Bathtub failure curve

The bathtub curve is the typical shape of a product's failure rate over its service life when you look at a large population of units. The name comes from that shape. The curve has three segments: early failures, also called infant mortality, then a long stretch of constant random failures, and finally wear-out. Infant mortality reflects the quality of the manufacturing process. These failures tend to show up very soon after installation, and the failure rate falls over time. Manufacturers reduce them with in-factory testing. In the flat middle segment, failures are random and the rate stays roughly constant. This segment reflects how reliable the design is in its operating environment, and it is the part of the curve that mean time between failures describes. Getting through infant mortality does not mean a unit will last until wear-out. A low rate of random failures continues throughout the useful life. Wear-out makes the failure rate rise at the end of a product's design life, and this sets its useful service life. In life-data analysis, the Weibull shape parameter shows which segment applies: values below one mean infant mortality, one means random failures (the exponential distribution), and values above one mean wear-out.

<!-- review notes

• sentence 3: support check: partial - u3 supports the manufacturing-quality link, the early timing and the reduction through factory testing, and u4 says the early-life failure rate decreases with time. But 'manufacturers reduce them' is generalized from Enphase's own testing. -->