What Is RMS (Root Mean Square) in Vibration Analysis?
RMS — Root Mean Square — is the industry-standard statistical method for quantifying the energy content and destructive capability of mechanical vibration in rotating machinery. The calculation squares every sample value of a vibration signal, takes the mean of those squared values, then takes the square root, yielding a single number that represents the signal’s true energy equivalent and correlates directly with component fatigue and wear. In practical vibration analysis, RMS velocity in mm/s is the headline figure you compare against international severity limits — which is exactly why it is the first number most engineers look at on a machine.
1. What Is RMS Vibration Analysis and Why Does It Matter?
RMS vibration analysis is the standard way to turn a complex, constantly changing vibration waveform into one physically meaningful number. RMS squares every sample value of the signal, calculates the mean of those squared values, then takes the square root, producing a value that represents the signal’s true energy equivalent and correlates directly with component fatigue and wear.
Mathematically, the RMS calculation follows three discrete steps. First, every instantaneous sample value of the vibration waveform is squared — this eliminates negative values and weights larger amplitudes more heavily. Second, the arithmetic mean of all squared values is computed over the measurement period. Third, the square root of that mean is taken. The result is analogous to the DC value that would deliver the same heating or power dissipation — making RMS the most physically meaningful single-number descriptor of vibration severity available to maintenance engineers.
For a discrete signal of N samples x1, x2 … xN, the RMS value is:
xRMS = √[ ( x1² + x2² + … + xN² ) / N ]
For a continuous waveform x(t) over a period T, it is the square root of the mean of x(t)² integrated over T — the “root of the mean of the squares,” which is where the name comes from.
This energy-based interpretation is what separates RMS from simpler metrics such as Peak or rectified Average. Per ISO 20816-1, RMS velocity expressed in mm/s is the primary parameter for evaluating machinery vibration severity across virtually all classes of rotating equipment. Facilities that adopt RMS-based trending as part of a structured predictive maintenance program typically report a 25–30% reduction in unplanned downtime, according to a 2022 Deloitte study on predictive-maintenance ROI.
2. Why Is RMS the Preferred Vibration Measurement Over Peak or Average?
RMS vibration analysis is preferred because it is the only single-number metric that directly represents the total energy content of a vibration signal, making it the most reliable indicator of a machine’s continuous running condition and the basis for all major international severity standards — including the modern ISO 20816 series and the legacy ISO 10816 it replaced.
There are four principal reasons why condition-monitoring professionals rely on RMS over alternative amplitude metrics:
- Direct energy correlation. The destructive power of vibration is proportional to energy, not to instantaneous peaks. RMS captures total energy across the entire waveform, which correlates with bearing fatigue-life calculations (per ISO 281) and with structural fatigue curves.
- Whole-waveform consideration. A Peak measurement captures only a single maximum point. RMS processes every sample in the measurement window, producing a stable, repeatable value with typical test–retest variability below ±2% under consistent operating conditions.
- Robustness against random impacts. A transient shock — such as debris passing through a pump — can inflate a Peak reading by 300% or more without reflecting any real change in machine health. The RMS value, being a statistical average, absorbs such events with minimal distortion, reducing false-alarm rates by an estimated 40–60% compared with Peak-based alarming.
- International standard compliance. ISO 20816-1 through 20816-9, API 670, and VDI 2056 all define alarm and trip thresholds in RMS velocity (mm/s or in/s). Using RMS allows direct benchmarking against these globally accepted limits.
3. The Difference Between RMS, Peak, and Peak-to-Peak Vibration Values
For a pure sine wave, RMS equals Peak divided by √2 (approximately 0.707 × Peak), and Peak-to-Peak equals 2 × Peak. However, real-world machinery vibration is never a pure sine wave; the ratio of Peak to RMS — called the Crest Factor — varies with signal complexity and serves as an independent diagnostic indicator of impulsive defects such as bearing spalling. A clean sinusoid carries its energy evenly, so its peaks stay close to its RMS; a signal full of sharp impacts spikes far above its RMS, and that excess is exactly what the Crest Factor measures.
