Whenever machinery runs, some amount of vibration is expected. It is when that movement changes that it deserves attention. These changes can include a new frequency spike, rising vibration amplitude, or an unusual sound in a bearing. These are hints that a fault might be developing.
A practical way to find those changes is vibration analysis. This is how it works: Vibration data is collected and compared with operating speed, historical data, and equipment design. At that point, analysts can compare anomalies associated with mechanical failures like imbalance or misalignment. The irregularities can also be associated with bearing defects or looseness. Vibration analysis is an effective predictive technique that helps maximize equipment reliability and minimize unscheduled downtime.
What Is Vibration Analysis?
Vibration analysis is the process of measuring and interpreting machine vibrations to assess machine health. A vibration sensor captures movement from a bearing housing, motor, pump, gearbox, or other component. Vibration analyzers then display the recorded vibration signal as a time waveform, frequency spectrum, trend, or another diagnostic view.
Readings may be expressed as displacement, velocity, or acceleration. Displacement sensors can pick out shaft movement. Velocity sensors are commonly used to evaluate overall machinery vibration. Accelerometers cover wider frequency ranges and are well suited to bearing condition monitoring.
A complete diagnostic picture doesn’t come from a single tool, but from the collection of readings. Analysts consider vibration amplitude, vibration frequency, rotational speed, load, sensor location, and the machine’s construction before deciding whether a pattern indicates a fault.
How Does Vibration Analysis Work?
The gathering of data under known operating conditions is the first step in vibration analysis. Vibration sensors are placed at repeatable measurement points to record the machine’s operating state, its load, and speed.
For route-based inspections, portable vibration analysis tools are useful. Online condition monitoring systems automatically gather data. Critical assets benefit from continuous vibration monitoring.
Raw data first appears in the time domain, where amplitude is plotted against time. Frequency domain analysis separates that same vibration signal into its individual frequency components. Comparing these views with baseline and historical data helps analysts distinguish normal operation from developing mechanical faults.
Good results depend on consistent vibration measurement. Changing the sensor position, sampling rate, signal lengths, or operating condition can make two otherwise valid readings difficult to compare.
Common Vibration Analysis Methods
Vibration analysts rarely rely on one method. The right approach depends on the machine, its rotational speed, the suspected fault, and whether vibration occurs steadily, intermittently, or only during startup.
| Method | What it shows | Common uses |
| Frequency analysis | Amplitude at each vibration frequency | Imbalance, misalignment, looseness, gears |
| Time waveform analysis | Changes in the vibration signal over time | Impacts, rubbing, looseness, reciprocating machinery |
| Phase analysis | Timing relationship between vibration signals | Alignment, resonance, structural movement |
| Envelope analysis | Repetitive high-frequency impacts | Early bearing defects in rolling element bearings |
| Order analysis | Vibration relative to shaft speed | Variable-speed rotating machinery |
| Modal analysis | A structure’s natural frequencies and mode shapes | Resonance and structural integrity investigations |
Frequency (FFT) Analysis
Fast Fourier transform, or FFT, analysis converts a time waveform into a vibration spectrum. This mathematical frequency analysis operation shows how much vibration is present at different frequencies. Peaks can then be compared with the fundamental frequency of shaft rotation, its harmonics, gear-mesh frequencies, and calculated bearing fault frequencies.
The fast Fourier transform FFT is widely used because it makes repeating patterns easier to recognize. An FFT may show imbalance at once per revolution, misalignment at one or two times rotational speed, or multiple harmonics associated with looseness.
It isn’t possible to get fine frequency resolution from a short time capture. Recording duration directly dictates FFT bin width (𝚫f = 1/T); if the time window is too short, closely spaced spectral peaks blur together into a single hump. When comparing two different spectrum plots, technicians must always verify that both used identical sample lengths before trusting the peak separation.
“Power spectral density” also expresses signal power distributed across frequency. It is particularly useful when analyzing random vibration signals rather than strong, repeating tones.
Time Waveform Analysis
Time domain analysis shows the original vibration signal as it was measured. Instead of separating the reading into frequencies, it reveals impacts, modulation, clipping, rubbing, and changes that happen during each machine cycle.
A time waveform is useful when a frequency spectrum looks inconclusive. It can expose irregular impacts from damaged gear teeth, looseness, or rolling element bearings. The peak and RMS value also help describe severity, although neither should be interpreted without considering the machine and its operating condition.
