A nutation drive with double circular-arc spiral bevel gears transmits motion through a small-angle wobbling motion of an external bevel gear relative to an internal bevel gear. The arrangement combines a large reduction ratio, a short axial envelope, and high load capacity, which makes it attractive for aerospace and robotic applications. The same compactness, however, restricts heat removal. Sliding between the meshing flanks generates frictional heat, and the resulting rise in gear body temperature can accelerate fatigue, change backlash through thermal distortion, and increase the risk of scuffing or seizure. The bulk temperature discussed here is the quasi-steady temperature of the gear body under repeated meshing, not the short-lived flash temperature at individual asperities.
The analysis combines tooth contact analysis, friction-heat estimation, and a single-tooth finite-element model. A control-variable approach is then used to isolate the effects of module, helix angle, nutation angle, and lubricating-oil temperature. The external bevel gear is the meshing gear examined; the internal gear provides the mating flank surface.
Baseline geometry and thermal problem
The baseline gear pair has a module of 2.5 mm, a helix angle of 25°, and a nutation angle of 5°. The external gear has 30 teeth and a pitch cone angle of 51.04°, while the internal gear has 32 teeth and a pitch cone angle of 123.96°. Both members have a pitch cone distance of 60 mm. The gear material is 45 steel, and the baseline environment temperature is 100 °C. These values define the reference case for the parameter study.
| Parameter | External bevel gear | Internal bevel gear |
|---|---|---|
| Helix angle β (°) | 25 | 25 |
| Nutation angle ε (°) | 5 | 5 |
| Number of teeth | 30 | 32 |
| Pitch cone angle (°) | 51.04 | 123.96 |
| Pitch cone distance (mm) | 60 | 60 |
| Module m (mm) | 2.5 | 2.5 |
In a compact nutation reducer, the gear pair is enclosed by other components, so convective cooling is limited. Heat generated by flank sliding is therefore not removed as readily as in an open or well-ventilated gearbox. The design question is not only how hot the gear becomes under one operating point, but how the bulk field responds to geometry and lubricant parameters that can be selected during design.
Contact path and frictional heat input
At any instant, the two meshing flanks must share the same position vector and normal vector in the fixed coordinate frame:
$$
\mathbf{r}_{m1}(\alpha_1,\theta_1,\phi_1)
=
\mathbf{r}_{m2}(\alpha_2,\theta_2,\phi_2),
\qquad
\mathbf{n}_{m1}(\alpha_1,\theta_1,\phi_1)
=
\mathbf{n}_{m2}(\alpha_2,\theta_2,\phi_2)
$$
Here, rm1 and rm2 are the position vectors of the external and internal gear flanks in the fixed frame, nm1 and nm2 are the corresponding normal vectors, α1 and α2 are the circular-arc profile parameters, θ1 and θ2 are the tooth-direction parameters, and φ1 and φ2 are the flank rotation angles. Advancing φ1 and solving the contact equations gives the contact points. Fifteen points were selected from the heel to the toe and used as the path of the moving heat source.
Only sliding friction is retained in the thermal load; rolling friction and elastic-plastic deformation are neglected because their contributions are small. The friction heat is estimated as
$$
Q=fP_{\max}v_{12}
$$
where f is the sliding friction coefficient, taken as 0.05, Pmax is the maximum contact stress, and v12 is the relative sliding velocity. The analysis assumes that the meshing gears absorb all of the generated friction heat and that the heat is divided equally between the two gears, so the partition coefficient is β′ = 0.5. The average heat flux entering each gear at a contact point is
$$
q_1=\beta’\frac{bQ\omega_1}{2\pi v_1},
\qquad
q_2=(1-\beta’)\frac{bQ\omega_2}{2\pi v_2}
$$
where b is the contact half-width, ω1 and ω2 are the angular velocities, and v1 and v2 are the tangential velocities at the contact point. Because the contact stress and sliding velocity vary along the path, the heat flux is not uniform from heel to toe.
Single-tooth thermal model and boundary conditions
The gear rotates much faster than the gear body can change its temperature distribution. Each tooth is therefore assumed to reach the same bulk temperature field, and a single-tooth model is used. With no internal heat source, steady heat conduction in the tooth is governed by
$$
\lambda\left(
\frac{\partial^2 T}{\partial x^2}
+
\frac{\partial^2 T}{\partial y^2}
+
\frac{\partial^2 T}{\partial z^2}
\right)=0
$$
where λ is the thermal conductivity of the gear material and T is the bulk temperature. On the meshing flank, conduction balances convection and the input friction flux:
$$
-\lambda\frac{\partial T}{\partial n}
=
\alpha_t(T_f-T_\tau)-q
$$
On the tooth tip, tooth root, and non-meshing flanks, the boundary condition is
$$
-\lambda\frac{\partial T}{\partial n}
=
\alpha’_t(T_f-T_\tau)
$$
and on the gear end faces it is
$$
-\lambda\frac{\partial T}{\partial n}
=
\alpha_s(T_f-T_\tau)
$$
Here, αt, α′t, and αs are the convective heat-transfer coefficients of the meshing flank, non-meshing flank or tip/root, and end face, respectively; Tf is the environment temperature; Tτ is the local gear body temperature; and n is the outward normal direction. The bottom surface of the tooth is treated as adiabatic, and the pitch-section faces use periodic conditions with equal temperature and equal normal gradient. The heat-transfer coefficients used in the model are listed in Table 2; all other surfaces have negligible convection and are treated as zero.
| Boundary | Coefficient | Value (W/(m²·K)) |
|---|---|---|
| Meshing flank | αt | 2198.4 |
| Non-meshing flank, tip, and root | α′t | 1681.3 |
| End face | αs | 5043.8 |
The moving heat source is applied sequentially over the 15 contact points in an ANSYS/APDL model to obtain the bulk temperature field of the external bevel gear.
