What is condition-based maintenance?

Condition-based maintenance triggers work when a measured indicator of equipment health, vibration, temperature, pressure, oil condition, current, crosses a defined threshold. Maintenance happens because the machine's condition says it should, not because a calendar interval came due.

What is the difference between condition-based and predictive maintenance?

CBM reacts to condition: a threshold is crossed, work is triggered. Predictive maintenance adds forecasting: it trends the condition data to estimate when failure will occur. CBM answers whether the machine is degrading now; PdM answers when it will fail. PdM costs more in tooling and skill.

Is condition-based maintenance better than preventive maintenance?

For critical assets with measurable degradation, usually yes, you stop over-maintaining healthy machines and stop being surprised between intervals. For assets where degradation is not measurable, or where failure is cheap, a fixed preventive interval or deliberate run-to-failure is the better economic answer. It is a per-asset decision.

Does condition-based maintenance require sensors?

No. Permanent sensors are one way to get condition data, but a structured inspection route, a technician or operator measuring vibration, temperature, or pressure with handheld tools against defined limits, is legitimate CBM. Sensors buy frequency and consistency; the strategy is the threshold, not the hardware.

What parameters are monitored in CBM?

The common ones are vibration on rotating equipment, temperature on electrical and mechanical components, oil condition and particle counts on lubricated systems, pressure and differential pressure on fluid systems and filters, and current draw on motors. Pick the parameter per failure mode, not per what a sensor kit happens to include.

How do you set condition monitoring thresholds?

Start from a healthy-state baseline measured under normal load, then set two limits: an alert level that triggers inspection and an action level that triggers a work order. Use manufacturer specs, published severity standards for the parameter, and your own failure history, and expect to tune the limits after the first quarter of false alarms or missed catches.

Condition-Based Maintenance: Between Preventive and Predictive

Act when the machine crosses a threshold, no calendar guessing, no forecasting models. Here is where CBM fits and how to set it up.

Condition-based maintenance (CBM) is a strategy that triggers maintenance when a measured indicator of equipment health, vibration, temperature, pressure, oil condition, current draw, crosses a defined threshold. Work happens because the machine's condition says it should, not because a calendar interval elapsed. CBM measures and reacts; it does not forecast failure dates.

That last distinction is why CBM deserves its own name. It is the practical middle rung between preventive maintenance, which guesses with intervals, and predictive maintenance, which invests in forecasting. For many plants and many assets, the middle rung is the right place to stop.

How does CBM differ from preventive and predictive maintenance?

Preventive maintenance acts on a schedule, CBM acts on a threshold, and predictive maintenance acts on a forecast. All three aim to intercept failure before it happens; they differ in what triggers the work and what data they demand.

	Preventive (PM)	Condition-based (CBM)	Predictive (PdM)
Trigger	Fixed interval: calendar time or usage (hours, cycles)	Measured parameter crosses a set threshold	Trend analysis forecasts a failure window
Question it answers	Is it due?	Is it degrading now?	When will it fail?
Data required	Asset list and a schedule	Condition readings plus a defined limit per parameter	Condition history, baselines, trending, analysis capability
Typical tools	CMMS calendar, meter readings	Sensors or inspection routes, alarm limits	Vibration spectra, oil labs, thermography, models
Waste profile	Over-maintains healthy machines; misses early failures	Little wasted work; warning window can be short	Least wasted work; highest setup and skill cost
Cost to run	Low	Moderate	Highest

Two honest notes on the boundaries. First, the industry blurs them: a monthly vibration route with alarm limits is CBM; the same route with spectrum trending and remaining-life estimates is PdM. Second, CBM does not require sensors everywhere, a structured operator inspection (look, listen, feel, measure) with defined pass/fail limits is legitimate CBM, and it is how TPM programs get condition data at zero instrument cost.

Which assets should get CBM? A decision flow

Assign strategies per asset, by failure consequence and by whether condition is measurable at all. The flow below is the short version of the logic in our equipment reliability guide.

A per-asset strategy decision flow. CBM is the answer for critical assets with measurable degradation where full predictive investment is not yet justified.

The pattern that falls out: run-to-failure for cheap redundant assets, a PM schedule where degradation is not measurable or the asset is only moderately critical, CBM for critical assets with measurable condition, and PdM reserved for the handful of assets where forecasting pays for itself.

How do you set up CBM? Five steps

Pick the parameter per failure mode. Start from the failure modes that actually hurt you (work-order and downtime history), then choose the measurable signal that precedes each: bearing wear → vibration; loose connections → temperature; hydraulic contamination → oil particle counts; filter loading → differential pressure.
Establish the baseline. Measure the healthy state under normal operating load. A threshold without a baseline is a guess. Where machines are already connected, weeks of history often exist before you start.
Set two thresholds: alert and act. An alert level that opens an inspection, and an action level that opens a work order. Sources for limits: manufacturer specs, ISO guidance for the parameter (for example, vibration severity standards), and your own history. Expect to tune them, the first thresholds will be wrong in one direction or the other.
Wire the threshold to a work order, not an email. The reading must land where it triggers action: a work order with the asset, reading, and limit attached. An alarm that goes to an inbox is where CBM programs go to die. This is a workflow-automation problem as much as a sensing problem, the event-triggers-action pattern described on our platform overview.
Review threshold performance quarterly. Count false alarms and missed catches. Tighten limits that let a failure through; loosen limits that cry wolf. Track the result in your maintenance KPIs unplanned downtime and emergency-work percentage should trend down on CBM-covered assets.

What does a condition threshold look like in practice?

Here is the shape of it: a parameter trending over weeks, an alert limit that buys inspection time, and an action limit that triggers the planned repair before functional failure.

A hypothetical bearing-vibration trend against alert and action limits. The two-threshold pattern buys inspection time before committing crew hours.

Note what CBM did not do here: it never estimated a failure date. It said degradation started, then said degradation reached the point where action is cheaper than waiting. For most assets, that is all the sophistication the economics justify.

What does CBM pay, honestly?

Condition-driven maintenance shares the documented economics of predictive approaches in the U.S. Department of Energy's FEMP O&M guidance, maintained by Pacific Northwest National Laboratory: 8–12% savings over a preventive-only program and opportunities that can exceed 30–40% versus heavily reactive maintenance (PNNL, O&M Best Practices: Maintenance Approaches). The same guidance is blunt about the catch: the savings require a functioning program, baselines kept current, thresholds tuned, findings acted on.

The costs are lower than full PdM but not zero: sensors or inspection-route time, threshold engineering, and the data plumbing to move readings from machine to work order. The labor market raises the stakes on getting this right, the U.S. Bureau of Labor Statistics projects 13% employment growth for industrial machinery mechanics and maintenance workers from 2024 to 2034 much faster than average (BLS), which means technician hours will keep getting more expensive and harder to fill. Strategies that stop spending those hours on healthy machines are labor strategies.

Where does CBM lead?

A CBM program that runs well for a year produces exactly the asset history, baselines, trends, catches, misses, that makes a later predictive maintenance investment rational instead of speculative. Measure the improvement with MTBF on covered assets, and keep the foundation honest: CBM on top of a broken PM schedule fixes nothing.