The dynamic performance of transmission systems, such as those found in high-speed printing presses, is fundamentally governed by nonlinear factors, with gear backlash being a primary source of vibration and instability. This article investigates the intricate influence of tooth flank clearance on the nonlinear dynamic characteristics of a helical gear transmission system. Traditional constant-backlash models often fail to capture the true complexity arising from microscopic surface topography. To address this, we introduce fractal theory to develop a refined model of time-varying backlash based on the Weierstrass-Mandelbrot (W-M) function. This model is integrated with Hertzian contact theory and the lumped mass method to formulate a comprehensive multi-degree-of-freedom nonlinear dynamic model for the helical gear system. A virtual prototype is subsequently constructed using SolidWorks and Adams to perform comparative simulations. The analysis focuses on system responses under various operational states—namely no-load and rated load (150 N) conditions at input speeds of 12,000 r/h and 18,000 r/h—while considering different fractal backlashes. The fractal dimension (D) is identified as the critical parameter defining the fractal gap and is thus treated as the control variable to examine its impact on dynamic stability. Findings reveal that under high-speed, no-load conditions, a lower fractal dimension exacerbates nonlinear vibrations. Conversely, under load, a higher fractal dimension contributes to improved system stability but can intensify high-frequency disturbances at elevated speeds. A further, practical insight emerges: a fractal dimension of D=1.5 appears to offer a favorable balance between achieving dynamic stability and managing associated manufacturing costs. This study provides a robust theoretical foundation for analyzing the dynamic behavior of helical gear systems and offers significant guidance for the optimal design of tooth flank clearance.
The core nonlinearity of gear systems is often encapsulated in the backlash function, which is typically modeled symmetrically for simplification. The piecewise function describing the relationship between the transmitted force/motion and the relative displacement across the gear mesh is given by:
$$ f(x) = \begin{cases} x – b, & x > b \\ 0, & -b \leq x \leq b \\ x + b, & x < -b \end{cases} $$
where $b$ represents the constant magnitude of the gear backlash. To incorporate the effect of surface roughness, the W-M fractal function is employed to model the microscopic topography of the contacting gear teeth surfaces. The surface profile height $Z(x)$ is expressed as:
$$ Z(x) = G^{(D-1)} \sum_{n=n_l}^{\infty} \frac{\cos(2\pi \gamma^n x / L)}{\gamma^{(2-D)n}}, \quad 1 < D < 2 $$
where $D$ is the fractal dimension, $G$ is the fractal roughness parameter, $\gamma$ is a scaling constant, $n$ is the frequency index, and $L$ is the sample length. The effective time-varying backlash $b_m(t)$, accounting for the combined surface profiles of the driving and driven helical gears, becomes:
$$ b_m(t) = b_0 – [Z_1(x) + Z_2(x)] = b_0 – \sum_{i=1}^{2} \left[ G_i^{(D_i-1)} \sum_{n=n_l}^{\infty} \frac{\cos(2\pi \gamma^n x / L)}{\gamma^{(2-D_i)n}} \right] $$
Here, $b_0$ is the nominal design backlash. Substituting $b_m(t)$ into the original backlash function $f(x)$ yields the modified, fractal-based nonlinear gap function $f'(y)$:
$$ f'(y) = \begin{cases} y – \left[ b_0 – \sum_{i=1}^{2} \left( G_i^{(D_i-1)} \sum_{n=n_l}^{\infty} \frac{\cos(2\pi \gamma^n x / L)}{\gamma^{(2-D_i)n}} \right) \right], & y > b_m \\ 0, & -b_m \leq y \leq b_m \\ y + \left[ b_0 – \sum_{i=1}^{2} \left( G_i^{(D_i-1)} \sum_{n=n_l}^{\infty} \frac{\cos(2\pi \gamma^n x / L)}{\gamma^{(2-D_i)n}} \right) \right], & y < -b_m \end{cases} $$
This fractal backlash model is then integrated into a 6-degree-of-freedom dynamic model for a single-stage helical gear pair using the lumped parameter method. The degrees of freedom are defined for both the pinion (1) and gear (2):
$$ \{ x_1, y_1, \theta_1, x_2, y_2, \theta_2 \} $$
where $x_i$ and $y_i$ are the translational displacements in the horizontal and vertical directions, and $\theta_i$ is the torsional displacement around the shaft axis. The equations of motion are derived considering time-varying mesh stiffness $k_m(t)$, mesh damping $c_m$, bearing stiffness ($k_{xi}$, $k_{yi}$), bearing damping ($c_{xi}$, $c_{yi}$), and the nonlinear fractal backlash function. The dynamic transmission error $\delta$ along the line of action is a key variable:
$$ \delta = R_1 \theta_1 – R_2 \theta_2 + (x_1 – x_2)\sin\psi + (y_1 – y_2)\cos\psi – e(t) $$
where $R_i$ are the base circle radii, $\psi$ is the helix angle at the base circle, $\beta$ is the operating pressure angle, and $e(t)$ is the static transmission error. The nonlinear mesh force $F_m$ is then:
$$ F_m = c_m \dot{\delta} + k_m(t) f'(\delta) $$
The comprehensive system dynamics are described by the following set of differential equations:
$$
\begin{aligned}
m_1 \ddot{x}_1 + c_{x1} \dot{x}_1 + k_{x1} x_1 &= -F_m [\cos\beta \sin\psi – \mu (\sin\beta \cos\psi \pm \sin\psi)] \\
m_1 \ddot{y}_1 + c_{y1} \dot{y}_1 + k_{y1} y_1 &= -F_m [\cos\beta \cos\psi + \mu (\sin\beta \sin\psi \mp \cos\psi)] \\
m_2 \ddot{x}_2 + c_{x2} \dot{x}_2 + k_{x2} x_2 &= F_m [\cos\beta \sin\psi – \mu (\sin\beta \cos\psi \pm \sin\psi)] \\
m_2 \ddot{y}_2 + c_{y2} \dot{y}_2 + k_{y2} y_2 &= F_m [\cos\beta \cos\psi + \mu (\sin\beta \sin\psi \mp \cos\psi)] \\
I_1 \ddot{\theta}_1 &= T_1 – R_1 [F_m \cos\beta – \mu F_m (\sin\beta \pm S_{11} \pm S_{12})] \\
I_2 \ddot{\theta}_2 &= -T_2 + R_2 [F_m \cos\beta + \mu F_m (\sin\beta \pm S_{21} \pm S_{22})]
\end{aligned}
$$
The parameters for the subject helical gear pair are detailed in the table below.
| Gear | Number of Teeth | Module (mm) | Pressure Angle (°) | Helix Angle (°) | Face Width (mm) | Mass (kg) | Moment of Inertia (kg·m²) | Bearing Stiffness (10⁸ N/m) | Bearing Damping (10⁴ N·s/m) | Mesh Damping (10³ N·s/m) |
|---|---|---|---|---|---|---|---|---|---|---|
| Pinion | 88 | 3.5 | 15 | 18 | 50 | 25 | 0.4 | 6 | 2.8 | 2.5 |
| Gear | 88 | 3.5 | 15 | 18 | 50 | 25 | 0.4 | 6 | 2.8 | 2.5 |
The relationship between the calculated average backlash for different fractal dimensions and the required center distance for assembly is critical for building the virtual prototype. This relationship is established through the gear mesh equation for helical gears considering backlash $b_{tb}$:
$$ \frac{2(x_1 + x_2)}{z_1 + z_2} \tan\alpha_t + \text{inv}\alpha_t = \text{inv}\alpha_w’ + \frac{b_{tb}}{m_n (z_1 + z_2) \cos\alpha_t} $$
where $x_i$ are profile shift coefficients, $z_i$ are tooth numbers, $m_n$ is the normal module, $\alpha_t$ is the transverse pressure angle, and $\alpha_w’$ is the operating pressure angle. The corresponding center distance $a_w$ is then $a_w = \frac{m_n (z_1 + z_2) \cos\alpha_t}{2 \cos\alpha_w’}$. The calculated values are summarized below.
