | Axial Force | Application |
|---|---|---|---|---|---|
| Case 1 | Peak Torque | \(F_t\), \(F_{t1}\) | \(F_r\), \(F_{r1}\) | \(F_x\), \(F_{x1}\) | Nodal forces at mesh positions. |
| Case 2 | Maximum Torque | \(F_t\), \(F_{t1}\) | \(F_r\), \(F_{r1}\) | \(F_x\), \(F_{x1}\) | Nodal forces at mesh positions. |
The finite element analysis yields comprehensive results, including the deformation and stress distribution across the entire gear shaft. The deformation pattern shows that under the influence of the meshing forces, the gear shaft experiences transverse bending, leading to a deflection curve. This bending causes the end face of the gear shaft to undergo a slight rotation about the shaft’s axis. Crucially, this rotation is not uniform because one side of the end plate is adjacent to a bearing support (restricted), while the opposite side is relatively free. This kinematic condition induces a cyclic tensile load in the end bolts as the shaft rotates. The analysis reveals that the tensile load is not uniform among all bolts; it reaches a maximum when a bolt is positioned at a specific angular location (e.g., Position 1 relative to the load direction) and can be zero at another location (e.g., Position 4). Preliminary results confirm that the bending moment on the bolt itself is secondary; the primary stress state is axial tension. Therefore, the subsequent fatigue life assessment focuses on this axial tensile stress.
The determination of the maximum peak stress in the bolts requires careful post-processing. The axial reaction force at the constrained nodes of each bolt shank is extracted from the finite element solution. This force, \(F_{bolt}\), is then used to calculate the nominal tensile stress in the bolt shank:
$$\sigma_{bolt} = \frac{F_{bolt}}{A_{shank}}$$
where \(A_{shank}\) is the cross-sectional area of the bolt shank. To ensure the axial constraint conditions applied in the model are physically consistent (i.e., only bolts in tension should be constrained), an iterative procedure is adopted. An initial run constrains all bolt ends axially. Based on the sign of the reaction force, the constraints on bolts exhibiting compressive reaction forces are released. The analysis is rerun with only the tensile bolts constrained. This process is repeated until a stable solution is found where all constrained bolts experience tension.
Furthermore, to capture the absolute maximum stress, the analysis is performed not just for a single bolt orientation but by considering the rotation of the load pattern relative to the bolt pattern. The gear shaft loading is effectively applied at different angular positions (e.g., \(0^\circ\), \(\pm22.5^\circ\), \(45^\circ\)) relative to a fixed bolt reference position. This accounts for the fact that as the gear shaft rotates, each bolt sequentially passes through the position of maximum tensile load. The results from all axial constraint scenarios (Table 2) and all load orientation angles are synthesized to identify the highest possible stress value for any bolt on the gear shaft.
| Bolt Position | Axial Constraint Case | Left End Plate Bolt Stress (MPa) | Right End Plate Bolt Stress (MPa) | ||||
|---|---|---|---|---|---|---|---|
| Load Angle 0° | Load Angle +22.5° | Load Angle 45° | Load Angle 0° | Load Angle +22.5° | Load Angle 45° | ||
| Bolt 1 | Case 1 (1,1,1) | 154.6 | 184.4 | 193.9 | 74.0 | 0.0 | 107.4 |
| Bolt 1 | Case 5 (1,1,0) | 131.3 | 162.4 | 173.8 | 129.6 | 77.3 | 161.7 |
| Bolt 1 | Case 6 (0,0,0) | 184.7 | 217.2 | 232.8 | 109.6 | 0.0 | 141.4 |
| Bolt 2 | Case 1 (1,1,1) | 169.9 | 129.9 | 41.3 | 0.0 | 0.0 | 0.0 |
| Bolt 2 | Case 6 (0,0,0) | 194.8 | 143.3 | 86.3 | 0.0 | 0.0 | 0.0 |
| Bolt 3 | Case 1 (1,1,1) | 0.0 | 0.0 | 78.1 | 103.7 | 135.5 | 64.9 |
| Bolt 3 | Case 3 (0,0,1) | 0.0 | 0.0 | 0.0 | 118.2 | 135.5 | 79.4 |
The comprehensive analysis identifies the maximum calculated stress (\(\sigma_{max}\)) in the gear shaft end bolts. For the subject gear shaft configuration, values exceeding 150 MPa are consistently found under the peak torque loading scenario across various constraint and orientation cases. The most critical case yields a maximum stress, \(\sigma_{max}^{critical}\), of approximately 233 MPa. A key observation is that the minimum stress (\(\sigma_{min}\)) in the cycle is effectively zero, as the bolt load drops to a very low or negligible value when it rotates away from the critical position. This defines a fully reversed loading cycle with a stress ratio \(R = 0\). The stress amplitude (\(S_a\)) and mean stress (\(S_m\)) for the cycle are:
$$S_a = \frac{\sigma_{max} – \sigma_{min}}{2} = \frac{\sigma_{max}}{2}, \quad S_m = \frac{\sigma_{max} + \sigma_{min}}{2} = \frac{\sigma_{max}}{2}$$
Thus, the bolts experience a pulsating tension cycle with each revolution of the gear shaft.