Baseline bulk temperature field
The calculated field is not uniform. The highest temperature appears in the central part of the meshing contact zone, where the friction heat flux is concentrated. The tooth tip has better cooling conditions, and the gear end faces receive less heat and lose heat readily, so these regions remain relatively cool. Heat conduction spreads the temperature from the contact centre outward, and along the face width the bulk temperature follows a roughly parabolic profile. The convex and concave flanks show the same qualitative pattern.
This distribution matters because the temperature gradient, rather than a single average value, drives thermal distortion. A hot contact centre combined with cooler end faces can alter the local contact pattern and clearance, which in turn affects vibration, noise, and fatigue behaviour. The bulk field therefore provides a useful link between frictional heating and mechanical response.
Parametric influences on bulk temperature
The maximum bulk temperature on the convex flank was used to compare cases. Each parameter was varied independently from the baseline, and the resulting changes are summarized in Table 3. The percentages for module, helix angle, and nutation angle are referenced to the lowest tested value of each parameter, not to the baseline design point.
| Parameter | Tested values | Change in maximum bulk temperature | Reference case |
|---|---|---|---|
| Module m (mm) | 2.0, 2.25, 2.5, 3.0, 3.5 | 0, −2.1%, −3.5%, −7.8%, −10.5% | m = 2.0 mm |
| Helix angle β (°) | 18, 20, 25, 30, 35 | 0, +4.6%, +7.2%, +11.2%, +22.2% | β = 18° |
| Nutation angle ε (°) | 3, 4, 5, 6, 7 | 0, −10.3%, −23.6%, −12.9%, −0.74% | ε = 3° |
| Lubricating-oil temperature | +10 °C increments | +10 °C per +10 °C | Linear slope ≈ 1 |
Module
Increasing the module lowers the bulk temperature at all points on the tooth. Relative to the 2.0 mm case, the maximum temperature on the convex flank decreased by 2.1% at 2.25 mm, 3.5% at 2.5 mm, 7.8% at 3.0 mm, and 10.5% at 3.5 mm. The mechanism is both geometric and thermal. A larger module changes the local curvature radius, reduces the contact pressure and the friction heat flux, and increases the convective heat-transfer coefficient of the tooth. The net effect is a cooler tooth surface and body.
The module is therefore a coupled design variable. Increasing it can reduce thermal risk, but it also increases gear size and mass and changes the bending-strength balance of the pair. The thermal benefit should be weighed against the compactness requirement that makes the nutation drive attractive in the first place.
Helix angle
The bulk temperature rises monotonically with helix angle in the tested range. Relative to 18°, the maximum temperature increased by 4.6% at 20°, 7.2% at 25°, 11.2% at 30°, and 22.2% at 35°. The penalty becomes particularly steep between 30° and 35°.
A larger helix angle increases the axial force during operation, which raises the contact pressure and the friction heat flux at the mesh. The result is a higher tooth temperature. This does not mean that a small helix angle is always preferable, because the helix angle also affects contact ratio, load sharing, and transmission smoothness. It does mean that thermal management becomes more demanding as the helix angle is increased for other design objectives.
Nutation angle
The nutation angle produces a non-monotonic response. Relative to 3°, the maximum bulk temperature fell by 10.3% at 4° and by 23.6% at 5°. It then rose again, although it remained 12.9% below the 3° case at 6° and 0.74% below it at 7°. Within the tested range, the thermal minimum occurred at 5°.
Two competing mechanisms explain this behaviour. On one hand, increasing the nutation angle raises the entrainment velocity at the inlet region, drawing more lubricating oil into the meshing gap and removing more heat. On the other hand, a larger nutation angle increases the curvature at the contact point, increases the contact stress, and reduces the convective heat-transfer coefficient, all of which tend to raise the tooth temperature. From 3° to 5°, the increased oil entrainment dominates. At 6° and 7°, the contact-stress increase and the reduced convection become dominant, so the temperature rises again. The existence of a minimum means that simply increasing the nutation angle is not a reliable cooling strategy.
Lubricating-oil temperature
The effect of lubricating-oil temperature is approximately linear. For every 10 °C increase in oil temperature, the overall gear bulk temperature increased by about 10 °C. In the tested range, the slope is close to unity:
$$
T_{\mathrm{bulk}}\approx T_{\mathrm{oil}}+C
$$
where C is a constant offset for the given geometry and operating condition. Lowering the oil supply temperature therefore shifts the entire bulk field downward by nearly the same amount. This makes oil temperature a direct and strong thermal-management variable. It does not, however, change the shape of the temperature distribution: the contact centre remains the hottest region, and the end faces remain cooler.
Design implications and limits of the model
The results point to three practical conclusions. First, module and helix angle influence bulk temperature through contact pressure, local curvature, and heat-transfer area. Selecting these parameters only for strength or compactness can leave a significant thermal penalty. Second, the nutation angle has a non-monotonic effect, so a thermal optimum should be identified together with contact stress and oil entrainment rather than assumed to improve monotonically. Third, lubricating-oil temperature acts as a nearly one-for-one offset on the bulk field, making coolant temperature an effective but shape-preserving control measure.
The field is hottest at the contact centre and cooler at the end faces, so cooling provision and oil-jet placement should account for this distribution rather than treating the whole gear as isothermal. The numerical model uses a single tooth, a fixed heat partition, and prescribed boundary heat-transfer coefficients. Absolute temperatures therefore depend on those inputs, while the reported trends are the more transferable design information. Case-specific thermal design should confirm the trends with boundary conditions and operating data representative of the actual reducer.