Probabilistic Fracture Models for Predicting the Strength of Notched Composites

Abstract

This paper presents two probabilistic fracture models for predicting the tensile strength of filamentary composites containing geometric discontinuities such as holes or cracks. The statistical fracture model considers the case of a constant load while the stochastic model deals with monotonically increasing loads. Both models use the Weibull distribution of fiber strength and elastic fiber/matrix properties to calculate the number of broken fibers near the crack tip in the 0-plies as a function of applied loads for different probability levels. Using the probabilistic models, the notched strength of (±45/02)s boron/aluminum composites with various crack sizes have been predicted. These results agree well with existing experimental data. In addition, the relationship between fracture stress and notch size is found to be governed by a power law, as previously suggested by Mar and Lin using a deterministic approach.