Frequency Distributions and Bayesian Techniques for Estimating Performance in Composite Materials

摘要

Composite materials deteriorate from exposure to the environment over time. Key design parameters, such as rupture stress and Young's modulus, decay as the materials undergo freeze-thaw cycles. Unfortunately, the extreme variability of these effects require structures using these materials to be routinely inspected. A good analytial tool could help decision-makers estimate the strength of these materials as a vunction of exposure to freeze-thaw cycles. This paper examines regression models and frequency distributions that could represent freeze-thaw data. The authors then select one distribution and show how it may be employed along with Bayesian techniques to help estimate either a composite material's present strength or the number of freeze-thaw cycles sustained by the material. The visual perspective of a distribution is itself a way of presenting information.