More Pitfalls of Accelerated Tests

• Autores: William Q. Meeker, Georgios Sarakakis, Athanasios Gerokostopoulos
• Localización: Journal of quality technology: A quarterly journal of methods applications and related topics, ISSN 0022-4065, Nº. 3, 2013, págs. 213-222
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • As product-development cycles become shorter and companies demand more rapid achievement of reliability goals, it is becoming more and more important to use quantitative accelerated life tests (ALT) to predict and improve reliability. Today there is an abundance of methods to plan and analyze accelerated tests, but there are also many pitfalls. This paper identifies some major problems and concerns in conducting and interpreting the results of accelerated tests. We identify various pitfalls, such as using an equal unit allocation at all levels of an accelerating variable, using unnecessarily complex testing and data analysis schemes, attempting to predict life from a HALT, using ALT at the system level, not having adequate time to- failure information, using extreme extrapolation, ignoring the impact of idle