Spiral bevel gears transmit torque between intersecting shafts, most commonly at a 90° shaft angle, where compactness, smooth motion, and high load capacity are required. Their flanks are point-contact conjugate surfaces: the contact ellipse, contact path, contact force, and transmission error depend on tooth compliance, mounting stiffness, and the transient load history. Tooth contact analysis (TCA) predicts the no-load contact pattern, while loaded tooth contact analysis (LTCA) adds compliance and load effects. Neither method is intended to resolve the short, high-frequency impact that occurs when a gear pair is accelerated from rest. Explicit dynamic finite element analysis is useful for that start-up window because it marches the contact solution in time and resolves the transient contact force, wheel speed, and vibration response.
The geometric basis for spiral bevel gear design was established by local conjugate theory and the Gleason TCA procedure; later LTCA models incorporated tooth deformation and load distribution. The present analysis extends that quasi-static basis to the transient regime by using a flexible-body explicit contact model of a complete spiral bevel gear pair.
Gear Pair and Finite Element Model
The analyzed pair consists of a 20-tooth right-hand pinion driving a 59-tooth left-hand wheel at a 90° shaft angle. The basic design parameters are listed in Table 1. The tooth surfaces were generated from the design parameters, projected onto an axial plane, and converted into a solid model. Because the flanks are strongly twisted, a hexahedral mesh was produced through solid partitioning, two-dimensional meshing, and local refinement before import into ABAQUS/Explicit.
| 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 pinion flank was defined as the master surface and the wheel flank as the slave surface. Contact was modeled as small-sliding, surface-to-surface contact between flexible bodies. Kinematic coupling constraints were applied at reference points on the gear axes, because rotational speed and torque cannot be applied directly to a hexahedral mesh without rotational degrees of freedom. The pinion speed was prescribed at its reference point, and a resisting torque of 500 N·m was applied to the wheel. Unless stated otherwise, the speed ramp was 0.001 s and the mass-proportional damping coefficient was α = 500.
Damping was represented by Rayleigh damping:
$$ C = \alpha M + \beta K $$
where \(C\), \(M\), and \(K\) are the damping, mass, and stiffness matrices, and \(\alpha\) and \(\beta\) are the mass- and stiffness-proportional coefficients. For the low-frequency contact modes relevant to gear meshing, the stiffness-proportional term was neglected, so the parameter study varied α.
Inertia load was varied by changing the density ratio of the gear bodies at fixed geometry. The inertial force follows:
$$ F = m a = m \omega^2 r $$
where \(m\) is mass, \(\omega\) is angular speed, and \(r\) is radius. For a fixed gear geometry, increasing density increases \(m\) and therefore the inertia load during acceleration.
Speed and Acceleration Ramp
With the acceleration time fixed at 0.001 s, damping at α = 500, and wheel torque at 500 N·m, the pinion speed was set to 50, 100, and 150 rad/s. The resulting wheel-speed and contact-force responses are summarized in Table 2.
| Pinion speed (rad/s) | Steady wheel speed (rad/s) | Time to steady wheel speed (s) | Contact-force peak (kN) | Time to steady contact force (s) | Steady contact force (kN) |
|---|---|---|---|---|---|
| 50 | 7.75 | 0.003 | 79.08 | 0.0051 | 7.56 |
| 100 | 20.35 | 0.007 | 144.83 | 0.0082 | 37.57 |
| 150 | 32.76 | 0.005 | 208.51 | 0.0132 | 82.77 |
The steady wheel speed increased with pinion speed, as expected for a driven gear. The settling time was not monotonic: it lengthened when the pinion speed rose from 50 to 100 rad/s, but shortened when the speed rose from 100 to 150 rad/s. Once the pinion speed exceeded a threshold, the higher contact force and shorter relative ramp apparently promoted faster settling. The contact-force peak and steady contact force both increased strongly with speed: the peak rose from 79.08 to 144.83 to 208.51 kN, and the steady value rose from 7.56 to 37.57 to 82.77 kN. The higher-speed cases also showed longer and higher-frequency contact-force fluctuation.
At low speed, the contact force may be insufficient to maintain continuous meshing, so the pair can repeatedly separate and re-enter contact. This is a design caution for lightly loaded or low-speed operation: stable transmission requires a contact force large enough to keep the flanks engaged through the mesh cycle.
In a second sweep, the pinion speed was fixed at 100 rad/s and the acceleration time was set to 0.001, 0.002, and 0.005 s. The wheel reached a steady speed of approximately 20.45 rad/s in all three cases, but the transient differed markedly (Table 3).
| Acceleration time (s) | Time to steady wheel speed (s) | Contact-force peak (kN) | Steady contact force (kN) |
|---|---|---|---|
| 0.001 | 0.0072 | 79.08 | 5.2 |
| 0.002 | 0.0081 | 42.57 | 5.2 |
| 0.005 | 0.010 | 20.06 | 5.2 |
A longer acceleration ramp reduced the contact-force peak from 79.08 to 42.57 to 20.06 kN and lowered the fluctuation frequency. The steady contact force remained approximately 5.2 kN. The absolute time to steady wheel speed increased with the ramp length, because the ramp itself occupies more time, but the severity of the initial meshing shock was substantially reduced. A slower start-up is therefore an effective way to improve meshing quality, at the cost of a longer transient.
