The study of spur gear transmission systems, fundamental to countless mechanical assemblies, persistently centers on understanding and mitigating vibration and noise. A critical aspect of this endeavor involves accurately modeling the nonlinear interactions that govern system behavior. Among the most significant nonlinearities are those arising from tooth backlash and time-varying mesh stiffness. Backlash, the intentional clearance between mating teeth, is essential for lubrication and thermal expansion but introduces complex dynamics. Under varying operational conditions such as high speed or light load, this clearance can lead to phenomena like teeth separation (loss of contact) and back-side contact, collectively termed gear rattling. Transitions between these meshing states, often accompanied by impacts due to velocity differences, are primary sources of undesirable vibration and acoustic emission. Furthermore, friction at the tooth interface, influenced by lubrication regimes, plays a substantial role in energy dissipation and dynamic response. Therefore, constructing a comprehensive dynamic model that accounts for friction and distinctly captures the multi-state meshing behavior of spur gears is imperative for precise analysis and effective design.
This article presents a detailed nonlinear dynamic model for a spur gear pair that explicitly formulates the equations of motion for three distinct meshing states: drive-side contact, teeth separation, and back-side contact. The model integrates time-varying mesh stiffness, static transmission error, damping, and, crucially, a friction model that spans different lubrication conditions. Unlike conventional models that use a piecewise linear function for the backlash nonlinearity, our approach derives the governing equations based on the kinematic and force analysis specific to each meshing state of the spur gears. We then investigate the influence of different lubrication-based friction models—elastohydrodynamic lubrication (EHL), mixed lubrication (ML), and dry friction—on the system’s dynamic transmission error and dynamic mesh force. Finally, by defining specific Poincaré mapping sections, we analyze the evolution of meshing states and the resulting impact vibrations as the system load varies.
1. Multi-State Meshing Dynamic Model for Spur Gears
We begin by considering a simplified torsional vibration model of a spur gear pair, assuming rigid supports. The system is reduced to a two-degree-of-freedom model representing the rotational motions of the driving and driven gears.
Where $T_p$ and $T_g$ are input and load torques, $I_p$ and $I_g$ are mass moments of inertia, $\theta_p$ and $\theta_g$ are angular displacements, $R_{bp}$ and $R_{bg}$ are base circle radii, $c_g$ is linear mesh damping, and $2\bar{D}$ is the total backlash. The time-varying mesh stiffness is $k(t) = k_{av} + k_a \cos(\omega_h t)$, where $k_{av}$ is average stiffness, $k_a$ is stiffness fluctuation, and $\omega_h$ is mesh frequency. The static transmission error excitation is $e(t) = E_a \omega_h^2 \cos(\omega_h t)$, where $E_a$ is error amplitude. For analysis, we consider geometrically identical spur gears with parameters listed in Table 1.
| Geometric Parameter | Pinion (Driver) | Gear (Driven) |
|---|---|---|
| Number of Teeth, $z$ | 40 | 40 |
| Module, $m$ (mm) | 3 | 3 |
| Pressure Angle, $\alpha$ (deg) | 20 | 20 |
| Face Width, $b$ (mm) | 48 | 48 |
| Elastic Modulus, $E$ (GPa) | 206 | 206 |
| Poisson’s Ratio, $\nu$ | 0.3 | 0.3 |
The relative displacement along the line of action is defined as $\bar{x} = R_{bp}\theta_p – R_{bg}\theta_g – e(t)$. The dynamic mesh force on the drive-side, when contact is maintained, is given by:
$$F_N = k(t) (\bar{x} – \bar{D}) + c_g (\dot{\bar{x}})$$
We now derive the equations for each state.
