I present a complete numerical framework for the loaded tooth contact analysis of a face gear pair that is modified to achieve localized line contact. My aim is to improve the load capacity of the face gear pair, suppress edge contact, and preserve a high computational efficiency for iterative design. The face gear transmission has many attractive properties, including high contact ratio, no axial force on the pinion, and good torque splitting. However, the actual meshing state of a face gear pair is sensitive to manufacturing errors and assembly errors. Such errors often push the contact path toward the tooth tip or the inner radius, producing edge contact and stress concentration. In addition, the generated face gear tooth surface has a gradually changing profile, and its topology is further complicated by any modification. Therefore, a purely analytical loaded contact model for a modified face gear pair is difficult to construct. Finite element analysis can provide accurate results, but it is computationally expensive and inconvenient for parametric optimization. I therefore propose a localized line contact modification method and an influence-coefficient-based numerical model for the face gear pair. The model uses a digital contact line representation, a prescribed contact path, and a superposition of modification amounts on discrete nodes. It then computes the tooth surface compliance by combining global finite element deformation with a modified Boussinesq contact deformation. The final contact stress and load distribution are obtained by solving deformation compatibility equations together with the torque equilibrium condition. The results show that the prescribed contact path agrees with the simulated contact path, that the contact stress magnitude and distribution agree well with the finite element reference, and that the computational efficiency is significantly improved. Compared with a point contact face gear pair, the localized line contact face gear pair reduces the maximum contact stress and avoids edge contact.
Keywords: face gear; modification method; meshing performance; loaded tooth contact analysis; numerical method; face gear pair; localized line contact; contact stress; load distribution; edge contact; influence coefficient method; deformation compatibility; torque equilibrium.
Nomenclature
| Symbol | Meaning |
|---|---|
| \(\mathbf{r}_s\) | Tool tooth surface position vector |
| \(\mathbf{r}_2\) | Standard face gear tooth surface position vector |
| \(\mathbf{r}_2’\) | Modified face gear tooth surface position vector |
| \(u_b\) | Tooth width coordinate of the generating tool |
| \(\theta_b\) | Involute angle parameter |
| \(\theta_{bs}\) | Half angle of the tool tooth space |
| \(\phi_b\) | Generating motion angle |
| \(\gamma_m\) | Shaft angle between the tool and the face gear |
| \(m_{bs}\) | Tool module parameter |
| \(\mathbf{M}_{2b}\) | Transformation matrix from tool frame to face gear frame |
| \(\mathbf{n}_2\) | Unit normal vector of the face gear tooth surface |
| \(\omega(y)\) | Modification amount along a contact line |
| \(w(x,y)\) | Modification amount for off-path contact lines |
| \(\mathbf{K}\) | Global stiffness matrix |
| \(\mathbf{F}\) | Nodal force vector |
| \(\boldsymbol{\eta}\) | Nodal displacement vector |
| \(f_{mk-p}\) | Total deformation influence coefficient from point \(m\) to point \(k\) on the pinion |
| \(f_{mk-g}\) | Total deformation influence coefficient from point \(m\) to point \(k\) on the face gear |
| \(T\) | Transmitted torque |
| \(\delta\) | Rigid body approach |
| \(F_m\) | Discrete contact force at point \(m\) |
| \(v_m\) | Torque coefficient at point \(m\) |
1. Geometry of the Standard Face Gear Tooth Surface
I begin with the standard face gear tooth surface because the modification method is applied directly to the digitally discretized contact lines of the face gear. The face gear is generated by a shaper cutter whose geometry matches the pinion. The cutter and the face gear perform a generating motion according to a prescribed ratio. The envelope of the cutter tooth surface family forms the theoretical face gear tooth surface. The coordinate systems include fixed frames attached to the machine frame and rotating frames attached to the face gear and the cutter. This arrangement is standard for face gear generation, but I express it in a form suitable for numerical discretization.
