What Can Vibration Tell Engineers About a Rotating System?
Every rotating machine vibrates. A perfectly balanced, perfectly aligned machine running on perfect bearings still generates vibration — from the forces inherent in rotation itself. However, when something changes — a bearing begins to wear, a shaft goes out of balance, a gear tooth develops a crack — the vibration changes with it. Vibration analysis engineering reads those changes. It turns what most people experience as unwanted noise and movement into precise, actionable engineering information.
Furthermore, vibration analysis is not simply a maintenance tool. Used well, it is a design discipline — one that informs how rotating systems are engineered from the outset, not just how engineers monitor them in service.
Why Rotating Machines Vibrate
Vibration in rotating machinery has several distinct sources. Understanding each one is the starting point for effective analysis.
Mass imbalance occurs when the centre of mass of a rotating component does not coincide with its centre of rotation. Even a small imbalance generates a centrifugal force that rotates with the shaft. This produces vibration at exactly the shaft’s running speed. Furthermore, its magnitude increases with the square of rotational speed — making imbalance increasingly significant as speeds rise. It is one of the most common vibration sources in rotating machinery and one of the most straightforward to diagnose.
Misalignment generates vibration when two coupled shafts are not correctly aligned — whether angularly, in parallel offset or in a combination of both. It typically produces vibration at one and two times the running speed. Furthermore, it imposes additional loads on bearings, seals and couplings — accelerating wear and reducing service life significantly. As covered in our article on laser alignment systems, precision alignment is one of the most effective interventions available to improve rotating machinery reliability.
Bearing defects produce vibration at specific frequencies related to the bearing geometry and running speed. These include the ball pass frequencies of the inner and outer races, ball spin frequency and cage frequency. Each produces a characteristic signature that engineers identify in the vibration spectrum. Consequently, engineers can detect bearing defects long before they produce audible noise or measurable performance degradation — giving maintenance teams the opportunity to plan interventions before failure occurs.
Gear Mesh and Structural Sources
Gear mesh vibration occurs at the frequency at which teeth engage — the gear mesh frequency. This equals the number of teeth multiplied by the shaft rotational speed. Tooth wear, chipping or profile errors all alter the gear mesh signature in characteristic ways. Moreover, sidebands around the mesh frequency indicate modulation effects — often associated with eccentricity, shaft misalignment or localised tooth damage.
Natural Frequencies and Resonance
Every structure has natural frequencies — the frequencies at which it vibrates when disturbed. When a forcing frequency from a rotating machine coincides with a natural frequency of the supporting structure or the machine itself, resonance occurs. At resonance, vibration amplitude amplifies dramatically. Sometimes it rises by a factor of ten or more above the off-resonance response.
Resonance is one of the most dangerous conditions in rotating machinery. It turns minor imbalance or misalignment forces into severe structural vibration. Furthermore, it can occur at specific speeds within the operating range — making certain speeds unsafe to run through or dwell at. Therefore, identifying natural frequencies and understanding how they relate to operating speeds is a fundamental part of rotating machinery engineering.
Modal analysis identifies the natural frequencies and mode shapes of a structure. Engineers apply it during the design phase — using FEA to predict natural frequencies before hardware is built — and during commissioning, using measured vibration data to confirm that predictions match reality. Consequently, modal analysis bridges the gap between design analysis and real-world behaviour.
The Frequency Spectrum — Reading the Vibration Signature
Raw vibration data is a time-domain signal — amplitude varying with time. Converting that signal to the frequency domain using a Fast Fourier Transform produces a spectrum. This is a plot of vibration amplitude against frequency. The frequency spectrum is the primary tool of vibration analysis.
Each component of a rotating machine contributes energy at specific frequencies in the spectrum. Shaft running speed, its harmonics, bearing defect frequencies, gear mesh frequencies and structural natural frequencies all appear as peaks at predictable locations. An experienced analyst reads the spectrum the way a doctor reads an ECG. They identify what is normal for this machine, at this speed, under these conditions — and recognise what has changed.
Furthermore, trend analysis adds a time dimension. Engineers record spectra at regular intervals and compare them to reveal developing faults — a growing peak at a bearing defect frequency, an increasing harmonic of running speed or a new sideband around gear mesh frequency. As a result, vibration trend analysis transforms periodic monitoring into a genuine predictive maintenance tool.
Vibration in the Design Process
Vibration analysis adds most value when it informs design — not just when it diagnoses problems in machines already built. A test rig designed without considering its natural frequencies may resonate under test loads. This generates data corruption, structural fatigue or outright failure. A rotating machine designed without vibration analysis may produce unacceptable noise and vibration in service — despite meeting all static stress requirements.
At CNR, vibration analysis and FEA-based modal analysis form part of the mechanical design process for rotating systems and bespoke test rigs. CNR predicts natural frequencies before hardware is committed. Engineers check operating speed ranges against structural resonances. Furthermore, engineers define instrumentation strategies for vibration measurement as part of the test programme — ensuring that test data is clean, interpretable and genuinely useful for engineering decisions.
Vibration Measurement and Instrumentation
Effective vibration analysis depends on good measurement. Accelerometers are the most widely used vibration sensors. They measure acceleration directly and cover a wide frequency range. Velocity sensors measure the rate of change of displacement — well suited to the mid-frequency range most relevant to rotating machinery faults. Displacement sensors — eddy current probes in particular — measure shaft motion directly and are standard in large rotating machinery with journal bearings.
Sensor mounting matters as much as sensor selection. A poorly mounted accelerometer introduces measurement errors that corrupt the spectrum and produce misleading results. Furthermore, measurement location — typically on bearing housings, in the load direction — determines which fault signatures are most clearly captured. Therefore, vibration measurement is an engineering discipline in itself. It requires the same careful specification as any other instrumentation system.
What Vibration Analysis Tells Engineers
Vibration analysis is ultimately a communication medium. It translates the mechanical behaviour of a rotating system — something invisible and intangible — into quantified, interpretable engineering data. That data tells engineers where problems are developing, what is causing them and how urgently they need addressing.
At CNR, vibration analysis capability spans design-phase modal analysis, test rig vibration characterisation and rotating machinery assessment across aerospace, automotive, defence, energy and research programmes. Whether the challenge involves a new rotating system that must avoid resonance in service or an existing machine whose vibration signature needs interpreting, that engineering depth is available from day one.
Note: This article is for general information only Image Credits: AI


