The accurate prediction of time-varying mesh stiffness (TVMS) is fundamental for analyzing dynamic characteristics, vibration, and noise in gear transmission systems. For standard spur gears with ideal involute profiles, the TVMS can be effectively calculated using analytical methods like the potential energy method. However, in practical applications, tooth flank modifications such as lead crowning are intentionally applied, and assembly errors like misalignment inevitably occur. These factors transform the nominal line contact into localized conjugate contact, leading to complex elliptical contact patterns under load. This significantly alters the load distribution and stiffness characteristics, rendering conventional analytical models inadequate. This study proposes a comprehensive and improved analytical framework for calculating the TVMS of spur gears considering both lead crown modification and axial misalignment.
The proposed method integrates several key techniques: a mathematical model for tooth flank errors, Tooth Contact Analysis (TCA) and Loaded Tooth Contact Analysis (LTCA) to determine instantaneous contact features, and an enhanced stiffness model based on the slice theory and potential energy method that accounts for inter-slice coupling effects. The accuracy of the model is validated against three-dimensional finite element analysis (FEA) results, and the influence of different crowning amounts and misalignment angles on mesh stiffness excitation is systematically investigated.
Mathematical Model of Tooth Flank Deviations
The total deviation on the tooth flank of spur gears is a superposition of intentional modifications and unintentional errors. The lead crown modification is typically described by a parabolic function along the tooth width. The crowning amount \( C_{ci} \) at any point along the face width is given by:
$$C_{ci} = C_c \left( \frac{|b_i – b_0|}{B/2} \right)^s$$
where \( C_c \) is the maximum crowning amount, \( b_0 \) is the coordinate of the crowning symmetry point, \( b_i \) is the coordinate of the point of interest, \( B \) is the face width, and \( s \) is the curve exponent (typically 2 for parabolic crowning).
Axial misalignment, often resulting from assembly or shaft deflection, causes an angular offset between the gear axes. The misalignment error \( C_{mi} \) at a given transverse section can be expressed as a function of angular misalignment angles:
$$C_{mi} = b_i \left[ (\theta_{x1} – \theta_{x2})\cos\psi_{12} + (\theta_{y1} – \theta_{y2})\sin\psi_{12} \right] / \cos\beta_b$$
Here, \( \theta_{x}, \theta_{y} \) are misalignment angles, \( \psi_{12} \) is the gear pair axis angle, and \( \beta_b \) is the base helix angle (zero for spur gears).
Therefore, the total composite deviation \( E_i \) for a tooth pair, considering modifications on both pinion and gear, is:
$$E_i = (C_{ci}^{(p)} + C_{ci}^{(g)}) + C_{mi}$$
Tooth Contact Analysis under Load
For modified or misaligned spur gears, the contact becomes localized. Determining the instantaneous contact point, path, and especially the elliptical contact area under load is crucial for accurate stiffness calculation. This is achieved through a two-step process: unloaded TCA and Loaded TCA (LTCA).
First, the modified tooth surface is represented parametrically, for instance, using a B-spline surface: \( \mathbf{r}^{(i)}(u, v) \), where \( i=1,2 \) denotes the pinion and gear. The condition for contact between two conjugate surfaces requires that their position vectors and unit normals coincide at the contact point in a fixed coordinate system. This leads to a system of five nonlinear equations with six unknowns \( (u_1, v_1, \phi_1, u_2, v_2, \phi_2) \), where \( \phi \) denotes the rotation angle. By fixing one parameter (e.g., \( v_2 \)), the system can be solved iteratively to find the contact point for a given roll angle, and the contact path is traced by varying this parameter.
Under load, the contact point expands into an elliptical area. The dimensions and orientation of this ellipse are governed by the principal curvatures and directions of the two tooth surfaces at the contact point. The relative curvature in an arbitrary direction \( \mu \) is found using Euler’s formula:
$$k_n^{(i)}(\mu) = k_I^{(i)}\cos^2\mu + k_{II}^{(i)}\sin^2\mu$$
The directions of the major and minor axes of the contact ellipse correspond to the directions of minimum and maximum relative curvature, respectively. The semi-major axis \( a \) and semi-minor axis \( b \) of the contact ellipse are determined by solving the classical Hertzian contact equations:
$$\frac{a}{b} = \left( \frac{R_{II}}{R_{I}} \right)^{1/2} = \frac{\mathcal{K}(e) – \mathcal{E}(e)}{\mathcal{E}(e) – (1-e^2)\mathcal{K}(e)}$$
$$(ab)^{1/2} = \left( \frac{3F_n R_e}{2E^*} f(e) \right)^{1/3}$$
where \( R_I, R_{II} \) are the principal relative radii of curvature, \( F_n \) is the normal load, \( R_e = \sqrt{R_I R_{II}} \), \( E^* \) is the equivalent elastic modulus, \( e = \sqrt{1-(b/a)^2} \) is the ellipticity, and \( \mathcal{K}(e) \), \( \mathcal{E}(e) \) are the complete elliptic integrals of the first and second kind. The function \( f(e) \) is defined as \( f(e) = (2/\pi)[(\mathcal{K}(e)-\mathcal{E}(e))/(e^2 \mathcal{K}(e))]^{1/3} \).
