The P-F curve represents the interval between a potential failure condition that a defined task can identify, point P, and the loss of the performance standard required for the function, point F. It helps answer a practical question: after a control detects degradation, is there enough time to confirm the diagnosis, plan the intervention, and act before functional failure?
The interval does not belong to the equipment alone. It depends on the failure mode, the technique used, the detection threshold, the operating condition, and the criterion that defines the function. For the same bearing, vibration, ultrasound, oil analysis, temperature, and sensory inspection can produce different P points.
What points P and F mean
Potential failure, P, is an identifiable condition indicating that a failure mode is developing. The point appears when a task and a criterion can distinguish that condition from normal variation. Changing the technique, frequency, or sensitivity changes when the signal can be recognized.
Functional failure, F, occurs when the item no longer meets the performance standard required in its operating context. A pump may still be turning and already be in functional failure if it cannot deliver minimum flow or pressure. A motor may run and fail to provide the required availability because of repeated trips. Point F does not have to coincide with catastrophic breakdown.
The NASA Reliability-Centered Maintenance Guide defines functional failure from performance requirements and shows that degraded components may continue to operate without the system function being lost. The guide also recommends adjusting tasks and intervals to location, application, environment, and the condition's proximity to the functional limit.
The P-F interval is an estimate, not a constant
The P-F interval is the time between possible detection of the potential failure and functional failure. It may change with load, speed, contamination, temperature, start-up regime, installation quality, and damage progression. Even identical assets can have different intervals.
For that reason, a value taken from a book, supplier, or similar asset is an initial hypothesis. The routine needs to revise that hypothesis using trends, inspection findings, removed parts, and the actual time between alert and intervention.
Three lead times need to fit within the decision:
- time to detection: how long can pass between the onset of an identifiable condition and the next reading;
- confirmation time: how long the team needs to repeat the measurement, rule out error, and close the diagnosis;
- response time: how long it takes to plan, obtain materials, negotiate the window, and execute the work.
A route may have an interval shorter than P-F and still be insufficient. If detection occurs near the end of the interval and the response takes longer than the time remaining, the task does not protect the function.
How to set the inspection frequency
There is no universal fraction of the P-F interval that fits every failure mode. An initial frequency needs to account for uncertainty, consequence, detection capability, and response time. The process is:
- define the function and minimum performance standard;
- select the failure mode to be monitored;
- choose a technique capable of identifying a specific potential condition;
- estimate the shortest plausible P-F interval instead of using only the average;
- subtract the time required to confirm and intervene;
- define the route with a margin compatible with the consequence;
- shorten or extend the interval as trends and accumulated evidence change.
After an alert, the frequency may need to increase. The NASA guide recommends reducing the interval when the trend indicates that failure is approaching and allows it to be increased when readings stabilize, always with an assessment of risk and uncertainty.
Techniques do not occupy a fixed order on the curve
A list that places oil analysis, vibration, ultrasound, thermography, and sensory inspection in one universal sequence creates false precision. Each technique observes a phenomenon. Point P appears when that phenomenon becomes detectable for the failure mode being analyzed.
| Technique | Phenomenon observed | Application example | Interpretation limit |
|---|---|---|---|
| Vibration | change in dynamic response | bearing defect, misalignment, imbalance, and looseness | lead time depends on sensor, measurement point, collection frequency, load, and failure mode |
| Ultrasound | high-frequency acoustic energy | friction, insufficient lubrication, leakage, and electrical discharge | an early signal for one phenomenon may not exist for another |
| Oil analysis | contamination, lubricant properties, and wear particles | gearbox, hydraulic system, and lubricated engine | sampling method, collection point, and laboratory practices affect the conclusion |
| Thermography | surface temperature distribution | electrical connection, overload, friction, and thermal insulation | emissivity, load, line of sight, and environment can mask the signal |
| Sensory inspection | noise, odor, temperature, vibration, and visible condition | leakage, looseness, overheating, and perceptible change | it depends on access, a defined standard, and consistent recording, and does not replace measurement when measurement is required |
The choice combines failure mode, technical applicability, detection confidence, cost, and response time. Techniques may complement one another. An ultrasound signal can guide lubrication, while vibration tracks damage progression in the same bearing. The sequence needs to be demonstrated on that asset rather than assumed from the tool category.