Phase Analysis
Phase analysis measures the relative timing of vibration peaks between two sensors. Instead of just reading overall vibration severity, shifting phase angles reveal whether a shaft is bouncing from imbalance or angularly forcing a couple from misalignment. It also shows if a machine frame is rocking as a rigid body or twisting apart under dynamic load.
Analysts may also use phase data when forced vibration is exciting one of the machine’s natural frequencies. Finding the natural frequency—and understanding how the machine responds near it—is important when resonance threatens structural integrity or causes repeated component failures.
Envelope Analysis
When bearing damage is in its early stages, it issues brief impacts that can be disguised under lower-frequency machine vibrations. Envelope analysis works by filtering the high-frequency response and processing it so that the impacts are apparent.
Analysts compare the resulting vibration patterns with calculated fault frequencies for the inner race, outer race, rolling elements, and cage. Envelope analysis, bearing condition monitoring, and shock pulse monitoring can therefore provide early fault detection before a defect causes excessive vibration in the rest of the machine.
Mechanical Faults Detected by Vibration Analysis
A change in vibration won’t automatically tell you what caused it. However, the location and pattern of the change can get you closer. It might be:
- Rotor imbalance, shaft misalignment, mechanical looseness, bent shafts, resonance, rubbing, gear damage, and bearing defects
- Structural movement, weak foundations, soft foot, damaged supports, belt problems, cavitation, and faults in reciprocating machinery
For example, imbalance often produces a dominant peak at rotational speed. Looseness may create several harmonics, while bearing damage can produce high-frequency impacts around known bearing fault frequencies. Similar patterns can have different causes, so operating observations and inspection history still matter..
Applications of Vibration Analysis Across Industries
People mostly think of vibration analysis in terms of rotating machinery, motors, pumps, fans, compressors, turbines, and gearboxes. It’s not limited to that, though. Conveyors, machine tools, generators, cooling towers, and other equipment that relies on controlled mechanical movement get checked this way too.
Utilities lean on vibration monitoring to check generation and distribution equipment. Manufacturing plants use it on their production machinery. Then there’s oil and gas, mining, marine, food processing, and water facilities: all of them apply it to critical pumps and drive systems.
Vibration analysis has structural uses too. Modal analysis can pin down a component or frame’s natural frequencies, and order analysis helps make sense of speed-related behavior on equipment that runs at variable speeds. That same data can also point to conditions that are dragging down asset performance or driving up energy use.
Benefits of Vibration Analysis
The main benefit is additional time. If a developing problem is found early, maintenance teams can investigate it during normal operations and schedule work before equipment failure dictates the timing.
A well-run condition monitoring program can support:
- Earlier fault detection, so fewer unexpected shutdowns
- Better repair planning, reduced maintenance costs, and improved asset performance
- Longer equipment life and stronger equipment reliability
- Less secondary damage, lower energy consumption, and safer operation
Vibration analysis won’t eliminate maintenance costs, but it does cut down on unnecessary work and helps keep unplanned downtime in check. Rather than swapping out parts simply because the calendar says it’s time, teams get to work from what the machine’s actual condition is telling them.
Vibration Analysis vs Vibration Monitoring
Vibration monitoring, at its essence, is collecting and trending vibration measurements. A step further, vibration analysis is interpreting those measurements and figuring out what they mean for machine behavior.
A monitoring system might flag something like a rising RMS value or a jump in overall velocity. That tells you something’s changed, but not necessarily why. To get to the why, an analyst has to dig into the time waveform, the vibration spectrum, operating speed, process conditions, past readings, whatever it takes before actually recommending action.
Route-based monitoring is often sufficient for noncritical machinery. Continuous vibration monitoring is better suited to machines that operate around the clock, change condition quickly, or carry a high consequence of failure. In either case, analyzing vibration data is what turns readings into a maintenance decision.
Combining Vibration Analysis with Motion Amplification
Traditional sensors measure vibration at specific points. While excellent in its precision, this can miss the bigger picture, like how the machine, base, piping, and surrounding structure are actually moving relative to each other.
Motion amplification takes a different approach. It records the asset with a high-resolution camera and blows up small movements that you’d never catch with the naked eye. Watch the resulting video, and you can spot looseness, resonance, support movement, shaft or frame deflection, and connected components interacting in ways you wouldn’t expect.
