The pursuit of higher power density, efficiency, and reliability in modern mechanical transmissions, particularly in demanding sectors like automotive, aerospace, and energy, places stringent performance requirements on core components such as gears. Among various gear types, helical gears are widely favored for their smoother engagement, higher load capacity, and reduced noise compared to spur gears. However, these components are frequently subjected to harsh operating conditions that can lead to failures like pitting, spalling, and cracking. To enhance their surface hardness, wear resistance, and fatigue life, heat treatment processes like quenching are indispensable. While effective for strengthening, quenching introduces significant residual stresses within the gear material. These internal, self-equilibrating stresses, if not properly accounted for, can profoundly influence the gear’s meshing behavior, load distribution, and ultimately, its dynamic response. The time-varying meshing stiffness (TVMS) is a fundamental excitation parameter in gear dynamics, dictating vibration and noise characteristics. Therefore, accurately quantifying the TVMS of a helical gear pair while considering the pre-existing residual stress field from manufacturing is crucial for a realistic prediction of system dynamics and long-term durability.
This article presents a comprehensive numerical investigation into the effects of quenching-induced residual stress on the TVMS and consequent dynamic loads in a helical gear transmission system. The analysis is structured around a coupled thermo-mechanical and dynamic modeling framework. First, a direct coupled thermal-stress analysis simulates the quenching process to obtain the spatial distribution of residual stress within the gear teeth. This stress field is then imported as an initial condition into a detailed finite element model of the meshing helical gear pair. Using a node-to-node engagement method, the TVMS under the influence of residual stress is calculated. Finally, a lumped-parameter dynamic model of the geared system is established and solved to reveal the impact of the altered stiffness on vibration and dynamic loads. The findings provide quantitative insights into how manufacturing-induced stresses can modify the functional performance of helical gears.
Theoretical Foundation
Finite Element Theory for Residual Stress Calculation
The generation of residual stress during quenching is a classic thermo-mechanical coupling problem. The stress-strain evolution is driven by the transient temperature field and the associated phase transformations. The fundamental finite element equilibrium equation governing the stress state is:
$$ \int_V \mathbf{B}^T \boldsymbol{\sigma} \, dV = \mathbf{P} $$
where \(\mathbf{B}\) is the strain-displacement transformation matrix, \(\boldsymbol{\sigma}\) is the Cauchy stress tensor, \(V\) is the volume, and \(\mathbf{P}\) is the vector of equivalent nodal forces. The total strain \(\boldsymbol{\varepsilon}\) is decomposed into elastic, plastic, and thermal components:
$$ \boldsymbol{\varepsilon} = \boldsymbol{\varepsilon}^e + \boldsymbol{\varepsilon}^p + \boldsymbol{\varepsilon}^{th} $$
The elastic strain \(\boldsymbol{\varepsilon}^e\) is related to stress by Hooke’s law, \(\boldsymbol{\sigma} = \mathbf{D} \boldsymbol{\varepsilon}^e\), where \(\mathbf{D}\) is the elasticity matrix. For an incremental thermo-elasto-plastic analysis, the stress increment can be expressed as:
$$ \Delta \boldsymbol{\sigma} = \mathbf{D}_{ep} \Delta \boldsymbol{\varepsilon} – \mathbf{h} \Delta T $$
Here, \(\mathbf{D}_{ep}\) is the temperature-dependent elasto-plastic stiffness matrix, \(\mathbf{h}\) is a thermal stress vector, and \(\Delta T\) is the temperature increment. Substituting into the equilibrium equation yields the discretized system equation for the displacement increment \(\Delta \mathbf{u}\):
$$ \int_V \mathbf{B}^T \mathbf{D}_{ep} \mathbf{B} \Delta \mathbf{u} \, dV = \Delta \mathbf{P} + \int_V \mathbf{B}^T \mathbf{h} \Delta T \, dV $$
Solving this equation throughout the cooling history, accounting for temperature-dependent material properties and phase transformation kinetics, yields the final residual stress field locked within the helical gear.
