As an essential power transmission element in machinery, the operational health of spur gears directly impacts the reliability and efficiency of entire mechanical systems. Tooth surface wear stands as one of the most prevalent failure modes for spur gears. Progressive, non-uniform wear alters the tooth profile geometry and the distribution of load across the meshing teeth. Beyond degrading transmission accuracy and mechanical efficiency, these geometrical changes inevitably modify the intrinsic dynamic characteristics—specifically, the modal properties—of the gear body. When the system’s excitation frequency coincides with or approaches these altered natural frequencies, resonant vibration can be triggered, leading to amplified noise, accelerated fatigue, and potential catastrophic failure. Therefore, a thorough investigation into the patterns of non-uniform wear and its subsequent effect on modal parameters is crucial. This understanding provides a foundational theory for mitigating wear, optimizing gear design against vibration and noise, and developing robust condition monitoring and predictive maintenance strategies for gearboxes. This article delves into the mechanisms of wear formation under quasi-static conditions and systematically analyzes its impact on the vibration modes and natural frequencies of spur gears.
The analysis of spur gears under wear conditions necessitates a robust theoretical framework to model the material loss process. For this purpose, the Archard wear model, integrated with Hertzian contact theory, provides a widely accepted methodology for predicting wear volume under sliding contact conditions, making it particularly suitable for studying the wear of spur gear teeth surfaces.
The fundamental Archard wear equation relates the volumetric wear to the sliding distance, normal load, material hardness, and a dimensionless wear coefficient. The equation is expressed as:
$$ \frac{V}{s} = k \frac{W}{H} $$
where \( V \) is the wear volume, \( s \) is the total sliding distance, \( W \) is the normal load, \( H \) is the hardness of the softer contacting surface, and \( k \) is the dimensionless wear coefficient. For gear analysis, it is often more practical to consider wear depth. By considering the contact area, the incremental wear depth \( dh \) at a point over an incremental sliding distance \( ds \) can be written as:
$$ dh = k \frac{p(x)}{H} ds $$
Here, \( p(x) \) is the contact pressure at the location of interest. The accurate determination of the three key parameters—wear coefficient \( k \), sliding distance \( s \), and contact pressure \( p(x) \)—is essential for predicting the wear distribution on the flanks of spur gears.
The wear coefficient \( k \) is not a constant but depends on a complex interplay of factors including material pairing, lubrication regime, surface roughness, and operating conditions. Empirical models are often employed. One such model, based on statistical regression of experimental data, expresses \( k \) as a function of dimensionless groups:
$$ k = 3.987 \times 10^{29} \cdot E_e \cdot L^{1.219} \cdot G^{-7.377} \cdot S^{1.589} $$
The dimensionless parameters are defined as follows:
$$ L = \frac{W}{E_e R_e}, \quad G = \alpha_0 E_e, \quad S = \frac{R_{rms}}{R_e} $$
where \( E_e \) is the equivalent elastic modulus, \( R_e \) is the equivalent radius of curvature at the contact point, \( \alpha_0 \) is the pressure-viscosity coefficient of the lubricant, and \( R_{rms} \) is the composite root-mean-square surface roughness. This formulation highlights that the wear coefficient increases with load \( L \) and surface roughness \( S \), but decreases significantly with a higher lubricant parameter \( G \).
The sliding distance is derived from the kinematics of meshing spur gears. The sliding velocity varies along the path of contact. The specific sliding or slide-to-roll ratio for the pinion (driver) and gear (driven) is given by:
$$ \lambda_1 = \left| 1 – \frac{V_{t2}}{V_{t1}} \right|, \quad \lambda_2 = \left| 1 – \frac{V_{t1}}{V_{t2}} \right| $$
where \( V_{t1} \) and \( V_{t2} \) are the tangential velocities of the pinion and gear at the meshing point, respectively. For a single engagement over the contact zone width \( 2a \), the sliding distances for the two surfaces are:
$$ s_1 = 2a \lambda_1, \quad s_2 = 2a \lambda_2 $$
These equations show that sliding is zero at the pitch point (pure rolling) and maximum towards the tips and roots of the teeth for spur gears.