Improved Time-Varying Mesh Stiffness Model
The core of the proposed method is an enhanced analytical stiffness model. The tooth is discretized into a series of independent thin slices along the face width. For spur gears with localized contact, only slices within the predicted contact ellipse contribute to the mesh stiffness.
1. Stiffness of a Single Slice
Each slice is modeled as a non-uniform cantilever beam. The normal force \( F \) acting on a slice is resolved into bending \( F_b \) and axial \( F_a \) components relative to the tooth centerline at the potential contact point \( (x_{ps}, h_{ps}) \). Using the potential energy method, the bending \( dU_b \), shear \( dU_s \), and axial compressive \( dU_a \) strain energies for a slice of thickness \( db \) are calculated by integrating along the tooth profile from the root to the contact point. The corresponding stiffness components for a slice are derived as:
$$\frac{1}{dK_b} = \int_{0}^{x_{ps}} \frac{3[ (x_{ps}-x)\cos\alpha – h_{ps}\sin\alpha]^2}{2E h_x^3 db} dx$$
$$\frac{1}{dK_s} = \int_{0}^{x_{ps}} \frac{1.2\cos^2\alpha}{2G h_x db} dx$$
$$\frac{1}{dK_a} = \int_{0}^{x_{ps}} \frac{\sin^2\alpha}{2E h_x db} dx$$
where \( E \) and \( G \) are Young’s and shear moduli, \( \alpha \) is the pressure angle at the contact point, and \( h_x \) is the variable tooth thickness. The nonlinear Hertzian contact stiffness for a slice is given by an empirical formula:
$$dK_h = \frac{1.275 E_e^{0.9} db^{0.8} F_i^{0.1}}{(1/R_I + 1/R_{II})^{0.1}}$$
where \( F_i \) is the load on the i-th slice.
2. Inter-Slice Coupling and System Stiffness
A critical improvement over the traditional slice method is accounting for the coupling between adjacent slices due to uneven load distribution. This coupling is modeled by introducing coupling spring stiffness \( K_{ci(i+1)} \) between slices \( i \) and \( i+1 \). It is approximated based on the geometric mean of the stiffness of the two adjacent slices:
$$K_{ci(i+1)} = \frac{2}{\left(\frac{1}{dK_{ti}} + \frac{1}{dK_{t(i+1)}}\right)} \times \left( \frac{2.75}{b} \right)^2 db^2$$
where \( dK_{ti} = 1/(1/dK_b + 1/dK_s + 1/dK_a)_i \) is the total geometric stiffness of slice \( i \).
The stiffness matrix \( \mathbf{K} \) for a tooth pair is assembled, incorporating all slice stiffnesses \( dK_{ti}^{(p)}, dK_{ti}^{(g)}, dK_{hi} \) and coupling stiffnesses \( K_{ci(i+1)}^{(p)}, K_{ci(i+1)}^{(g)} \). The load-deformation relationship for the slices under load \( \mathbf{F} \) is:
$$\mathbf{F} = \mathbf{K} \boldsymbol{\delta}$$
The total deflection of the tooth pair \( \delta^{(pg)} \) is the maximum slice deflection. The total stiffness of the tooth pair without considering the gear body is then \( K^{(pg)} = F / \delta^{(pg)} \).
3. Gear Body (Foundation) Stiffness and Multi-Tooth Engagement
The gear body compliance significantly affects TVMS. An analytical formula for foundation stiffness \( K_f \) is used. For single tooth engagement, the mesh stiffness \( K_e \) is the series combination of pinion tooth pair stiffness, gear tooth pair stiffness, and both foundation stiffnesses:
$$\frac{1}{K_e} = \frac{1}{K^{(pg)}} + \frac{1}{K_f^{(p)}} + \frac{1}{K_f^{(g)}}$$
During double-tooth engagement, a key modification is applied. Traditional models sum the stiffness of two tooth pairs in parallel, treating their foundations as independent. However, in reality, the two teeth share the same gear body. Therefore, a correction factor \( \lambda \) (where \( \lambda > 1 \)) is introduced to modify the foundation stiffness for multi-tooth engagement: \( K_{tf} = \lambda K_f \). The value of \( \lambda \) can be determined via finite element analysis. For a double-tooth engagement zone, the total mesh stiffness \( K_{mesh} \) is calculated as:
$$\frac{1}{K_{mesh}} = \sum_{j=1}^{2} \left( \frac{1}{K^{(pg)}_j} + \frac{1}{\lambda K_f^{(p)}} + \frac{1}{\lambda K_f^{(g)}} \right)^{-1}$$
The total error vector \( \mathbf{E} \) for both tooth pairs, calculated from the deviation model, is incorporated into the load distribution iteration between the two meshing tooth pairs to find the static equilibrium.
