Vibration Analysis

How to Set Vibration Alarm Thresholds

A practical guide to setting advisory and danger alarm levels for overall vibration and fault-specific frequencies, using ISO 10816 zone boundaries and baseline-relative methods.

The Core Problem

An alarm set too low generates constant false alarms and trains operators to ignore them. An alarm set too high lets real faults go undetected until they are critical. The right alarm level depends on the machine's baseline, the consequence of failure, and how fast faults typically progress on that asset class.

Vibration Severity (mm/s RMS) Machine Class (Group 1: rigid mount, above 300 kW) 0 2.3 4.5 11.2 >11.2 Zone A Normal operation. No action. Log as baseline. (0 to 2.3 mm/s) Zone B Long-term operation acceptable. Monitor trend. (2.3 to 4.5 mm/s) Zone C Restricted operation. Plan maintenance. Increase monitoring frequency. (4.5 to 11.2 mm/s) Zone D Avoid. Imminent machine damage. Shut down or escalate. (above 11.2 mm/s)

ISO 10816-3 vibration severity zones. Zone boundaries vary by machine class and mounting. Values shown for Group 1 (large machines, rigid mount).

Two Methods for Setting Alarm Levels

Vibration alarm thresholds come from two different approaches. Neither is universally correct. Most programs use both, applied to different situations.

Absolute Level Alarms (ISO 10816)

ISO 10816 defines fixed velocity thresholds in mm/s RMS that apply across machine categories. These thresholds work well for initial commissioning, when you have no historical baseline, or for comparing machines of similar type across sites.

The weakness of absolute alarms is that they ignore the machine's individual characteristics. A pump that vibrates at 4.5 mm/s under normal conditions will trigger a Zone C alarm at ISO thresholds even when nothing is wrong. The alarm is technically correct and practically useless.

Baseline-Relative Alarms

Baseline-relative alarms anchor to the machine's own healthy vibration level. The alarm fires when the signal deviates significantly from established normal behavior, regardless of the absolute value.

This approach is essential for machines with high inherent vibration levels and for frequency-specific alarms on individual fault frequencies. A bearing outer race defect frequency at 2.4x its baseline is meaningful; 2.4 mm/s absolute is not.

Need to check severity against ISO zones? Use the Vibration Severity Calculator.

Open Calculator

ISO 10816-3 Zone Boundaries

ISO 10816-3 covers industrial machines with power above 15 kW and shaft heights between 160 mm and 315 mm. It defines two machine groups and four severity zones.

Group 1: Large machines (typically 300 kW to 50 MW) on rigid or heavy foundations.
Group 2: Medium machines (15 kW to 300 kW) on flexible or light foundations, including most pumps, compressors, and fans in industrial settings.

Zone Description Group 1 (mm/s RMS) Group 1 (in/s RMS) Group 2 (mm/s RMS) Group 2 (in/s RMS) Recommended Action
A New or recently overhauled machinery 0 to 2.3 0 to 0.09 0 to 3.5 0 to 0.14 No action. Log as baseline.
B Acceptable for unrestricted long-term operation 2.3 to 4.5 0.09 to 0.18 3.5 to 7.1 0.14 to 0.28 Monitor trend. No intervention needed unless rising.
C Unsatisfactory for continuous operation 4.5 to 11.2 0.18 to 0.44 7.1 to 18.0 0.28 to 0.71 Plan maintenance. Increase monitoring frequency. Do not ignore.
D Sufficient severity to damage the machine Above 11.2 Above 0.44 Above 18.0 Above 0.71 Immediate action. Shut down or escalate based on criticality.

Zone C is the advisory trigger. Zone D is the danger trigger. Most programs set their alert alarm at the A/B boundary and their danger alarm at the C/D boundary, then adjust based on operational experience.

Baseline-Relative Method for Frequency-Specific Alarms

Overall velocity captures total vibration energy but masks what is happening at specific fault frequencies. A bearing with a developing outer race defect can produce a strong spectral peak at its fault frequency while the overall level barely moves.

For fault-specific alarms, you need to track the amplitude at individual frequencies: bearing fault frequencies (BPFO, BPFI, BSF, FTF), gear mesh frequency, shaft running speed (1x), and its harmonics.

Starting Thresholds

The most widely used starting point for baseline-relative alarms:

These are starting points, not fixed rules. A machine with a six-month window between inspections may need a tighter advisory threshold so faults are caught before the next scheduled visit. A machine checked weekly can tolerate a slightly wider band before escalating.

Establishing a Reliable Baseline

A baseline measured once on a machine that was already degraded is worse than no baseline at all. Collect baseline data after a confirmed overhaul or bearing replacement, under representative load and speed conditions. Average several readings taken over two to four weeks. A single measurement is not a baseline.

How Fault Progression Rate Affects Threshold Choice

Not all faults develop at the same speed. Threshold strategy must account for how fast a failure mode typically moves from detectable to critical on each asset class.

Fast-Progressing Faults

Bearing spalling on a high-speed rotor, lubrication failure, or imbalance on a light-duty fan can progress from detectable to catastrophic in days or hours. For these:

Slow-Progressing Faults

Misalignment, soft foot, and early-stage looseness often develop over weeks or months. For these:

Practical Alarm Threshold Reference Table

The table below consolidates starting thresholds across common monitoring parameters. Tighten each based on asset criticality and confirmed fault history.

