RCF

Rollng contact

Fatigue & twin-disc rig analysis

Interactive rig

Rolling contact fatigue &
twin disc simulation

H. K. D. H. Bhadeshia

Rolling contact fatigue (RCF) is a cyclic stress-induced localized failure mechanism generated in non-conformal curved bodies under intense cyclic normal and shear pressures (0.8 to 3.5 GPa). Operating across railways, wind turbine planetary bearings, and gear sets, the twin disc apparatus provides the definitive method to decouple contact mechanics, slide-to-roll kinematics, and the fierce competition between surface wear and subsurface spalling.

Interactive twin disc test workstation

● Rig Operational

Adjust normal force ($w = W/L$), disc rotation speeds ($U_1, U_2$), traction coefficient ($\mu$), and roughness ($\sigma$) to view live kinematic contact slippage, Hertzian contact dimensions, and subsurface stress profiles.

Operational Inputs

800 kN/m
100 kN/m 2500 kN/m
1200 RPM (3.14 m/s)
1150 RPM (3.01 m/s)
0.08
0.01 (Hydrodynamic) 0.45 (Dry Rail Slip)
0.35 μm

Disc Kinematics & Contact Patch

EHL Regime (Full Film)
Contact Width: 2b = 0.76 mm
Peak Pressure (p₀) 1.64 GPa
Slide-to-Roll Ratio +4.26 %
Film Ratio (Λ) 3.82 (EHL)

Subsurface Shear Stress

Stress vs. Depth (z / b)

Max Shear (τ_max): 492 MPa
Depth (z_crit): 0.786 b (0.30 mm)

Subsurface initiation zone active. Plastic micro-straining accumulates around metallurgical inclusions.

Wear versus rolling contact fatigue

Twin disc testing evaluates wear and fatigue simultaneously, measuring how material loss competes against crack advancement.

Active test rig regime: Equilibrium / controlled competition

Under moderate slide-to-roll ratio (4.26%) and elastohydrodynamic lubrication, low surface wear rate allows subsurface crack propagation to compete with mild running-in asperity flattening.

Dominant risk Subsurface inclusion spalling
01

"Magic wear rate"

In heavy railway track engineering, mild surface wear acts as a protective mechanism. If the rate of surface wear exceeds the crack initiation and propagation velocity, microscopic cracks are continually machined off before they can kink downward and trigger catastrophic spalling.

Target condition: balanced wear removes micro-cracks without excessive geometry degradation.
02

Low-wear / spalling trap

When friction modifiers or ultra-hard clean steels suppress surface wear, micro-cracks formed at the surface or subsurface inclusions do not get polished out. Under millions of subsequent cycles, cyclic shear stresses drive deep crack propagation, detaching massive metal flakes (spalls).

Failure consequence: deep pit craters, squats, and complete rail/bearing fractures.
03

Severe wear & scuffing

Under high Slide-to-Roll Ratios ($SRR > 15\%$) or boundary lubrication ($\Lambda < 1$), frictional heating and metallic adhesion trigger severe adhesive transfer or severe abrasive grinding. While fatigue spalls never form, components fail prematurely through profile loss and dimension tolerance breaches.

Failure consequence: rapid radial profile shrinkage, noise, and loss of involute/wheel profile.

Experimental parameter comparison: wear vs. RCF Testing

ASTM G99 / G181 / ISO Standards Reference
Test objective Slide-roll ratio (SRR) Lubrication regime (Λ) Test duration Key measurements
Wear assessment High (> 10% – 50%) Dry / boundary (Λ < 1) 10⁴ to 2×10⁵ revolutions Disc mass loss ($\Delta m$), specific wear coefficient ($K$), optical profilometry profile reduction.
RCF & spalling endurance Low (0% – 5%) Full film EHL (Λ > 3) 10⁶ to 5×10⁷ revolutions Vibration spikes via accelerometer, crack section metallography, $L_{10}$ cycles to first spall.
Wear-RCF interaction Moderate (2% – 8%) Mixed film (1.0 < Λ < 2.5) 5×10⁵ to 5×10⁶ revolutions Micro-pit area percentage, crack depth distribution vs. radial wear step height.

Hertzian iine contact & subsurface stress fields

Analytical framework governing the elastic stress concentrations inside two rotating cylindrical discs.

1. Reduced radius & modulus

For cylindrical discs with radii $R_1, R_2$, Young's moduli $E_1, E_2$, and Poisson's ratios $\nu_1, \nu_2$:

1 / R* = 1 / R₁ + 1 / R₂
1 / E* = (1 - ν₁²)/E₁ + (1 - ν₂²)/E₂

R* denotes the equivalent radius; E* is the effective contact modulus.

