Cepstrum Analysis in Vibration Diagnostics

Devices

Portable balancer & Vibration analyzer Balanset-1A

€1,975.00 + VAT (if applicable)

Parts

Vibration sensor

€90.00 + VAT (if applicable)

Parts

Optical Sensor (Laser Tachometer)

€124.00 + VAT (if applicable)

Devices

Balanset-4

€6,803.00 + VAT (if applicable)

Parts

Magnetic Stand Insize-60-kgf

€46.00 + VAT (if applicable)

Parts

Reflective tape

€10.00 + VAT (if applicable)

Devices

Dynamic balancer “Balanset-1A” OEM

€1,735.00 + VAT (if applicable)

Cepstrum analysis is an advanced signal-processing technique that reveals periodic structure within a frequency spectrum. The name “cepstrum” is an anagram of “spectrum,” and that wordplay captures its nature exactly: it is effectively the “spectrum of a spectrum.” It is computed by taking the logarithm of a frequency spectrum and then performing an inverse Fourier transform on the result, a step that collapses repeating patterns — families of harmonics or sidebands — into single, easily read peaks that can be hard to pick out in the raw spectrum. For complex machinery such as gearboxes it brings a clarity that ordinary FFT analysis often cannot.

In a cepstrum plot the x-axis is called quefrency (an anagram of frequency) and carries units of time. Peaks along this axis, called rahmonics, give the period — in seconds — of the repeating patterns present in the original spectrum. The deliberately rearranged vocabulary (cepstrum, quefrency, rahmonics) is a standing reminder that the technique works in a domain one transform removed from the familiar one.

1. Why Use Cepstrum Analysis?

A standard FFT spectrum is excellent for identifying individual frequency components, but it can become cluttered and hard to read when a fault generates many harmonics and sidebands at once. Cepstrum analysis cuts through that clutter by consolidating an entire family of evenly spaced frequencies into one clear peak. Its primary uses are:

• Detecting harmonic families: it identifies a fundamental frequency and its harmonics even when the fundamental itself is weak or absent in the spectrum.
• Identifying sideband families: it excels at finding sidebands that are low in amplitude and buried in noise, clearly showing their presence and measuring their spacing.
• Separating source and path effects: in some applications it helps separate the vibration source signal from the structural response of the machine that colours it.
• Echo detection: it can pick out echoes or reflections within a signal.

The key idea is one of conversion: a regular spacing in the frequency domain — say, sidebands every 30 Hz — becomes a single position in the quefrency domain (here, a rahmonic at 1/30 = 0.033 s). Many scattered peaks of varying height thus reduce to one measurable feature.

2. Key Applications in Machine Diagnostics

2.1 Gearbox Diagnostics

This is the most common and most powerful application. A damaged gear tooth modulates the gear mesh frequency (GMF), creating sidebands around the GMF peak spaced at the rotational speed of the faulty gear. In a gearbox with several shafts and gear pairs, the spectrum becomes a confusing mixture of different GMFs and their sidebands. The cepstrum cuts through that complexity:

• A peak at the quefrency corresponding to a gear’s rotational period (1 / RPM) is a clear indicator of a fault on that specific gear, pinpointing the offending shaft rather than just confirming “a gear problem.”
• The amplitude of that cepstrum peak can be trended to monitor how the gear wear progresses over time.

It complements rather than replaces direct spectral work: a Gear Mesh Frequency calculator tells you which mesh and sideband frequencies to expect, and the cepstrum then confirms which family is actually growing. Both feed into a fuller diagnosis of gear defects.

2.2 Rolling-Element Bearing Analysis

Bearing defects also generate sidebands. A defect on the inner race, for instance, creates sidebands spaced at shaft speed around the inner-race defect frequency (BPFI) and its harmonics. The cepstrum helps confirm these patterns, especially when they are not obvious in the spectrum. In practice it works alongside the predicted bearing fault frequencies — readily obtained from a Bearing Defect Frequency calculator — and is frequently paired with envelope analysis, which demodulates the high-frequency impacts that bearing faults excite.

2.3 Turbomachinery Analysis

In turbines and compressors, cepstrum can identify blade-pass frequency harmonics and help diagnose blade damage or aerodynamic problems, where many closely spaced blade-related harmonics would otherwise crowd the spectrum.

3. How to Interpret a Cepstrum Plot

A disciplined reading proceeds in four steps:

1. Calculate rotational periods first: before looking at the cepstrum, work out the time periods of the main rotating components. For a shaft at 1800 RPM (30 Hz) the period is 1/30 = 0.033 s. A Harmonic Frequency calculator speeds up the RPM-to-Hz conversions for every shaft in the train.
2. Look for peaks at known periods: examine the cepstrum for significant rahmonics that line up with those calculated periods, since a peak at a known period points straight to a known component.
3. Identify harmonic structure: look for peaks at integer multiples of a fundamental quefrency, which indicate strong harmonic families in the original spectrum.
4. Trend the amplitudes: monitor the height of the cepstrum peaks over time — a rising amplitude signals a worsening condition, making the cepstrum peak a compact health indicator for trending.

4. Where Cepstrum Fits in a Diagnostic Toolkit

Cepstrum analysis is powerful but demands experience to apply well; it is best treated as one specialised instrument within a broader programme of vibration diagnostics rather than a stand-alone answer. The usual workflow is to start with the spectrum and spectral analysis, reach for the cepstrum when dense families of sidebands or harmonics obscure the picture, and confirm bearing impacts with envelope methods. Most faults that cepstrum exposes — gear-tooth and bearing defects — are diagnostic findings rather than balancing problems, so the cepstrum sits in the analysis stage that precedes any corrective action. Where the underlying issue turns out to be unbalance at running speed, a portable analyser such as the Balanset-1A measures the 1× amplitude and phase needed to correct it on site, while the cepstrum stays focused on the gear and bearing faults it diagnoses best. For complex machinery, that combination delivers diagnostic clarity that spectrum analysis alone cannot match.

← Back to Main Index