| Metric | Definition | Relation to Sine-Wave Peak | Best Use Case | Standard Reference |
|---|---|---|---|---|
| RMS | Square root of the mean of squared values | 0.707 × Peak | Overall machine-health trending, severity classification | ISO 20816 (formerly ISO 10816) |
| Peak (0-to-Peak) | Maximum absolute amplitude | 1.0 × Peak | Short-duration impact detection, clearance checks | API 670 (shaft displacement) |
| Peak-to-Peak | Total swing from negative to positive maximum | 2.0 × Peak | Shaft displacement, orbit analysis | API 670, ISO 7919 |
| Average (rectified) | Mean of the rectified signal | 0.637 × Peak | Legacy instruments only — rarely used today | Historical / obsolete |
The choice of metric is not academic: alarm limits, trend charts, and acceptance reports are only comparable when everyone uses the same descriptor. A reading quoted as “5 mm/s” means very different things as RMS, Peak, or Peak-to-Peak, so always state which one you mean. For a side-by-side treatment of all three descriptors see the glossary entry on vibration amplitude, and when you need to move between them quickly the Vibration Unit Converter handles the mm/s ↔ µm ↔ g conversions for you.
3.1 What Is the Crest Factor and Why Does It Matter?
The Crest Factor is the ratio of Peak amplitude to RMS amplitude. For a pure sine wave, the Crest Factor is exactly √2 ≈ 1.414. A Crest Factor exceeding 3.0 in a vibration measurement strongly suggests the presence of repetitive impacts — a hallmark of early-stage rolling-element bearing defects, gear-tooth damage, or cavitation. Monitoring the Crest Factor alongside RMS adds a powerful diagnostic dimension:
- Rising Crest Factor with stable RMS indicates emerging localised damage — sharp impacts are appearing on top of an otherwise unchanged energy level (classic early spalling).
- Rising RMS with stable Crest Factor indicates distributed or progressing wear — the whole energy level is climbing while the waveform shape stays the same.
4. Should I Use RMS Velocity, Acceleration, or Displacement?
For general-purpose machine condition monitoring in the 10 Hz–1,000 Hz frequency range — which covers the vast majority of rotating-machinery faults — RMS velocity in mm/s is the industry-standard parameter, as specified by ISO 20816. RMS acceleration is preferred above 1,000 Hz (for example, high-frequency bearing-defect detection), while RMS displacement is used below 10 Hz for slow-speed machinery.
| Parameter | Optimal Frequency Range | Unit (SI / Imperial) | Typical Application |
|---|---|---|---|
| RMS Displacement | < 10 Hz | µm / mils | Slow-speed machines (< 600 RPM), shaft proximity probes |
| RMS Velocity | 10 Hz – 1,000 Hz | mm/s / in/s | General machine health, ISO 20816 severity, most rotating equipment |
| RMS Acceleration | > 1,000 Hz | g / m/s² | High-frequency bearing enveloping, gearbox analysis, ultrasonic detection |
The reason RMS velocity dominates the mid-frequency band is physical: velocity is proportional to vibration energy across a wide frequency range, giving roughly equal weight to low- and high-frequency fault components. Displacement over-emphasises low frequencies, while acceleration over-emphasises high frequencies. A robust strategy is to trend RMS velocity for overall severity and add high-frequency techniques — such as envelope analysis or ultrasonic measurement above 20 kHz — to catch the earliest stages of bearing degradation, often 3–6 months before changes appear in conventional vibration spectra. If you already work in one unit and need another, the mm/s-to-m/s² acceleration converter bridges velocity and acceleration directly.
5. How Is RMS Applied in Predictive Maintenance Programs?
RMS vibration analysis forms the backbone of condition-monitoring and predictive-maintenance (PdM) programs by providing trendable, standards-referenced severity values that enable condition-based maintenance decisions. When RMS velocity readings are collected at regular intervals and compared against ISO 20816 alarm thresholds, maintenance teams can detect deterioration weeks or months before failure and schedule repairs during planned outages.
A typical implementation follows these steps:
- Baseline establishment. Collect RMS velocity measurements on all monitored bearings and housings immediately after commissioning or after a known-good overhaul, and store them as the baseline. Record operating speed, load, and temperature.
- Threshold assignment. Apply ISO 20816 vibration-severity zones (A through D) appropriate to the machine class, or establish statistical baselines using 3× the baseline RMS value as an Alert threshold and 6× as a Danger threshold.
- Trend monitoring. Collect measurements on a route-based schedule — typically every 28–30 days for critical assets, quarterly for non-critical. Plot RMS values over time.
- Alarm response. When a reading exceeds the Alert threshold, increase measurement frequency and perform detailed spectral analysis to identify the fault type.
- Root-cause analysis. Use spectral data, phase analysis, and complementary technologies (ultrasound, thermography, oil analysis) to confirm the fault — distinguishing unbalance, misalignment, and looseness — and to estimate remaining useful life.