2. Strip half-width & peak pressure

For a normal line load w = W / L, the half-contact strip width b and peak pressure p₀ are:

b = √( 4 · w · R* / (π · E*) )
p₀ = 2 · w / (π · b) = √( (w · E*) / (π · R*) )

Semi-elliptical pressure distribution: p(x) = p₀ · √(1 - x²/b²).

3. Critical subsurface depths

Maximum cyclic shear triggers internal plastic micro-slip and inclusion debonding:

Maximum principal shear: τ_max ≈ 0.300 · p₀
Critical depth: z_crit ≈ 0.786 · b
Reversing orthogonal shear: τ_zx ≈ ±0.256 · p₀ at z ≈ 0.50 b

Pure rolling keeps the surface in compressive principal stress, insulating it from tensile cracking.

Contact Width 2b p(x) Hertzian τ_max @ z ≈ 0.78b Traction q(x) = μ·p(x)

Surface traction shifts the stress field

When traction (μ) rises due to sliding friction or traction/braking torque, the surface experiences a superimposed tangential shear: q(x) = μ · p(x). As friction climbs beyond μ > 0.25, the location of peak shear stress migrates rapidly from deep subsurface ($z \approx 0.78b$) directly onto the contact surface. This causes an abrupt transition from subsurface spalling to surface-initiated micro-pitting, squats, and head-checks.

RCF damage s & microstructural changes

Inspection of crack incubation, propagation trajectories, and structural alteration in the steel matrix.

Non-metallic inclusion (Al₂O₃) White etching wings

Inclusion-initiated fatigue

Under high film ratios ($\Lambda > 3$), asperity contact is eliminated. Failure initiates internally around rigid non-metallic inclusions (aluminum oxides, titanium carbonitrides). Micro-cracks spread horizontally along the maximum shear band ($z \approx 0.78b$) before branching toward the surface, detaching large flakes.

Typical depth: 0.3 – 1.5 mm below the contact surface.
Hydraulic wedge crack (15°-30°)

Asperity fatigue & fluid wedging

Prevalent in mixed and boundary lubrication ($\Lambda < 1.5$). Asperities contact directly, shearing the uppermost 10–20 μm layer. Lubricant is driven into shallow angled cracks ($15^\circ - 30^\circ$) by the passing contact roll, acting as a pressurized fluid wedge that forces the crack tip forward.

Driven by: High SRR, low viscosity oil, rough surfaces.
Nanocrystalline Ferrite Microstructure

White etching cracks

A premature failure mode in wind turbine generator bearings and planet wheels, failing at 5–15% of design $L_{10h}$ lifespan. Metallurgical examination reveals micro-cracks bordered by ultrafine nanocrystalline ferrite grains (10–30 nm) that etch white under nital, catalyzed by hydrogen migration and electrical currents.

Catalysts: dynamic slip, stray currents, hydrogen embrittlement.

Lubricant film thickness & specific film ratio ($\Lambda$)

How fluid entrainment pressure determines whether the contact behaves hydrodynamically or suffers asperity contact.

In counterformal rolling contacts, the central minimum lubricant film thickness h_min is governed by hydrodynamic entrainment velocity U = (U₁ + U₂) / 2 and piezoviscous rheology through the Dowson-Higginson formulation:

h_min / R* = 2.65 · G^0.54 · U_dim^0.70 · W_dim^(-0.13)

Resistance to surface-initiated micro-pitting vs. subsurface spalling is governed by the Specific Film Thickness, or Lambda ratio ($\Lambda$):

Λ = h_min / √( R_q1² + R_q2² ) = h_min / σ Dimensionless
Lambda (Λ) Regime Contact interaction Wear / RCF mode
< 1.0 Boundary Complete asperity contact Severe abrasive/adhesive wear, scuffing, surface micro-cracks.
1.0 – 3.0 Mixed EHL Partial asperity contact Micro-pitting, gray frosting, hydraulic wedge crack propagation.
> 3.0 Full film EHL Complete fluid separation Subsurface initiated spalling around deep inclusion sites.

Interactive lambda evaluator

Film Thickness (h_min): 0.85 μm
Surface Roughness Disc 1 (Rq1): 0.25 μm
Surface Roughness Disc 2 (Rq2): 0.30 μm
Computed film ratio (Λ)
2.18
Mixed lubrication regime

Partial asperity interaction. Elevated probability of micro-pitting and surface distress.