Understanding Vibration Analysis (VA)

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)

Vibration Analysis (VA) is the technical discipline of measuring, processing, and interpreting the vibration signatures of rotating machinery to reveal its mechanical condition. It is the working core of vibration diagnostics and a cornerstone of modern predictive maintenance. Every running machine radiates a small amount of vibration; vibration analysis treats that signal as a language, decoding it to detect faults and to identify their nature, location, and severity long before they become failures.

1. Definition: What is Vibration Analysis?

At its simplest, vibration analysis is the systematic study of how a machine moves while it runs. A healthy machine produces a stable, low-level vibration pattern; a developing fault changes that pattern in characteristic ways. By capturing the motion with a sensor and examining it in the right domain, an analyst can separate a benign signature from a warning sign and assign that warning to a specific cause — unbalance, misalignment, a failing bearing, or a gear defect.

Because it sees inside the machine without stopping or opening it, vibration analysis is fundamentally a non-intrusive technique. That is what makes it so valuable for condition monitoring: a single measurement, taken in seconds at operating speed, can confirm health or flag a problem on equipment that must stay in production.

2. Analysis vs. Monitoring: Diagnosing the Cause

The terms vibration monitoring and vibration analysis are often used together, but they answer two different questions. Vibration monitoring watches the overall level over time and detects that something has changed — it is a surveillance role, trending a single number across many machines and raising a flag when a reading drifts from its history. Analysis takes over from there to determine why.

Put plainly: monitoring detects the change; analysis diagnoses its cause. Where a monitoring system might report only that velocity at a bearing has doubled, the analyst opens the frequency spectrum and the time waveform to decide whether that rise is unbalance, a loosening foot, or the first stage of a bearing defect. The two activities are complementary halves of one programme — monitoring narrows the population of suspect machines to a handful, and analysis resolves each of those into a named, actionable fault.

3. The Core of Vibration Analysis: the FFT

While many techniques exist, modern vibration analysis is built upon the Fast Fourier Transform (FFT). The FFT is a highly efficient algorithm that takes a complex time waveform — a wiggly trace of displacement, velocity, or acceleration against time that is very difficult to interpret by eye — and deconstructs it into its individual frequency components.

The result is a spectrum: a graph that plots the amplitude of vibration against each specific frequency present in the signal. This spectrum is the analyst’s most powerful tool, because different mechanical and electrical faults appear as distinct patterns and peaks on it. The logic is direct: nearly every fault excites a frequency tied to a physical event in the machine, so unbalance shows at 1× running speed, misalignment adds energy at 2×, and rolling-element defects appear at their own bearing fault frequencies. Reading those peaks is the essence of spectral analysis.

4. Reading the Spectrum: Characteristic Fault Frequencies

The diagnostic power of vibration analysis comes from the fact that each common fault excites vibration at a predictable frequency, expressed as a multiple of running speed (1× = once per revolution). Recognising where energy appears in the spectrum is what turns a measurement into a diagnosis. The most important signatures are:

• Unbalance — dominant 1×. A heavy spot rotates with the shaft and produces a single, strong peak at exactly running speed, largely in the radial direction. A clean 1× peak that grows over time is the classic signature of unbalance.
• Misalignment — strong 2× (often with 1× and 3×). Misalignment between coupled shafts typically raises a prominent peak at twice running speed, frequently with significant axial vibration — a key distinction from unbalance, which is mainly radial.
• Mechanical looseness — a series of running-speed harmonics. Looseness generates a row of harmonics (1×, 2×, 3×, 4× and beyond), and sometimes half-order (0.5×) components, because the non-linear joint clips and distorts the waveform.
• Rolling-element bearing defects — non-synchronous bearing fault frequencies. A flaw on the outer race, inner race, rolling element, or cage produces vibration at a calculable, non-integer multiple of running speed — the bearing fault frequencies. Early defects are weak and ride on a high-frequency carrier, so they are best exposed by envelope (demodulation) analysis.
• Gears — gear-mesh frequency and sidebands. A gear pair vibrates at its gear-mesh frequency (number of teeth × shaft speed). A worn or cracked tooth modulates that peak, producing sidebands spaced at the faulty shaft’s running speed on either side of the mesh frequency.
• Electrical faults — twice line frequency. Problems in induction motors, such as an air-gap or rotor-bar issue, characteristically place energy at twice the electrical supply (line) frequency, distinguishing them from purely mechanical sources.

Because these relationships scale with speed, an analyst working on a variable-speed machine often switches to order analysis, which expresses the spectrum in orders (multiples of running speed) rather than absolute hertz so the fault peaks stay locked in place as the machine accelerates.

5. Key Techniques in Vibration Analysis

Vibration analysis is not a single activity but a collection of specialised techniques, each providing a different view of the machine’s health. A skilled analyst combines several rather than relying on one:

• Overall Level Monitoring: the simplest form of VA, where a single value — usually RMS velocity representing the total vibrational energy — is trended over time. A sharp increase signals a problem but does not reveal its cause; it is a tripwire, not a diagnosis.
• Spectral Analysis: detailed examination of the FFT spectrum to identify the frequencies of vibration and so diagnose the root cause, distinguishing unbalance from misalignment, looseness, or electrical issues.
• Time Waveform Analysis: direct analysis of the raw signal over time, particularly useful for identifying transient events, impacts, and certain non-linear behaviours that are not always clear in the spectrum.
• Phase Analysis: measurement of the relative timing between a vibration signal and a reference point such as a once-per-revolution pulse. Phase is indispensable for single-shot balancing, for confirming misalignment, and for telling apart faults that look identical in amplitude alone.
• Envelope Analysis: a signal-processing technique that demodulates the high-frequency carrier to expose low-energy, repetitive impacts characteristic of early-stage rolling-element bearing and gear faults.
• Modal Analysis and ODS Analysis: advanced methods used to understand the structural vibration characteristics of a machine or its foundation, chiefly to identify and solve resonance problems.
• Order Analysis: an adaptation of spectral analysis for machines that change speed, presenting the spectrum in terms of “orders” (multiples of running speed) instead of absolute frequency in Hz.

