Vibration and noise are fundamental issues in geared transmission systems. In recent years, sliding friction at the gear mesh interface has been identified as a significant contributor to these phenomena. This friction-induced excitation not only acts as a source of vibration and acoustic noise but also accelerates various failure modes such as pitting, scuffing, and crack initiation, ultimately impacting the fatigue life and reliability of the system. Therefore, a comprehensive understanding of the coupled dynamics of friction and vibration in spur gears is crucial for optimal design aimed at vibration suppression and noise reduction.
Traditional dynamic models for spur gears often simplify or neglect the time-varying nature of friction, focusing primarily on time-varying mesh stiffness and transmission error. However, the friction force is inherently dynamic, influenced by the instantaneous relative sliding velocity, which is itself modulated by the system’s vibrational response. This creates a complex feedback loop between friction and vibration. This article establishes a refined lumped-parameter model for a spur gear pair that explicitly incorporates this time-varying friction coefficient. The model is used to numerically investigate the influence of friction on key dynamic indicators, namely the Dynamic Transmission Error (DTE) and the Dynamic Mesh Force (DMF), under various operating conditions.
1. Dynamic Modeling of Spur Gears with Friction
A six-degree-of-freedom (6-DOF) lumped parameter model is formulated for a spur gear pair, considering motions along the line-of-action (LOA, y-direction), the off-line-of-action (OLOA, x-direction), and torsional rotation. The schematic represents the system with pinion (1) and gear (2), where $k^i_x$, $k^i_y$, $c^i_x$, $c^i_y$ are the bearing stiffness and damping in the x and y directions, $I_i$ and $m_i$ are the mass moments of inertia and masses, and $r_{bi}$ are the base circle radii. The gear mesh is characterized by a time-varying mesh stiffness $k_m(t)$, mesh damping $c_m$, and static transmission error $e_m(t)$.
The governing equations of motion are derived by considering forces from the elastic mesh deflection, damping, and the tangential friction forces. The friction force acts in the OLOA direction and introduces an additional moment on the gears about their centers. The differential equations are:
For the pinion (1):
$$I_1 \ddot{\theta}_1 + r_{b1} \left[ k_m(t) (r_{b1}\theta_1 – r_{b2}\theta_2 + y_1 – y_2 – e_m) + c_m (r_{b1}\dot{\theta}_1 – r_{b2}\dot{\theta}_2 + \dot{y}_1 – \dot{y}_2 – \dot{e}_m) \right] = T_1 + \sum_{i=1}^{N} L_{pi} \mu(t) F_{mi}(t)$$
$$m_1 \ddot{y}_1 + c^1_y \dot{y}_1 + k^1_y y_1 = -F_m(t)$$
$$m_1 \ddot{x}_1 + c^1_x \dot{x}_1 + k^1_x x_1 = \sum_{i=1}^{N} \mu(t) F_{mi}(t)$$
For the gear (2):
$$I_2 \ddot{\theta}_2 – r_{b2} \left[ k_m(t) (r_{b1}\theta_1 – r_{b2}\theta_2 + y_1 – y_2 – e_m) + c_m (r_{b1}\dot{\theta}_1 – r_{b2}\dot{\theta}_2 + \dot{y}_1 – \dot{y}_2 – \dot{e}_m) \right] = -T_2 – \sum_{i=1}^{N} L_{gi} \mu(t) F_{mi}(t)$$
$$m_2 \ddot{y}_2 + c^2_y \dot{y}_2 + k^2_y y_2 = F_m(t)$$
$$m_2 \ddot{x}_2 + c^2_x \dot{x}_2 + k^2_x x_2 = -\sum_{i=1}^{N} \mu(t) F_{mi}(t)$$
In these equations, $T_1$ and $T_2$ are the input and output torques, $N$ is the number of tooth pairs in contact, and $L_{pi}$ and $L_{gi}$ are the moment arms of the friction force on the pinion and gear for the i-th tooth pair, respectively. The term $F_{mi}(t)$ represents the dynamic mesh force on the i-th tooth pair, defined as:
$$F_{mi}(t) = k_{mi}(t) \delta_i(t) + c_{mi} \dot{\delta}_i(t)$$
where $\delta_i(t)$ is the relative deflection along the LOA for that pair.