Calculation of Time-Varying Meshing Stiffness
The TVMS of a helical gear pair is calculated using a detailed finite element approach combined with a node-based engagement principle. For a given angular position, the contact zone between mating tooth surfaces is identified. The meshing action is simulated under load, and for each node \(i\) within this contact area, the local contact stiffness \(k_i\) is defined as the ratio of the normal contact force \(F_i\) to the corresponding local elastic deformation \(\delta_i\):
$$ k_i = \frac{F_i}{\delta_i} $$
The deformation \(\delta_i\) primarily comprises Hertzian contact deformation and tooth bending/shearing deformation. The single-tooth mesh stiffness for the pinion \(k_p\) and gear \(k_g\) are obtained by summing the stiffness contributions of all \(m\) contact nodes on their respective teeth:
$$ k_p = \sum_{i=1}^{m} k_i^{(p)}, \quad k_g = \sum_{i=1}^{m} k_i^{(g)} $$
For a given tooth pair \(j\), the combined mesh stiffness is calculated as these two springs in series:
$$ k_j = \frac{k_p \cdot k_g}{k_p + k_g} $$
Finally, if \(t\) tooth pairs are in simultaneous contact at that instant (governed by the total contact ratio of the helical gear), the total TVMS \(k_t\) for the gear pair is the sum of the stiffness of all engaged pairs:
$$ k_t = \sum_{j=1}^{t} k_j $$
This process is repeated over a full mesh cycle to obtain the periodic TVMS curve. When residual stress is considered, it is applied as an initial pre-stress field in the finite element model, altering the local material compliance and thus affecting both \(F_i\) and \(\delta_i\) in the calculation of \(k_i\).
Numerical Modeling of the Helical Gear System
Gear Geometry and Material
The study focuses on a pair of carburized helical gears made of low-alloy steel 20CrMnTi, a common material for high-strength gears. The key geometric parameters are summarized in the table below.
| Parameter | Pinion (Driver) | Gear (Driven) |
|---|---|---|
| Number of Teeth, \(z\) | 41 | 145 |
| Normal Module, \(m_n\) (mm) | 2 | 2 |
| Normal Pressure Angle, \(\alpha_n\) (°) | 20 | 20 |
| Helix Angle, \(\beta\) (°) | 21.56 | 21.56 |
| Hand of Helix | Left | Right |
| Profile Shift Coefficient, \(x\) | +0.345 | -0.345 |
| Face Width, \(b\) (mm) | 50 | 45 |
| Center Distance, \(a\) (mm) | 200 | |
The total contact ratio for this helical gear pair is approximately 4.2, indicating that the transmission alternates between 4-tooth and 5-tooth contact zones during operation.
Quenching Process Simulation for Residual Stress
A finite element model of a single helical gear tooth segment is created for the quenching simulation. The model assumes cyclic symmetry. The initial temperature of the gear is set to 850°C (austenitizing temperature). The quenching process involves two stages: first, oil quenching until the gear cools to 80°C, followed by air cooling to room temperature (25°C). The heat transfer coefficients for oil and air cooling are defined based on established data. The material properties of 20CrMnTi, including thermal conductivity, specific heat, density, and coefficient of thermal expansion, are defined as temperature-dependent functions. The phase transformation kinetics and the associated volumetric changes are also incorporated to accurately capture the generation of transformation-induced stresses. The simulation solves the coupled thermal and mechanical fields over time to obtain the final, stabilized residual stress distribution.
Finite Element Model for Meshing Stiffness Analysis
A three-dimensional finite element model containing 11 teeth of the pinion and 12 teeth of the gear is constructed to analyze a complete mesh cycle. The calculated mesh cycle corresponds to a pinion rotation of 8.78°. The refined residual stress field from the quenching simulation is mapped onto the gear bodies as an initial state. The material is modeled as linear elastic with a Young’s modulus of 210 GPa and a Poisson’s ratio of 0.278 for the stiffness calculation. A surface-to-surface contact formulation with a friction coefficient of 0.15 is defined between the mating tooth flanks. The pinion’s inner bore is coupled to a reference point where a rotational speed of 20,900 rpm is applied, while the gear’s inner bore is coupled to a reference point where a resisting torque of 2,424 N·m is applied. Both load and speed are ramped up gradually to simulate quasi-static engagement and minimize inertial effects for stiffness extraction. The model is solved to obtain the contact forces and displacements at each engagement position.
