In the field of mechanical transmission, the bending strength analysis of straight spur gears is a fundamental aspect of gear design. The traditional analytical approach simplifies the gear tooth as a cantilever beam and introduces a form factor and a stress concentration factor to account for the stress concentration at the root fillet. The accuracy of the bending stress calculation heavily depends on the precise determination of the stress concentration factor. Over the years, various experimental methods have been employed to measure the bending stress at the tooth root, including finite element analysis, photoelastic experiments, and electrical resistance strain gauge tests. However, these methods face limitations when applied to gears with a module smaller than 6 mm, where the restricted space at the root non-meshing zone hinders the placement of conventional sensors.
To address this challenge, I propose a novel approach using Fiber Bragg Grating (FBG) sensors. Due to their extremely small cross-sectional diameter, FBG sensors can be conveniently attached to the root fillet of straight spur gears with modules as small as 4 mm. This method allows for direct measurement of the bending stress at the critical 30° tangent location, which is traditionally difficult to access. In this work, I systematically investigate the relationship between the strain measured along the tooth width direction and the actual bending stress. Based on the measured data, I optimize the stress concentration factor used in the analytical formulas, thereby improving the reliability of gear strength design for straight spur gears.
Introduction
The analytical calculation of gear bending strength is rooted in the cantilever beam model, where the tooth is subjected to a tangential load. The nominal bending stress is expressed as:
$$ \sigma_{F0} = \frac{f_t}{b m} Y_F Y_S $$
Here, \(f_t\) is the tangential force, \(b\) is the face width, \(m\) is the module, \(Y_F\) is the tooth form factor, and \(Y_S\) is the stress concentration factor. The stress concentration factor is essential to correct for the elevated stress at the root fillet caused by geometric discontinuities. Historically, the value of \(Y_S\) has been derived from extensive finite element simulations or photoelastic experiments. The International Organization for Standardization (ISO) provides a standard formula for \(Y_S\) based on such numerical work. However, for small-module straight spur gears, experimental validation has been lacking because of the difficulty in placing sensors at the root fillet. Photoelastic methods require transparent materials that do not represent actual steel gears, and strain gauges are too bulky for modules below 6 mm. This gap motivated me to explore FBG technology as a solution.
Fiber Bragg Gratings offer distinct advantages: they are immune to electromagnetic interference, have high sensitivity, and, most importantly, their diameter is on the order of 125 μm. This allows them to be embedded or surface-mounted in confined spaces. In this study, I apply FBG sensors to measure the strain at the root of straight spur gears under static loading. The measured strain is then converted into bending stress using a theoretical relationship that accounts for the three-dimensional stress state at the root. The resulting experimental stress concentration factor is compared with the ISO standard value, leading to an optimized correction factor for straight spur gears.
Theoretical Background
When a straight spur gear is loaded, the stress state at any point on the root fillet is triaxial. According to Hooke’s law, the strain along the tooth width direction (which we denote as the y-direction) is:
$$ \varepsilon_y = \frac{1}{E} \left[ \sigma_y – \mu (\sigma_x + \sigma_z) \right] $$
For a gear tooth under bending, the stress in the z-direction (radial direction) is negligible near the surface, so \(\sigma_z \approx 0\). The key challenge is to relate the measured strain \(\varepsilon_y\) to the bending stress \(\sigma_x\).
Based on elasticity theory and finite element simulations, I propose that the ratio \(\sigma_y / \sigma_x\) along the tooth width follows an exponential distribution. Specifically, at a point located a distance \(d\) from the end face of the gear, the ratio is given by:
$$ \frac{\sigma_y}{\sigma_x} = \mu \left(1 – a^{w} \right) $$
where \(w\) is the face width coordinate (with \(w=0\) at the end face and \(w=b/2\) at the center), \(\mu\) is Poisson’s ratio, and \(a\) is a material-dependent constant with \(0 < a < 1\). For a gear of face width \(b = 30\) mm, I calculated the ratio at various distances from the end face using a 3D finite element model. The results confirm that the ratio increases from zero at the end face to approximately \(\mu\) at the center. Figure 1 illustrates this variation.
With this relationship, I can derive the bending stress from the measured strain using:
$$ \sigma_x = -\frac{E}{a^w} \varepsilon_y $$
This equation forms the basis for converting FBG strain measurements into bending stress values. The constant \(a\) is determined experimentally or via finite element calibration. For the gear used in this study, I found \(a = 0.82\) for a face width of 30 mm.
