The pf curve is a drawing of how an asset fails over time: it runs normally, a potential failure becomes detectable at point P, the condition worsens, and functional failure comes at point F. The time between P and F is the P-F interval, and it is the reason inspections have the intervals they do: to catch the asset between the two points, an inspection has to happen more often than the P-F interval. This page explains what the curve shows, how it decides an inspection frequency, and where it is a real measurement and where it is a sketch on a whiteboard.
What the curve shows
A bearing does not fail all at once. Vibration rises, then noise, then heat, then it seizes. The point at which the first of those can be detected with the tools you have is P; the seizure is F. The curve draws condition falling from P to F over time, and the shape is usually steep at the end, which is why the last warning sign is not the one to plan around. Every asset with a wear-out failure mode has a curve; the P-F interval is different for each mode and each detection method.
How it decides an inspection interval
If the P-F interval for a bearing detected by a hand on the housing is a week, a monthly inspection will miss most of the failures. If it is three months detected by vibration monitoring, a monthly check catches them with time to plan the repair. The rule is that the inspection interval must be shorter than the P-F interval, with enough margin to act; half the interval is the usual working rule. That is where the intervals on a preventive schedule come from, whether or not anyone drew the curve.
Where it is honest
When P and F are known from history: the work orders say how long before failure the symptom was first recorded. When the detection method is fixed and its sensitivity is known. When the failure mode is one that develops, wear, corrosion, fouling, rather than one that arrives without warning. On those assets the curve sets the inspection interval and the history tests it.
Where it is a drawing
On assets with no recorded history, where P and F are guesses. On random failures, electronics and some seals, where there is no P to find and inspection is theatre. And on any curve presented with precise numbers nobody measured. The curve is most useful as a way of asking whether an inspection can catch the failure at all, and the honest answer for a good share of the register is no, which moves those assets to a preventive replacement or to run to failure with a spare on the shelf.
Questions people ask about pf curve
Is the P-F interval the same as MTBF?
No. MTBF is how often the asset fails; the P-F interval is how long the warning lasts before it does. An asset can have a long MTBF and a short P-F interval, which makes it hard to inspect usefully.
Do I need condition monitoring to use the curve?
No. The detection method can be a technician's senses on a walk-round; it only moves P later and shortens the interval. The curve is about the relationship, not the instrument.
How do I find the P-F interval for my equipment?
From your own history: when was the symptom first recorded on a work order, and when did the asset fail. A CMMS with findings recorded gives it; a log with checked, ok, does not.