Oil film thickness in high-speed roller bearings under combined loads
Roller bearings in high-speed aircraft gearboxes operate under a demanding combination of centrifugal force, high rolling speed, radial loading, axial thrust and intermittent sliding. The lubricant film separating the contacting surfaces is therefore much more than a thin layer of oil. Its minimum thickness, pressure distribution and shear behaviour determine friction, power loss, wear and the ability of the bearing to survive the full torque and speed envelope. A gearbox with efficient gears can still lose performance if bearing drag and lubrication losses are not controlled.
A defensible methodology for predicting oil film thickness in high-speed roller bearings under combined loads links the mechanical duty, lubricant properties, elastohydrodynamic behaviour, oil-supply conditions and physical test evidence. This approach supports the OPTIMIZE Project’s broader goal of reducing losses in geared aircraft engines while improving durability, weight and power density. It also replaces judgement based on a single lubricant figure with a traceable engineering process suitable for design-of-experiments studies, tolerance analysis and aerospace qualification.
From gearbox duty to bearing contact
The starting point is a bearing load spectrum rather than a catalogue load rating. Gear torque, shaft speed, gear-mesh forces, rotor imbalance, housing deflection and bearing preload all vary through the operating cycle. These quantities establish roller and raceway contact pressure, while centrifugal migration can change lubricant distribution inside the bearing. Shaft and housing stiffness are especially important in a hyperstatic gearbox, where several constrained contact paths may prevent a shaft from moving as freely as an isolated bearing model would assume.
Radial and axial loads must be treated together because their interaction changes the loaded rolling path. A modest axial component can shift the load zone towards an edge, increase local contact stress and reduce the effective lubricant film. The model should therefore represent roller skew, misalignment, cage motion and local sliding alongside nominal Hertzian pressure. Static structural analysis provides the load field, but transient duty data are needed to assess how rapidly that field changes and whether starvation or heat build-up becomes likely.
Gear geometry also belongs within this system. The OPTIMIZE Project’s gear-tooth tip-relief findings demonstrate how contact conditions can change across torque and speed. That information should feed the bearing-duty calculation rather than being assessed separately, since altered gear-mesh loading can affect shaft deflection, lubricant movement and the conditions seen by the rolling elements.
Constructing an elastohydrodynamic film model
At the core of the calculation is an elastohydrodynamic lubrication model that couples fluid pressure, elastic deformation of the rolling bodies and raceways, lubricant viscosity and surface roughness. Reynolds’ equation describes pressure generation, while the pressure–viscosity relationship accounts for the strong increase in lubricant viscosity under high contact pressure. Temperature changes viscosity and elastic properties, so a credible model needs an energy balance rather than a fixed oil-temperature input.
A classical line-contact solution can provide a useful baseline for cylindrical and tapered roller contacts, but it should be extended when point contact, edge loading, misaligned rollers or varying surface topography dominate. The important outputs include minimum film thickness, peak pressure, pressure gradients and shear stress. Minimum film thickness alone is insufficient: engineers should also calculate the film parameter, commonly expressed as minimum film thickness divided by the composite RMS roughness of the two surfaces. A low film parameter indicates that asperities are likely to share the load and increases the risk of scuffing.
Uncertainty must be built into the model rather than added after calculation. Relevant variables include roller and raceway dimensions, clearance or preload, viscosity–temperature coefficients, pressure–viscosity behaviour, roughness, applied load, speed, feed rate and oil-supply pressure. Monte Carlo trials can expose combinations that produce thin films or high shear. Like the random outcomes displayed in low-stakes online slots, these trials sample uncertain inputs; unlike a game, every bearing case must remain physically constrained, numerically converged and traceable to a justified input range.
Connecting oil supply with thermal behaviour
Oil reaching the bearing must have enough momentum and flow to enter the rolling contacts at the required entrainment speed. A supply system that looks adequate under static conditions can fail once roller speed increases, the lubricant viscosity falls or centrifugal effects move oil towards the outer race. The OPTIMIZE Project’s oil-jet supply-pressure study is relevant because it demonstrates how supply pressure influences gear-mesh engagement and churning losses. Bearing analysis should apply the same systems thinking to jet targeting, delivery pressure, back-pressure and the interaction between external spray and splash lubrication.