Alarm Type Metric Starting Advisory Starting Danger When to Tighten
Overall vibration, Group 1 Velocity RMS (mm/s) 2.3 mm/s (A/B boundary) 11.2 mm/s (C/D boundary) High-speed or high-criticality machines
Overall vibration, Group 2 Velocity RMS (mm/s) 3.5 mm/s (A/B boundary) 18.0 mm/s (C/D boundary) Machines near or above Zone B at baseline
Fault-specific frequency Spectral amplitude at fault frequency 1.5x baseline 2.5x baseline Fast-progressing faults; critical assets; short inspection windows
1x running speed (imbalance) Amplitude at 1x RPM 1.5x baseline 2.5x baseline Precision-balanced rotors; overhung loads
Bearing temperature Surface or RTD temperature Baseline +10°C Baseline +20°C Grease-lubricated bearings; sealed units
Motor winding temperature Winding RTD or thermistor Rated class limit minus 10°C Rated class limit Ambient temperature extremes; VFD-driven motors

The Risk of Fixed Thresholds Without Baselines

Applying ISO 10816 absolute thresholds uniformly across all machines in a plant is the most common alarm configuration mistake. It creates two problems simultaneously.

Machines that run inherently rough, such as reciprocating compressors, heavily loaded gearboxes, or poorly isolated pumps, will sit in Zone C or D permanently. Every reading triggers an alarm. Operators learn to ignore it. When an actual fault develops, the alarm is already firing and provides no new information.

Precision machines or lightly loaded motors that normally run in Zone A will trigger an alarm only when something is severely wrong. The Zone B to Zone C step represents a large change in vibration energy. By the time the alarm fires, the fault is already advanced.

The solution is not to find a middle ground on the absolute threshold. It is to use baseline-relative alarms for machines where absolute thresholds do not fit, and reserve ISO thresholds for commissioning checks and cross-site comparisons.

Handling Machines with High Inherent Vibration

Reciprocating compressors, agitators, and high-speed machines above 3,000 RPM often produce baseline vibration levels that sit inside ISO Zone C under perfectly healthy operating conditions. The ISO zones were not designed for these asset classes.

For these machines:

High-speed machines also require attention to measurement frequency range. Standard velocity sensors are calibrated for 10 Hz to 1,000 Hz. Faults at frequencies above 1,000 Hz, such as those common in high-speed spindles or small-bore bearings, require acceleration measurements and alarm thresholds in g rather than mm/s.

Refining Thresholds Over Time

Initial thresholds are educated guesses. The first confirmed fault event on each machine is calibration data.

When an alarm fires and a subsequent inspection confirms a fault, record the vibration level at the time of the alarm. If the level at confirmation was significantly above the alarm threshold, the advisory was set too high; tighten it. If the level at confirmation was close to the advisory threshold, the setting is reasonable.

When an alarm fires and no fault is found, record the conditions. Was the machine under unusual load? Was a process variable outside its normal range? These are candidates for nuisance alarms, addressable through hysteresis, confirmation delays, or process-adjusted baselines.

A program that never reviews alarm history will generate the same false alarms indefinitely. Review alarm events quarterly, cross-reference against confirmed fault findings, and adjust thresholds accordingly.

Temperature Alarms as a Secondary Signal

Bearing and motor temperature alarms complement vibration alarms but respond differently to faults. Temperature rise typically lags vibration change, sometimes by hours.

Starting alarm levels for temperature:

Absolute upper limits from bearing manufacturer specifications also apply. If a bearing is rated to 80°C continuous and the ambient is 30°C, the danger alarm cannot be set higher than 80°C regardless of what the baseline-relative method suggests.

When both vibration and temperature exceed advisory thresholds simultaneously, escalate urgency. Co-occurring alerts are more predictive than either signal alone.

Frequently Asked Questions

What are the ISO 10816 vibration alarm zones?

ISO 10816-3 defines four zones: Zone A (new machinery, no action), Zone B (normal long-term operation, monitor), Zone C (unsatisfactory for continuous operation, action required), and Zone D (sufficient severity to cause machine damage, immediate shutdown). Zone boundaries differ for Group 1 and Group 2 machines.

What is a good starting point for baseline-relative vibration alarms?

150% of the established baseline RMS velocity for an advisory alarm and 250% for a danger alarm. Tighten these for fast-progressing faults, high-criticality assets, or short inspection intervals. Refine using confirmed fault events as calibration data.

How do you set alarms for machines with inherently high vibration?

Use baseline-relative alarms anchored to the machine's own healthy baseline rather than ISO absolute thresholds. Set advisory at 150% of baseline and danger at 250%. This responds to deviation from normal behavior rather than an absolute level that doesn't fit the machine's operating conditions.

How do temperature alarms relate to vibration monitoring?

Temperature alarms complement vibration alarms for bearing and motor monitoring. A common starting point is baseline +10°C for advisory and baseline +20°C for danger. Temperature rise often lags vibration changes. If both signals exceed advisory levels simultaneously, escalate urgency.

How do you avoid alarm fatigue in vibration monitoring?

Alarm fatigue comes from thresholds set too low or applied globally without baselines. Reduce it by using baseline-relative alarms calibrated to each machine, applying hysteresis so alarms clear only when vibration drops below a secondary threshold, requiring the level to hold above threshold for multiple consecutive readings before alerting, and reviewing false alarm events quarterly to adjust specific thresholds.

Continuous Vibration Monitoring with Automated Alarms

Tractian sensors collect data continuously and apply adaptive alarm thresholds per machine, removing the guesswork from threshold configuration.

See Mechanical Monitoring
The alarm thresholds and zone boundaries on this page are starting points based on ISO 10816-3 and common industry practice. Apply engineering judgment based on your specific machines, operating conditions, and failure history. Consult the relevant ISO standard and equipment manufacturer specifications before finalizing alarm settings for safety-critical applications.