The two methods aren’t really answering the same question. Vibration analysis tells you the frequencies and amplitudes showing up at a given sensor location. Motion amplification shows you where the movement’s actually happening across the whole scene you can see. Put them together, and troubleshooting gets a lot faster in cases where the vibration data says there’s a problem but the physical source isn’t obvious.
IRISS provides motion amplification and equipment reliability services for complex mechanical and structural investigations.
Best Practices for Effective Vibration Analysis
Getting measurements repeatable first is the foundation everything else builds on. Technicians should mark where the sensors go, write down the operating conditions, and not switch up the direction or mounting method between inspections on the same route. And if they can grab baseline data while the machine is still healthy, later readings will actually mean something by comparison.
Sampling settings need to match the equipment and whatever problem’s suspected. The measurement has to cover the right frequency ranges, but not at the cost of losing the resolution needed to tell nearby peaks apart. It’s also worth confirming machine speed, bearing geometry, gear ratios, and anything else that feeds into calculating expected fault frequencies.
More than any single number on its own, the real insight is in trends. A machine that’s always read a certain way might not need much attention even if the number looks high, while a smaller reading that has shifted quickly could be the one worth concern. Alarm limits, how often inspections take place, and how closely the data is reviewed shouldn’t follow one rule for every asset; they should reflect how critical the equipment actually is.
And don’t stop at the graph. Whatever the vibration data suggests, it’s worth checking against maintenance history, a visual inspection, process data, ultrasound, thermography, oil analysis, or motion amplification, whatever fits the situation.
Frequently Asked Questions
What is vibration analysis?
Vibration analysis comes down to measuring movement in machinery or structures and assessing what it means. Technicians pull patterns from that data, including things like imbalance, looseness, misalignment, bearing damage, resonance, and other mechanical faults, to get a read on how healthy the machine actually is.
How does vibration analysis work?
A sensor picks up mechanical movement and turns it into an electrical vibration signal. From there, analysis tools can display it in the time domain or convert it over to the frequency domain instead. From there, analysts check the results against rotational speed, expected fault frequencies, baseline readings, and whatever historical data’s on hand.
How do you perform vibration analysis?
Vibration analysis starts with choosing measurement points and collecting data under conditions that can be reproduced next time around. From there, it’s a matter of working through the overall vibration amplitude, the time waveform, and the frequency spectrum. If something unusual turns up, a peak or an impact that stands out, it gets checked against machine speed, bearing geometry, and prior readings before an actual diagnosis can be made.
What tools are used for vibration analysis?
Common vibration analysis tools include accelerometers, velocity sensors, displacement sensors, portable data collectors, vibration analyzers, and analysis software. Online condition monitoring systems can collect readings continuously. Tachometers are also useful when frequency or order analysis must be tied to shaft speed.
What mechanical faults can vibration analysis detect?
Vibration analysis can indicate imbalance, misalignment, looseness, resonance, bent shafts, rubbing, damaged gears, cavitation, and bearing defects. It may also uncover movement in foundations, supports, or piping. Additional testing is sometimes needed because more than one fault can produce a similar vibration pattern.
What are the main vibration analysis methods?
The main methods are fast Fourier transform analysis, time waveform analysis, phase analysis, and envelope analysis. Depending on the asset, vibration analysts may also use order analysis, power spectral density, shock pulse monitoring, operating deflection shape testing, or modal analysis.
What is the difference between vibration analysis and vibration monitoring?
Vibration monitoring means collecting the data and trending it. Analysis follows, actually explaining what those numbers mean. Monitoring might reveal that a level has risen, but it takes analysis, or digging into the signal and the operating context, to pin down the likely mechanical cause.
How often should vibration analysis be performed?
There’s no one-size-fits-all interval here. How often an operation inspects comes down to asset criticality, operating conditions, failure history, and how fast a fault could actually develop. Some machines are simply checked on scheduled routes, but critical equipment might need continuous vibration monitoring instead. Whatever the interval, there needs to be enough room to act before a developing fault turns into an actual failure.
Key Takeaways
- Vibration analysis turns machine movement into data that can reveal developing mechanical problems.
- FFT, time waveform, phase, and envelope analysis provide different views of the same machine behavior.
- Consistent sensor placement, operating information, baselines, and trending are essential for dependable results.
- Combining vibration data with motion amplification can make complex mechanical and structural movement easier to diagnose.
