Order tracking analysis for gearbox vibration in flyover
Using order tracking analysis to identify gearbox vibration sources in flyover requires more than finding a large peak in a vibration spectrum. A flyover is a changing operating condition, often combining high rotational speed, climb or transit load, and several internal forces, so a conventional frequency plot can smear useful information. Order tracking converts the signal into a shaft-referenced view, allowing engineers to follow behaviour whose frequency rises and falls with gearbox speed.
The OPTIMIZE Project treats this as an engineering evidence problem rather than a simple fault-isolation exercise. Order tracking establishes which orders are present, how their amplitudes evolve and whether their timing matches a candidate component. Interpretation still needs geometry, lubrication data, drive speed and test context. For an Australian research or maintenance team, the result can support design decisions, test planning and continued-airworthiness investigations without treating a narrow spectral clue as proof of failure. A structured approach keeps the diagnosis tied to observable evidence and the condition of the engine.
Defining the flyover operating profile
During a flyover profile, shaft speed and torque are rarely constant. Engines accelerate, gear stages change load, and a clutch or free-running section may briefly separate the input from the output. If “flyover” has a local meaning within a test programme, the first step is to translate it into a repeatable envelope: rated speed, acceleration rate, load schedule, lubrication flow and the exact interval retained. This removes ambiguity and gives the order-tracking analysis a stable reference.
A fixed-frequency Fourier transform assumes that the machinery is reasonably steady during each record. As rpm changes, one component can spread across several bins, while nearby gear-mesh and bearing frequencies move at different rates. Plotting the complete climb against time can therefore hide the narrow signature of a defect. Engineers often call this broadening or spectral smear; it is a consequence of analysing a non-stationary event with stationary assumptions.
Order analysis solves that timing problem by expressing frequency relative to shaft rotation. A once-per-revolution tachometer or encoder supplies the reference, while accelerometers capture the response. Matching each vibration sample to a shaft angle lets analysts resample the record at equal angular intervals. Components then remain at identifiable orders even while absolute frequency changes. The technical documentation is a useful reference for connecting the measurement plan to the project’s broader gearbox design and validation workflow.
Building a trustworthy order-tracking dataset
Useful data begins at the casing, not only in the software. Triaxial accelerometers should be mounted firmly on bearing housings or structural ribs near the candidate path, with consistent sensor orientation across repeated runs. At least one independent speed reference is needed for each rotating shaft whose behaviour is under investigation. Electrical mains noise, accelerometer resonance, cable movement and weak mounting can mimic mechanical lines, so pre-test checks and a documented rigid-surface mounting method are essential.
Signal processing then follows a controlled chain: detect the speed pulse, calculate shaft angle, interpolate vibration channels against that angle, and apply an order-domain transform. Anti-alias filtering and a sample rate high enough for the expected tooth-mesh content prevent misleading folds. Analysts should keep the raw speed and vibration data, processing settings and angular reference used for every figure. This makes results repeatable when another engineer reviews the evidence during an arvo handover or compares runs from a later campaign.
The finished output should include a spectrum at stable operating points, an order-versus-time map through the flyover, and amplitude and phase trends for selected peaks. Sidebands, harmonics and coherence between shafts often reveal modulation that a single average spectrum suppresses. Waterfall or ridge plots are useful for following a component that grows during acceleration, while narrowband order analysis can expose bearing excitation that remains hidden in a broadband reading.
Matching vibration orders to gearbox mechanisms
Candidate sources can be grouped by the order they create. Shaft imbalance generally follows the first rotational order of its shaft; angular misalignment often adds axial content and a second-order component; gear tooth error excites the mesh order and sidebands. Bearing defects may appear at geometric defect frequencies, often with harmonics or speed sidebands. Looseness, rubbing and structural resonance can modify the same bands. Order tracking therefore narrows the field, but a matching number does not establish a unique cause.
Location and direction provide the next layer of evidence. A peak strongest near one bearing housing may support a local source, while phase differences across housings can show how a disturbance propagates through the housing. In a hyperstatic gearbox, built-up load sharing means that tooth contact, housing stiffness and bearing support cannot be separated from the spectrum alone. Design-of-experiments and sensitivity methods can test how those relationships change with speed, load and geometry. The GenOptics resources can support parameter studies that connect an observed order with influential design choices.