Model Verification and Results Analysis
The proposed method is validated against 3D nonlinear finite element analysis. The parameters of the example spur gear pair are listed below.
| Parameter | Pinion | Gear |
|---|---|---|
| Module (mm) | 2 | 2 |
| Number of Teeth | 25 | 30 |
| Face Width (mm) | 20 | 20 |
| Pressure Angle (°) | 20 | 20 |
| Young’s Modulus (GPa) | 210 | 210 |
| Poisson’s Ratio | 0.3 | 0.3 |
| Crowning Amount (mm) | 0.01 (Variable) | |
| Misalignment Angle (°) | 0.01 (Variable) | |
Validation for Ideal and Modified Spur Gears
First, the model is tested for ideal spur gears (zero crowning and misalignment). The results from the proposed method, a traditional accumulated integral potential energy method (which neglects slice coupling), and FEA are in excellent agreement, with a mean stiffness error of about 2.1% against FEA. For this case, slice coupling is negligible as the load is uniformly distributed.
Second, for a spur gear pair with 0.01 mm crowning and 0.01° misalignment, the TCA/LTCA results for contact path and ellipse are in good agreement with FEA contact pressure plots. The TVMS results are compared in the figure below. The proposed method shows a mean stiffness error of 3% against FEA, while the traditional slice method underestimates stiffness by about 4.2%. This confirms that accounting for inter-slice coupling improves accuracy for gears with localized contact.
| Method | Mean TVMS (kN/mm) | Crowning 0.01mm, Misalignment 0.01° | Error vs. FEA |
|---|---|---|
| Proposed Method | 339 | +3.0% |
| Traditional Slice Method | 315 | -4.2% |
| Finite Element Analysis (FEA) | 329 | Reference |
Compared to the ideal gear case (mean stiffness ~483 kN/mm), the introduced deviations cause a substantial reduction (approximately 30%) in mesh stiffness, highlighting their significant impact on the gear system’s excitation.
Influence of Misalignment and Crowning
The influence of varying misalignment angle and crowning amount on the TVMS of spur gears is analyzed systematically.
Effect of Misalignment: As the misalignment angle increases (e.g., from 0.01° to 0.04°), the contact ellipse shifts towards one end of the tooth face, reducing the effective contact length. This leads to a significant decrease in the overall mesh stiffness. The relationship is non-linear, emphasizing the necessity of including misalignment in dynamic models for accurate prediction.
Effect of Crowning Amount: Increasing the lead crown modification (e.g., from 10 μm to 25 μm) increases the localized curvature of the tooth flank. This results in a smaller contact ellipse under the same load, reducing the number of loaded slices and consequently decreasing the geometric, Hertzian, and total mesh stiffness. Similar to misalignment, the effect is pronounced and non-linear.
These analyses demonstrate that both parameters are critical factors in gear mesh stiffness excitation and cannot be represented by simple linear correction factors.
Conclusion
This study presents a generalized and improved analytical method for calculating the time-varying mesh stiffness of spur gears with tooth flank deviations, specifically lead crowning and axial misalignment. The method successfully integrates error modeling, loaded tooth contact analysis, and an enhanced slice-based stiffness model that incorporates inter-slice coupling effects and a corrected foundation stiffness model for multi-tooth engagement.
The key findings are:
- The proposed TCA and LTCA procedure effectively predicts the contact path and elliptical contact area for spur gears under localized contact conditions.
- For gears with perfect involute profiles, slice coupling is negligible. However, for gears with crowning and misalignment, accounting for this coupling improves the TVMS calculation accuracy, yielding a 3% error against FEA compared to 4.2% for the uncoupled model.
- Both axial misalignment and lead crowning significantly reduce the mesh stiffness of spur gears in a substantial and non-linear manner. Ignoring these factors can lead to inaccurate dynamic predictions.
The developed analytical framework provides an efficient and accurate tool for evaluating the stiffness excitation of modified spur gears, forming a solid theoretical basis for the dynamic design and optimization of gear transmission systems. The methodology can be further extended to analyze helical gears or bevel gears under similar localized contact conditions.