| Fractal Dimension (D) | Calculated Backlash (mm) | Required Center Distance (mm) |
|---|---|---|
| 1.1 | 5.5094 | 323.860 |
| 1.3 | 5.7967 | 323.886 |
| 1.5 | 6.2937 | 323.938 |
| 1.7 | 7.1635 | 324.088 |
| 1.9 | 8.6869 | 324.264 |
The virtual prototype for the helical gear system was implemented in Adams. Contact between the helical gear teeth was modeled using the impact function, which computes the normal force $F_n$ as:
$$ F_n = \begin{cases} k \cdot \delta^e + \text{step}(\delta, 0, 0, d_{\text{max}}, C_{\text{max}}) \cdot \dot{\delta}, & \delta > 0 \\ 0, & \delta \leq 0 \end{cases} $$
Here, $k$ is the contact stiffness derived from Hertzian theory, $\delta$ is penetration depth, $e$ is the force exponent (typically 1.5), $d_{\text{max}}$ is the maximum penetration for full damping, and $C_{\text{max}}$ is the maximum damping coefficient. The contact stiffness $k$ for the helical gear pair is given by:
$$ k = \frac{4}{3} R^{1/2} E^* $$
where $R$ is the equivalent radius of curvature at the contact point and $E^*$ is the combined elastic modulus of the gear materials. Simulations were conducted for three primary operational cases to analyze the dynamic response of the helical gear system under the influence of fractal backlash.
Case 1: No-Load Condition at 12,000 r/h. Under no external load, the system’s response is primarily driven by inertia and the nonlinearities from the fractal backlash. Angular velocity fluctuations were most pronounced for lower fractal dimensions (D=1.1, 1.3), indicating stronger transient and nonlinear responses. As D increased to 1.7 and 1.9, the angular velocity stabilized more quickly, though high-frequency oscillations were still present. The angular acceleration and mesh force responses mirrored this trend, with lower D values showing larger impulsive peaks during startup and engagement, signifying more severe impacts. The torque response on the driven helical gear also exhibited higher amplitude fluctuations for lower D, confirming that a smoother fractal surface (higher D) can mitigate self-excited vibrations in an unloaded state.
Case 2: 150 N Load Condition at 12,000 r/h. The application of a 150 N load significantly altered the dynamics of the helical gear system. The load provided a constant resistive force, which dampened some of the free vibration seen in the no-load case. Interestingly, the system with higher fractal dimensions (D=1.7, 1.9) demonstrated superior stability under load. Their angular velocity traces settled into steady-state with smaller amplitude oscillations compared to the low-D models. The mesh force plots showed that while the mean force level increased due to the load, the dynamic fluctuation around this mean was better controlled for higher D values. The torque output became more consistent. This suggests that for a loaded helical gear transmission, a higher fractal dimension (representing a more complex but potentially smoother average profile) can enhance dynamic stability by improving load distribution and damping nonlinear oscillations.
Case 3: High-Speed, No-Load Condition at 18,000 r/h. Increasing the rotational speed of the helical gear system to 18,000 r/h intensified all dynamic responses. The inertial effects and the frequency of engagement events increased. At this elevated speed, the benefits of higher fractal dimensions became mixed. While angular velocity stability improved with increasing D, the angular acceleration and mesh force responses told a different story. For D=1.7 and 1.9, the mesh force and angular acceleration plots revealed bursts of high-frequency, high-amplitude oscillations. This indicates that at very high speeds, the microscopic surface roughness characterized by a high fractal dimension can induce high-frequency contact disturbances and exacerbate dynamic shocks, even if the gross motion seems stable. The lower-D systems, while having larger low-frequency swings, did not exhibit these intense high-frequency spikes to the same degree.
In conclusion, the dynamic behavior of a helical gear system is profoundly and nonlinearly influenced by the fractal characteristics of its tooth flank clearance. The fractal dimension serves as a critical design parameter that mediates this influence. Under no-load conditions, particularly at high speeds, a lower fractal dimension can lead to pronounced low-frequency instability, whereas a very high dimension may provoke high-frequency shocks. Under typical operational loads, a higher fractal dimension generally promotes stability by ensuring more continuous contact and better damping. The simulation results point to an optimal range, exemplified by D=1.5 in this study, which balances the trade-off between dynamic smoothness and the practical manufacturing cost associated with achieving very high surface fractal dimensions. This research establishes a fractal-based framework for modeling gear backlash that more accurately reflects physical reality, providing valuable insights for the design and analysis of high-performance helical gear transmissions aimed at minimizing vibration and noise.