Statistical Characterization of Graphite Fiber for Prediction of Composite Structure Reliability.

摘要

The increased use of composite materials in Navy Applications brings with it the need to know quantitatively the reliability of composite structures. Traditional methods of reliability prediction cannot be used. Therefore, analytical modeling is required, and experimental methods with which to assess the model are needed. This research refined experimental methods and obtained benchmark statistical data on the strength of graphite fibers. With the Chain-of-Bundles local load sharing model, the data were then used as a guide to perform a parametric influence study on the sensitivity of composite structure reliability to the statistical distribution of its constituent graphite fibers. The results indicate a strong influence by the lower tail of the single fiber statistical distribution, which has important implications for design, acceptance testing, and other Navy procurement functions.