1.1 Drive-Side Meshing State
In this state, the pinion’s drive-side flank pushes the gear’s drive-side flank. The condition for this state is $\bar{x} > \bar{D}$ and $F_N > 0$. Applying Newton’s second law and considering friction forces $F_{fp}$ and $F_{fg}$ yields:
$$I_p \ddot{\theta}_p + F_N R_{bp} + F_{fp} S_{mp}(t) = T_p$$
$$I_g \ddot{\theta}_g – F_N R_{bg} – F_{fg} S_{mg}(t) = -T_g$$
The friction forces are proportional to the normal force and the time-varying friction coefficient $\mu_m(t)$: $F_{fp} = F_{fg} = \lambda_m(t) \mu_m(t) F_N$. The direction coefficient $\lambda_m(t) = \text{sgn}(v_{ms}(t))$ depends on the sliding velocity $v_{ms}(t)$ at the mesh point. The friction arms $S_{mp}(t)$ and $S_{mg}(t)$ are geometric functions of the contact position along the path of contact. Combining and non-dimensionalizing these equations leads to the drive-side equation of motion for the spur gear pair:
$$\ddot{x} + [1 + g_m(t) \lambda_m(t) \mu_m(t)][(1 + \kappa \cos(\omega \tau)) (x – D) + \xi \dot{x}] = F + \varepsilon \omega^2 \cos(\omega \tau)$$
Here, $x=\bar{x}/D_c$ is dimensionless displacement, $\tau$ is dimensionless time, $D=\bar{D}/D_c$ is dimensionless backlash, $\xi$ is damping ratio, $F$ is dimensionless load, $\kappa$ is stiffness fluctuation ratio, $\varepsilon$ is error excitation ratio, and $g_m(t)$ is the equivalent dimensionless friction arm. $\omega = \omega_h / \omega_n$ is the dimensionless mesh frequency, with $\omega_n = \sqrt{k_{av}/m_e}$ as the natural frequency and $m_e$ as the equivalent mass.
1.2 Teeth Separation State
Separation occurs when the dynamic mesh force drops to zero and the relative displacement is within the backlash zone. The conditions are $| \bar{x} | < \bar{D}$ and $F_N = 0$. In this state, the spur gears rotate independently with no mesh force or friction between them:
$$I_p \ddot{\theta}_p = T_p$$
$$I_g \ddot{\theta}_g = -T_g$$
The corresponding non-dimensional equation for the relative coordinate is:
$$\ddot{x} = F + \varepsilon \omega^2 \cos(\omega \tau)$$
1.3 Back-Side Meshing State
Under light loads, back-side contact can occur, where the gear’s back-side flank drives the pinion’s back-side flank. The conditions are $\bar{x} < -\bar{D}$ and $F_N < 0$. The direction of the normal force reverses. The equations of motion become:
$$I_p \ddot{\theta}_p – F_N R_{bp} – F_{fp} S_{ip}(t) = T_p$$
$$I_g \ddot{\theta}_g + F_N R_{bg} + F_{fg} S_{ig}(t) = -T_g$$
Here, $F_N$ uses the same stiffness and damping but with a negative relative displacement argument: $F_N = k(t) (-\bar{x} – \bar{D}) + c_g (-\dot{\bar{x}})$. The friction forces are $F_{fp} = F_{fg} = \lambda_i(t) \mu_i(t) |F_N|$, where $\lambda_i(t)$ and $\mu_i(t)$ are the direction coefficient and friction coefficient for back-side contact, and $S_{ip}(t), S_{ig}(t)$ are the corresponding friction arms. The non-dimensional back-side meshing equation for the spur gear system is:
$$\ddot{x} – [1 + g_i(t) \lambda_i(t) \mu_i(t)][(1 + \kappa \cos(\omega \tau)) (-x – D) – \xi \dot{x}] = F + \varepsilon \omega^2 \cos(\omega \tau)$$
1.4 Impact Conditions
Transitions between meshing states at the backlash boundaries ($x = \pm D$) are treated as instantaneous impacts. A coefficient of restitution $R$ ($0<r<1$) $$\dot{x}^+="R" $\dot{x}^+$="" $\dot{x}^-$="" \dot{x}^-,="" \quad="" \text{when}="" after="" and="" are="" before="" d$$="" during="" energy="" impact.
The complete piecewise-smooth dynamical system for the spur gear pair with multi-state meshing and friction is thus:
$$
\begin{cases}
\ddot{x} + [1 + g_m(t) \lambda_m(t) \mu_m(t)]H_m – F_h = 0, & x > D, \quad (Drive\text{-}side) \\
\ddot{x} – F_h = 0, & |x| < D, \quad (Separation) \\
\ddot{x} – [1 + g_i(t) \lambda_i(t) \mu_i(t)]H_i – F_h = 0, & x < -D, \quad (Back\text{-}side) \\
\dot{x}^+ = R \dot{x}^-, & x = \pm D, \quad (Impact)
\end{cases}
$$
where $H_m = (1+\kappa \cos(\omega \tau))(x-D) + \xi \dot{x}$, $H_i = (1+\kappa \cos(\omega \tau))(-x-D) – \xi \dot{x}$, and $F_h = F + \varepsilon \omega^2 \cos(\omega \tau)$.
2. Friction Models for Different Lubrication Regimes in Spur Gears
The friction coefficient $\mu(t)$ is not constant but varies with the instantaneous operating conditions at the tooth contact of the spur gears. It depends heavily on the lubrication regime, which is characterized by the film thickness parameter $\Lambda$, the ratio of central lubricant film thickness $H_c$ to the composite surface roughness.
