7.1 Introduction: Why Contact Stress Governs Machine Failures
Gears, rolling bearings, cam-follower pairs, railroad wheels, and bolted joints fail not from bulk bending or tension but from localized contact stresses at the interface between two bodies. Hertzian contact theory (1882) predicts:
- Contact area much smaller than body dimensions.
- Peak pressures reaching hundreds of MPa to several GPa.
- Subsurface stress maxima that drive fatigue crack initiation below the surface.
Understanding contact mechanics is essential for:
- Bearing and gear design β subsurface fatigue (spalling, pitting).
- Fretting and wear (Chapter 8) β surface damage at contact interfaces.
- Failure analysis β interpreting spall morphology and crack origin depth.
- Tribology β bridging to film thickness and friction (EHL).
7.2 Hertzian Contact: Fundamental Assumptions
Hertz developed his theory for non-conformal (counter-formal) contact between two elastic bodies with:
- Linear elastic material behavior (Hooke's law, small strains).
- Frictionless contact (no tangential tractions at the interface in the classical formulation).
- Smooth surfaces (no roughness β real contact addressed in Section 7.8).
- Small contact zone relative to body curvatures (half-space approximation).
- Non-conforming geometry β bodies touch at a point (sphere-plane) or line (cylinder-plane) before loading.
When these assumptions are violated (conformal contact, plasticity, roughness, sliding friction), Hertz theory must be modified or supplemented.
7.3 Reduced Modulus $E^*$
When two elastic bodies with moduli $E_1$, $E_2$ and Poisson's ratios $\nu_1$, $\nu_2$ come into contact, the combined stiffness is captured by the reduced modulus:
$$\boxed{\frac{1}{E^*} = \frac{1-\nu_1^2}{E_1} + \frac{1-\nu_2^2}{E_2}}$$
7.3.1 Identical materials
For two bodies of the same material ($E_1 = E_2 = E$, $\nu_1 = \nu_2 = \nu$):
$$E^* = \frac{E}{2(1-\nu^2)}$$
For steel ($E = 210$ GPa, $\nu = 0.3$): $E^* = 210/(2 \times 0.91) = 115.4$ GPa.
7.3.2 Rigid body limit
If body 2 is rigid ($E_2 \to \infty$): $E^* = E_1/(1-\nu_1^2)$.
7.4 Sphere on Flat (Point Contact)
A sphere of radius $R$ pressed against a flat plane with normal force $P$ creates a circular contact patch of radius $a$:
7.4.1 Contact radius
$$\boxed{a = \left(\frac{3PR}{2E^*}\right)^{1/3}}$$
For sphere-on-sphere contact, use the equivalent radius $R = R_1 R_2 / (R_1 + R_2)$.
7.4.2 Peak (Hertz) pressure
$$\boxed{p_0 = \frac{3P}{2\pi a^2}}$$
The pressure distribution is semi-ellipsoidal:
$$p(r) = p_0 \sqrt{1 - \left(\frac{r}{a}\right)^2}, \qquad 0 \le r \le a$$
7.4.3 Approach (displacement)
The relative approach of the bodies:
$$\delta = \frac{a^2}{R} = \frac{3P}{2E^* a}$$
7.5 Line Contact: Cylinder on Flat (or Two Cylinders)
A cylinder of radius $R$ and length $L$ pressed against a flat (or another cylinder) creates a rectangular contact strip of half-width $a$:
7.5.1 Contact half-width
$$\boxed{a = \sqrt{\frac{4PR}{\pi E^* L}}}$$
7.5.2 Peak pressure
$$\boxed{p_0 = \frac{2P}{\pi a L}}$$
Pressure distribution (plane strain):
$$p(x) = p_0 \sqrt{1 - \left(\frac{x}{a}\right)^2}$$
7.5.3 Gear tooth contact
Spur gear teeth are approximated as cylinders in line contact with equivalent radius:
$$R_{\mathrm{eq}} = \frac{R_1 R_2}{R_1 + R_2}$$
where $R_1$, $R_2$ are the radii of curvature at the pitch point. AGMA and ISO gear standards use Hertzian contact stress for surface durability rating.
7.6 Subsurface Stress Field and Maximum Shear
The Hertzian pressure at the surface creates a subsurface stress field. For point contact (sphere on flat), along the axis of symmetry ($x = y = 0$):
7.6.1 Principal stresses vs depth $z$
$$\sigma_z = -p_0 \left[1 - \frac{z^3}{(a^2 + z^2)^{3/2}}\right]$$
$$\sigma_r = \sigma_\theta = -\frac{p_0}{2}\left[1 - \frac{z}{\sqrt{a^2 + z^2}} - \frac{z^3}{(a^2 + z^2)^{3/2}}\right]$$
7.6.2 Maximum orthogonal shear stress
$$\boxed{\tau_{\max} = 0.31 \, p_0 \quad \text{at} \quad z \approx 0.48 \, a}$$
This is the classic result: fatigue cracks in bearings and gears initiate below the surface at a depth of roughly half the contact radius, not at the surface.
7.6.3 von Mises stress
Maximum von Mises stress occurs at $z \approx 0.48a$ for $\nu = 0.3$, with value $\approx 0.32 p_0$.
7.6.4 Line contact (cylinder)
$\tau_{\max} \approx 0.30 p_0$ at $z \approx 0.78a$ β slightly deeper than point contact.