The cutter tooth surface can be written as
$$
\mathbf{r}_s(u_b,\theta_b)=
\begin{bmatrix}
r_b\left[\sin(\theta_{bs}+\theta_b)-\theta_b\cos(\theta_{bs}+\theta_b)\right]\\
-r_b\left[\cos(\theta_{bs}+\theta_b)+\theta_b\sin(\theta_{bs}+\theta_b)\right]\\
u_b\\
1
\end{bmatrix}
$$
where \(r_b\) is the base radius of the cutter, \(u_b\) is the tooth width coordinate, \(\theta_b\) is the involute parameter, and \(\theta_{bs}\) is the half angle of the tooth space. The transformation from the cutter frame to the face gear frame is
$$
\mathbf{M}_{2b}(\phi_b)=
\begin{bmatrix}
\cos\phi_b\cos\phi_m & \cos\phi_b\sin\phi_m\sin\gamma_m & \sin\phi_b\cos\gamma_m & 0\\
-\sin\phi_b\cos\phi_m & -\sin\phi_b\sin\phi_m\sin\gamma_m & \cos\phi_b\cos\gamma_m & 0\\
-\sin\phi_m\cos\gamma_m & \cos\phi_m\cos\gamma_m & \sin\gamma_m & 0\\
0 & 0 & 0 & 1
\end{bmatrix}
$$
where \(\phi_b\) is the generating angle and \(\gamma_m\) is the shaft angle between the cutter and the face gear. The meshing equation that relates the generating motion to the tooth surface parameters is
$$
f(u_b,\theta_b,\phi_b)
=
(1-\cos\gamma_m)\sin\phi_b\, r_b
–
\cos\gamma_m\, u_b
+
m_{bs}\cos(\theta_{bs}+\theta_b)
=
0
$$
The standard face gear tooth surface is then obtained by eliminating \(\phi_b\) from the meshing equation:
$$
\mathbf{r}_2(u_b,\theta_b)=
\mathbf{M}_{2b}(\phi_b)\mathbf{r}_b(u_b,\theta_b),
\qquad
f(u_b,\theta_b,\phi_b)=0
$$
This surface is the basis for my discretization. I sample the contact lines of the face gear tooth surface. Each contact line is a spatial curve. For a standard face gear pair, the theoretical contact is a line contact. However, the line contact is sensitive to assembly errors and may produce edge contact at the tooth tip and the inner radius. A point contact face gear pair is more tolerant to errors, but it reduces the contact area and increases the contact stress. My modification method seeks a compromise: a localized line contact that retains a controllable contact line length while avoiding the tooth edges.
2. Localized Line Contact Modification Method
I modify the face gear tooth surface along the digitally discretized contact lines. The modification is defined by two parts. The first part creates a finite contact line segment along each contact line. The second part ensures that contact lines away from the prescribed contact path do not participate in contact. This produces a localized line contact face gear pair. I prescribe a contact path on the face gear tooth surface. The path starts at \(m_1\) and ends at \(m_2\). The midpoint is \(M\). I construct a parabolic curve through the start point and the end point. The projection of this curve onto the face gear tooth surface provides the prescribed contact path. This path is also used as the midpoints of the localized contact lines.
