Condition based maintenance: the triggers, the limits every condition task needs, planning from the trend, and turning a crossed limit into planned work
Condition based maintenance as a maintenance strategy: when it pays, the kinds of trigger from measured limits to inspection findings, the alert, action and minimum limits every condition task needs, planning the work from the trend, converting time based tasks, and making a crossed limit become a planned work order in the CMMS. With a drawing of limits and a trend projection.
Key takeaways
- Condition based maintenance does the work when the equipment's condition says it is needed, rather than on a fixed interval. It suits failure modes that give a measurable warning long enough to plan the work.
- Every condition task needs a measure, an interval shorter than the warning time, and limits: an alert limit to look closer, an action limit to do the work, and a limit not to run past.
- The trend matters more than the reading. Projecting the trend to the action limit lets the work be planned into a shutdown weeks ahead instead of raised as urgent.
- CBM only saves money when a crossed limit becomes a planned work order. Measurements that sit in a report and never reach the CMMS are condition monitoring without the maintenance.
- Converting a time based task to condition based starts from the failure mode it addresses, and keeps a longer time based backstop where part of the risk cannot be measured.
A site replaces the gearboxes on its conveyor drives every two years, whatever their condition. Most come out with years of life left. Then one fails eight months after its overhaul, because a bearing was damaged when it was refitted, and nothing between overhauls would have shown it. The time based task cost money on the gearboxes that were fine and did nothing for the one that was not.
This guide covers condition based maintenance as a maintenance strategy: what it is and when it pays, the kinds of trigger, the limits every condition task needs, planning from the trend, converting time based tasks, and making a crossed limit become planned work in the CMMS. The measurement techniques themselves are in the condition monitoring guide. The standards named are reference points.
Limits and the trend
A condition task measures something that changes as a failure develops. On its own a reading says only where the equipment is today. Three limits turn it into a decision. The alert limit says something has changed: look closer, measure more often, start planning. The action limit says do the work at the next planned opportunity. The minimum, or stop limit, says do not run past here.
The trend is what makes the work plannable. A wear rate taken from the last few readings, projected forward, says when the action limit will be reached. If a planned shutdown falls before that date, the work goes into it with weeks of notice for parts, crews and access. Without the trend, the action limit arrives as a surprise and the work becomes urgent, which is most of what CBM was meant to avoid.
A reading says where the equipment is. A trend says when the work is due. Plan from the trend.
What triggers the work
Condition based work can be triggered in several ways, and a good program uses all of them. Measured limits and trends cover the failure modes that can be measured. Inspection findings from operator rounds and maintenance routes cover what a person can see, hear and feel. Counters cover wear that tracks usage closely enough to replace measurement.
| Trigger | What it is | Examples |
|---|---|---|
| Measured limit | A measurement crosses a set limit | Liner or belt cover thickness, filter differential pressure, bearing temperature, oil particle count |
| Trend projection | The trend of measurements says when a limit will be reached | Wear rate projected to the action limit, to fit the work into a planned shutdown |
| Inspection finding | A person finds a defect on a route or a task | A cracked weld, a leaking seal, a damaged idler found on an inspection |
| Counter or usage | Run hours, starts, tonnes or cycles reach a limit | Liner change by tonnes crushed, overhaul by operating hours |
| Process signal | A process value shows the equipment's condition | Pump flow against head for impeller wear, motor current for a dragging conveyor |
The limits every condition task needs
A condition task without limits produces data and no decisions. Each measure on each asset needs its limits written into the task, with the response to each, so the person reading the result knows what to do without asking. The action limit has to leave enough remaining life for the work to be planned, scheduled and done, which ties it to the P-F interval and to the lead time for parts.
| Limit | What it means | The response |
|---|---|---|
| Alert | Something has changed. Look closer or measure more often | Increase the inspection frequency, confirm the reading, plan the work |
| Action | Do the work, at the next planned opportunity before failure | A work order raised and scheduled, parts and labour confirmed |
| Stop or minimum | Do not run past here | Equipment taken out of service, or run under a risk assessment until it can be |
When CBM pays, and when it does not
CBM is worth it for failure modes that develop gradually, show a measurable sign, give enough warning to plan, and cost enough when they fail to justify the measuring. Bearings, gears, liners, belts, filters, lubricants, insulation and many wear surfaces fit. Failure modes that give no warning, such as many electronic failures, or give a warning too short to act on, do not, and suit time based replacement, redesign, or a failure-finding task for hidden failures.