According to a 2023 McKinsey report on industrial analytics, organisations with mature PdM programs built on standardised vibration metrics such as RMS velocity achieve 10–20% reduction in overall maintenance costs and 50–70% fewer unexpected breakdowns.
5.1 Measuring RMS Velocity in the Field
On assembled machines, the overall RMS velocity is read directly from a sensor mounted on the bearing housing, and the same instrument that reports severity can usually also balance the rotor that is causing the vibration. A portable two-channel analyser such as the Balanset-1A measures RMS velocity on each bearing, displays the vibration spectrum so you can see which frequency is contributing the energy, and reports the broadband value you compare against the ISO 20816 zones. Because it works in the machine’s own bearings at operating speed — across an FFT range from roughly 5 Hz up to 1,000 Hz — it captures the true running condition, then lets you correct an unbalance on the spot and confirm the RMS velocity has dropped back into Zone A or B. That closes the loop from “the number is too high” to “the number is fixed” without a trip to a balancing machine.
6. ISO 20816 Vibration-Severity Zones for RMS Velocity
ISO 20816 — the modern standard that superseded ISO 10816 and the long-withdrawn ISO 2372 — classifies machinery vibration severity into four zones: A (good), B (acceptable), C (alert), and D (danger), based on broadband RMS velocity in mm/s. The exact thresholds depend on machine class, foundation type, and power rating, but the following table shows representative values for Group 1 large machines (Class III/IV) as a practical reference.
| Zone | Condition | RMS Velocity (mm/s) — Rigid Foundation | RMS Velocity (mm/s) — Flexible Foundation | Recommended Action |
|---|---|---|---|---|
| A | Dobro | 0 – 2.3 | 0 – 3.5 | Normal operation |
| B | Acceptable | 2.3 – 4.5 | 3.5 – 7.1 | Acceptable for long-term operation |
| C | Alert | 4.5 – 7.1 | 7.1 – 11.2 | Restricted operation; plan maintenance |
| D | Danger | > 7.1 | > 11.2 | Immediate shutdown risk; urgent action |
Zone boundaries are evaluated on the highest broadband RMS velocity measured at any monitoring point, so a single bad bearing is enough to push a machine into a worse zone. To assign a measured value to its zone for a specific machine group and mounting, the ISO 20816-1 Zone Evaluation tool applies the correct boundaries automatically, and the ISO 10816 / 20816 severity chart gives a quick at-a-glance reference.
7. Worked Example: How Do You Calculate RMS From a Vibration Signal?
To calculate the RMS value of a discrete vibration signal, square each sample, compute the mean of those squares, and take the square root. For example, given five instantaneous velocity readings of 3.0, −4.0, 2.5, −1.0, and 5.0 mm/s, the RMS velocity is approximately 3.39 mm/s — which would place this machine in Zone B (Acceptable) per ISO 20816 on a rigid foundation.
Step-by-step calculation:
- Square each sample: 9.0, 16.0, 6.25, 1.0, 25.0
- Calculate the mean of squares: (9.0 + 16.0 + 6.25 + 1.0 + 25.0) / 5 = 57.25 / 5 = 11.45
- Take the square root: √11.45 ≈ 3.385 mm/s RMS
Notice that the simple arithmetic mean of the five raw readings is just (3.0 − 4.0 + 2.5 − 1.0 + 5.0) / 5 = 1.1 mm/s — far lower, because the negative swings cancel the positive ones. Squaring first is precisely what prevents that cancellation and makes RMS represent real energy. In practice, portable data collectors and online monitoring systems perform this calculation automatically on thousands of samples per second, delivering RMS values with high statistical confidence. When the input is a frequency spectrum rather than a raw time waveform, the overall RMS is found by combining the RMS of each spectral line in quadrature (the root-sum-of-squares) — the job done by the Overall Vibration Level calculator (RMS from spectrum).
8. The Most Common Mistakes in RMS Vibration Measurement
The most common mistakes in RMS vibration analysis are sensor-mounting errors, incorrect frequency-range selection, inadequate averaging time, and comparing RMS values measured under different operating conditions. Any of these errors can produce misleading trends that either mask real faults or trigger false alarms, undermining confidence in the predictive-maintenance program.
- Poor sensor mounting. A loosely attached accelerometer can attenuate high-frequency signals by 50% or more above 2 kHz, producing artificially low RMS acceleration readings. Always use stud-mounted or high-quality magnetic mounts on clean, flat surfaces — see the guidance on correct sensor mounting.