Dynamic Model of the Gear Transmission System
To investigate the dynamic consequences of the altered TVMS, a 10-degree-of-freedom (DOF) lumped-parameter model of the geared rotor system is established. The model accounts for translational vibrations of both the pinion and gear in three orthogonal directions (\(x\), \(y\), \(z\)) and rotational vibrations about the shaft axis (\(z\)) and the axis perpendicular to the shaft and parallel to the gear face (\(y\)). The equations of motion are:
$$ \begin{aligned}
m_1 \ddot{x}_1 + c_{1x} \dot{x}_1 + k_{1x} x_1 &= -F_x \\
m_1 \ddot{y}_1 + c_{1y} \dot{y}_1 + k_{1y} y_1 &= -F_y \\
m_1 \ddot{z}_1 + c_{1z} \dot{z}_1 + k_{1z} z_1 &= -F_z \\
J_{1z} \ddot{\theta}_{1z} + F_z r_{b1} &= T_1 \\
J_{1y} \ddot{\theta}_{1y} + c_{1y\theta} \dot{\theta}_{1y} + k_{1y\theta} \theta_{1y} &= F_z R_1 \\
m_2 \ddot{x}_2 + c_{2x} \dot{x}_2 + k_{2x} x_2 &= F_x \\
m_2 \ddot{y}_2 + c_{2y} \dot{y}_2 + k_{2y} y_2 &= F_y \\
m_2 \ddot{z}_2 + c_{2z} \dot{z}_2 + k_{2z} z_2 &= F_z \\
J_{2z} \ddot{\theta}_{2z} + F_z r_{b2} &= -T_2 \\
J_{2y} \ddot{\theta}_{2y} + c_{2y\theta} \dot{\theta}_{2y} + k_{2y\theta} \theta_{2y} &= F_z R_2
\end{aligned} $$
Here, subscripts \(1\) and \(2\) denote the pinion and gear, respectively. \(m_i\) are masses, \(J_{iz}\) and \(J_{iy}\) are moments of inertia, \(k_{ix}\), \(k_{iy}\), \(k_{iz}\), \(k_{iy\theta}\) are supporting stiffnesses, \(c_{ix}\), \(c_{iy}\), \(c_{iz}\), \(c_{iy\theta}\) are damping coefficients, \(T_i\) are torques, and \(r_{bi}\) are base circle radii. \(F_x\), \(F_y\), and \(F_z\) represent the dynamic mesh forces in the axial, radial, and tangential (circumferential) directions, which are governed by the TVMS \(k_t(t)\) and the relative displacement along the line of action \(u\):
$$ F_z = k_t(t) u + c_m \dot{u} $$
The mesh damping \(c_m\) is given by \(c_m = 2 \zeta \sqrt{k_t m_e}\), where \(\zeta\) is the damping ratio (taken as 0.07) and \(m_e\) is the equivalent mass \(\left( \frac{1}{m_1} + \frac{1}{m_2} \right)^{-1}\). The system parameters are listed below.
| Parameter | Pinion | Gear |
|---|---|---|
| Mass, \(m\) (kg) | 2.74 | 11.15 |
| Polar Moment of Inertia, \(J_z\) (kg·m²) | 0.0028 | 0.14 |
| Bending Moment of Inertia, \(J_y\) (kg·m²) | – | – |
| Support Stiffness, \(k_x\), \(k_y\) (N/m) | 2.3×10⁸ | 3.7×10⁸ |
| Support Stiffness, \(k_z\) (N/m) | 1.5×10⁸ | 2.0×10⁸ |
| Torsional Stiffness, \(k_{y\theta}\) (N/m) | 1.5×10⁶ | 2.0×10⁶ |
| Damping, \(c_x\), \(c_y\), \(c_z\) (N·s/m) | 1.5×10³ | 5.9×10³ |
The TVMS functions \(k_t(t)\) obtained from the FE analysis, both with and without residual stress, are fed into this dynamic model. The equations are solved numerically using the Runge-Kutta method to obtain the time-domain vibration responses and the corresponding dynamic loads.
