Maintenance & ReliabilityFree worksheetPDF, Excel, Word exportManufacturing

Condition Monitoring Alarm Limit Worksheet (Mean plus Sigma)

Fixed limits from standards are a good ceiling but they are blunt. A statistical baseline for each measurement point lets you detect change early. This worksheet computes alert and alarm limits from baseline readings and checks them against your absolute limits.

Download:

Alert (mean + 2 sd)	Alarm (mean + 3 sd)	Alarm headroom to absolute limit
2.3	2.55	1.95
1.9	2.1	2.4
0.8	0.9	1.1
69	72.5	12.5
45.4	47.6	4.4

Highest alarm value

72.5

Smallest headroom

1.1

Collect baselines in normal, healthy, loaded operation. Where headroom is negative the statistical alarm sits above the absolute limit, so use the absolute limit instead. Mixing different operating states in one baseline inflates the standard deviation.

Automatic alarm limits from your data

Optional: discuss computing limits from historian data and raising ERP work requests.

Optional. The tool and every export work without this. No spam.

How to use it

Add a row for each measurement point. Record at least 20 to 30 healthy readings per point, then enter the mean and standard deviation from your spreadsheet or analysis software. The sheet calculates alert at mean plus two standard deviations and alarm at mean plus three. Enter the absolute limit from your standard or supplier and read the headroom. Export to Excel to load the thresholds into your monitoring system.

The method

For roughly normal data, about 97.7% of healthy readings fall below mean plus two sigma and 99.87% below mean plus three sigma, so crossings are rare in a healthy machine and meaningful when they occur. This is the logic of a Shewhart control chart applied to condition data. Vibration and acceleration data are often skewed, so a log transform or percentile approach can be better if the standard deviation is large relative to the mean.

Use one baseline per operating state (speed and load)
Recalculate after a repair, re-alignment or lubricant change
Cap statistical limits at the absolute limit

Common mistakes

Building the baseline from a period when the machine was already degrading hides the very change you want to catch. Another is applying one limit to a machine that runs at several speeds. Setting alarms too tight causes nuisance alerts that train the team to ignore them, so review the alert rate after the first month and adjust.

Baseline taken from unhealthy operation
Too few samples for a stable standard deviation
Ignoring trend slope and looking only at limits

Frequently Asked Questions

How many readings do I need for a baseline?

At least 20 to 30 readings in stable, healthy operation give a usable mean and standard deviation. Fewer readings make the limits unstable and likely to cause false alarms.

Why mean plus 2 and 3 sigma?

For normally distributed data they correspond to roughly 2.3% and 0.13% exceedance for healthy machines, which gives an early alert and a stronger alarm with a low false positive rate.

Should I use the standard limits instead?

Use both. Standards such as ISO 10816 / 20816 set absolute severity ceilings. Statistical limits detect change from the machine's own healthy normal, usually well below the ceiling.

What if my data is not normal?

Use percentiles, for example the 95th and 99th percentile of baseline readings, or analyze the log of the readings. Many vibration measures are skewed to the right.

Netray can compute baselines and rolling alarm limits from your historian or route data and raise a work request in your ERP when a point crosses a limit.

Talk to an Expert Explore More Free Tools

Related Tools

Maintenance & Reliability

More Maintenance & Reliability tools

See all

Back to All Tools