Applied example: bearing on a critical fan
This instructional example uses hypothetical figures to demonstrate the decision. They are not bearing benchmarks or route recommendations.
Function: the fan must maintain minimum process air flow throughout the production campaign.
Functional failure: flow below the limit or fan unavailability outside a scheduled window.
Failure mode analyzed: degradation of the drive-end bearing.
Condition monitoring task: vibration readings at defined points and load condition, with trending by bands associated with the bearing and technical confirmation when the level crosses the alert criterion.
The available history indicates that, in the few comparable cases, the period between the first confirmed alert and loss of function ranged from 8 to 14 weeks. The team needs 3 weeks to confirm the diagnosis, obtain the bearing, and negotiate a safe work window.
With a 6-week route, the worst case is inadequate. The condition may begin immediately after a reading, be found 6 weeks later, and leave only 2 weeks before the functional limit, less than the required response time. An initial route every 2 weeks provides more detection opportunities within the shortest observed P-F interval and preserves margin to confirm and act.
After the alert, continuing at a 2-week frequency without assessing the trend would also be insufficient. If the slope increases, the team shortens the interval, confirms the finding with a complementary technique when applicable, and brings the intervention forward. If the signal stabilizes and the cause is understood, engineering may revise the decision.
The case demonstrates the logic. The final interval depends on consequence, actual data, detection capability, and the plant's work window.
P-F curve and predictive maintenance
The curve helps determine whether a condition monitoring task finds degradation with useful lead time. It does not make every measurement predictive or turn an alert into an automatic diagnosis. The task needs to be tied to a failure mode, produce an interpretable signal, and trigger a defined action.
The predictive maintenance guide explains the differences among condition monitoring, trend development, and intervention scheduling.
P-F curve and bathtub curve
The bathtub curve represents a hypothesis about the failure rate across the life of a population of items. The P-F curve follows the progression of one failure mode from a detectable potential condition to loss of function. They operate at different scales and answer different questions.
What RCM decides beyond the P-F interval
RCM starts with functions, functional failures, failure modes, and consequences. The P-F interval enters when a condition monitoring task is technically applicable. The decision still needs to verify whether the task is effective in reducing risk or consequence and whether a feasible action exists after the alert.
The NASA guide organizes time-directed tasks, condition-directed tasks, and failure-finding tasks, as well as the conscious decision to run some equipment to failure. The choice depends on function and consequence, not only on the ability to measure a variable.
For the complete process, see the RCM guide in Portuguese and the maintenance reliability engineering guide.
How to bring P-F into the maintenance plan
- record the function, performance standard, and failure mode;
- document the technique, measurement point, operating condition, and alert criterion;
- store the P-F estimate as a range and identify its source;
- include confirmation, material, access, and work window lead time in the margin;
- define the action required at each alert level;
- revise the interval with every finding, intervention, and removed part.
The preventive maintenance plan template in Portuguese helps organize assets, tasks, intervals, ownership, and execution criteria. The approved strategy and intervals remain in SAP PM. PM Run schedules and executes the resulting work orders without replacing the engineering decision.
Frequently Asked Questions
What is the P-F curve?
It represents the interval between a potential failure condition identifiable by a defined task and the loss of the performance standard required for the function.
Is point P the physical start of degradation?
Not necessarily. Degradation may begin earlier. Point P marks a potential condition that the selected technique and criterion can identify.
Does functional failure always mean that the equipment has stopped?
No. Functional failure means that required performance is no longer met. Insufficient flow, repeated trips, or inability to operate for the required duration may characterize it before complete breakdown.
Which technique detects failure first?
It depends on the failure mode and application. Vibration, ultrasound, oil analysis, thermography, and sensory inspection observe different phenomena. Lead time needs to be demonstrated on the asset and operating condition being analyzed.
How should the inspection frequency be set?
Use the shortest plausible P-F interval, confirmation and response time, technique confidence, and failure consequence. Start conservatively and adjust using trends and evidence from the asset.
Does P-F alone define the RCM strategy?
No. It helps evaluate condition monitoring tasks. RCM also considers function, functional failure, failure mode, consequence, and whether each task is applicable and effective.
Go deeper into the application
The PM Run Advanced Study Group, GEA, holds invitation-only sessions focused on customers and partners. Customers and partners can request an invitation from their account executive. Companies that are not yet customers can contact the commercial team to learn about the initiative.
