Spiral bevel gears transmit torque between intersecting shafts in a compact layout, and their smooth motion makes them common in high-speed, heavy-load drivetrains. Their tooth flanks form a point-contact conjugate surface, and the mesh itself is nonlinear: contact force, contact location, and elastic deformation change continuously along the path of contact. During run-up, the pair is especially sensitive to meshing impact because the pinion accelerates from rest while the wheel resists through a load torque. A transient explicit finite-element formulation is well suited to this problem because it resolves moving contact and inertial loading without assuming a fixed contact pattern. The analysis below examines a 20/59-tooth spiral bevel gear pair and isolates four inputs: pinion speed, acceleration ramp time, mesh damping, and inertia load. Wheel angular velocity and contact force are used as contact-performance indicators, while vibration displacement and acceleration at the mesh node characterize the initial impact.
Model Geometry, Mesh, and Loading
The gear pair had 20 pinion teeth and 59 wheel teeth, a module of 4.1 mm, and a 90° shaft angle. The pinion was right-handed and the wheel left-handed, with a 20° pressure angle and a 35° spiral angle. The basic geometry is summarized in Table 1.
Table 1. Basic parameters of the spiral bevel gear pair.
| Parameter | Pinion | Wheel |
|---|---|---|
| Number of teeth | 20 | 59 |
| Module (mm) | 4.1 | 4.1 |
| Outer diameter (mm) | 91.28 | 243.23 |
| Pressure angle (°) | 20 | 20 |
| Spiral angle (°) | 35 | 35 |
| Shaft angle (°) | 90 | 90 |
| Hand | Right | Left |
The tooth surfaces were generated from discrete coordinates projected onto an axial plane and then assembled into a solid model. Because the flanks are strongly twisted, direct hexahedral meshing in ABAQUS was impractical. Hypermesh was used for solid partitioning, two-dimensional meshing, and local refinement before the hexahedral mesh was exported for the transient analysis. The pinion flank served as the master contact surface and the wheel flank as the slave surface. Contact was defined as flexible-body surface-to-surface with small sliding. A reference point at each gear-axis centroid was kinematically coupled to the gear body so that rotational speed and torque could be applied to a hexahedral mesh whose nodes do not carry rotational degrees of freedom. The pinion received the speed input, and the wheel received a resisting torque; the baseline wheel load was 500 N·m.
Pinion Speed and Contact Response
Three runs used pinion speeds of 50, 100, and 150 rad/s with a 0.001 s acceleration ramp, a mass-proportional damping coefficient of 500, and a wheel load torque of 500 N·m. Table 2 lists the wheel-speed and contact-force metrics.
Table 2. Effect of pinion speed on wheel-speed and contact-force response.
| Pinion speed (rad/s) | Wheel settling time (s) | Stable wheel speed (rad/s) | Contact-force peak (kN) | Force settling time (s) | Stable contact force (kN) |
|---|---|---|---|---|---|
| 50 | 0.003 | 7.75 | 79.08 | 0.0051 | 7.56 |
| 100 | 0.007 | 20.35 | 144.83 | 0.0082 | 37.57 |
| 150 | 0.005 | 32.76 | 208.51 | 0.0132 | 82.77 |
Wheel speed reached stable values of 7.75, 20.35, and 32.76 rad/s as pinion speed increased. The settling time was non-monotonic: 0.003 s at 50 rad/s, 0.007 s at 100 rad/s, and 0.005 s at 150 rad/s. Increasing speed therefore does not simply prolong the transient; above a threshold, the higher contact force can bring the wheel to a stable speed sooner. Contact force followed a clearer monotonic trend. Peaks rose from 79.08 to 144.83 to 208.51 kN, and stable force rose from 7.56 to 37.57 to 82.77 kN. Higher speed also produced a longer force fluctuation interval before stabilization and a higher fluctuation frequency. Pinion speed thus changes both the excitation amplitude and the contact load carried in steady mesh. At low speed, insufficient contact force can allow the pair to pass repeatedly between separation and re-entry.
