Models and Analysis of Degradation Data      

Sergio Yañez, Ronald Andrés Granada & Mario Cesar Jaramillo       

 

Abstract

Degradation is a weakness that eventually can cause failure (e.g. car tire wear). When it is possible to measure degradation, such measures often provide more information than failure-time data for purposes of assessing and improving product reliability. This is a paper which mainly pretends to divulge techniques that had been developed by Meeker & Escobar (1998). We think it is worth to make this topics known, because they are in the research frontier of the Reliability Theory (Lawless 2000). We compare in this work the explicit degradation models with the approximate degradation analysis. The explicit degradation model requires specific models developed by engineers and physical scientists, which are treated as mixed models with random effects. To obtain ML estimates we use S-PLUS following Pinheiro & Bates (2000), and also use bootstrap confidence intervals.

 

Key words: Reliability theory, Degradation data, boostrap, Mixed effects models, Random effects.

 

PDF (Spanish)