Infant Mortality
Definition
The early-life period of a component or system characterized by a higher-than-normal failure rate that decreases over time (Weibull beta < 1). Infant mortality failures are caused by latent defects introduced during manufacturing, installation, or commissioning — they are not age-related degradation but rather quality problems that manifest early. Common causes include material defects, welding flaws, installation errors, contamination, and incorrect settings.
Context
Infant mortality drives the need for commissioning quality assurance, burn-in testing, and early-life inspection programs. In reliability analysis, infant mortality failure data should ideally be separated from steady-state data because it represents a different failure population. Mixing the two can distort MTBF calculations and Weibull parameter estimates.
Examples
- A batch of new control valves shows 15% failure in the first 6 months (infant mortality), dropping to 2%/year thereafter. Investigation reveals incorrect spring settings from the factory.
- Post-turnaround monitoring detects an elevated failure rate in the first 30 days after restart, attributed to infant mortality from reinstated equipment — gasket leaks, incorrect torquing, and foreign material left in systems.
Applicable Standards
- IEC 61649:2008
- IEC 60300-3-1
Related Terms
A graph of the hazard rate versus time that describes the typical life cycle of a population of items, characterized by three distinct phases: (1) Infant Mortality — a decreasing failure rate period caused by manufacturing defects, installation errors, or design weaknesses; (2) Useful Life — a period of approximately constant, low failure rate representing random failures; (3) Wear-Out — an increasing failure rate period caused by age-related degradation mechanisms.
Wear-OutThe late-life period of a component or system characterized by an increasing failure rate over time (Weibull beta > 1), caused by progressive degradation mechanisms such as fatigue, corrosion, erosion, creep, and material aging. Wear-out failures are predictable and age-related, making them amenable to time-based or condition-based preventive maintenance strategies.
Random FailureA failure that occurs with approximately constant probability per unit time, independent of the item's age or accumulated operating time (Weibull beta approximately 1). Random failures are caused by unpredictable combinations of stress and strength, external events, or operating conditions that cannot be anticipated by age-based inspection or maintenance. They are equally likely to occur at any point during the useful life period.
Shape Parameter (Beta)(Beta)The Weibull shape parameter (beta, also written as b or k) that determines the form of the failure distribution and the nature of the failure rate over time. Beta < 1 indicates decreasing failure rate (infant mortality); beta = 1 indicates constant failure rate (random failures, equivalent to exponential distribution); beta > 1 indicates increasing failure rate (wear-out). The value of beta reveals the physics of the dominant failure mechanism.
Weibull DistributionA continuous probability distribution widely used in reliability engineering to model the time-to-failure of components and systems. The two-parameter Weibull distribution is defined by the shape parameter (beta) and scale parameter (eta). Its versatility lies in its ability to model increasing, constant, or decreasing failure rates depending on the shape parameter value, making it suitable for all phases of the bathtub curve.
Where this appears in Reliatic
This concept is used in the Reliability Analytics module.