Published Online: 21 October 2004
Page Count: 6
Senior Scientist, National Renewable Energy Laboratory, Golden, CO
(Received 16 March 2004; accepted 20 March 2004)
Two approaches are commonly used to derive correlations between in-service and accelerated exposure test results. When detailed degradation mechanisms are well understood, a deterministic formalism can be applied in which a precise damage function model is utilized. If failure mechanisms are not known or multiple mechanisms interact in a complicated manner that makes it difficult and tedious to treat them explicitly, a probabilistic procedure can be used. Observed failures are fit to appropriate life distributions to obtain expressions for related failure rates. A third approach uses a phenomenological methodology. This procedure is similar to the deterministic approach in that damage functions are hypothesized, except that they are based on macroscopic observations and effects, rather than on microscopic mechanisms. The phenomenological approach is used herein to describe results accurately from a number of highly accelerated exposure tests of organic materials. The ensuing damage function models are then used to predict real-world behavior. Excellent agreement is demonstrated between these predictions and actual measured data, thereby validating the phenomenological approach and providing a very useful way to estimate service life of organic-based materials.
Paper ID: JTE12600