The shape
The failure rate is plotted on the vertical axis and operating time on the horizontal axis. The figure resembles the profile of a bathtub: a side sloping downward, a long flat bottom and a side that rises.
The three regions
- Early failures: the rate is high at first and decreases. This corresponds to manufacturing, installation or design defects that show up early. They are known as infant mortality.
- Useful life: the rate is low and approximately constant. Failures are random, caused by occasional overloads or external events. This is the region where MTBF makes sense.
- Wear-out: the rate rises because aging mechanisms, such as fatigue, corrosion or material degradation, bring components to the end of their life.
What it is used for
It guides decisions. For the early phase, screening and burn-in are used. For useful life, the design and overload protection are improved. For wear-out, preventive replacement is planned before the rate shoots up, and the warranty or service period is defined.
How it relates to distributions
The Weibull distribution illustrates the idea: depending on the value of its shape parameter, the failure rate decreases, stays constant or increases. The bathtub curve can be seen as a combination of populations with different behaviors.
Nuances
The curve is a conceptual model, not a law. Many products do not show the three phases clearly. Some modern electronic components have practically no wear-out during their service life; others, such as mechanical components, have almost no flat region. The shape depends on the product, the environment and the stress level.
How to use it well
It is advisable to identify from real data which phase the product is in before applying a measure. Mixing units of different ages in the same analysis can hide the wear-out phase. Interpretation must be supported by failure data and analysis of their cause.