The parabolic path can be expressed as
$$
\mathbf{u}(t)=
\mathbf{M}
+
t\hat{\mathbf{u}}
+
\left(
\frac{d}{t_1^2}t^2+d
\right)\hat{\mathbf{a}},
\qquad
t\in[-t_1,t_1]
$$
where \(\hat{\mathbf{u}}=(\mathbf{m}_2-\mathbf{m}_1)/\|\mathbf{m}_2-\mathbf{m}_1\|\), \(\hat{\mathbf{a}}\) is perpendicular to \(\hat{\mathbf{u}}\), \(t_1=\hat{\mathbf{u}}\cdot(\mathbf{m}_1-\mathbf{M})\), and \(d\) is the vertex height. The prescribed contact path is then projected onto the face gear tooth surface. Along each contact line, I superimpose a piecewise modification function. This function is zero over a central segment and parabolic outside the central segment. The central segment defines the localized line contact length. The piecewise function is
$$
\omega(y)=
\begin{cases}
A_1(y-y_{a1})^2, & y<y_{a1}\\ $$="" &="" 0,="" <p="" \end{cases}="" a_2(y-y_{a2})^2,="" a_3(y-y_{a3})^2,="" y\ge="" y\le="" y_{a1}\le="" y_{a2}<yHere \(A_1,A_2,A_3\) are modification coefficients. The unmodified central region is controlled by the contact line length coefficient \(t\), which determines the locations \(y_{a2}\) and \(y_{a3}\). For contact lines that are not on the prescribed contact path, I superimpose an additional modification over the entire contact line. This additional modification varies along the tooth width direction and is written as
$$
w(x,y)=
\begin{cases}
B_1(y-y_{m1})^2, & yy_{m2}
\end{cases}
$$
The final modified face gear tooth surface is
$$
\mathbf{r}_2′(u_b,\theta_b)=
\mathbf{r}_2(u_b,\theta_b)
+
\left[\omega(y)+w(x,y)\right]
\mathbf{n}_2(u_b,\theta_b)
$$
where \(\mathbf{n}_2(u_b,\theta_b)\) is the unit normal vector of the standard face gear tooth surface:
$$
\mathbf{n}_2(u_b,\theta_b)=
\frac{
\frac{\partial \mathbf{r}_2}{\partial u_b}
\times
\frac{\partial \mathbf{r}_2}{\partial \theta_b}
}{
\left\|
\frac{\partial \mathbf{r}_2}{\partial u_b}
\times
\frac{\partial \mathbf{r}_2}{\partial \theta_b}
\right\|
}
$$
I use this modified surface to build the numerical contact model. The modification is applied on the digital nodes of the contact lines, so the resulting surface is smooth enough for contact analysis and flexible enough for parametric studies. The localized line contact face gear pair is thus defined by the prescribed contact path, the contact line length coefficient, and the modification coefficients. The point contact and full line contact cases can be recovered as limiting cases.
3. Global Deformation of the Face Gear Tooth
The contact deformation of a face gear pair includes global deformation, such as bending and shear, and local contact deformation. Because the face gear tooth has a complex geometry, I compute the global deformation by the finite element method. I discretize the modified face gear tooth and the mating pinion tooth into eight-node hexahedral isoparametric elements. I apply full constraints to the bottom region of each tooth according to the actual boundary conditions of the face gear transmission. The element stiffness matrix is
$$
\mathbf{k}^e=
\int_{-1}^{1}
\int_{-1}^{1}
\int_{-1}^{1}
\mathbf{B}^T
\mathbf{D}
\mathbf{B}
\left|\mathbf{J}\right|
d\xi\,d\eta\,d\zeta
$$
where \(\mathbf{D}\) is the elasticity matrix, \(\mathbf{J}\) is the Jacobian matrix, and \(\mathbf{B}\) is the strain-displacement matrix. The matrix \(\mathbf{B}\) is assembled from submatrices \(\mathbf{B}_i\):
$$
\mathbf{B}=
\begin{bmatrix}
\mathbf{B}_1 & \mathbf{B}_2 & \cdots & \mathbf{B}_8
\end{bmatrix}
$$
$$
\mathbf{B}_i=
\begin{bmatrix}
\partial N_i/\partial x & 0 & 0\\
0 & \partial N_i/\partial y & 0\\