Elastohydrodynamic Lubrication (EHL): Occurs at $\Lambda \geq 4$. The friction coefficient is modeled by a semi-empirical relation:
$$\mu_{E}(t) = e^{f(R_i, P_{hi}, \eta_M, R_{aavg})} P_{hi}^{b_2} |R_i(t)|^{b_3} \left( \frac{v_e(t)}{2} \right)^{b_6} \eta_M^{b_7} \rho_r(t)^{b_8}$$
where $f(\cdot) = b_1 + b_4 |R_i| P_{hi} \log(\eta_M) + b_5 e^{-|R_i P_{hi} \log(\eta_M)|} + b_9 e^{R_{aavg}}$. Here, $R_i(t)=2v_s(t)/v_e(t)$ is slide-to-roll ratio, $P_{hi}(t)$ is maximum Hertzian pressure, $v_e(t)$ is entrainment velocity, $\rho_r(t)$ is relative radius of curvature, $\eta_M$ is dynamic viscosity, and $b_1$-$b_9$ are empirical constants.
Mixed Lubrication (ML): Occurs at $1 < \Lambda < 4$. The friction coefficient is given by:
$$\mu_{M}(t) = 0.0127 \times \frac{1.13}{1.13 – R_{aavg}} \log_{10} \left( \frac{29700 f_e}{\eta_M |v_s(t)| v_e^2(t)} \right)$$
where $f_e$ is the normal load per unit face width and $R_{aavg}$ is the average surface roughness.
Boundary/Dry Friction: For very small $\Lambda$, the friction coefficient is approximated as a constant $\mu_B$.
The central film thickness $H_c(t)$ for calculating $\Lambda$ is given by the well-known Hamrock-Dowson formula:
$$H_c(t) = 3.06 G^{0.56} U(t)^{0.69} W(t)^{-0.10}$$
where $G$, $U(t)$, and $W(t)$ are dimensionless material, speed, and load parameters, respectively.
3. Dynamic Analysis Under Different Lubrication Schemes
We numerically integrate the system equations using a fourth-order Runge-Kutta algorithm. The dynamic mesh force, a key response metric, is calculated as $F_m = H_m$ for drive-side contact and $F_m = -H_i$ for back-side contact, both including their respective friction terms $[1+g(t)\lambda(t)\mu(t)]$.
3.1 Influence of Lubrication on System Response
We first analyze the effect of different friction models—No Friction, EHL, ML, and Dry Friction—on the spur gear dynamics for a fixed load ($F=0.1$). The dimensionless parameters are: $\kappa=0.25$, $\xi=0.05$, $\varepsilon=0.12$, $D=1.0$, $\omega=0.2$.
Case 1: Pure Drive-Side Contact (No Impacts). For this relatively high load, the system remains in perpetual drive-side contact ($x>D$, $F_m>0$). The results show:
- Dynamic Mesh Force: Friction increases the vibration amplitude of $F_m$. The EHL model causes the smallest increase, the ML model a more significant increase, and the Dry Friction model the largest amplitude amplification. This underscores the importance of maintaining a healthy lubricant film in spur gears to minimize dynamic load fluctuations.
- Phase Trajectory: The phase portrait ($x$ vs. $\dot{x}$) is a smooth closed orbit for the frictionless case. Introducing friction perturbs this orbit. The perturbation is mildest for EHL, more pronounced for ML, and most severe for Dry Friction, correlating directly with the magnitude of the friction coefficient introduced by each model.
Case 2: Drive-Side Impact (Periodic Separation). At a slightly lower load, the system exhibits periodic loss of contact and subsequent drive-side impacts ($x=D$). The dynamic mesh force $F_m$ periodically drops to zero.
- The influence of different lubrication schemes on the amplitude of $F_m$ during contact phases follows the same trend as in Case 1: EHL < ML < Dry Friction.
- However, the effect on the overall phase trajectory is less distinct compared to Case 1. The recurring separation events create a discontinuous phase flow, and the perturbations caused by different friction models during the brief contact periods do not dramatically alter the global orbit shape. The impact dynamics dominate the trajectory morphology.
Case 3: Back-Side Impact (Periodic Separation & Back-Side Contact). Under light load, the system experiences a complex sequence of drive-side contact, separation, back-side contact, and back-side impacts ($x=-D$). The dynamic mesh force $F_m$ oscillates between positive, zero, and negative values.
- The trends for $F_m$ amplitude
- The phase trajectory
Key Observation: The influence of the friction model on the dynamic mesh force is consistently significant across all meshing states. Its influence on the global phase trajectory is most pronounced when the spur gear system is in permanent contact (no impacts). When separation and impacts occur, the trajectory is largely governed by the switching dynamics, diluting the visible effect of different friction coefficients on the orbit’s geometry.