The contact pressure distribution is calculated using Hertzian theory for line contact, which models the meshing teeth as two equivalent cylinders. The semi-contact width \( a \) and the parabolic pressure distribution \( p(x) \) across the width \( -a \le x \le a \) are:
$$ a = \sqrt{ \frac{4 P R_e}{\pi E_e} }, \quad p(x) = \frac{2W}{\pi a^2} \sqrt{a^2 – x^2} $$
The maximum contact pressure, \( p_0 = \frac{2W}{\pi a} \), occurs at the center of the contact zone. The equivalent radius \( R_e \) is determined from the curvatures of the tooth profiles at the instant of contact: \( \frac{1}{R_e} = \frac{1}{R_1} \pm \frac{1}{R_2} \) (where the sign depends on external or internal contact). For external spur gears, it is \( \frac{1}{R_e} = \frac{1}{R_1} + \frac{1}{R_2} \).
A critical aspect for spur gears is the load sharing between tooth pairs as the contact moves through the single and double tooth contact zones along the line of action. The total transmitted load \( F_n \) is shared between two pairs in the double contact region. After initial wear occurs, the profile deviations \( h_p \) and \( h_g \) (wear depths on pinion and gear) affect the load distribution. The modified load on each tooth pair can be estimated using compatibility conditions related to gear mesh stiffness \( C_T \). If \( F_{n1} \) and \( F_{n2} \) are the loads on the first and second contacting pair in the double zone, and \( d_{ck} \) is the distance from the pitch point, the relations are:
$$ F_{n1} = \frac{F_n}{2} – \frac{C_T}{2} \left[ (h_g(d_{ck}) + h_p(d_{ck})) – (h_g(d_{ck}+p_b) + h_p(d_{ck}+p_b)) \right] $$
$$ F_{n2} = \frac{F_n}{2} + \frac{C_T}{2} \left[ (h_g(d_{ck}) + h_p(d_{ck})) – (h_g(d_{ck}+p_b) + h_p(d_{ck}+p_b)) \right] $$
where \( p_b \) is the base pitch. This iterative coupling between wear depth and load distribution is essential for accurate long-term wear simulation in spur gears.
| Parameter | Symbol | Value |
|---|---|---|
| Module | m | 6 mm |
| Number of Teeth (Pinion/Gear) | z1 / z2 | 17 / 43 |
| Pressure Angle | α | 20° |
| Face Width | b | 50 mm |
| Center Distance | A | 180 mm |
| Young’s Modulus | E | 206 GPa |
| Poisson’s Ratio | ν | 0.3 |
| Surface Roughness, Ra | – | 0.2 μm |
| Mesh Stiffness (Unit Width) | C_T | 1.0 × 10^4 N/mm² |
| Input Torque | T_p | 2400 Nm |
| Input Speed | n | 1440 rpm |
| Pressure-Viscosity Coefficient | α_0 | 3.0 × 10⁻⁸ m²/N |
Using the theoretical framework and the parameters for a representative spur gear pair (Table 1), a numerical simulation of the quasi-static wear process was performed. The wear depth was updated iteratively until a specified maximum wear threshold was reached on the pinion tooth.
The calculated wear coefficient distribution along the path of contact reveals a distinctive pattern. It generally decreases from the root to the tip but exhibits a noticeable peak in the region where the meshing transitions between double and single tooth contact. This spike is directly attributed to the sudden change in the shared load, which increases the contact pressure on the remaining tooth pair, thereby elevating the local wear coefficient according to the empirical model.