0 & 0 & \partial N_i/\partial z\\
\partial N_i/\partial y & \partial N_i/\partial x & 0\\
0 & \partial N_i/\partial z & \partial N_i/\partial y\\
\partial N_i/\partial z & 0 & \partial N_i/\partial x
\end{bmatrix}
$$
I use the B-Bar method to reduce volumetric locking and shear locking. The global stiffness matrix is assembled as
$$
\mathbf{K}=
\sum_e \mathbf{k}^e
$$
For a nodal force vector \(\mathbf{F}\), the nodal displacement vector \(\boldsymbol{\eta}\) is obtained from
$$
\mathbf{K}\boldsymbol{\eta}=\mathbf{F}
$$
For the tooth surface, each node has three force components and three displacement components. If I apply a unit force at node \(m\) along the \(x\) direction, then the displacement at any node \(p\) along the \(x\) direction is an influence coefficient. Repeating this process for all nodes and all three directions gives the global compliance matrix of the tooth surface. For a point that is not a node, I use the shape functions to distribute the force to the eight nodes of the element. The equivalent nodal forces are
$$
F_k=
N_k(\xi,\eta,\zeta)F_\epsilon,
\qquad
k=1,2,\ldots,8
$$
Since the force is applied on the tooth surface, the nodes on the loaded surface have much larger shape function values than the nodes on the opposite surface. I therefore neglect the equivalent forces on the non-loaded nodes. The displacement at any point on the tooth surface is obtained by superposition:
$$
\eta_{p}^{x}(u_b,\theta_b)=
\sum_{m}
\sum_{k}
F_k
\eta_{m-p}^{x-x}
N_k(u_b,\theta_b)
$$
For numerical implementation, I interpolate the influence coefficients over the tooth surface using the generation parameters \(u_b\) and \(\theta_b\). This yields a continuous compliance representation for the global deformation of the modified face gear tooth.
4. Contact Deformation and Modified Boussinesq Influence Coefficients
The local contact deformation is computed from a half-space force-displacement relation. For point contact, the classical Boussinesq formula gives the normal displacement as
$$
W=
\frac{1-\nu^2}{\pi E}
\iint_{\Omega}
\frac{p(x,y)}
{\left|S_X-S_{X’}\right|}
dx\,dy
$$
where \(\nu\) is Poisson’s ratio, \(E\) is the elastic modulus, \(p(x,y)\) is the contact pressure, and \(S_X\) and \(S_{X’}\) are the source and observation points. This formula is accurate for a point contact or a straight contact line on a flat surface. However, the contact line of a localized line contact face gear pair is a spatial curve. The tooth surface is a complex spatial surface. Therefore, I modify the Boussinesq formula by constructing a surface along the contact line. I take the tangent vector of the contact line and the normal vector of the tooth surface to generate a local surface:
$$
\mathbf{r}(v,u_b)=
\boldsymbol{\alpha}(v)
+
u_b\boldsymbol{\beta}(v)
$$
where \(\boldsymbol{\alpha}(v)\) is a point on the contact line, and \(\boldsymbol{\beta}(v)\) is the normal vector at that point. The modified Boussinesq influence coefficient is
$$
W=
\frac{1-\nu^2}{\pi E}
\iint_{\Omega}
\frac{
\mathbf{p}(u_b,v)\cdot
\mathbf{n}_{X1}(u_b,v)
}{
\left|
S_X(u_b,v)-S_{X’}(u_b,v)
\right|
}
du_b\,dv
$$
where the normal vector of the constructed surface is
$$
\mathbf{n}_{X1}(u_b,v)=
\frac{\partial \mathbf{r}}{\partial v}
\times
\frac{\partial \mathbf{r}}{\partial u_b}
$$
This modification allows the contact deformation to follow the curvature of the spatial contact line. It is essential for a localized line contact face gear pair because the contact line is not straight and the tooth surface curvature changes along the path.