Film formation and temperature form a feedback loop. Greater shear generates heat, a warmer lubricant has lower viscosity, and the lower viscosity can reduce film thickness and increase shear again. Heat also changes clearance, preload and material properties. Designers should evaluate this loop across the complete operating profile, including warm-up, continuous cruise, high-power transients and hot-soak periods. Optimising the minimum film at one nominal operating point may therefore worsen the thermal margin elsewhere.
Starvation, aeration and cavitation require explicit checks. None can be inferred reliably from a full-flow assumption. This matters in Australia, where aircraft may operate in the extreme summer heat of Alice Springs, the humidity of Darwin or long coastal sectors around Perth and Brisbane. Ambient conditions influence heat soak and turnaround time even when onboard systems control lubricant temperature. Regional routes, FIFO operations and charter services across the Pilbara also create a practical requirement for reliable lubrication: a marginal bearing should not depend on perfect oil delivery every time the aircraft is dispatched.
Validating predictions against measured behaviour
Validation should proceed from simple checks towards representative high-speed testing. Material properties and viscosity data can first be verified in laboratory measurements, followed by rolling-contact tests, instrumented bearing rigs and finally gearbox-level trials. Direct film-thickness measurement is uncommon, so the evidence will often come from friction torque, bearing temperature, vibration signatures, electrical contact resistance, wear debris and post-test surface inspection. Specialist suppliers such as Cavendish Instruments can be consulted when identifying suitable friction, wear and tribology instrumentation for each stage.
The Australian operating environment shapes both test planning and maintenance acceptance. Specialist fixtures, replacement sensors and certified lubricants may need to travel between Melbourne, Brisbane, Adelaide or Perth, so procurement and calibration lead times belong in the project plan. For civil maintenance, predictions may inform work performed under a CASA-approved maintenance organisation’s controlled procedures, including the requirements of CASR Part 145. Defence applications also demand compatibility with Royal Australian Air Force sustainment practices. In Australian engineering shorthand, the result must be “fair dinkum” and the assumptions must “keep it honest”; that means documented evidence, not simply a convincing simulation.
Measured data should be compared with calculated film pressure, shear and temperature using predefined acceptance criteria. Manufacturing tolerances, surface finish, assembly preload and lubricant variability require statistical treatment rather than a single nominal case. If the model reproduces average friction but misses the worst credible combination, it remains incomplete. Conversely, a conservative model with quantified uncertainty is more useful than a precise-looking prediction that cannot be verified. Test articles, calibration records and model versions should therefore be stored together, allowing later maintenance data to close the loop.
From validated prediction to maintenance decision
A reliable implementation keeps the engineering chain connected from gearbox requirements to maintenance outcomes. The following actions provide a practical route from early simulation to operational use:
- Define the complete torque, speed, thrust and thermal duty cycle before selecting bearing dimensions or lubricant parameters.
- Characterise the lubricant with measured viscosity, density and pressure–viscosity data across the expected temperature range.
- Use coupled elastohydrodynamic and thermal calculations that account for combined loading, roughness, preload and local roller behaviour.
- Model oil delivery, entrainment and starvation explicitly, including the operating conditions that produce the thinnest film.
- Validate the model in stages with material, component, full-bearing and gearbox tests, using both direct and indirect measurements.
- Record input ranges, tolerances, convergence settings and uncertainty so predicted margins can be updated when hardware or lubricant specifications change.
Teams can act now by reviewing the OPTIMIZE Project’s technical videos, objectives and methodology, then bringing representative bearing duties, lubricant data and planned rig tests into a joint design-of-experiments workshop. Australian research organisations, component suppliers, maintenance providers and aerospace programs can use that evidence to establish a shared, measurable route towards lower gearbox losses and more dependable high-speed propulsion systems.