The sliding distance distribution is asymmetric between the pinion and gear of the spur gear pair. For both, the sliding distance is high near the root and tip, falling to zero at the pitch point. However, the magnitude of sliding is significantly greater for the pinion than for the gear, especially in the root region. This is because the pinion tooth has a smaller curvature radius, leading to a higher slide-to-roll ratio. A minor discontinuity in sliding distance is also observed at the meshing transition points due to the change in the instantaneous contact half-width.
| Characteristic | Pinion (Driver) | Gear (Driven) |
|---|---|---|
| Maximum Wear Depth Location | Root (Dedendum) | Root (Dedendum) |
| Relative Wear Magnitude (Root vs. Tip) | Root >> Tip | Root > Tip |
| Wear at Pitch Point | ~0 (Pure Rolling) | ~0 (Pure Rolling) |
| Wear at Meshing Transition Zone | Discontinuous Jump | Discontinuous Jump |
| Cycles to Reach Threshold | 7.67 million | 3.03 million |
The final wear profile, synthesized from the calculated wear coefficient, sliding distance, and contact pressure, shows a clear non-uniform pattern (Table 2). The most severe wear occurs at the root of the pinion tooth. This is a consequence of the combined effect of high sliding distance and a relatively high wear coefficient in that region. While the gear tooth also experiences significant root wear, it is less severe than on the pinion due to its lower number of stress cycles (the gear rotates slower). Wear at the tooth tip (addendum) is present but notably less than at the root. As predicted, the pitch point region shows negligible wear. A sharp, discontinuous change in wear depth is evident at the transition between the single and double tooth contact zones, directly resulting from the corresponding jumps in load and contact conditions.
To assess the impact of this non-uniform wear on the dynamic behavior, a modal analysis of the spur gear body was conducted. Modal analysis determines the natural frequencies and corresponding mode shapes—the inherent vibration patterns—of a structure. The undamped free-vibration equation governing the motion is derived from finite element theory:
$$ [M]\{\ddot{x}(t)\} + [K]\{x(t)\} = 0 $$
where \([M]\) is the global mass matrix, \([K]\) is the global stiffness matrix, and \(\{x(t)\}\) is the displacement vector. The solution to this eigenvalue problem yields the natural frequencies \( \omega_n \) (where \( f_n = \omega_n / 2\pi \)) and the mode shapes \(\{\phi_n\}\):
$$ \left( [K] – \omega_n^2 [M] \right) \{\phi_n\} = 0 $$
Wear affects these modal parameters by locally reducing the mass and altering the stiffness distribution of the gear tooth. The net shift in natural frequency depends on which effect dominates.
To perform the analysis, three-dimensional models of the spur gear were created: one with the nominal perfect involute profile and another with the simulated non-uniform wear profile applied to the tooth flanks. The worn geometry was generated by scaling the simulated wear depth distribution to a maximum permissible value (based on a fraction of tooth thickness) and modifying the tooth profile accordingly using spline curves. Both models were then discretized using a fine mesh of tetrahedral elements, with material properties assigned (Steel, E=206 GPa, ν=0.3, ρ=7850 kg/m³). A fixed constraint was applied to the gear bore. The Lanczos eigenvalue extraction method was used to compute the first ten natural frequencies and mode shapes for both the unworn and worn configurations of the spur gear.
| Mode Order | Natural Frequency – Unworn (Hz) | Natural Frequency – Worn (Hz) | Frequency Shift (Hz) | Percent Increase (%) |
|---|---|---|---|---|
| 1 | 2639.9 | 2780.9 | +141.0 | 5.34 |
| 2 | 2806.1 | 2886.1 | +80.0 | 2.85 |
| 3 | 2818.6 | 2960.4 | +141.8 | 5.03 |
| 4 | 3329.4 | 3497.0 | +167.6 | 5.04 |
| 5 | 3726.8 | 3915.6 | +188.8 | 5.07 |
| 6 | 4579.0 | 4664.7 | +85.7 | 1.87 |
| 7 | 6171.9 | 6460.7 | +288.8 | 4.68 |
| 8 | 6295.6 | 6622.0 | +326.4 | 5.18 |
| 9 | 6789.3 | 7142.9 | +353.6 | 5.21 |
| 10 | 7287.3 | 7572.0 | +284.7 | 3.91 |
The results, summarized in Table 3, reveal a clear and consistent trend: all calculated natural frequencies of the spur gear increase after the introduction of non-uniform tooth surface wear. The magnitude of the increase varies across different mode orders. While lower-order modes (1-6) show significant increases, the higher-order modes (7-10) generally exhibit the largest absolute and often relative frequency shifts. For instance, the 9th mode frequency increases by over 350 Hz (5.21%), while the 6th mode increases by about 86 Hz (1.87%). This indicates that the effect of wear is more pronounced on modes that involve higher deformation energy in the tooth region.
