Invited Session: Accelerated Testing
Operational Statistical Lifetime Models for Engineering Design & Accelerated Testing
Stephen E. Chick
We present a methodology for directly connecting assumptions about failure mechanisms of engineering components (such as relations between stress and failure probability) with statistical lifetime models. This leads to statistical likelihoods and parameters which are operationally defined in terms of engineering observables. The work is motivated by a need to perform probabilistic design and reduce costs related to accelerated lifetime tests. Applications to simple mechanical and electrical systems will be presented.
[Stephen E. Chick, Dept. of Industrial Engineering, Univ. of Michigan, 1205 Beal Avenue, Ann Arbor, Michigan 48109-2117 USA; Stephen.E.Chick@umich.edu ]
Accelerated Degradation Test Modeling & Analysis
William Q. Meeker
Design of high reliability systems generally requires that the individual system components have extremely high reliability, even after long periods of time. With short product development times, reliability tests must be conducted with severe time constraints. Frequently few or no failures occur during such tests. Thus, it is difficult to assess reliability with traditional life tests that record only failure times. For some components degradation measures can be taken over time. A relationship between component failure and amount of degradation makes it possible to use degradation models and data to make inferences and predictions about the failure time distribution.
This talk describes a number of generic degradation reliability models that correspond to physical-failure mechanisms. The connection between degradation reliability models and failure-time reliability models is described and acceleration is modeled by having an accelerating factor or factors affect chemical reaction rate constants. Approximate maximum likelihood estimation is used to estimate model parameters from the mixed-effect nonlinear regression model. Simulation-based methods are used to compute estimates and confidence intervals for quantities of interest (like activation energy and failure probabilities). Finally we use several numerical examples to compare the results of accelerated degradation analysis and analysis with traditional accelerated life test time-to-failure data.
[William Q. Meeker, Dept. of Statistics, Iowa State Univ., Ames, Iowa 50010 USA; email@example.com ]
Date created: 6/5/2001