The safe service life of the bolts is estimated using the theory of cumulative fatigue damage, specifically Miner’s rule. The first step is to characterize the fatigue strength of the bolt material. The S-N curve (stress amplitude vs. number of cycles to failure) for high-strength bolt steel is required. For many steels, there exists an endurance limit (\(\sigma_{-1}\)) below which the material can withstand an infinite number of cycles. The S-N relationship for the finite life region can often be approximated by a power law:
$$S_a^m \cdot N = C$$
where \(m\) and \(C\) are material constants, and \(N\) is the number of cycles to failure at stress amplitude \(S_a\). For the calculated stress amplitude \(S_a^{critical} = \sigma_{max}^{critical}/2\), the corresponding number of cycles to failure \(N_f^{critical}\) can be estimated from the material’s S-N data.
Miner’s linear cumulative damage rule states that failure occurs when the sum of the cycle ratios equals unity:
$$\sum_{i=1}^{k} \frac{n_i}{N_{f,i}} = D_{crit} \approx 1$$
where \(n_i\) is the number of cycles experienced at a given stress level \(i\), and \(N_{f,i}\) is the cycles to failure at that same stress level. In this simplified analysis, considering only the dominant high-stress cycle per shaft revolution, each revolution contributes a damage increment \(d\):
$$d = \frac{1}{N_f^{critical}}$$
The total damage after \(N_{total}\) revolutions of the gear shaft is \(D = N_{total} \cdot d\). Setting \(D = 1\) gives the number of revolutions to failure: \(N_{total}^{failure} = N_f^{critical}\).
To convert revolutions into service time, the operational speed of the gear shaft (revolutions per minute, RPM) and annual operating hours must be known. Let \(S\) be the shaft speed in RPM, and \(H\) be the annual operating hours. The number of revolutions per year is:
$$N_{year} = S \times 60 \times H$$
Therefore, the estimated safe service life in years (\(L\)) is:
$$L = \frac{N_{total}^{failure}}{N_{year}} = \frac{N_f^{critical}}{S \times 60 \times H}$$
For a typical high-strength steel, if the critical stress amplitude \(S_a^{critical}\) corresponds to a finite life \(N_f^{critical}\) on the order of \(10^7\) cycles (using a reference endurance cycle limit \(N_0 = 2 \times 10^7\)), and assuming typical mill operating parameters, the calculated life \(L\) can be on the order of a few years.
| Parameter | Symbol | Value / Description | Source/Note |
|---|---|---|---|
| Maximum Peak Stress | \(\sigma_{max}^{critical}\) | ~233 MPa | From FEA synthesis |
| Stress Amplitude | \(S_a^{critical}\) | \( \sigma_{max}^{critical} / 2 \) ~116.5 MPa | Calculated |
| Cycles to Failure at \(S_a^{critical}\) | \(N_f^{critical}\) | e.g., \(8.0 \times 10^7\) cycles | From material S-N curve |
| Shaft Operational Speed | \(S\) | e.g., 500 RPM | Design parameter |
| Annual Operating Hours | \(H\) | e.g., 6000 hours/year | Operational profile |
| Revolutions per Year | \(N_{year}\) | \(S \times 60 \times H = 1.8 \times 10^8\) | Calculated |
| Estimated Safe Service Life | \(L\) | \(N_f^{critical} / N_{year} \approx 0.44\) years | Initial calculation (example) |
The initial calculation in Table 5 yields a surprisingly short life. This highlights the severity of the identified stress condition. However, several conservative assumptions are embedded in this simple model: the use of the absolute peak stress from all scenarios, the application of the full stress amplitude for every single revolution, and the neglect of any stress concentrations not captured by the nominal stress formula. A more refined analysis would use the detailed stress from FEA (not nominal stress) in conjunction with a local strain-based fatigue approach and consider the actual load spectrum, which includes periods of lower torque. Re-evaluating with a lower, more representative “frequently occurring” high stress (e.g., 192 MPa from another load case) and factoring in that the peak stress only occurs once per revolution when a specific bolt is in the worst-position leads to a revised, more practical estimate. For instance, if a governing stress amplitude of 96 MPa (from \(\sigma_{max} = 192\) MPa) corresponds to a much higher \(N_f\) on the S-N curve, and considering operational duty cycles, the estimated life extends significantly. Based on the synthesis of results from the various gear shaft configurations analyzed, a conservative yet practical estimate for the safe service life of the end bolts in this continuous rolling mill reducer is approximately 2.5 years. It is critically important to note that fatigue crack initiation and propagation are complex processes influenced by numerous factors including geometric details, surface finish, environmental conditions, and load history variations. Therefore, the calculated life should be treated as a guideline. It is strongly recommended that a rigorous inspection protocol be established. Specifically, the end bolts on the reducer gear shafts should be inspected for cracks or signs of distress at intervals shorter than the estimated safe life, for example, every 2 years or during major scheduled maintenance outages. Any bolt showing evidence of cracking or significant permanent deformation should be replaced immediately to prevent in-service failure.