The primary physical explanation for this frequency increase lies in the interplay between mass reduction and stiffness change. Wear removes material from the tooth surfaces, thereby reducing the overall mass of the spur gear. Simultaneously, it reduces the effective bending and contact stiffness of the teeth by thinning them. However, for the moderate, simulated wear levels in this study, the reduction in localized mass has a stronger influence on the eigenvalue \( \omega_n^2 \) than the reduction in localized stiffness. Since \( \omega_n \) is proportional to \( \sqrt{K/M} \), a greater percentage decrease in \( M \) relative to \( K \) leads to an increase in the natural frequency. A noticeable “jump” in the frequency curve occurs between modes 6 and 7, which corresponds to a transition in the type of mode shape from predominantly torsional/axial patterns to more complex radial and umbrella-type patterns involving significant tooth bending.
Regarding the mode shapes, the analysis shows that the fundamental character of the vibration modes for the spur gear remains largely unchanged by wear. The gear exhibits a rich variety of deformation patterns, including:
- Torsional/Rocking Modes (e.g., Modes 1 & 2): Rotation of the gear body about an axis in its plane.
- Axial “Umbrella” Modes (e.g., Modes 3, 4, 5, 6): Out-of-phase axial displacement of opposing sides of the gear rim, resembling an opening/closing umbrella. These can be 1st, 2nd, or higher-order circumferential patterns.
- Radial Modes (e.g., Modes 7 & 8): In-phase radial expansion and contraction of the gear rim, where the rim deforms into a polygonal shape (e.g., 2-lobe, 3-lobe).
- Coupled Radial-Axial Modes (e.g., Modes 9 & 10): Complex shapes combining radial polygon deformation with axial waviness, representing higher-order bending of the teeth and web.
While the spatial pattern (shape) of these modes is visually similar before and after wear, the amplitudes and precise nodal lines may shift slightly due to the altered mass and stiffness distribution. The most significant deformation in modes involving tooth bending consistently occurs at the tooth tips, which are the least constrained regions.
In conclusion, this integrated study on spur gears, combining quasi-static wear modeling and modal analysis, yields important insights. The numerical wear simulation successfully predicts the characteristic non-uniform wear pattern observed in spur gears: maximum wear at the pinion root, significant wear at the gear root, lesser wear at the addenda, negligible wear at the pitch point, and discontinuities at meshing transition zones. This pattern is a direct result of the variations in sliding distance, contact pressure, and load sharing inherent to the meshing process of spur gears.
Subsequent modal analysis of the worn spur gear geometry demonstrates that such non-uniform wear has a measurable and systematic impact on the dynamic properties. The most significant finding is a consistent increase in all natural frequencies, with higher-order modes often showing the greatest sensitivity. This shift occurs because the mass-removal effect of wear dominates over the stiffness-reduction effect for the analyzed wear depth. Importantly, while the frequencies change, the fundamental types of mode shapes (torsional, axial, radial) remain identifiable.
The practical implications are substantial for designers and maintenance engineers working with spur gears. The widening of the natural frequency band due to wear means that a spur gear’s resonance conditions will evolve over its operational lifetime. A system designed to avoid resonance with a new gear’s frequencies may eventually encounter resonance as the gear wears. Therefore, for high-reliability applications, dynamic design and diagnostic protocols should account for this progressive change in modal characteristics. Monitoring shifts in the frequency spectrum of gear vibration could potentially serve as an indicator of advancing wear severity. Future work could extend this analysis to dynamic wear models, the effects of wear on geared system response (rather than just the isolated gear), and experimental validation of the predicted frequency shifts.