Bearing axial preload and gear train rotordynamic stability
The OPTIMIZE Project treats an aircraft-engine gearbox as an integrated engineering system. A lighter gear web or improved tooth form delivers useful gains only when the rotating assembly stays aligned, controlled and durable. Bearing axial preload is one of the less visible variables in that balance. It is the clamping action applied through opposed bearings to limit relative axial movement between the shaft and bearing rings, and it differs from gear-tooth contact preload.
Bearing axial preload affects the rotordynamic stability of the gear train by changing support stiffness, load paths and shaft alignment under gyroscopic and axial loads. A modest change in installed clamping force can move a critical speed, redistribute gear-mesh forces or alter vibration growth at operating speed. The engineering task is to establish a safe, repeatable preload range across hot oil, high rotational speed, assembly variation and the full flight envelope.
How preload changes the support system
Preload needs a defined force path: the axial force from one bearing must react through the casing and the opposing bearing. With angular-contact ball bearings or tapered-roller units, a controlled axial force holds the rolling contacts at their intended geometry and limits axial float. Installed preload depends on bearing clearance, contact angle, applied load, fits and the stiffness of the shaft and housing. It is therefore an assembly outcome that must be measured, rather than a value obtained merely by tightening a locknut.
The consequences extend beyond the bearing race. Preload raises axial stiffness directly and can change lateral effective support stiffness through contact-angle and alignment effects. A shaft held with consistent clamping force can exhibit a more repeatable relationship between deflection, bearing reactions and gear-mesh loading. This consistency becomes valuable when gyroscopic moments, reversing thrust and helical-gear forces create a demanding combination of radial and axial loads.
The operating machine also changes the initial assembly condition. Depending on bearing type, centrifugal growth of the rings, differential thermal expansion and the selected fits can alter clamping during spool-up. A preload that is satisfactory on the bench may become excessive at high speed or insufficient during a transient. Analysis must therefore represent preload as a function of speed, temperature, load and installation tolerance, using nonlinear bearing load-deflection data where the behaviour changes appreciably.
The risks of low and excessive clamping
Insufficient clamping allows axial travel and weakens the control over bearing alignment. As the shaft moves under thrust or gyroscopic loading, the loaded portion of a rolling contact can shift towards its edge, changing contact angle, friction and load distribution between opposing bearings. The gear-mesh force can then act through a less predictable support stiffness. Small movements may also excite flexible shaft modes that remain quiet in an unloaded, unloaded-preload test arrangement.
Excessive clamping creates a different set of risks. Higher bearing loads increase rolling-contact stress and may raise frictional losses, lubricant temperature and the rate of wear. An over-constrained bearing stack can also reduce the compliance needed for controlled self-alignment, making modest housing or shaft misalignment more damaging. Thermal growth then has fewer available paths for accommodation, allowing parasitic axial loads to develop in the casing and gear train.
The useful design range must account for both extremes. The lower limit should prevent harmful free travel under the largest expected combined axial and gyroscopic load. The upper limit should retain acceptable contact stress, thermal margin, lubrication conditions and damping. A practical target is set with a tolerance band around it, followed by verification on representative hardware. Assembly torque alone cannot establish this range because friction at threads and contact surfaces can vary considerably.
Critical speeds, damping and vibration growth
A credible rotordynamic assessment couples the rotor, bearings, gear mesh and supporting structure. Shaft and disc dynamics provide the inertia and gyroscopic terms; bearing stiffness and damping represent the support; mesh stiffness and phasing represent gear forcing; and housing flexibility determines where those forces ultimately react. Preload affects several parts of this model at once. A model based on one equivalent spring can conceal the mode shapes, damping changes and load-sharing shifts that matter to stability.
Moving a critical speed away from a range of operating speeds can reduce vibration amplification, but speed placement alone does not prove stability. A high response near a resonance may remain bounded if damping is positive and the excitation is limited. Preload can alter damping through rolling-contact behaviour, friction, lubricant effects and structural flexibility. It can also support stability by maintaining contact alignment and limiting some forms of slip, while excessive clamping can increase heat and introduce friction-related excitation. The sign and size of the effect depend on the specific bearing and gearbox arrangement.
