In this study, I present a comprehensive finite element contact analysis of a novel roller enveloping end face engagement worm drive. This type of worm gears is characterized by a worm that is generated by the envelope of roller teeth mounted on a worm wheel. The roller-tooth geometry replaces traditional sliding contact with rolling contact, which significantly reduces friction and wear. The worm gears in this work are designed to achieve high load-carrying capacity and improved efficiency, making them suitable for heavy-duty precision applications. I employ the ANSYS Workbench platform to perform a transient structural analysis, considering the actual meshing conditions and material properties. The results confirm that the roller enveloping end face engagement worm gears exhibit line contact, with up to five pairs of teeth in simultaneous engagement for a single worm segment. The contact stress distribution along the tooth flank follows a characteristic U
shape, with the maximum stress occurring near the root of the worm wheel tooth. The analysis also reveals that the equivalent stress on the worm is consistently higher than that on the worm wheel, and the deformation is largest at the outermost periphery of the worm. These findings provide a solid foundation for further development and optimization of high-performance worm gears.
1. Introduction
The roller enveloping end face engagement worm drive is an innovative transmission system that replaces conventional worm gear tooth surfaces with cylindrical rollers. In traditional worm gears, the sliding motion between the worm and worm wheel leads to significant power loss, heat generation, and rapid wear. By incorporating rollers as the worm wheel teeth, the friction mode changes from sliding to rolling, dramatically reducing these adverse effects. Furthermore, the enveloping generation process creates a line contact pattern that can distribute loads more evenly across multiple teeth. My objective is to investigate the contact behavior of this new type of worm gears using finite element analysis. The study focuses on the stress distribution, deformation characteristics, and load-sharing among the simultaneously engaged tooth pairs. I hope to demonstrate that the roller enveloping end face engagement worm gears possess superior load-carrying capacity compared to conventional designs, and to provide quantitative data for future design improvements.
2. Mathematical Modeling of the Worm Gear Tooth Surfaces
2.1 Coordinate Systems and Kinematic Relationships
To establish the tooth surface geometry of the roller enveloping end face engagement worm gears, I define a set of coordinate systems based on the differential geometry and the enveloping principle. Figure 1 (not shown) illustrates the arrangement: a fixed coordinate system attached to the worm, a fixed system attached to the worm wheel, and moving coordinate systems that rotate with their respective bodies. The rotation angles of the worm and worm wheel are denoted by \(\phi_1\) and \(\phi_2\), respectively, with the transmission ratio \(i_{12} = \omega_1/\omega_2 = z_2/z_1\), where \(z_1\) is the number of worm threads and \(z_2\) is the number of worm wheel rollers. The center distance is \(a\). A roller tooth is represented as a cylinder, and its axial line is fixed in the worm wheel coordinate system. The envelope of this cylinder family, as the worm wheel rotates, generates the worm thread surface.
2.2 Worm Surface Equation
According to the meshing theory, the worm surface is the envelope of the roller tooth surfaces. Let the roller tooth surface be described in its own coordinate system \(\sigma_0\) with position vector \(\mathbf{r}_0 = [x_0,\; y_0,\; z_0]^T\). After applying the coordinate transformations from \(\sigma_0\) to the moving worm coordinate system \(\sigma_{1′}\), and imposing the meshing condition, the worm surface equation becomes:
$$
\mathbf{r}_{1′} =
\begin{bmatrix}
x_1 \\ y_1 \\ z_1
\end{bmatrix}
=
\begin{bmatrix}
-\cos\phi_1\cos\phi_2(a_2-z_0) + \cos\phi_1\sin\phi_2 x_0 – y_0\sin\phi_1 + a\cos\phi_1 \\
\sin\phi_1\cos\phi_2(a_2-z_0) – \sin\phi_1\sin\phi_2 x_0 – y_0\cos\phi_1 – a\sin\phi_1 \\
-\sin\phi_2(a_2-z_0) – \cos\phi_2 x_0
\end{bmatrix}
$$
where \(a_2\), \(b_2\), \(c_2\) are the coordinates of the roller center in the worm wheel fixed system. The relationship \(\phi_2 = i_{21}\phi_1\) holds, with \(i_{21} = z_1/z_2\). The meshing condition \(\Phi = 0\) yields the envelope, but for brevity the full parametric system is not expanded here. The resulting worm surface is continuous and smooth, providing line contact with the roller teeth.
3. Finite Element Model Development
3.1 Model Geometry and Import
I generated the worm and worm wheel geometries using the tooth surface equations derived above. Starting from MATLAB, I computed a dense set of points on the worm spiral, then exported them to CREO Parametric to create solid models. The worm wheel was modeled as a disc with cylindrical roller teeth. After assembly, I established a seamless connection between CREO and ANSYS Workbench through the Geometry interface. To reduce computational cost, I extracted only one worm segment and the five adjacent roller teeth that are simultaneously in contact under typical operating conditions. This subassembly represents the critical meshing region of the worm gears.
| Parameter | Symbol | Value | Unit |
|---|---|---|---|
| Center distance | a | 125 | mm |
| Number of worm threads | z₁ | 1 | – |
| Number of worm wheel teeth (rollers) | z₂ | 25 | – |
| Roller radius | R | 11 | mm |
| Throat diameter coefficient | k₁ | 0.3 | – |
| Addendum coefficient | h*a | 0.8 | – |
| Dedendum coefficient | h*f | 0.8 | – |
| Clearance coefficient | c* | 0.2 | – |
3.2 Material Properties and Contact Definition
Both the worm and the worm wheel were made from 45# steel (equivalent to AISI 1045). The material properties used in the simulation are listed in Table 2. I set the contact type to frictionless because the rolling friction is negligible compared to the normal contact stresses, and the objective is to evaluate the Hertzian contact behavior. A no-separation contact condition was applied to allow sliding without penetration.
