Overall vibration values are efficient screening tools, but complex faults often require indicators that emphasise impacts, modulation, repetition and frequency-band behaviour. Advanced indices condense these patterns into values that can be trended over time.
Impulsiveness and resonant bands
Envelope kurtosis highlights repetitive impacts after the signal has been conditioned around a resonant band. Spectral kurtosis helps identify frequency regions containing non-stationary impulsive energy. These indicators are commonly relevant to rolling-element bearing and gear-tooth defects, but they are also sensitive to process impacts and poor mounting.
Modulation and periodic structure
Sideband density describes the concentration of modulation components around dominant frequencies. Harmonic consistency evaluates whether expected harmonic relationships remain stable. A cepstrum-based gear index can emphasise regularly spaced spectral components associated with gear mesh and modulation.
Cyclostationary behaviour
Cyclostationary analysis looks for statistical properties that repeat with the machine cycle. It can reveal periodic fault behaviour even when the raw signal appears noisy or the operating process introduces strong background energy.
Responsible interpretation
- Trend indicators against a healthy baseline under comparable speed and load.
- Correlate changes with overall vibration, harmonics, temperature and maintenance history.
- Use machine geometry and known frequencies to test a diagnosis.
- Treat firmware-dependent indices as decision support rather than an automatic root-cause verdict.
Technical basis: PyXis Smart 3-Axis Vibration Sensor datasheet TTN-DS-2026-002, issued 19 July 2026. Confirm the current hardware, firmware and controlled document revision before implementation.