Acceleration Ramp and Meshing Impact
With pinion speed fixed at 100 rad/s, acceleration times of 0.001, 0.002, and 0.005 s were compared while other settings were held unchanged. The response metrics are given in Table 3.
Table 3. Effect of acceleration ramp time on wheel-speed and contact-force response.
| Acceleration time (s) | Wheel settling time (s) | Stable wheel speed (rad/s) | Contact-force peak (kN) | Stable contact force (kN) |
|---|---|---|---|---|
| 0.001 | 0.0072 | ≈20.45 | 79.08 | ≈5.2 |
| 0.002 | 0.0081 | ≈20.45 | 42.57 | ≈5.2 |
| 0.005 | 0.010 | ≈20.45 | 20.06 | ≈5.2 |
All three cases settled near 20.45 rad/s, but longer ramps reduced the wheel-speed overshoot and the frequency of the initial oscillation. Contact-force peaks fell from 79.08 to 42.57 to 20.06 kN, while the stable force remained near 5.2 kN. Settling time increased modestly, from 0.0072 to 0.0081 to 0.010 s, but the contact-force trace became smoother. A longer acceleration ramp therefore lowers the mesh-entry impact without materially changing the steady contact load. In drive systems where the motor can be ramped, stretching the acceleration interval is a direct way to reduce the first contact transient.
Damping and the Low-Frequency Contact Mode
Rayleigh damping was represented as a linear combination of the mass and stiffness matrices:
$$ C = \alpha M + \beta K $$
where \(C\), \(M\), and \(K\) are the damping, mass, and stiffness matrices, \(\alpha\) is the mass-proportional coefficient, and \(\beta\) is the stiffness-proportional coefficient. Gear contact is a low-frequency oscillation problem, so the stiffness-proportional term was neglected and the mass coefficient was varied. Three models used \(\alpha = 100\), 500, and 1000, with a pinion speed of 100 rad/s, a 0.001 s ramp, and a 500 N·m wheel torque.
Table 4. Effect of mass-proportional damping on wheel-speed and contact-force response.
| Mass coefficient α | Wheel settling time (s) | Stable wheel speed (rad/s) | Contact-force peak (kN) | Force settling time (s) | Stable contact force (kN) |
|---|---|---|---|---|---|
| 100 | 0.016 | 20.41 | 165.28 | 0.017 | 6.01 |
| 500 | 0.007 | 20.41 | 174.57 | 0.012 | 30.89 |
| 1000 | 0.005 | 20.41 | 191.32 | 0.005 | 55.03 |
Wheel speed settled at 0.016, 0.007, and 0.005 s, all near 20.41 rad/s. Higher damping reduced both the amplitude and frequency of wheel-speed oscillation. Contact-force peaks, however, rose from 165.28 to 174.57 to 191.32 kN, successive increases of 5.62% and 9.59%. The corresponding force settling times fell from 0.017 to 0.012 to 0.005 s, successive reductions of 29.4% and 58.3%, while stable contact force rose from 6.01 to 30.89 to 55.03 kN. Damping therefore acts as a trade-off: it suppresses low-frequency speed oscillation and shortens the transient, but it also changes the contact-load path and raises the force level. The damping coefficient should be selected against both vibration and tooth-load limits rather than as a single-objective parameter.
Inertia Load and Material Density
The inertia load of a rotating gear follows:
$$ F = m a = m \omega^2 r $$
For a fixed geometry, the radius \(r\) is fixed and the mass \(m\) is controlled by material density. Three relative density ratios of 1:5, 1:1, and 5:1 were compared with a pinion speed of 100 rad/s, a 0.001 s ramp, a 500 N·m wheel torque, and \(\alpha = 500\).