Results and Discussion
Distribution of Quenching Residual Stress
The quenching simulation reveals the characteristic distribution of residual stress in the helical gear tooth. The evolution is complex, governed by competing thermal shrinkage and phase transformation expansion. Initially, rapid surface cooling causes thermal contraction, placing the surface in tension and the core in compression. Subsequently, martensitic transformation begins at the surface, causing volumetric expansion that reverses the stress state. The final, stabilized residual stress field exhibits a compressive layer at the surface and subsurface regions, transitioning to tensile stress in the core. This “outside-in compression” profile is beneficial for fatigue resistance as it hinders crack initiation at the highly stressed surface. The magnitude of these stresses is significant, often reaching hundreds of MPa, confirming that they constitute a non-negligible pre-load condition for subsequent meshing analysis.
| Location | Stress Component | Approximate Magnitude (MPa) | Sign |
|---|---|---|---|
| Tooth Surface/Subsurface (A₂) | S₁₁ (Hoop) | 200 – 350 | Compressive (-) |
| S₂₂ (Axial) | 150 – 300 | Compressive (-) | |
| S₃₃ (Radial) | 50 – 150 | Compressive (-) | |
| Gear Core (B) | S₁₁ (Hoop) | 100 – 200 | Tensile (+) |
| S₂₂ (Axial) | 80 – 180 | Tensile (+) | |
| S₃₃ (Radial) | 20 – 100 | Tensile (+) |
Influence of Residual Stress on Contact and Deformation
The presence of the residual stress field alters the local stiffness of the tooth material. Analysis of the meshing FE model shows that the nominal contact force distribution along the tooth face is only slightly modified by residual stress, with a minor tendency for contact force to decrease slightly at highly loaded nodes. The more pronounced effect is observed in the tooth bending deformation. The bending deformation curve over a mesh cycle shows characteristic dips corresponding to 4-tooth contact zones and peaks corresponding to 5-tooth contact zones. Under the influence of residual stress, the bending deformation increases in the 4-tooth contact regions and decreases in the 5-tooth contact regions. This can be attributed to the complex interaction between the applied bending moment from the mesh load and the pre-existing stress field, which either softens or stiffens the tooth root region depending on the superposition of stresses. In contrast, the local Hertzian contact deformation remains largely unaffected by the bulk residual stresses.
Effect on Time-Varying Meshing Stiffness
The changes in deformation directly translate to changes in the calculated TVMS. The periodic pattern of the TVMS for the helical gear pair is clearly observed, with higher stiffness values during 5-tooth contact and lower values during 4-tooth contact. A comparison between the TVMS calculated with and without considering the quenching residual stress reveals distinct differences:
- Amplitude Modulation: The overall waveform of the TVMS is modulated by the residual stress. The peaks in the 5-tooth contact zone become more pronounced, indicating a higher maximum mesh stiffness.
- Increased Fluctuations: The transition regions, particularly from 5-tooth to 4-tooth contact, exhibit heightened fluctuations or “ripples” in stiffness when residual stress is present. This suggests that the stress field exacerbates the non-uniform load-sharing during the engagement/disengagement of tooth pairs.
- Mean Stiffness Change: The average value of the TVMS over one cycle shows a slight increase when residual stress is accounted for. The calculated values are summarized below.
| Condition | Average TVMS, \( \bar{k}_t \) (N/mm) | Peak TVMS in 5-tooth zone (N/mm) | Fluctuation Intensity* |
|---|---|---|---|
| Without Residual Stress | 6.051 × 10⁵ | ~6.35 × 10⁵ | Baseline |
| With Quenching Residual Stress | 6.065 × 10⁵ | ~6.50 × 10⁵ | Increased |
*A qualitative measure of high-frequency variations superimposed on the main stiffness curve.
