Identification of Cumulative Fiber Breaking Process in a CFRP by Stochastic Model

摘要

Reinforcing fibers used for long fiber composites, such as carbon and glass, are superior in strength, whereas they have a large statistical variation in strength. Because of this variation, it is difficult to predict the composite strength from deterministic models such as the rule of mixture, and therefore many theoretical models of composite strength have been proposed from the probabilistic point of view1-5). Chain-of-bundles model is one of the representative models, which is close to the concept of the weakest link theory, but each link consists of a fiber bundle with a constant length such as the ineffective length. In the fiber bundle, the fiber breaking process starts from weaker fibers in strength, in case the influence of stress concentration is not taken into account around the fiber-break points. This case is applicable for the fiber breaking process of a ceramic matrix composite, in which the interface between the fiber and matrix is relatively weakly formed. Fibers would be broken dispersively, and the number of fiber-breaks would be accumulated. On the other hand, the stress concentration works strongly if the interface is strongly formed. Fibers would successively be broken by stress concentration, not from weaker fibers. The number of fiber-breaks would also be lowered. Thus, conventional theoretical models have been developed including such two assumed points of view. Conventional theoretical models are mathematically excellent, while the authors consider that the prediction of fiber breaking process might be developed using experimental data.