6. Time Waveform vs. Spectrum: Two Views of One Signal

The spectrum is powerful, but it is a derived view — the FFT assumes the signal repeats and averages energy into frequency bins, which can hide brief, irregular events. The raw time waveform preserves what the spectrum smooths away, and the two are read together rather than in isolation.

The waveform is the better view for short-lived impacts, rubs, and beating between two close frequencies, and for judging whether a signal is sinusoidal (typical of unbalance) or sharp and impulsive (typical of looseness or a bearing defect). A practical workflow is to use the spectrum to identify which frequencies carry energy, then return to the waveform to see how that energy is delivered — smoothly, in periodic spikes, or as random transients. Combining both domains is what separates a confident diagnosis from a guess based on a single peak.

7. The Vibration Analysis Workflow

A repeatable diagnosis follows a consistent sequence rather than a single reading:

• Gather machine context. Note running speed, bearing types, number of gear teeth, drive arrangement, and load. The fault frequencies above cannot be located in the spectrum without these basic facts.
• Mount the sensor correctly. An accelerometer fixed firmly to the bearing housing, at the same point each time, in the right measurement direction, is the foundation of repeatable data.
• Acquire overall level, spectrum, waveform and phase. Capture a few seconds at operating speed, with a tachometer reference where 1× phase is needed.
• Compare against history and limits. Set the reading against the machine’s trend and against recognised severity zones (see below). A change relative to the machine’s own baseline is often more revealing than an absolute limit.
• Diagnose, then act. Match the peaks to a fault, confirm with the waveform and phase, then recommend the correction — alignment, tightening, bearing replacement, or field balancing.

8. How the Measurement is Made in the Field

In practice an analyst attaches an accelerometer to the bearing housing, records a few seconds of data at operating speed, and lets the instrument compute the spectrum and overall level on the spot. For balancing work a second piece of information is essential — the phase reference — supplied by a tachometer pulse once per revolution. A portable two-channel instrument such as the Balanset-1A performs exactly this workflow: it measures amplitude and phase, builds the FFT spectrum, and supports on-site single- and two-plane balancing without disassembly. Because the reading is taken in the machine’s own bearings under real load, it captures the true running condition rather than a bench approximation.

9. Applications and Benefits

Vibration analysis is applied across virtually every industry that uses rotating equipment, including manufacturing, power generation, oil and gas, water utilities, pulp and paper, marine propulsion, and transport. Severity judgements are usually anchored to recognised limits — most commonly the ISO 20816 series (which superseded the older ISO 10816), defining acceptance zones from “good” to “unacceptable” by machine class.

The benefits of a well-implemented programme are substantial:

• Increased Uptime: detecting faults early lets maintenance be scheduled before a catastrophic failure, avoiding unplanned downtime.
• Enhanced Safety: prevents equipment failures that could endanger personnel.
• Reduced Maintenance Costs: eliminates unnecessary “preventive” work on healthy machines and limits repair costs by catching problems before extensive secondary damage occurs.
• Improved Asset Reliability: moves maintenance from a reactive or calendar-based model to a condition-based approach, maximising the life and performance of machinery.

10. Frequently Asked Questions

What is the difference between vibration analysis and vibration monitoring?
Monitoring trends the overall level to detect that a machine’s condition has changed across many machines at once; analysis then examines the spectrum, waveform and phase on a flagged machine to diagnose why. Monitoring narrows the field; analysis names the fault. See vibration monitoring.

What does the FFT spectrum show?
The FFT converts the raw time waveform into a spectrum of amplitude versus frequency. Because each fault excites a characteristic frequency — 1× for unbalance, 2× for misalignment, bearing fault frequencies for defective bearings — the position of the peaks identifies the cause.

Which frequency indicates unbalance versus misalignment?
Unbalance shows a dominant peak at 1× running speed, mostly radial. Misalignment typically raises a strong 2× peak and is usually accompanied by noticeable axial vibration, which is the practical way to tell the two apart.

What equipment is needed for vibration analysis?
At minimum, an accelerometer and an instrument capable of computing the FFT spectrum and overall level. For balancing and phase-based diagnosis you also need a tachometer reference; a two-channel vibration analyzer such as the Balanset-1A combines all of these in one portable unit.

How accurate is vibration analysis at predicting failure?
On most rotating machinery it reliably detects developing faults weeks or months ahead of failure, especially when readings are trended against a stable baseline. Accuracy depends on consistent sensor mounting, correct machine data, and combining spectrum, waveform and phase rather than relying on a single number.

Can vibration analysis be done without stopping the machine?
Yes. It is a non-intrusive technique performed at operating speed, which is precisely why it suits production equipment that cannot be taken offline for inspection.

---

← Back to Main Index