Probabilistic Fracture Mechanics: A Validation of Predictive Capability
Abstract
A two-phase task was undertaken to the Probabilistic Fracture Analysis Code (PROFRANC) developed under this project. Phase one consisted in predicting deterministically the outcome of a subset of a larger number of experiments in which variability in geometry and material parameters was purposely minimized. The purpose of this phase was to verify that PROFRANC could predict nearly deterministic events accurately. This phase was shown to be highly successful. This verification was based on experimental results which had to be obtained within this project due to a paucity of comprehensive mixed-mode fracture propagation data in the open literature. In Phase two all currently available data involving inherent uncertainties in some material and geometrical parameters was assembled in a probabilistic framework and subsequently compared to the probabilistic predictions of PROFRANC qualitatively and quantitatively. Theses comparisons were shown to be very successful. The quantitative comparison was performed by hypothesis testing, which is a mathematical rule deciding whether to except or reject PROFRANC predictions using experimental data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1990
- Accession Number
- ADA228877
Entities
People
- Anthony R. Ingraffea
- Mircea Grigoriu
Organizations
- Cornell University College of Engineering