The Dynamic Transmission Error (DTE) along the line of action is a critical response variable, defined as:
$$\delta(t) = r_{b1}\theta_1(t) – r_{b2}\theta_2(t) + y_1(t) – y_2(t)$$
The total Dynamic Mesh Force (DMF) is the sum of forces from all contacting pairs: $F_m(t) = \sum F_{mi}(t)$.
2. Modeling of Time-Varying Parameters
2.1 Time-Varying Mesh Stiffness $k_m(t)$
The mesh stiffness of spur gears varies periodically with gear rotation due to the changing number of teeth in contact and the fluctuation in the contact line length. A common approach is to model it as a rectangular wave for simplicity in analytical studies, or more accurately using potential energy methods or finite element analysis to obtain a periodic function. For this analysis, $k_m(t)$ is represented as a periodic function with fundamental frequency equal to the gear mesh frequency $f_m = n z / 60$, where $n$ is the rotational speed (RPM) and $z$ is the number of teeth. The stiffness is higher in the double-tooth contact region and lower in the single-tooth contact region.
2.2 Time-Varying Friction Coefficient $\mu(t)$
The friction coefficient at the tooth interface is not constant. It depends on lubrication regime, surface roughness, load, and most importantly for dynamics, the instantaneous relative sliding velocity. A widely accepted model based on elastohydrodynamic lubrication (EHL) theory is employed here. The model calculates the friction coefficient as:
$$
\mu(t) = e^{f(S_R(t), P_h, \nu_0, S)} \cdot P_h^{b2} \cdot |S_R(t)|^{b3} \cdot V_{r}^{b6} \cdot \nu_0^{b7} \cdot R^{b8}
$$
where the function $f$ is given by:
$$
f(S_R(t), P_h, \nu_0, S) = b_1 + b_4 |S_R(t)| P_h \log_{10}(\nu_0) + b_5 e^{-|S_R(t)| P_h |8(\nu_0)} + b_9 S
$$
The key parameters are:
- $S_R(t)$: Slide-to-roll ratio, $S_R(t) = 2 v_s(t) / v_r(t)$. Here, $v_s(t)$ is the sliding velocity and $v_r(t)$ is the rolling velocity, both varying along the path of contact.
- $P_h$: Maximum Hertzian contact pressure, $P_h = \sqrt{ W’ E’ / (2 \pi R) }$.
- $W’$: Load per unit face width.
- $E’$: Equivalent elastic modulus, $2/E’ = (1-\nu_1^2)/E_1 + (1-\nu_2^2)/E_2$.
- $R$: Equivalent radius of curvature at the contact point, $R = R_1 R_2 / (R_1 + R_2)$.
- $\nu_0$: Dynamic viscosity of the lubricant at inlet temperature.
- $S$: Composite root mean square (RMS) surface roughness.
- $b_1$ to $b_9$: Empirical coefficients specific to the lubricant. A representative set is: $b_1$ = -8.916465, $b_2$ = 1.03303, $b_3$ = 1.036077, $b_4$ = -0.354068, $b_5$ = 2.812084, $b_6$ = -0.100601, $b_7$ = 0.752755, $b_8$ = -0.390958, $b_9$ = 0.620305.
This model captures the essential feature: the friction coefficient is high near the start and end of the mesh (high slide-to-roll ratio), drops to a minimum around the pitch point where pure rolling occurs ($S_R \approx 0$), and its magnitude is sensitive to load, speed, and surface finish.