- Wrong frequency band. Measuring RMS velocity in a 2 Hz–100 Hz band when the standard calls for 10 Hz–1,000 Hz produces non-comparable results. Always verify that the band-pass filter settings match the applicable standard.
- Insufficient averaging time. RMS values computed from very short time records (< 1 second) are statistically unstable. For machines running at 1,500 RPM (25 Hz), a minimum of 4–8 complete shaft revolutions — approximately 0.16–0.32 seconds — is needed, though 1–2 seconds is preferred for higher confidence.
- Inconsistent operating conditions. RMS vibration varies with speed and load. Comparing a measurement taken at 80% load with a baseline at 100% load can show a false improvement. Always document and normalise for operating conditions.
- Confusing overall RMS with narrowband RMS. Overall (broadband) RMS includes energy from all frequencies, while narrowband RMS isolates a specific frequency range. Both are useful, but they must not be confused when trending or alarming.
9. Frequently Asked Questions About RMS Vibration Analysis
9.1 What does RMS stand for in vibration analysis?
RMS stands for Root Mean Square. It is a statistical calculation that produces a single value representing the effective energy of a vibration signal by squaring all samples, averaging those squares, and taking the square root. RMS is the most widely used amplitude metric in machinery vibration analysis because it directly correlates with the signal’s energy content and destructive potential.
9.2 How do you convert RMS to Peak vibration?
For a pure sine wave only, Peak = RMS × √2 ≈ RMS × 1.414. For real-world machinery signals containing multiple frequencies and impacts, this simple conversion is inaccurate. The actual ratio (the Crest Factor) depends on signal complexity and can range from 1.4 to above 5.0. Always measure both values directly rather than converting — and never confuse a calculated peak with a measured true peak.
9.3 What is a good RMS vibration level for a motor?
Per ISO 20816, an RMS velocity below 2.3 mm/s (0.09 in/s) on a rigidly mounted large industrial motor places it in Zone A (good condition). Values between 2.3 and 4.5 mm/s are acceptable for long-term operation (Zone B). Above 4.5 mm/s, remedial action should be planned. Specific thresholds vary by machine class and mounting type.
9.4 Why is RMS velocity preferred over RMS acceleration for general monitoring?
RMS velocity gives approximately equal weight to fault frequencies across the 10 Hz–1,000 Hz range, which covers most common machinery defects — including unbalance, misalignment, looseness, and bearing wear. RMS acceleration overweights high frequencies, which can mask low-frequency faults. ISO 20816 specifies RMS velocity as the primary severity metric for this reason.
9.5 Can RMS vibration analysis detect bearing faults?
Yes, but with limitations. Overall RMS velocity detects moderate-to-advanced bearing damage that raises broadband energy. Early-stage bearing defects — such as micro-pitting — produce high-frequency impulsive signals that may not significantly change the overall RMS. For early detection, combine RMS velocity trending with high-frequency techniques such as enveloping (demodulation), the shock-pulse method, or ultrasonic monitoring, and watch the Crest Factor for the first sign of impacts.
9.6 What is the difference between ISO 10816 and ISO 20816?
ISO 20816 is the modern replacement for ISO 10816. Both define vibration-severity zones based on RMS velocity. The key difference is that ISO 20816 consolidates and updates the several parts of the older standard, incorporates lessons from more than 20 years of field experience, and introduces refined zone boundaries for certain machine types. ISO 20816-1:2016 replaced ISO 10816-1:1995, and the older ISO 2372 was withdrawn long before that; migration across all parts of the family is ongoing.
9.7 How often should RMS vibration measurements be taken?
For critical rotating assets, industry best practice is monthly route-based RMS measurements as a minimum. High-criticality machines benefit from continuous online monitoring with measurement intervals of seconds to minutes. Non-critical equipment can be measured quarterly. The measurement frequency should increase immediately whenever a reading exceeds the Alert threshold or when operating conditions change significantly.
9.8 What tools are needed for RMS vibration analysis?
At minimum, you need a calibrated accelerometer, a data collector or vibration analyser capable of computing RMS in the correct frequency band, and trending software. A portable two-channel instrument that combines RMS-velocity measurement with single- and two-plane balancing — such as the Balanset-1A — lets the same engineer both assess severity against ISO 20816 and correct the underlying unbalance, which is why field teams favour an all-in-one analyser over separate measure-only and balance-only devices.