5. Contact Area Estimation and Total Compliance
The contact area of a face gear pair is not known in advance. I estimate the initial contact area on a plane perpendicular to the normal at the prescribed contact point. On this plane, I define an elliptical region. The long axis is set to the contact line length. The short axis is estimated from the ratio of the principal curvatures:
$$
\frac{a}{b}=
\left(
\frac{|B|}{|A|}
\right)^{1/2}
$$
where \(A\) and \(B\) are the principal curvatures of the two tooth surfaces. The initial contact area is intentionally overestimated so that the iterative solver can remove non-contacting regions. I map the points on the contact plane back to the exact tooth surfaces. For each point \(k\) on the contact plane, I compute the total deformation influence coefficient from a unit force at point \(m\):
$$
f_{mk-p}=
f_{m\eta-p}
+
f_{mw-p}
$$
where \(f_{m\eta-p}\) is the global deformation coefficient and \(f_{mw-p}\) is the contact deformation coefficient. The global coefficient is obtained from the finite element compliance matrix, and the contact coefficient is obtained from the modified Boussinesq formula. For a general force direction, the influence coefficient is transformed by the surface normal components:
$$
f_{mk-p}=
\begin{bmatrix}
n_x & n_y & n_z
\end{bmatrix}
\begin{bmatrix}
\eta_{xx} & \eta_{xy} & \eta_{xz}\\
\eta_{yx} & \eta_{yy} & \eta_{yz}\\
\eta_{zx} & \eta_{zy} & \eta_{zz}
\end{bmatrix}
\begin{bmatrix}
n_x\\
n_y\\
n_z
\end{bmatrix}
$$
The same operation is performed for the mating pinion tooth. The total deformation coefficient for the contact pair is the sum of the pinion and face gear contributions:
$$
f_{mk}=
f_{mk-p}
+
f_{mk-g}
$$
This total compliance matrix is the core of the loaded tooth contact analysis. It includes the global bending and shear deformation of both teeth and the local contact deformation of the face gear pair.
6. Deformation Compatibility and Torque Equilibrium
Under load, the contact points in the contact region deform. Let \(\delta_p\) and \(\delta_g\) be the deformations of the pinion and face gear at a contact pair. The total maximum deformation is
$$
\delta=
\delta_p+\delta_g
$$
For any contact point \(k\), the separation distance before loading is \(d_k\). After loading, the deformation compatibility condition is
$$
d_k+
\delta_{k-p}+
\delta_{k-g}
=
\delta
$$
where \(\delta_{k-p}\) and \(\delta_{k-g}\) are the deformations of the pinion and face gear at point \(k\). These deformations are expressed as
$$
\delta_{k-p}=
\sum_{m=1}^{n}
F_m f_{mk-p}
$$
$$
\delta_{k-g}=
\sum_{m=1}^{n}
F_m f_{mk-g}
$$
where \(F_m\) is the discrete contact force at point \(m\), and \(n\) is the number of contact points in the candidate contact region. The discrete force is related to the average pressure \(P_m\) on the rectangular cell \(m\) by
$$
F_m=
P_m s_m
$$
where \(s_m\) is the area of the cell. The static torque equilibrium condition is
$$
\sum_{m=1}^{n}
F_m v_m
=
T
$$
where \(T\) is the transmitted torque and \(v_m\) is the torque coefficient:
$$
v_m=
\mathbf{n}_m
\times
\mathbf{r}_m
$$
Here \(\mathbf{r}_m\) is the position vector of point \(m\). Combining the deformation compatibility equations and the torque equilibrium condition gives the linear system
$$
\begin{bmatrix}
\mathbf{f} & \mathbf{I}\\
\mathbf{v}^T & 0
\end{bmatrix}
\begin{bmatrix}
\mathbf{F}\\
\boldsymbol{\delta}
\end{bmatrix}
=
\begin{bmatrix}
\mathbf{d}\\
T
\end{bmatrix}
$$
where \(\mathbf{f}\) is the total compliance matrix, \(\mathbf{I}\) is the identity vector, \(\mathbf{F}\) is the discrete force vector, \(\boldsymbol{\delta}\) is the rigid body approach, \(\mathbf{d}\) is the initial separation vector, and \(\mathbf{v}\) is the torque coefficient vector. The compliance matrix is
$$
\mathbf{f}=
\begin{bmatrix}
f_{11} & f_{12} & \cdots & f_{1n}\\
f_{21} & f_{22} & \cdots & f_{2n}\\
\vdots & \vdots & \ddots & \vdots\\
f_{n1} & f_{n2} & \cdots & f_{nn}
\end{bmatrix}
$$
Because the initial contact area is overestimated, the first solution of the linear system may contain negative forces. Negative forces are not physical in a unilateral contact problem. I therefore set all negative forces to zero and remove the corresponding cells from the contact region. I repeat the solution until all forces are non-negative. The remaining cells define the actual contact area. This iterative procedure is robust and converges quickly for the localized line contact face gear pair.