The choice is made failure mode by failure mode, which is what FMECA and RCM under SAE JA1011 are for. The failure patterns explain why so many time based tasks achieve little: most failure modes are not age related, so replacing on a schedule removes good parts without preventing failures.
Converting time based tasks
Most sites move to CBM by converting existing time based tasks rather than starting again. Each conversion follows the same reasoning, and the time based task is often kept at a longer interval as a backstop for the parts of the risk that cannot be measured.
| Step | The question | Example |
|---|---|---|
| Find the failure mode | Which failure mode does the time based task address? | Overhaul gearbox every 12 months addresses bearing and gear wear |
| Check it is detectable | Is there a measurable sign of the failure developing? | Bearing wear shows in vibration and oil debris weeks before failure |
| Check the warning time | Is the P-F interval long enough to plan and do the work? | Weeks of warning against a two week lead time for a planned outage |
| Set the task and limits | What is measured, how often, and what the alert and action limits are | Monthly vibration and quarterly oil sample, with alert and action limits per measure |
| Change the plan | Replace the time based task, or keep a longer interval as a backstop | Overhaul on condition, with a five year backstop for parts not monitored |
| Check it worked | Did failures fall and work move from breakdown to planned? | Compare breakdowns and planned work on the asset before and after |
CBM in the CMMS
The measurement is only half of condition based maintenance. The other half is the work: a crossed limit has to become a work order, planned and scheduled like any other, and the work has to reset the condition record. In the CMMS, that means condition tasks with their readings recorded against the asset, limits that raise a work request when crossed, and counters that trigger usage based work. A program run from spreadsheets and reports that never reach the CMMS is condition monitoring without the maintenance.
Readings also need to be kept as data rather than as attachments. A thickness or vibration value stored against the asset and date can be trended, projected and reported. The same value in a PDF report can only be read. Where measurements come from online systems or the historian, the limits can be checked automatically and the work request raised without anyone typing it, which is the software side of CBM that asset performance management builds on.
Where CBM programs go wrong
- Measurements taken without limits, so nobody knows when to act.
- Limits set so close to failure that every crossing becomes an emergency.
- Readings kept in reports instead of against the asset, so trends are never seen.
- Crossed limits that never become work orders.
- Time based tasks kept unchanged alongside the new condition tasks, so the cost goes up instead of down.
- CBM applied to failure modes that give no warning.
- Limits never reviewed against what was found when the equipment was opened.
Common questions
What is condition based maintenance?
Condition based maintenance, or CBM, is maintenance done when the measured or inspected condition of the equipment shows it is needed, rather than at a fixed interval or after failure. A condition task measures something that changes as a failure develops, such as wear, vibration, temperature or oil contamination, and work is triggered when the measurement crosses a set limit or its trend says it soon will. EN 13306 defines it as preventive maintenance that includes assessment of physical conditions.
What is the difference between condition based and predictive maintenance?
The terms overlap and are often used as one. Condition based maintenance triggers work when a measured condition crosses a limit. Predictive maintenance, in the narrower sense, uses the trend or a model of the condition to forecast when the limit will be reached, so the work can be planned ahead. In practice a good CBM program does both: limits to trigger work, and trends to plan it.
What are examples of condition based maintenance?
Changing a filter when its differential pressure reaches a set value, replacing crusher liners or conveyor belt when measured thickness reaches a limit, changing gearbox oil when analysis shows contamination or degradation, planning a bearing replacement when vibration trends past an alert level, and repairing a defect found on an inspection route. Each has a measure, a limit and a planned response.
When is condition based maintenance worth it?
When the failure mode gives a measurable warning, the warning time is long enough to plan and do the work, the cost of measuring is less than the cost of the failures and unnecessary time based work it avoids, and someone will act on the results. Failure modes with no warning, or with a warning of hours, suit time based replacement, redesign or a failure-finding task instead.
How do you set condition based maintenance limits?
Start from the manufacturer's limits and the relevant standards, then the failure mode: the action limit must leave enough remaining life to plan, schedule and do the work, and the alert limit enough to watch the trend and prepare. Limits are refined from the site's own history, especially where readings crossed a limit and the equipment was found in good or poor condition when opened.
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)
- EN 13306:2017 Maintenance, maintenance terminology (CEN)
- SAE JA1011 Evaluation criteria for reliability-centered maintenance (RCM) processes (SAE)
- IEC 60300-3-11 Dependability management: application guide, reliability centred maintenance (IEC)
Related case studies and tools
- Predictive maintenance and condition monitoring (case study)
Related reading
Working through something like this?
See how we approach these initiatives, or tell us what you are dealing with.