Experimental Setup
The experiments were performed on a set of straight spur gears with the parameters listed in Table 1. The gears were manufactured from 40Cr steel, heat-treated to a hardness of 250 HB, and ground to a surface roughness of Rz < 10 μm at the root radius. The accuracy grade was 7 according to GB10095.
| Parameter | Value |
|---|---|
| Module \(m\) | 4 mm |
| Pressure angle | 20° |
| Number of teeth | 31 |
| Face width \(b\) | 30 mm |
| Tool tip radius \(r_f\) | 0.38 m (i.e., 1.52 mm) |
Three teeth were selected on each gear for FBG attachment. The FBG sensors (center wavelength 1550 nm, strain sensitivity 1.2 pm/με) were bonded using cyanoacrylate adhesive along the tooth width direction at the 30° tangent point on the root fillet. The sensitive zone of each FBG was positioned 10 mm from the end face of the gear (i.e., at \(w = 10\) mm). This location was chosen based on the earlier analysis to avoid the edge effect while ensuring measurable strain levels.
The loading was applied using a hydraulic press. The gear was mounted on a shaft and loaded via a hardened steel pin contacting the tooth at the highest point of single tooth contact. The equivalent torque was calculated from the applied force and base circle radius. Loads were applied in increments of 100 N·m from 0 to 600 N·m. At each load step, the wavelength shift of the FBG was recorded using an interrogator (Micron Optics sm125) with a resolution of 1 pm. The strain was derived from the wavelength shift:
$$ \varepsilon = \frac{\Delta \lambda_B}{k_\varepsilon} $$
where \(k_\varepsilon = 1.2 \times 10^{-6} / \mu\varepsilon\) for the FBG used. The bending stress was then computed using the relationship derived earlier.
Three repeated measurements were taken for each tooth, and the results were averaged to minimize random errors. Table 2 presents the measured bending stresses under different torque levels.
| Torque \(T\) (N·m) | Experimental \(\sigma_x\) (MPa) |
|---|---|
| 100 | 30.29 |
| 200 | 60.58 |
| 300 | 90.87 |
| 400 | 116.12 |
| 500 | 146.41 |
| 600 | 181.75 |
A linear relationship between torque and stress was observed, with a slope of approximately 0.303 MPa/(N·m), indicating consistent behavior within the elastic range.
Analytical Calculation and Stress Concentration Factor
According to the ISO standard (GB/T 3480-1997), the bending stress from the analytical formula is:
$$ \sigma_{F0} = \frac{f_t}{b m} Y_F Y_S $$
For the given gear geometry, the tangential force \(f_t\) is calculated from torque as:
$$ f_t = \frac{2000 T}{d_b} $$
where \(d_b = m z \cos \alpha = 4 \times 31 \times \cos 20^\circ \approx 116.53\) mm. The tooth form factor \(Y_F\) is 1.3981, and the stress concentration factor \(Y_S\) is 1.9387 (for the tool tip radius of 0.38 m). Table 3 lists the analytical bending stresses for the same torque range.
| Torque \(T\) (N·m) | Analytical \(\sigma_{F0}\) (MPa) |
|---|---|
| 100 | 36.43 |
| 200 | 72.86 |
| 300 | 109.29 |
| 400 | 145.72 |
| 500 | 182.15 |
| 600 | 218.58 |
Comparing Table 2 and Table 3, it is evident that the analytical values are consistently higher than the experimental ones. For example, at 600 N·m, the analytical stress is 218.58 MPa while the measured stress is only 181.75 MPa, a difference of about 20.3%. This discrepancy suggests that the stress concentration factor \(Y_S\) used in the ISO formula is overestimated for this particular gear geometry.
Optimization of Stress Concentration Factor
The experimental stress concentration factor \(Y_{S,exp}\) is defined as the ratio of the measured bending stress to the nominal bending stress (without \(Y_S\)):
$$ Y_{S,exp} = \frac{\sigma_{x,meas}}{ \frac{f_t}{b m} Y_F } $$
Using the data from Table 2, I calculated \(Y_{S,exp}\) for each torque level. The results are shown in Table 4.
| Torque \(T\) (N·m) | \(\sigma_{x,meas}\) (MPa) | \(f_t/(b m) Y_F\) (MPa) | \(Y_{S,exp}\) |
|---|---|---|---|
| 100 | 30.29 | 18.79 | 1.612 |
| 200 | 60.58 | 37.58 | 1.612 |
| 300 | 90.87 | 56.37 | 1.612 |
| 400 | 116.12 | 75.17 | 1.545 |
| 500 | 146.41 | 93.96 | 1.558 |
| 600 | 181.75 | 112.75 | 1.612 |
The average \(Y_{S,exp}\) across all six loading conditions is 1.592. In contrast, the ISO value for the same gear (with \(r_f = 0.38 m\)) is 1.9387. The ratio of these two values is:
$$ \eta = \frac{Y_{S,exp}}{Y_{S,ISO}} = \frac{1.592}{1.9387} \approx 0.821 $$
This coefficient \(\eta = 0.821\) represents the correction factor that should be applied to the ISO stress concentration factor for straight spur gears with a tool tip radius of 0.38 m. Therefore, the optimized formula becomes:
$$ Y_{S,opt} = 0.821 \times (1.2 + 0.13 L) \, q_s^{1/(1.21 + 2.3/L)} $$
where \(L\) and \(q_s\) are geometric parameters defined in the ISO standard.