7. Numerical Algorithm
I summarize the numerical algorithm in the following table. The algorithm is implemented in a computer program for the loaded tooth contact analysis of the modified face gear pair.
| Step | Operation | Output |
|---|---|---|
| 1 | Generate the standard face gear tooth surface from the meshing equation | \(\mathbf{r}_2(u_b,\theta_b)\) |
| 2 | Discretize the tooth surface into contact lines and nodes | Contact line data |
| 3 | Prescribe the contact path and construct the parabolic path | \(\mathbf{u}(t)\) |
| 4 | Superimpose the contact line modification and the off-path modification | \(\mathbf{r}_2′(u_b,\theta_b)\) |
| 5 | Build the finite element mesh of the pinion and face gear teeth | Element and node data |
| 6 | Assemble the global stiffness matrix and compute the compliance matrix | \(\mathbf{K}^{-1}\) |
| 7 | Compute the contact deformation by the modified Boussinesq formula | \(f_{mw-p}\) |
| 8 | Estimate the initial contact ellipse and map points to the tooth surfaces | Candidate contact points |
| 9 | Assemble the total compliance matrix | \(\mathbf{f}\) |
| 10 | Solve the deformation compatibility and torque equilibrium system | \(\mathbf{F}\), \(\boldsymbol{\delta}\) |
| 11 | Remove negative forces and update the contact area | Actual contact region |
| 12 | Repeat until all forces are non-negative | Contact stress and load distribution |
8. Validation of the Contact Path and Modification Method
I validate the localized line contact modification method by comparing the prescribed contact path with the simulated contact path. I use a face gear pair with the basic parameters listed in the following table. The pinion is the driving gear. I start from an initial pinion rotation angle and increase the angle with a constant step. For each step, I compute the contact ellipse on the face gear tooth surface. I rotate the contact ellipses back to the initial tooth surface and obtain a continuous contact pattern. The midpoint of the long axis of each contact ellipse forms the contact path. The simulated contact path agrees with the prescribed contact path. This confirms that the modification method can realize the intended contact path on the face gear tooth surface.
| Parameter | Value |
|---|---|
| Module | 4 mm |
| Pinion tooth number | 32 |
| Face gear tooth number | 90 |
| Face width | 15 mm |
| Pressure angle | 25 degrees |
| Contact line length coefficient \(t\) | 0.2 |
| Modification coefficients \(A_1,A_2\) | \(0.8\times 10^{-3}\) |
| Modification coefficients \(B_1,B_2\) | \(1.5\times 10^{-3}\) |
9. Error Sensitivity of Different Contact Types
I analyze the sensitivity of three contact types to assembly errors: full line contact, localized line contact, and point contact. The assembly errors include shaft angle error, shaft offset error, and axial misalignment error. I define the shaft angle error as \(\Delta\gamma\), the shaft offset error as \(\Delta q\), and the axial misalignment error as \(\Delta E\). For each error case, I compute the contact path on the face gear tooth surface. I use the Y-direction displacement of the contact path tip divided by the face width as an error sensitivity index.
