The bathtub curve is a picture almost every reliability engineer carries in their head: a graph of how the failure rate of a population of equipment changes over its lifetime. It gets its name from its shape, which dips steeply, flattens across a long middle, then rises again at the end - like the cross-section of a bathtub. This guide explains the three phases the curve describes, why they exist, and what each one implies for how you should treat equipment at that stage of life.
Bathtub Curve in one line: The bathtub curve is a graph of failure rate, or hazard rate, plotted against time for a population of similar components. It shows three phases: a high but falling infant-mortality period at the start, a long flat useful-life period where failures are random and roughly constant, and a rising wear-out period at the end - a shape that explains why early burn-in and end-of-life replacement both reduce failures.
The first phase is infant mortality, a period of high but rapidly decreasing failure rate just after equipment is put into service. Failures here come from defects present from the start: manufacturing flaws, damaged components, incorrect installation, or commissioning errors. As the weak units in the population fail and are removed, the surviving population is stronger, so the failure rate drops steeply. This is the descending left wall of the tub.
The second phase is the useful life, a long stretch where the failure rate is low and roughly constant. Failures still happen, but they are random - triggered by chance events like overload, contamination, or operator error rather than by age. Because the rate is flat, a unit that has survived to this point is no more likely to fail next week than it was last week, which is a counterintuitive but important property. This is the flat bottom of the tub.
The third phase is wear-out, where the failure rate climbs as accumulated fatigue, corrosion, erosion, and material degradation finally catch up with the population. Here age genuinely predicts failure, so failures cluster as more and more units reach the end of their physical life. This is the rising right wall of the tub.
Each phase argues for a different action. Infant mortality is the case for burn-in and commissioning testing: running new equipment under controlled stress before it goes into critical service weeds out the defective units early, so they fail on the test bench rather than in production. It is also the reason a brand-new install is not automatically more reliable than a proven one.
The flat useful-life phase is where age-based maintenance does the least good. If the failure rate is constant, replacing a healthy component on a calendar does not reduce the chance of a random failure - it just consumes labor and parts and risks introducing an infant-mortality fault with the new unit. This is precisely the insight that reliability-centered maintenance uses to reject unnecessary scheduled overhauls in favor of condition monitoring or run-to-failure.
The wear-out phase is the one case where time-based replacement is clearly justified. Once a component enters the rising portion of the curve, its remaining life is genuinely a function of age, so scheduling replacement before the failure rate climbs steeply pays off. A crucial caveat is that not every component actually has a pronounced wear-out phase within its service life, and studies of complex equipment have found that many failure patterns are dominated by randomness rather than a clean bathtub shape.
The bathtub curve is a statistical property of a population, so seeing it in your own fleet requires failure history across many similar assets. That history is exactly what accumulates when field equipment is monitored and its runtime, trips, and failures are logged over years. A cluster of early trips on freshly installed units versus a slow rise in faults on aging ones is the bathtub curve showing up in real operational records.
In a SCADA-monitored oil and gas operation, the runtime hours, alarm events, and downtime records a platform captures are the ingredients for estimating where a given class of equipment sits on the curve. If newly deployed pumps show a burst of alarms in their first weeks and then settle, that is the infant-mortality wall; if an aging population of motors starts logging more frequent overload trips, that is wear-out beginning. Merobix historizes runtime and alarm data continuously, so the raw failure timeline exists in the platform even though building the statistical curve from it is a separate reliability analysis.
The practical value is that the curve turns scattered failure events into a diagnosis. Knowing whether recent failures are infant mortality, random useful-life events, or the onset of wear-out tells an operator whether the fix is better commissioning, better operating discipline, or a planned replacement program - three very different responses to what can look at first like the same problem.
Because the plotted shape resembles the cross-section of a bathtub: a steep drop on the left as early defects are weeded out, a long flat bottom during useful life, and a rising wall on the right as wear-out sets in. The name is purely descriptive of the graph's silhouette, not of any physical process.
No. The bathtub is an idealized model, and real components show a range of patterns. Some have strong infant mortality with little wear-out, some show almost constant random failure, and studies of complex assemblies have found that many pieces of equipment do not have a pronounced age-related wear-out phase at all. The curve is a teaching tool and a starting hypothesis, not a universal law.
The flat useful-life region has a constant failure rate, meaning failures are random and not driven by age. In that region, replacing a healthy part on a fixed schedule does not lower the failure risk and can even add risk by introducing a new infant-mortality fault. This is why condition-based monitoring or accepted run-to-failure often beats calendar-based overhauls during useful life.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.