This confirms that ignoring manufacturing-induced residual stress leads to an underestimation of the mesh stiffness and its peak variations, which are critical excitations for dynamics.
Dynamic Response and Load Analysis
The modified TVMS profiles serve as the primary excitation in the 10-DOF dynamic model. The solution yields the dynamic loads on the helical gear teeth in three directions: circumferential (tangential), axial, and radial (normal to tooth face).
Time-Domain Analysis: The time-history plots of the dynamic mesh forces show that the system with residual stress experiences more severe load fluctuations. The peak-to-peak variation of the dynamic load increases significantly. For instance, the peak normal dynamic load increases by approximately 8.5% compared to the case without residual stress. The impact-like events during the transition between contact zones become more pronounced.
Frequency-Domain Analysis: The frequency spectra of the dynamic loads provide further insight. Key frequency components include the pinion rotational frequency (\(f_{s1} \approx 348 \, \text{Hz}\)), the gear rotational frequency (\(f_{s2} \approx 98 \, \text{Hz}\)), and the gear mesh frequency (\(f_m = f_{s1} \times z_1 \approx 14,281 \, \text{Hz}\)).
| Frequency Component | Direction | Amplitude without RS | Amplitude with RS | Change |
|---|---|---|---|---|
| Mesh Frequency, \(f_m\) | Normal | Baseline (1.0 p.u.) | ~1.30 p.u. | +30% |
| 2× Mesh Frequency, \(2f_m\) | Normal | Baseline (1.0 p.u.) | ~0.84 p.u. | -16% |
| Pinion Rotational, \(f_{s1}\) | Circumferential | Baseline (1.0 p.u.) | ~1.10 p.u. | +10% |
The table illustrates that residual stress not only increases the vibration energy at the fundamental mesh frequency but also redistributes energy among different harmonics. The significant increase (≈30%) in the amplitude at the mesh frequency for the normal dynamic load is particularly noteworthy, as this directly correlates with noise radiation and structural vibration. The changes in the rotational frequency components indicate that the residual stress also affects the low-frequency modulation of the gear motion.
Conclusion
This investigation establishes a integrated numerical framework to quantify the impact of quenching-induced residual stress on the dynamic performance of helical gear transmissions. The key findings are systematically concluded as follows:
- Residual Stress Field: The quenching process for a 20CrMnTi helical gear generates a significant internal stress field characterized by compressive stresses at the surface/subsurface and tensile stresses in the core. This field acts as a permanent pre-load state.
- Effect on Mesh Stiffness: The pre-existing residual stress modifies the local tooth compliance. It increases tooth bending deformation in low-contact-ratio zones (4-tooth contact) and decreases it in high-contact-ratio zones (5-tooth contact), while leaving contact deformation largely unchanged. This results in a measurable alteration of the time-varying meshing stiffness (TVMS), increasing its mean value, amplifying its peak values, and introducing more pronounced fluctuations during contact transitions.
- Consequence for Dynamics: The altered TVMS profile directly excites the geared system differently. The dynamic model reveals that residual stress leads to increased peak dynamic loads and greater load fluctuations in the time domain. In the frequency domain, it causes a substantial increase (e.g., ~30%) in the vibration amplitude at the fundamental mesh frequency, which is critical for noise and vibration analysis. The energy at other harmonics, such as the second mesh harmonic and rotational frequencies, is also redistributed.
In summary, neglecting the residual stress from manufacturing processes like quenching leads to an incomplete and potentially non-conservative assessment of a helical gear‘s meshing stiffness and dynamic behavior. The proposed methodology and results underscore the importance of incorporating manufacturing-induced effects into the design and analysis phase for high-performance helical gear systems to achieve accurate predictions of durability, noise, and vibration.