Accession Number:

ADA232634

Title:

Statistical Models and Inference Procedures for Structural and Materials Reliability

Descriptive Note:

Final rept. 1 Jul 1987-30 Nov 1990,

Corporate Author:

SOUTH CAROLINA UNIV COLUMBIA DEPT OF STATISTICS

Report Date:

1990-12-01

Pagination or Media Count:

10.0

Abstract:

The objectives of this project were to develop and understand probability and statistical models which adequately describe the reliability of composite materials. Prerequisite to the study of fibrous composites was the study of the failure characteristics of brittle fibers e.g. carbon fibers that are used in such composites. Thus, new approaches to the reliability of structures with respect to their strength to withstand an environmental stress were considered. More specifically, it was shown that the probability distribution for the tensile strength of brittle fibers had the IFR property and that is carried over to composite containing such fibers. A new probability distribution for the strength of brittle materials was developed and its characteristics were studied. Some general stress-strength models were also developed and applied to the failure of systems subject to cyclic loading. Involved in the failure of composite materials was the development of new graphical and inferential statistical methods for estimating mixtures of failure distributions or mixtures of hazard functions. The related program of competing risks models was also investigated.

Subject Categories:

  • Laminates and Composite Materials
  • Statistics and Probability

Distribution Statement:

APPROVED FOR PUBLIC RELEASE