3. System Parameters and Analysis Methodology
The analysis is performed on a representative spur gear pair. The key geometric and material parameters are listed in the table below.
| Parameter | Symbol | Pinion (1) | Gear (2) | Unit |
|---|---|---|---|---|
| Number of Teeth | $z$ | 16 | 24 | – |
| Module | $m_n$ | 4.5 | mm | |
| Pressure Angle | $\alpha$ | 20 | ° | |
| Face Width | $b$ | 20 | mm | |
| Young’s Modulus | $E$ | 210 | GPa | |
| Poisson’s Ratio | $\nu$ | 0.3 | – | |
| Mass Moment of Inertia | $I$ | Calculated | Calculated | kg·m² |
| Mass | $m$ | Calculated | Calculated | kg |
| Bearing Stiffness (x,y) | $k_x, k_y$ | 1.0e8 | N/m | |
| Damping Ratio | $\zeta$ | 0.07 | – | |
The dynamic equations are solved numerically using a time-step integration method (e.g., Runge-Kutta) to obtain the steady-state response. Different operating conditions are simulated by varying the input torque $T_1$ and the pinion rotational speed $\omega_1$. The surface roughness $S$ is also varied to study its effect as a proxy for different friction conditions.
| Case | Pinion Speed (RPM) | Input Torque (Nm) | RMS Roughness, S (μm) |
|---|---|---|---|
| A1 | 1000 | 1000 | 0.04 |
| A2 | 1000 | 1000 | 0.08 |
| A3 | 1000 | 1000 | 0.16 |
| B1 | 1500 | 1000 | 0.08 |
| B2 | 2000 | 1000 | 0.08 |
| C1 | 1000 | 1500 | 0.08 |
| C2 | 1000 | 2000 | 0.08 |
4. Results and Discussion: Influence of Friction on Spur Gear Dynamics
4.1 Characteristics of the Time-Varying Friction Coefficient
The calculated friction coefficient over one mesh cycle exhibits a characteristic “V” or “U” shape. It starts at a relatively high value at the initial contact point (dedendum of the driven gear, addendum of the driving gear), decreases as the contact point moves towards the pitch point, reaches a theoretical minimum near the pitch point (where sliding velocity is zero), and then increases again as the contact moves towards the end of the mesh. The model confirms that $\mu(t)$ is highly dependent on operating conditions.
Figure 2 (simulated) shows that for a constant torque, increasing the rotational speed generally leads to a decrease in the average and peak friction coefficient. This is attributed to better formation of an elastohydrodynamic lubricant film at higher rolling speeds, which separates the surfaces more effectively. Conversely, Figure 3 shows that for a constant speed, increasing the input torque (and thus the contact pressure $P_h$) results in an increase in the friction coefficient. Higher pressure thins the lubricant film, promoting more boundary/mixed lubrication and higher friction. Furthermore, increased surface roughness $S$ directly increases the calculated friction coefficient across the entire mesh cycle, as seen in the cases A1, A2, A3.
4.2 Effect on Dynamic Transmission Error (DTE)
The Dynamic Transmission Error is a primary source of vibration excitation in spur gears. The inclusion of friction modifies the DTE response significantly. Without friction, the DTE waveform typically follows the pattern of the time-varying mesh stiffness, with lower amplitude in the double-tooth contact zones. When friction is considered, a distinct disturbance is observed as the tooth pair passes through the pitch point.
This disturbance manifests as a small “kink” or phase shift in the DTE waveform at the pitch point. The physical explanation is as follows: Before the pitch point, the direction of the sliding velocity on the pinion tooth flank is opposite to its rolling direction, causing the friction force to oppose the motion. This friction force creates an additional torque on the pinion that counteracts the input torque. To maintain dynamic equilibrium and transmit the required load, the elastic mesh deflection (and thus DTE) must slightly increase. After crossing the pitch point, the sliding velocity direction reverses. The friction force now acts in the direction of motion on the pinion, assisting the input torque. Consequently, a smaller mesh deflection is needed to transmit the load, leading to a slight decrease in DTE. This transition causes the observed feature in the waveform.