Lubrication deserves particular attention because a flyover can reduce oil pressure or change oil temperature, viscosity and flow at the same time that bearing and gear loads rise. Aeration, cavitation, churning and poor sump scavenging can then create broadband energy or narrow bands that resemble mechanical resonance. The vibration map should be read beside oil-flow, temperature and pressure records, with inspection findings used to confirm or reject a hydraulic explanation.
Australian operating conditions give this analysis a broad application beyond the test bench. Qantas and Virgin Australia operate fleets that include A320neo and 737 aircraft, while regional turboprops connect major centres and smaller communities. Maintenance and overhaul teams around Melbourne, Brisbane and Perth support mixed aircraft and engine programmes. CASA-approved design organisations, Part 91 certificate holders and approved maintenance programmes provide the regulatory framework, so any conclusion affecting type approval or continued airworthiness must remain traceable to approved data and procedures.
Validating causes with simulation and testing
A defensible diagnosis moves from correlation to validation. Simulation can predict expected gear-mesh, shaft and bearing orders under the measured speed and load cases, then compare those predictions with experimental amplitudes and phase. Modal testing of the gearbox and sensor fixtures can reveal whether a suspected order is amplified by a natural frequency. A targeted physical check can then look for the predicted rubbing mark, looseness, tooth surface condition or bearing damage.
Design-of-experiments and sensitivity methods are valuable here because gearbox performance can depend on several interacting variables. A sequence that varies speed, load, lubricant condition and selected geometry parameters can show whether a vibration feature is robust, sensitive to manufacture or driven mainly by test repeatability. Tolerance studies are particularly important because small dimensional differences can alter mesh frequency, contact stiffness and load sharing without producing an obvious change in the gearbox’s external appearance.
Physical testing remains essential because models simplify complex interactions. The study of oil-return channel geometry shows how a hidden flow feature can alter sump scavenging in aero engines; the same systems-level reasoning applies to lubrication-induced vibration. Instrumented runs should combine vibration, shaft speed, torque, oil pressure and flow, and surface temperature. Repeat the suspected flyover point, inspect the sump and oil-return path, and record the assembly condition so the final explanation reflects the machine as tested rather than an idealised model.
Practical recommendations for flyover investigations
A useful order-tracking investigation should follow a consistent sequence. These recommendations reduce the chance that a speed artefact, installation problem or loose operating assumption is reported as a gearbox defect.
- Define the flyover envelope before testing, including speed, load, oil conditions, duration and repeat count, so every run maps to the same operating event.
- Keep sensor positions, orientations and surface treatments repeatable across test days, and verify mounting stiffness with a controlled impact or reference measurement.
- Retain raw tachometer and acceleration channels with calibrated time alignment, preserving enough resolution for the highest tooth-mesh order of interest.
- Compare the suspect run with stable low-load, steady high-load and nearby healthy runs to separate speed-dependent structure from condition-specific growth.
- Require independent corroboration from phase, bearing-load calculations, inspection or modal evidence before assigning a mechanism or initiating an engineering change.
The sequence is intentionally practical because order tracking is strongest when designed into the test, rather than added after a waveform has already been collected. It also gives quality teams a common language for comparing design iterations, hardware changes and service observations.
Turning findings into an engineering action
Turn the evidence into an engineering action. The next step is to select one repeatable flyover condition, define the component orders expected from the design, and collect synchronised vibration and speed data against that baseline. Have the analysis and test plan reviewed by specialists in rotating machinery, lubrication and structural response. If the pattern persists outside the expected operating envelope, schedule inspection or a controlled component trial rather than relying on a spectrum alone.
The OPTIMIZE Project links experimental analysis with design-of-experiments methods, simulation, tolerance assessment and physical testing, giving teams a practical route from vibration evidence to improved efficiency, durability, weight and power density. Start with its published objectives and methodology, define the flyover envelope, and request a focused analysis-and-test plan that ends in an evidence-based design or maintenance decision.