| Contact type | Error sensitivity | Edge contact risk | Contact area |
|---|---|---|---|
| Full line contact | Highest | Highest at tooth tip and inner radius | Largest |
| Localized line contact | Medium | Low within a certain error range | Large and controllable |
| Point contact | Lowest | Moderate at tooth tip | Smallest |
The results show that the contact paths of all three contact types shift under assembly errors. The point contact path extends toward the tooth tip, which can cause tooth tip edge contact. The full line contact path extends toward both the tooth tip and the inner radius, so its edge contact probability is higher than that of point contact. The localized line contact path also shifts, but the shift is constrained by the prescribed contact path. Within a certain error range, the localized line contact face gear pair completely avoids contact at the tooth edges. The sensitivity ranking from high to low is full line contact, localized line contact, and point contact. For the three error types, the sensitivity ranking from high to low is shaft angle error, shaft offset error, and axial misalignment error. The localized line contact method therefore provides a better error tolerance than full line contact and avoids the tooth tip edge contact that can occur in point contact.
10. Contact Stress Validation Against Finite Element Analysis
I verify the numerical loaded tooth contact analysis by comparing its results with a finite element reference. The material properties are the same in both models: elastic modulus 210 GPa and Poisson’s ratio 0.3. The finite element model uses hexahedral elements. I compute the maximum contact stress along the contact path and the contact stress distribution at several rotation angles. The following table summarizes the comparison.
| Rotation angle | Numerical maximum stress | Finite element maximum stress | Difference |
|---|---|---|---|
| Initial meshing | Reference value | Reference value | 7.4% |
| 0.15 rad | Reference value | Reference value | 1.2% |
| End of meshing | Reference value | Reference value | 1.1% |
| Overall maximum | Reference value | Reference value | 8.5% |
The maximum contact stress from the numerical method follows the same trend as the finite element result. The largest difference occurs near the beginning of the meshing cycle and is 8.5%. The contact stress distributions at the initial meshing position, at 0.15 rad, and at the end of meshing are consistent between the two methods. The maximum stress differences at these positions are 7.4%, 1.2%, and 1.1%. The meshing-in and meshing-out positions agree with the prescribed contact path. These results confirm that the proposed numerical method is accurate for the localized line contact face gear pair.
11. Loaded Contact Performance of Localized Line Contact and Point Contact
I compare the maximum contact stress of a point contact face gear pair and a localized line contact face gear pair. Before 0.3 rad, both contact types show the same trend: the maximum contact stress first increases and then decreases. After 0.3 rad, the localized line contact continues to decrease, while the point contact increases again. The increase in the point contact case indicates edge contact at the tooth tip. In the region without edge contact, that is, before 0.3 rad, the localized line contact reduces the maximum contact stress by an average of 17.4% compared with point contact. Over the entire rotation interval, including the edge contact region, the localized line contact reduces the maximum contact stress by an average of 30%. This demonstrates that the localized line contact face gear pair effectively suppresses edge contact and reduces stress concentration.
| Contact type | Average stress reduction before edge region | Average stress reduction over full interval | Edge contact behavior |
|---|---|---|---|
| Point contact | Baseline | Baseline | Tooth tip edge contact appears after 0.3 rad |
| Localized line contact | 17.4% lower | 30% lower | Edge contact avoided |
12. Computational Efficiency
I also compare the computational efficiency of the proposed numerical method with the finite element method. Both computations are performed on the same computer with an eight-core processor, 16 GB of memory, and a frequency of 3 GHz. The finite element analysis uses six parallel cores and a mass scaling factor of 100. For a single-tooth loaded contact analysis, the finite element method requires 4.51 hours. The numerical method requires 0.77 hours. The numerical method therefore improves the computational efficiency by approximately 83% while maintaining a comparable accuracy. This efficiency improvement is important for iterative modification optimization of the face gear pair, because many design candidates must be evaluated.
