Exploring the Failure Distributions from Discrete and Continuous Degradation Processes
碩士 === 東海大學 === 統計學系 === 90 === Abstract It is diffcult to asses reliability with traditional life tests for high reliability products that record only time to failure and usually have not many observations. When degradation measures can be taken over time, a relationship betw...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2002
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/88916291465259741724 |
| Summary: | 碩士 === 東海大學 === 統計學系 === 90 === Abstract
It is diffcult to asses reliability with traditional life tests for high reliability products that record only time to failure and usually have not many observations. When degradation measures can be taken over time, a relationship between component failure and degradation makes it possible to use degradation models to provide inferences about failure time. In this research, we explore the failure distributions with discrete and continuous degradation process. The mark point process and the Wiener process will be introduced.
For the needs of flexibility of the models to fit the data, three parameter models are considered such as extended generalized gamma (EGENG), exponential Weibull and generalize inverse Gaussian distributions. We employ probability plot and also Anderson-Darling test to identify the proper distribution for the 2024-T351 aluminum data. The data set was produced by the fatigue laboratory in the Department of Mechanical Engineering of national Taiwan University in year 2001. We use bootstrap method to give the confidence intervals of the p-h quantile of the best fitted distribution.
|
|---|