| Property | Value | Unit (SI) |
|---|---|---|
| Young’s modulus (E) | 2 × 10⁵ | MPa (2 × 10¹¹ Pa) |
| Shear modulus (G) | 7.6923 × 10⁴ | MPa (7.6923 × 10¹⁰ Pa) |
| Poisson’s ratio (ν) | 0.3 | – |
| Density (ρ) | 7850 | kg/m³ |
3.3 Mesh Generation
I used hexahedral dominant meshing for both the worm and the worm wheel. The initial mesh was generated with an automatic sizing method, and then a local refinement was applied to the contact region using the Refinement command. I iterated the mesh parameters until the element quality average (Element Quality Average) exceeded 0.8. The final model contained 133,658 elements and 225,151 nodes. Figure 2 shows the meshed assembly with refined contact zones.
3.4 Boundary Conditions and Loading
I performed a transient structural analysis using the Transient Structural module in ANSYS Workbench. The Lagrange multiplier method was employed for contact formulation. The worm was allowed to rotate only about its own axis, while all other degrees of freedom were constrained via remote displacement. A rotational velocity of 157 rad/s was applied to the worm inner surface to simulate a constant input speed. On the worm wheel, a resisting torque was applied at the inner bore surface to represent the output load. Four different input torque levels were studied: 15.60 N·m, 31.90 N·m, 38.99 N·m, and 46.76 N·m. The torque direction opposes the rotation driven by the worm. The simulation duration was 1 second, with a minimum time step of 0.0025 s to ensure convergence.
4. Results and Discussion
4.1 Equivalent Stress Distribution
Figures 3-6 (conceptual) present the von Mises stress contours for the worm and worm wheel under different input torques. I observed that the worm generally experiences higher equivalent stress than the worm wheel, because the worm thread has a smaller contact area and more severe bending. The maximum stress on the worm at the rated torque (31.90 N·m) was 64.115 MPa, while the worm wheel maximum was 36.32 MPa. As the torque increased, the stress magnitude rose proportionally. The stress distribution on the worm wheel was concentrated near the tooth root, indicating that root bending is the critical failure mode for the roller teeth in these worm gears.
4.2 Total Deformation
The total deformation of the worm gears increased with torque. The largest deformation always occurred at the outermost periphery of the worm, which is consistent with the overhung geometry. For example, when the input torque rose from 15.60 N·m to 46.76 N·m, the maximum deformation increased by 21.887%. This deformation is primarily elastic and reversible.
4.3 Contact Stress Along the Tooth Flank
I examined the contact stress on the five simultaneously engaged roller teeth (numbered from the entry side to the exit side). Figure 7 (conceptual) shows the contact stress distribution for each torque level. The maximum contact stress location gradually shifted from near the tooth tip (entry side) toward the tooth root (exit side). The peak contact stress on the exit-side tooth was significantly higher than on the entry-side tooth. Quantitatively, when the torque increased from 15.60 N·m to 46.76 N·m, the ratio of exit-side to entry-side maximum contact stress rose from 122.4% to 151.1%. This trend indicates that the exit side bears a larger portion of the load in the meshing cycle. The absolute maximum contact stress at the highest torque was 186.07 MPa, which is well below the yield strength of 45# steel, confirming the safety margin of the design.
4.4 Stress-Time History of Critical Nodes
I identified the node with the highest von Mises stress for each torque case (node IDs listed in Table 3). The stress-time curves (conceptual) show a steep rise during the first 0.1–0.3 seconds, followed by a plateau. For instance, at 15.60 N·m, the stress peaked at 38.582 MPa at t=0.2 s, then settled to about 33.5 MPa. At higher torques, the settling stress was proportionally higher. This behavior is typical of a transient contact problem where the initial impact is followed by stable rolling contact.
| Input torque (N·m) | Maximum stress node ID |
|---|---|
| 15.60 | 133323 |
| 31.90 | 12232 |
| 38.99 | 12232 |
| 46.76 | 12232 |
4.5 Effect of Load on Stress Distribution Along Contact Line
Taking the third meshing tooth pair as an example, I plotted the equivalent stress along the contact line for four load levels (0.5×, 1.0×, 1.25×, 1.5× rated torque). Figure 8 (conceptual) reveals a distinct U
-shaped distribution: the stress is highest at the two ends of the contact line (near the tooth root and tooth tip) and lower in the middle. As the load increased, the distribution became more uniform, but the peak near the root remained dominant. This confirms that root stress is the most critical for fatigue design of the worm gears.
The overall analysis demonstrates that the roller enveloping end face engagement worm gears exhibit excellent load-sharing characteristics, with five teeth in simultaneous contact. The line contact type reduces contact stress concentration compared to point contact in conventional worm gears. The maximum contact stress and equivalent stress are well within safe limits for 45# steel at the rated torque, indicating a high safety factor. The deformation pattern is predictable and manageable.
5. Conclusion
I have performed a finite element contact analysis of roller enveloping end face engagement worm gears using ANSYS Workbench. The key findings are:
- The proposed worm gears maintain line contact, with up to five tooth pairs simultaneously engaged, significantly enhancing load capacity.
- The worm experiences higher equivalent stress than the worm wheel, with the maximum stress located at the worm thread root.
- The contact stress distribution along the tooth flank is U-shaped, with the highest stress near the root of the exit-side tooth.
- Stress-time history shows an initial transient rise followed by steady-state contact, suitable for continuous operation.
- The maximum contact stress at 1.5× rated torque (186 MPa) is well below the material yield limit, confirming the robustness of the design.
These results provide essential guidance for the design optimization of high-load-capacity worm gears. Future work will include experimental validation and parametric studies to further improve efficiency and reduce weight.