Table 5. Effect of relative material density on wheel-speed and contact-force response.
| Density ratio | Wheel settling time (s) | Stable wheel speed (rad/s) | Contact-force peak (kN) | Force settling time (s) | Stable contact force (kN) |
|---|---|---|---|---|---|
| 1:5 | 0.007 | 20.41 | 12.41 | 0.137 | 6.42 |
| 1:1 | 0.012 | 20.41 | 53.04 | 0.151 | 25.27 |
| 5:1 | 0.018 | 20.41 | 324.74 | 0.168 | 197.01 |
All cases reached approximately 20.41 rad/s, but settling time lengthened from 0.007 to 0.012 to 0.018 s as the density ratio increased. Contact-force peaks rose sharply from 12.41 to 53.04 to 324.74 kN, and stable force rose from 6.42 to 25.27 to 197.01 kN. The force settling times were 0.137, 0.151, and 0.168 s. Higher inertia produced larger wheel-speed and contact-force excursions at a lower oscillation frequency. For this gear pair, low-density material is favorable when strength and fatigue requirements permit, because added mass raises the impact load and delays stable mesh.
Vibration at the Mesh Node
Using the 1:1 density model, a pinion speed of 100 rad/s, a 0.001 s ramp, a 500 N·m wheel torque, and \(\alpha = 500\), the vibration response was extracted at the mesh node in the \(x\), \(y\), and \(z\) directions. The \(z\) axis represents the axial direction, and \(x\) and \(y\) represent the radial directions. Table 6 summarizes the reported response levels.
Table 6. Vibration displacement and acceleration at the mesh node.
| Direction | Pinion displacement | Wheel displacement | Pinion acceleration | Wheel acceleration |
|---|---|---|---|---|
| Axial \(z\) | 31 μm | 8 μm | 15 m/s² | 4 m/s² |
| Radial \(x,y\) | 11–50 μm | 36–75 μm | 5.5–22.5 m/s² | 16–34 m/s² |
The pinion showed larger axial displacement and acceleration than the wheel, at 31 μm and 15 m/s² versus 8 μm and 4 m/s². In the radial direction the wheel responded more strongly, at 36–75 μm and 16–34 m/s², compared with 11–50 μm and 5.5–22.5 m/s² for the pinion. The pattern is consistent with a stiffer radial support path on the pinion: lower radial deflection gives lower radial vibration, while axial compliance dominates the pinion response. The acceleration trends follow the displacement trends. For mounting and bearing layout, the pinion axial direction and the wheel radial direction therefore deserve separate attention; a support design that targets only one direction will leave the other dominant response uncontrolled.
Transmission Error
The transmission error was evaluated from the angular difference scaled by the wheel base radius:
$$ TE = r_b(\theta_2-\theta_1) $$
where \(r_b\) is the wheel base radius and \(\theta_1,\theta_2\) are the angular displacements used in the contact solution. Under the same operating condition, the transmission error varied periodically and remained comparatively steady in the initial interval; after a period of operation it showed an increasing trend. For motion-transfer or positioning drives, this drift is more relevant to angular accuracy than the peak contact force alone.
Design Implications
- Pinion speed must be high enough to maintain contact force and avoid repeated separation and re-entry. Higher speed raises both peak and steady contact force, and it can lengthen the force-fluctuation interval even when wheel-speed settling becomes shorter.
- A longer acceleration ramp reduces the wheel-speed overshoot, the contact-force peak, and the frequency of the initial oscillation, while leaving the steady contact force nearly unchanged.
- Higher mass-proportional damping suppresses wheel-speed oscillation and shortens settling, but it raises the contact-force peak and stable force. Damping should be tuned together with tooth-load limits.
- Higher material density increases inertia load, contact-force peaks, and settling time. Low-density material is preferable where strength and fatigue life allow.
- The pinion dominates axial vibration, while the wheel dominates radial vibration. Bearing and housing stiffness should be selected with these direction-dependent responses in mind.