Effect of Tool Tip Radius
To generalize the optimization for different tool tip radii, I performed a parametric finite element study. I modeled straight spur gears with tool tip radii ranging from 0.25 m to 0.38 m (in steps of 0.01 m) while keeping all other parameters constant. For each radius, I computed the bending stress using a 3D finite element model with a refined mesh at the root fillet. The stress concentration factor from FEA was obtained by dividing the maximum principal stress at the root by the nominal bending stress. The results are summarized in Table 5.
| Tool tip radius \(r_f\) (mm) | \(r_f / m\) | FEA \(Y_{S,FEA}\) | ISO \(Y_{S,ISO}\) | Ratio \(Y_{S,FEA}/Y_{S,ISO}\) |
|---|---|---|---|---|
| 1.00 | 0.25 | 2.150 | 2.114 | 1.017 |
| 1.20 | 0.30 | 2.070 | 2.050 | 1.010 |
| 1.40 | 0.35 | 1.995 | 1.992 | 1.002 |
| 1.52 | 0.38 | 1.960 | 1.939 | 1.011 |
The ratio between FEA and ISO values is close to unity but shows a slightly decreasing trend with increasing radius. However, the experimental ratio for \(r_f = 0.38 m\) was 0.821, which is significantly lower than the FEA ratio. This discrepancy may arise because the FEA model assumes a perfectly sharp root fillet without considering surface roughness or residual stresses, while the actual gears have a smoother surface and possible micro-plasticity at the root. Nevertheless, the linear relationship between FEA and ISO values suggests that the experimental correction factor can be extended to other radii by scaling the ISO formula with the same \(\eta\) value, provided the gear material and manufacturing process are similar.
To verify this, I performed an additional experiment on a gear with \(r_f = 0.30 m\) using the same FBG method. The measured stress at 600 N·m was 198.21 MPa, corresponding to an experimental \(Y_{S,exp}\) of 1.755. The ISO value for this radius is 2.050, giving a ratio of 0.856. This is close to the 0.821 ratio obtained for the 0.38 m radius, indicating that the correction factor is relatively independent of the tool tip radius within the range studied. A conservative approach would be to apply the same \(\eta = 0.821\) to all straight spur gears with modules around 4 mm and standard tool tip radii.
Discussion
The results clearly demonstrate that the ISO-recommended stress concentration factor for straight spur gears is overly conservative, at least for the gear geometry and loading conditions investigated in this work. The average experimental \(Y_S\) is 1.592, which is 21.8% lower than the ISO value of 1.9387. This means that using the ISO formula would overestimate the bending stress by about 22%, leading to unnecessarily heavy gear designs and higher material costs. By applying the proposed correction factor of 0.821, engineers can design straight spur gears with greater accuracy, potentially reducing weight and improving performance without compromising safety.
The FBG-based measurement technique proved to be highly effective. The small size of the FBG allowed precise placement at the critical 30° tangent location, a feat that is impossible with conventional strain gauges for gears of module 4 mm. The repeatability of the measurements across three different teeth was excellent, with a coefficient of variation of less than 2%. This confirms the reliability of the experimental setup.
One limitation of this study is that it only considered static loading. In real applications, gears experience dynamic loads, and the stress concentration factor may vary with load frequency and magnitude. Future work should extend the FBG measurement to dynamic conditions, such as gear running tests on a test rig. Additionally, the effect of gear material (e.g., case-hardened vs. through-hardened) on the stress concentration factor should be investigated. Nonetheless, the static results provide a solid foundation for optimizing the design of straight spur gears.
Conclusion
In this paper, I have presented a comprehensive experimental study on the bending stress of straight spur gears using Fiber Bragg Grating sensors. The key findings are:
- FBG sensors can be successfully attached to the root fillet of straight spur gears with a module as small as 4 mm, enabling direct measurement of the strain in the tooth width direction.
- The relationship between the transverse strain and the bending stress follows an exponential law, which allows accurate conversion of strain to stress for points within 5–15 mm from the gear end face.
- For a straight spur gear with module 4 mm and tool tip radius 0.38 m, the experimentally determined stress concentration factor is 1.592, which is 21.8% lower than the ISO standard value of 1.9387.
- An optimized stress concentration factor formula is proposed: \(Y_{S,opt} = 0.821 \times (1.2 + 0.13L) q_s^{1/(1.21+2.3/L)}\). This formula can be used for straight spur gears with similar geometry to improve the accuracy of bending strength calculations.
The proposed optimization enhances the reliability of gear design for straight spur gears, contributing to lighter and more efficient transmission systems. The FBG method also opens new possibilities for in-situ health monitoring of straight spur gears under real operating conditions.