|
Reliabilitygrowthtestinginvolvestheselectionofoptimaldesignparameterstoenhanceaproduct’sreliability.This paperproposesasplitplotexperimentaldesignthataccommodatestherestrictiononrandomizationontheorderofexperimentalrunscausedbytheexperimentalnatureofacceleratedreliabilitytesting.Theproposedexperimental designprovidesstatisticallyrelevantsolutionsaboutthechoiceofdesignparameters,intermsoftheirreliabilityimpact, inamuchshortertime.Adegradationmodelthataidsinpredictingthefailuretimeforthegivenproblemfurther supplementsthisdiscussion. |
|