As shown in Figure 4, increasing the surface roughness (and thus friction coefficient) amplifies this disturbance. The amplitude of the DTE also shows a general increasing trend with higher friction, indicating that friction acts as an additional damping-like mechanism that can slightly alter the system’s effective stiffness and energy dissipation, affecting the resonance characteristics of the spur gear pair.
4.3 Effect on Dynamic Mesh Force (DMF) and Vibrational Velocity
The Dynamic Mesh Force is directly proportional to the DTE and the mesh stiffness ($DMF \approx k_m(t) \cdot DTE$ for the dominant elastic component). Therefore, the disturbances observed in the DTE are also present in the DMF. Figure 6 clearly shows oscillations or ripples superimposed on the primary DMF waveform, synchronized with the pitch point passage. The magnitude of these friction-induced ripples increases with higher friction coefficients (rougher surfaces).
The vibrational velocity, which is the time derivative of the DTE ($v(t) = \dot{\delta}(t)$), is particularly sensitive to these rapid changes. As seen in Figure 5, the velocity waveform exhibits sharp spikes or discontinuities at the instants corresponding to the pitch point transition in the DTE. This is a direct consequence of the sudden change in the slope of the DTE curve caused by the friction force reversal. These high-frequency velocity components can be significant contributors to the radiated noise from the gearbox, as sound power is often related to vibrational velocity.
4.4 Combined Effect of Speed and Load
The interaction between operating conditions and friction is complex. Higher speeds reduce the friction coefficient but increase the mesh frequency and inertias. Higher loads increase the friction coefficient and the mean DMF. The net effect on vibration levels depends on the proximity of the mesh frequency to system resonances. The following table summarizes the qualitative trends observed for the peak-to-peak DTE and DMF.
| Parameter Increase | Effect on Friction Coef. $\mu$ | Effect on DTE/DMF Amplitude | Effect on Pitch-Point Disturbance |
|---|---|---|---|
| Rotational Speed ($\omega$) | Decrease | Variable (resonance dependent) | Decrease (smoother) |
| Input Torque ($T$) | Increase | Increase | Increase (more pronounced) |
| Surface Roughness ($S$) | Increase | Increase | Increase (more pronounced) |
5. Conclusions
This analysis provides a detailed investigation into the dynamic interaction between sliding friction and vibration in spur gear pairs. A comprehensive 6-DOF model incorporating a physics-based, time-varying friction coefficient was developed and solved numerically. The key findings are:
- Friction is a Dynamic Modulator: The friction coefficient in spur gears is not static but varies significantly within a single mesh cycle, influenced by the instantaneous slide-to-roll ratio, which is itself affected by the system’s vibrational state.
- Pitch Point Signature: The reversal of the friction force direction at the pitch point imparts a distinct signature on the dynamic response. It causes a observable disturbance in the Dynamic Transmission Error waveform, leading to corresponding ripples in the Dynamic Mesh Force and sharp spikes in the vibrational velocity.
- Amplification Mechanism: Higher friction conditions, resulting from increased surface roughness or higher loads, amplify the amplitude of the DTE and DMF and make the pitch-point disturbance more pronounced. This can elevate overall vibration and noise levels.
- Design Implications: For the design of quiet and durable spur gears, minimizing friction is crucial. This can be achieved through:
- Optimized surface finish (lower $S$).
- Selection of effective lubricants and additives.
- Operating in conditions that promote full-film EHL (higher speeds, moderate loads).
- Potential use of surface treatments or coatings.
In conclusion, sliding friction is a critical, non-negligible excitation mechanism in spur gear dynamics. Accurately modeling its time-varying nature is essential for predicting high-frequency vibrations and noise, ultimately leading to better designs for gear systems with improved fatigue life and acoustic performance.