Common Errors in Statistics (and How to Avoid Them)

Chapter 14 briefly describes different types of reliability databases and issues of data quality. These databases provide nominal "handbook" reliability values for different components for different kinds of systems and are commonly used as inputs to system reliability models. These nominal values are obtained from different sources such as laboratory life tests and field data, and sometimes, expert opinion. As pointed out by the authors, for reasons of simplicity and tradition, the reliability of a component is often given as a single number (either mean time to failure or the hazard value) under the assumption of a constant hazard. The assumption of a constant hazard is usually unrealistic because complicated electronic components (like ICs) typically have a decreasing hazard and mechanical devices generally have an increasing hazard. With improved database technology and computerization of system reliability design processes, we can expect, in the future, continued evolution toward more realistic reliability databases and models.