3.2 Evolution of Meshing States with Varying Load
To systematically study the transitions between meshing states in spur gears, we define three Poincaré sections:
- $\Sigma_n$: Stroboscopic section at period $\frac{2\pi}{\omega}$.
- $\Sigma_p$: Section at drive-side impact condition $x = D$.
- $\Sigma_q$: Section at back-side impact condition $x = -D$.
A periodic motion can then be classified by the triple $n$-$p$-$q$, denoting the number of periodic points on each respective section.
We analyze the bifurcation behavior using the dimensionless load parameter $F$ as the control variable, employing the EHL friction model.
- High Load ($F > 0.11$): The system exhibits a $1$-$0$-$0$ motion. The phase trajectory is a smooth, period-1 limit cycle entirely in the $x>D$ region. The dynamic mesh force is always positive. This corresponds to stable, rattle-free operation of the spur gear pair with continuous drive-side contact.
- Medium Load ($F \approx 0.10$): As $F$ decreases, a grazing bifurcation occurs, leading to a $1$-$1$-$0$ motion. The spur gears now experience periodic single impacts on the drive-side, accompanied by phases of separation. The dynamic mesh force touches zero once per period.
- Decreasing Load ($F$ reducing further): The $1$-$1$-$0$ motion undergoes period-doubling bifurcations (e.g., to $2$-$2$-$0$) and eventually leads to chaotic motion. Within the chaotic attractor, the trajectory begins to reach the back-side boundary $x = -D$.
- Light Load ($F = 0.03$): The system operates in a chaotic regime characterized by $n$-$p$-$q$ motion. The phase portrait is a complex, scattered attractor spanning $x>D$, $|x|<d$, $x<-d$="" (back-side="" (drive-side="" (separation),="" and="" back-side="" both="" contact),="" drive-side="" dynamic="" force="" gear="" history="" impacts="" in="" indicating="" intervals="" irregular="" li="" mesh="" negative="" of="" positive="" rattling="" regions.="" severe="" shows="" spur="" the="" time="" transmission.
This progression clearly illustrates the load-dependent nature of nonlinear dynamics in spur gears: from smooth motion under heavy load, to periodic knocking under medium load, and finally to chaotic, severe rattling under light load conditions.
4. Conclusions
In this work, we have developed and analyzed a comprehensive nonlinear dynamic model for a spur gear pair that explicitly accounts for multi-state meshing and friction under various lubrication regimes. The key contributions and findings are summarized as follows:
- Model Formulation: A piecewise-smooth dynamic model was derived from first principles, separately formulating the equations of motion for the spur gear system in drive-side meshing, teeth separation, and back-side meshing states. This approach, combined with impact conditions at the backlash boundaries, provides a more physically accurate representation of gear rattling phenomena compared to models using a simple piecewise linear spring for backlash.
- Friction Integration: Realistic time-varying friction coefficients for Elastohydrodynamic Lubrication (EHL), Mixed Lubrication (ML), and Dry Friction conditions were incorporated into the multi-state framework. The model highlights how friction interacts distinctly with each meshing state.
- Effects of Lubrication Regime: Numerical simulations reveal that the lubrication regime significantly affects the dynamic response of spur gears.
- The amplitude of the dynamic mesh force is consistently amplified by friction, with the degree of amplification ranking as: Dry Friction > Mixed Lubrication > EHL. This underscores the critical role of maintaining an effective EHL film to minimize dynamic loads in spur gear systems.
- The influence on the system’s phase trajectory is most visible during permanent drive-side contact. When separation and impacts occur, the trajectory’s structure is dominated by the switching dynamics, though the friction model still quantitatively affects the response during contact intervals.
- Load-Dependent Meshing State Transitions: By employing specialized Poincaré mappings, the evolution of meshing states with decreasing load was elucidated for spur gears under EHL conditions.
- High loads promote stable, continuous drive-side contact.
- Moderate loads induce periodic teeth separation and drive-side impacts.
- Light loads lead to complex, often chaotic motion involving both separation and back-side contact, resulting in severe impact vibration (rattling).
The presented model offers a refined tool for analyzing the nonlinear dynamics of spur gear transmissions, particularly in regimes where backlash-induced multi-state meshing and friction are significant. It provides a foundation for future studies on optimizing gear design parameters, assessing the impact of tooth profile modifications, and developing control strategies to suppress undesirable rattling vibrations in spur gear systems operating under variable loads and speeds.
</r