Vibration Analysis

The foundational layer of vibration diagnostics, examining the raw amplitude of oscillations over time. Crest factor and RMS velocity values reveal overall machine health, while impulsive spikes in the waveform often indicate localized defects like gear tooth breakage or bearing race spalling. Time-domain analysis excels at detecting transient events that frequency-domain methods may obscure, such as rub and looseness signatures that appear as truncated or asymmetrical waveforms.

The Fast Fourier Transform converts complex time-domain signals into the frequency domain, revealing the distinct spectral fingerprints of rotating components. Each fault type generates characteristic frequency peaks:

• 1x RPM: Unbalance
• 2x RPM: Misalignment or looseness
• Harmonics of running speed: Mechanical looseness
• Non-synchronous peaks: Bearing defect frequencies (BPFO, BPFI, BSF, FTF) Spectral analysis enables precise fault isolation by matching observed peaks to calculated bearing defect frequencies based on geometry and rotational speed.

A specialized signal processing technique that extracts low-amplitude, high-frequency bearing impact signatures from the dominant low-frequency machinery vibrations. The process involves band-pass filtering around a structural resonance frequency, followed by Hilbert transform demodulation to reveal the repetition rate of impacts. This method detects incipient bearing faults—such as sub-surface fatigue spalls—that produce energy at ultrasonic frequencies (20–40 kHz) invisible to standard FFT analysis.

Using orthogonal proximity probes (X-Y configuration) to track the centerline motion of a rotating shaft within its bearing clearance. The resulting orbit plot reveals shaft precession patterns critical for diagnosing:

• Oil whirl/whip: Sub-synchronous instability in fluid-film bearings
• Rub conditions: Flattened or figure-eight orbits indicating rotor-stator contact
• Unbalance: Circular orbits with 1x RPM frequency
• Misalignment: Elongated elliptical orbits Orbit analysis is essential for turbomachinery and large rotating equipment where fluid-film bearing dynamics dominate.

Phase measurement correlates vibration signals with a fixed reference trigger (typically a keyphasor or tachometer pulse) to determine the relative timing of oscillations across multiple measurement points. This enables:

• Balancing: Determining the angular location of unbalance mass
• Modal analysis: Visualizing structural resonance modes
• Operational Deflection Shapes (ODS): Animating how a machine structure physically deforms under operating conditions Phase data transforms vibration from a scalar magnitude into a vector quantity with both amplitude and direction, essential for root cause identification.

The cepstrum—the inverse FFT of a logarithmic spectrum—excels at detecting periodic patterns within complex gearbox spectra where multiple meshing frequencies and sidebands overlap. It identifies harmonic families and sideband spacing that reveal:

• Gear tooth wear: Increased sideband amplitudes around gear mesh frequency
• Multiple fault separation: Distinguishing between shaft, gear, and bearing faults in a single spectrum
• Echo detection: Identifying reflected vibration paths in complex mechanical assemblies Cepstrum analysis is the definitive tool for diagnosing compound gearbox faults where traditional spectral analysis becomes ambiguous.