Experimental verification should therefore look beyond a single vibration limit. Run-up and coast-down sweeps can reveal synchronous peaks, subsynchronous whirl, gear-mesh sidebands and components associated with rolling-contact friction. Accelerometers on the casing and bearing supports should be paired with measurements of shaft motion, oil temperature and drive torque. The result is a clearer link between a calculated stability margin and the vibration behaviour observed on the actual rotor-bearing-gear system.
Tolerances, assembly and hyperstatic constraints
Preload is also vulnerable to manufacturing and assembly variation. Spacer thickness, housing bore geometry, bearing fit, shoulder squareness and clamp method all contribute to the installed force. Multiple bearings and constrained mounting points can make a shaft system hyperstatic, meaning that load distribution depends on the stiffness compatibility between components. Strong preload removes an unintended freedom for axial motion, yet it can introduce parasitic forces as the shaft, housing and bearings expand differently. Design tolerances and service clearances must be considered together.
Design of experiments provides a disciplined way to connect assembly inputs with dynamic outcomes. The project’s response-surface optimisation methods offer a useful model: vary spacer thickness, bearing fits, housing geometry and clamp method, then measure installed preload, critical speeds and vibration amplitudes. Simulation can screen the design space, while physical tests verify that the response surface matches reality.
For Australian supply chains, this work must also support material traceability. Specialist bearings and high-strength bearing steel are often imported, so customs records, supplier certificates and lot identification should connect to incoming inspection results. Correct dimensions alone do not guarantee identical clearance, fit or material performance. These records become especially valuable for aerospace manufacturing, maintenance and overhaul operations around Bankstown, New South Wales, where small assembly changes can otherwise be difficult to investigate.
Engineering checks for stable, buildable systems
Before hardware is built, the assessment should bring design intent, tolerances and test evidence into one control process. The following design checks treat preload as a variable with a measurable range rather than a nominal assembly number.
- Map installed preload against the mission envelope, including start-stop transients, take-off, cruise and thermal conditions that change bearing clearance and lubricant viscosity.
- Use a coupled rotor–bearing–gear model with measured bearing load-deflection data, housing compliance and time-varying mesh stiffness, rather than a single equivalent spring.
- Perform tolerance sensitivity studies on bearing fits, spacer thickness, housing bores, shoulder squareness and assembly preload to identify combinations that erode the stability margin.
- Balance the clamping force against rolling-contact stress, friction heat, lubricant flow and the stiffness needed for controlled self-alignment.
Verification then needs to distinguish a prediction error from an assembly error. A practical test and delivery checklist includes the following measures.
- Measure installed preload on representative hardware using axial displacement, load or torque measurements, and record the associated temperature and speed.
- Run loaded and unloaded speed sweeps, including run-up and coast-down, while tracking bearing and casing vibration, shaft motion, oil temperature and drive torque.
- Repeat the build across several units or lots so the report separates systematic modelling error from assembly scatter and bearing-to-bearing variation.
- Benchmark test and documentation practices against European guidance available through AA Europa resources, while retaining the evidence trail required for Australian approval.
From stable analysis to validated hardware
Australia’s resources and energy engineering base around Brisbane offers relevant expertise in rotating machinery and high-temperature systems. High-speed, load-representative gearbox test facilities remain limited compared with Europe and North America, so Australian programmes may need to combine university rigs, defence facilities and overseas test houses. Logistics, specialist hardware shipping and defence data-security requirements belong in the test plan from the outset, allowing a representative gearbox to be built around the instrumentation required to measure preload and vibration together.
For civil aircraft certification in Australia, the evidence must fit the certification basis agreed with CASA. Defence programmes add contractual, security and verification requirements, and an overseas test report may require careful acceptance by the relevant authority. Australian regional airline, defence and energy markets reward availability, predictable maintenance and fuel-efficient powerplants, so the results should preserve traceability from bearing batch and assembly parameters through to simulation cases and measured vibration data.
The practical outcome is a preload specification that remains stable across the assembly line and credible across the operating envelope. Engineers, researchers and industry partners can contact the project team to discuss simulation studies, tolerance experiments, physical testing and access to project videos.