7.7 Assumptions, Limitations, and Extensions
7.7.1 When Hertz theory fails
| Violation | Effect | Remedy |
|---|---|---|
| Plastic yielding ($p_0 > \approx 3 S_y$) | Flattening, larger $a$, lower $p_0$ | Hardness-based models, FEA with plasticity |
| Friction / sliding | Tangential tractions, $\tau_{\max}$ moves to surface | Cattaneo-Mindlin, FEA |
| Conformal contact (journal bearing) | Not a half-space | Full numerical solution |
| Rough surfaces | Real area $\ll$ nominal area | Greenwood-Williamson (Section 7.8) |
| Dynamic / impact loading | Rate effects, stress waves | Dynamic FEA |
| Layered materials (coatings) | Stress field modified | Layered elastic theory |
7.7.2 Effect of friction on subsurface stress
With friction coefficient $\mu$, the maximum shear stress shifts toward the surface and can exceed the frictionless value. This is why sliding contacts (gears under scuffing, cam followers) show surface-initiated failures rather than subsurface spalling.
7.7.3 Elastic-plastic contact
When $p_0$ exceeds $\approx 2.8 S_y$ (von Mises) or $\approx 3.2 S_y$ (Tresca), plastic flow begins. The contact area enlarges and peak pressure is capped at $\approx 2.8 S_y$ (hardness limit). Meyer hardness and indentation theory extend Hertz to the elastic-plastic regime.
7.8 Surface Roughness and the Plasticity Index
Real surfaces are not smooth. Greenwood and Williamson (1966) modeled contact between rough surfaces as asperity contact:
7.8.1 Composite roughness
$$\sigma = \sqrt{\sigma_1^2 + \sigma_2^2}$$
where $\sigma_1$, $\sigma_2$ are RMS roughness heights of the two surfaces.
7.8.2 Plasticity index
$$\boxed{\psi = \frac{E^* \sigma}{H}}$$
where $H$ is the hardness of the softer material.
- $\psi < 0.6$: elastic asperity contact (Hertz valid at asperity level).
- $\psi > 1.0$: plastic asperity contact (most asperities yield on first contact).
- $0.6 < \psi < 1.0$: mixed elastic-plastic.
7.8.3 Real contact area
The nominal contact area (Hertz prediction) overestimates the real area of asperity contact:
$$A_{\mathrm{real}} = \frac{P}{p_{\mathrm{asperity}}}$$
where $p_{\mathrm{asperity}}$ can approach hardness $H$ for plastic asperities. Real area is typically 1β10% of nominal area for machined surfaces.
7.8.4 Implications for design
- High $\psi$ (soft + rough): high wear, plastic deformation, poor fatigue life.
- Low $\psi$ (hard + smooth): elastic contact, Hertz predictions reliable, good fatigue resistance.
- Shot peening and superfinishing reduce $\sigma$, lowering $\psi$ and improving contact fatigue life.
7.9 Contact Fatigue: Spalling and Pitting
7.9.1 Spalling mechanism
- Cyclic Hertzian stress creates subsurface shear at $z \approx 0.48a$.
- Micro-cracks initiate at material defects (inclusions, carbides) at the subsurface shear maximum.
- Cracks propagate at $\approx 45Β°$ to the surface (Mode II shear).
- Crack reaches the surface β spall (flake of material detaches).
7.9.2 Pitting
Surface-initiated fatigue (from friction, asperity interaction, or lubricant contamination) creates pits β small craters at the surface. More common in gears with boundary lubrication.
7.9.3 Design against contact fatigue
- Limit $p_0$ below the material's contact fatigue strength (from bench tests or standards).
- Ensure adequate lubricant film (EHL β Chapter 8) to separate surfaces.
- Use clean steel (low inclusion content, vacuum remelt) for bearing applications.
- Case hardening (carburizing, nitriding) creates a hard surface with compressive residual stress and a tough core.
7.10 Design Procedure: Hertzian Contact Analysis
- Identify contact geometry: sphere-plane, cylinder-plane, or general ellipsoid.
- Compute equivalent radius $R_{\mathrm{eq}}$ and reduced modulus $E^*$.
- Calculate contact dimensions: $a$ (radius or half-width) from Hertz formulas.
- Compute peak pressure $p_0$.
- Check yield: $p_0 < 2.8 S_y$ (elastic contact) or $p_0 < $ allowable contact fatigue stress.
- Locate subsurface $\tau_{\max}$: $z \approx 0.48a$ (point) or $0.78a$ (line).
- Assess roughness: compute $\psi$. If $\psi > 1$, Hertz overestimates performance.
- Check lubrication: compute minimum film thickness (EHL, Ch.8). If $\lambda < 1$, boundary lubrication β expect surface-initiated failure.
7.11 Summary
| Contact Type | $a$ or contact size | $p_0$ | $\tau_{\max}$ location |
|---|---|---|---|
| Sphere on flat | $a = (3PR/2E^*)^{1/3}$ | $3P/(2\pi a^2)$ | $z = 0.48a$ |
| Cylinder on flat | $a = \sqrt{4PR/(\pi E^* L)}$ | $2P/(\pi a L)$ | $z = 0.78a$ |
| General ellipsoid | Numerical (Hertz general) | β | $z \approx 0.4$β$0.8a$ |
Design takeaway: Never size gears, bearings, or cams on nominal compressive stress ($P/A_{\mathrm{nominal}}$). Use Hertzian $p_0$, check subsurface shear, verify lubrication regime, and assess roughness via the plasticity index. Contact failures initiate subsurface β inspect for spalling at the predicted depth.