| Method | Computational time | Efficiency improvement |
|---|---|---|
| Finite element analysis | 4.51 h | Baseline |
| Proposed numerical method | 0.77 h | 83% |
13. Discussion
The localized line contact modification method offers a practical way to improve the load performance of a face gear pair. The method starts from the standard face gear tooth surface and modifies the contact lines instead of modifying the entire tooth surface. This keeps the modification localized and makes the contact path controllable. The prescribed contact path is defined by a parabolic curve between two points on the face gear tooth surface. The contact line length is controlled by a piecewise function. The off-path contact lines are modified so that they do not participate in contact. The result is a face gear pair that has a finite contact line segment on each engaged tooth pair. This localized line contact retains a larger contact area than point contact and avoids the edge contact that often occurs in full line contact.
The numerical loaded tooth contact analysis model combines finite element global deformation with a modified Boussinesq contact deformation. The finite element method captures the bending and shear deformation of the complex face gear tooth. The modified Boussinesq formula captures the local contact deformation along the spatial contact line. The influence coefficient method assembles the total compliance matrix without requiring a full finite element contact simulation for every load case. The deformation compatibility equations and the torque equilibrium condition are solved together. The iterative removal of negative forces determines the actual contact area. This numerical procedure is accurate, efficient, and suitable for parametric studies.
The error sensitivity analysis shows that the localized line contact face gear pair is less sensitive to assembly errors than a full line contact face gear pair and is more resistant to edge contact than a point contact face gear pair. The localized line contact face gear pair therefore combines the advantages of both contact types. It preserves a high load capacity and improves the tolerance to manufacturing and assembly errors. The maximum contact stress reduction of 30% over the full rotation interval is a significant improvement for the durability and reliability of the face gear pair.
14. Summary of Results
| Item | Result |
|---|---|
| Prescribed contact path | Agrees with simulated contact path |
| Maximum contact stress error | Less than 9% compared with finite element analysis |
| Contact stress distribution | Consistent with finite element analysis |
| Computational efficiency | 83% improvement over finite element analysis |
| Stress reduction before edge region | 17.4% lower than point contact |
| Stress reduction over full interval | 30% lower than point contact |
| Edge contact | Avoided by localized line contact |
| Error sensitivity ranking | Full line contact > localized line contact > point contact |
| Error type sensitivity ranking | Shaft angle error > shaft offset error > axial misalignment error |
15. Conclusions
I have presented a localized line contact modification method and a numerical loaded tooth contact analysis model for a face gear pair. The modification method uses the contact lines of the standard face gear tooth surface and a prescribed contact path. The modification amount is superimposed on discrete nodes along the contact line and along the prescribed path. This produces a face gear pair with localized line contact characteristics. The numerical model uses finite element analysis for global deformation and a modified Boussinesq formula for contact deformation. The influence coefficient method is used to build the tooth surface compliance matrix. The deformation compatibility equations and the torque equilibrium condition are solved together to obtain the contact stress and load distribution.
The simulated contact path agrees with the prescribed contact path. The numerical method computes contact stresses that are in good agreement with the finite element reference, with a maximum difference of less than 9%. The contact stress distribution is also consistent. The computational efficiency is improved by about 83% compared with the finite element method. The localized line contact face gear pair reduces the maximum contact stress by about 30% over the full rotation interval compared with a point contact face gear pair. It also avoids edge contact and the associated stress concentration. The sensitivity ranking of the contact types is full line contact, localized line contact, and point contact. The sensitivity ranking of the assembly errors is shaft angle error, shaft offset error, and axial misalignment error. The proposed method therefore provides an effective and efficient tool for the design and analysis of high-performance face gear pairs.
In future work, I will extend the model to multi-tooth contact and consider the coupling between the localized line contact and the dynamic behavior of the face gear transmission. I will also investigate the influence of manufacturing errors and thermal effects on the contact stress and load distribution of the modified face gear pair. The numerical framework can be integrated into an optimization loop to determine the best modification coefficients and contact path for a given application.