Damping and Inertia Load
The mass-proportional damping coefficient was varied at α = 100, 500, and 1000, with the pinion speed at 100 rad/s, acceleration time at 0.001 s, and wheel torque at 500 N·m. Table 4 summarizes the response.
| α | Time to steady wheel speed (s) | Contact-force peak (kN) | Time to steady contact force (s) | Steady contact force (kN) |
|---|---|---|---|---|
| 100 | 0.016 | 165.28 | 0.017 | 6.01 |
| 500 | 0.007 | 174.57 | 0.012 | 30.89 |
| 1000 | 0.005 | 191.32 | 0.005 | 55.03 |
Higher damping reduced the amplitude and frequency of wheel-speed fluctuation. The contact-force peak increased by 5.62% from α = 100 to 500 and by 9.59% from α = 500 to 1000. The time to steady contact force shortened by 29.4% and then by 58.3% over the same steps. The steady contact force rose from 6.01 to 30.89 to 55.03 kN. Damping therefore improves settling and suppresses speed oscillation, but it does so at the cost of a slightly higher contact-force peak and a higher steady contact force. It is not a free reduction in vibration; it redistributes the transient response.
Inertia load was varied through three density ratios: 1:5, 1:1, and 5:1. The pinion speed was 100 rad/s, acceleration time 0.001 s, wheel torque 500 N·m, and α = 500. The wheel reached a steady speed of 20.41 rad/s in all cases, but the transient depended strongly on density ratio (Table 5).
| Density ratio | Time to steady wheel speed (s) | Contact-force peak (kN) | Time to steady contact force (s) | Steady contact force (kN) |
|---|---|---|---|---|
| 1:5 | 0.007 | 12.41 | 0.137 | 6.42 |
| 1:1 | 0.012 | 53.04 | 0.151 | 25.27 |
| 5:1 | 0.018 | 324.74 | 0.168 | 197.01 |
Higher density ratio increased the inertia load, lengthened the time to steady wheel speed, and raised the contact-force peak from 12.41 to 53.04 to 324.74 kN. The steady contact force rose from 6.42 to 25.27 to 197.01 kN. The fluctuation frequency decreased as inertia increased, but the fluctuation peak and settling time increased. For applications that require fast transient response, unnecessarily dense gear materials or large rotating inertia should be avoided.
The absolute contact-force baselines differ among the parameter sweeps, so the tables are intended for within-sweep comparison of trends rather than for cross-sweep normalization of absolute force levels.
Vibration and Transmission Error
Vibration was evaluated at the meshing node for the baseline case: pinion speed 100 rad/s, acceleration time 0.001 s, wheel torque 500 N·m, α = 500, and density ratio 1:1. The z direction is axial, while x and y represent the radial directions. The displacement and acceleration responses are summarized in Table 6.
| Component | Axial displacement (μm) | Radial displacement (μm) | Axial acceleration (m/s²) | Radial acceleration (m/s²) |
|---|---|---|---|---|
| Pinion | 31 | 11–50 | 15 | 5.5–22.5 |
| Wheel | 8 | 36–75 | 4 | 16–34 |
The pinion showed larger axial vibration displacement and acceleration, while the wheel showed larger radial vibration displacement and acceleration. In the axial direction, the pinion displacement reached 31 μm and the axial acceleration 15 m/s², compared with 8 μm and 4 m/s² for the wheel. In the radial direction, the pinion displacement was 11–50 μm and the radial acceleration 5.5–22.5 m/s², while the wheel displacement was 36–75 μm and the radial acceleration 16–34 m/s². The acceleration response follows the same anisotropy as the displacement response.
This pattern is consistent with a pinion support that is relatively compliant axially and stiff radially, and a wheel support that is relatively stiff axially and compliant radially. The radial stiffness of the pinion limits its radial deformation, while the wheel’s greater radial compliance allows larger radial motion. For mounting and bearing design, the axial support of the pinion and the radial support of the wheel are therefore the more critical directions.
Transmission error was evaluated as the deviation between the actual and theoretical angular displacement of the wheel. In the analyzed window, the error oscillated periodically. It remained relatively stable immediately after start-up and then showed a tendency to increase. This indicates that the initial meshing impact is not the only source of transmission error; the evolving contact state and dynamic response also contribute.
Design Implications and Scope
The parameter study supports several practical conclusions for spiral bevel gear design and operation:
- Speed selection involves a trade-off. Higher pinion speed raises the steady wheel speed and the contact force, and it lengthens the high-frequency contact-force fluctuation. However, above a threshold, higher speed can shorten the settling time. Very low speed can produce insufficient contact force and repeated loss of contact.
- Start-up control matters. A longer acceleration ramp reduces the initial contact-force peak and the wheel-speed fluctuation, improving meshing quality, although the absolute settling time is extended by the ramp itself.
- Damping improves settling but raises contact force. Increasing the mass-proportional damping coefficient suppresses wheel-speed oscillation and shortens the transient, but it increases the contact-force peak and the steady contact force.
- Inertia should be minimized for fast response. Higher density ratio increases the inertia load, contact-force peak, and settling time. Dense materials and large rotating inertia are unfavorable when rapid transient response is required.
- Support stiffness should be matched to the vibration direction. The pinion is more sensitive axially, and the wheel is more sensitive radially. Bearing and mounting design should account for this anisotropy.
- Explicit finite element analysis complements TCA and LTCA. TCA and LTCA remain efficient for contact-pattern and load-sharing design iterations; explicit dynamics is appropriate for start-up impact, transient contact force, and vibration response.
The results apply to the analyzed 20/59-tooth pair, a 500 N·m wheel torque, acceleration times of 0.001–0.005 s, mass-proportional damping of α = 100–1000, and density ratios of 1:5–5:1. Extrapolation to other gear geometries, loads, support stiffnesses, or lubrication conditions requires separate validation.