Condition monitoring people actually act on
Plenty of operations collect condition data. Fewer act on it. What separates monitoring that changes decisions from monitoring that just fills a database.
Key takeaways
- Detection is cheap. The program lives or dies in the steps between an alert and a planned work order.
- Monitoring frequency and alarm limits are both P-F interval questions: detect early enough to plan, kit and schedule.
- Alerts get ignored for predictable reasons: false positives, no owner, no route to work, no failure mode behind the limit.
- Measure the program by conversion to planned work, achieved lead time and the trend in unplanned failures.
Detection is rarely the hard part any more. Sensors are cheap, vibration routes are established, and oil samples get taken on schedule. The hard part is the distance between a detected fault and a planned work order.
Programs that fail do not fail at the sensor. They fail in the handful of steps after it.
The interval that decides everything
The P-F interval is the time between the point a failure becomes detectable and the point it becomes a functional failure. Every design decision in a condition monitoring program is really a question about that interval.
Monitoring frequency has to be shorter than the P-F interval, or you will reliably detect faults too late to plan. Alarm limits have to trigger early enough in the interval to leave time to kit, schedule and execute. A perfect diagnosis delivered two days before failure has not bought you much.
Why alerts get ignored
- False positives. A channel that cries wolf gets muted, formally or informally, and then the real one goes unnoticed.
- No owner. An alert that lands in a shared inbox belongs to nobody.
- No route to work. If acting on an alert means arguing for schedule time, most alerts quietly become notes.
- No link to a failure mode. Alarms set on generic thresholds rather than the specific mode being managed produce noise that experienced people learn to discount.
Every ignored alert teaches the organisation to ignore the next one.
Choosing technique by failure mode
Technique selection follows the mode you are trying to catch, which is where condition monitoring and reliability analysis meet. Vibration analysis suits rolling element bearing and gear defects. Oil analysis catches wear debris and contamination trends. Thermography finds electrical faults and abnormal loading. Ultrasound picks up leaks and early bearing distress.
ISO 17359 is worth reading here as one well structured example: it sets out how to establish a program, from criticality and mode selection through to review, and points to the technique-specific standards underneath it. Many operations run equally good programs from their own internal standards.
Designing for action
Give every alert a named owner and a staged severity that maps to a response, not just a colour. Build the path from confirmed diagnosis to planned work order into the system rather than leaving it to goodwill, and make it auditable.
Then close the loop. Every alert that turns out to be real, and every one that does not, is information for tuning limits. Without that feedback the program cannot improve, and its credibility erodes at the speed of its false alarm rate.
Measuring whether it works
- Share of alerts that convert to planned work orders, which is the real output of the program.
- False positive rate per channel, tracked over time rather than argued about.
- Lead time achieved against the P-F interval for the mode being managed.
- Unplanned failures on monitored assets, which should be the number that falls.
Building a program that gets used
Rank by criticality
Target monitoring where failure consequence is highest, not where sensors are easiest to fit.
Select by failure mode
Match technique and interval to the specific modes you intend to catch.
Baseline and set limits
Per-asset baselines with staged alarms timed to fall early in the P-F interval.
Route alerts to work
Named owners and a built-in path from confirmed diagnosis to a planned work order.
Close the loop
Feed outcomes back into limits and technique choice so the program keeps earning trust.
Rank by criticality
Target monitoring where failure consequence is highest, not where sensors are easiest to fit.
Select by failure mode
Match technique and interval to the specific modes you intend to catch.
Baseline and set limits
Per-asset baselines with staged alarms timed to fall early in the P-F interval.
Route alerts to work
Named owners and a built-in path from confirmed diagnosis to a planned work order.
Close the loop
Feed outcomes back into limits and technique choice so the program keeps earning trust.
Common questions
Which condition monitoring technique should we start with?
The one matched to the failure modes on your critical assets. Vibration for bearings and gears, oil for wear and contamination, thermography for electrical and loading, ultrasound for leaks and early distress. The mode chooses the technique, not the catalogue.
How do we reduce false alarms?
Per-asset baselines instead of generic thresholds, staged severities, and a feedback loop that tunes limits from confirmed outcomes. False positive rate per channel is worth tracking as a first-class measure.
Continuous sensors or manual routes?
Decided by criticality and P-F interval. Short intervals and high consequence justify permanent sensing; longer intervals on lower consequence equipment are served well by routes. Most good programs run both.
Key terms
Plain-language definitions from our glossary for the concepts this article leans on.
Standards and further reading
- ISO 17359:2018 Condition monitoring and diagnostics of machines, general guidelines (ISO)
- ISO 13374-1 Condition monitoring data processing, communication and presentation (ISO)
- SAE JA1011 Evaluation criteria for RCM processes (SAE International)
- ISO 14224:2016 Collection and exchange of reliability and maintenance data for equipment (ISO)
- Best practices, metrics and guidelines for maintenance and reliability (SMRP)
- Operations and maintenance best practices guide (US DOE Federal Energy Management Program)
Related case studies and tools
- Predictive maintenance and condition monitoring (case study)
- Telemetry and monitoring for remote assets (case study)
- MTBF / MTTR Calculator (tool)
Related reading
Working through something like this?
See how we approach these initiatives, or tell us what you are dealing with.