Location

San Diego, CA

Start Date

1-1-1985 12:00 AM

Description

In the general problem of quantitative NDE, the majority of past approaches are based upon the questionable assumption that the dominant defect or flaw can be identified before the beginning of the main body of the analysis. This concept is fundamental to most (but not all) treatments of probabilistic failure prediction and accept/reject optimization. This oversimplification and associated logical tangles are obviated by a more comprehensive approach to defect characterization and probabilistic failure prediction, in which it is assumed in the pertinent stochastic models that the various significant types of defects occur in all possible numbers. In this paper, we present a version of such an approach that involves a single type of defect and a promising approximation methodology.

Book Title

Review of Progress in Quantitative Nondestructive Evaluation

Volume

4A

Chapter

Chapter 1: Ultrasonics

Section

Transducers and Signal Processing

Pages

241-252

DOI

10.1007/978-1-4615-9421-5_27

Language

en

File Format

Application/pdf

Share

COinS
 
Jan 1st, 12:00 AM

Analysis of Many-Defect Systems

San Diego, CA

In the general problem of quantitative NDE, the majority of past approaches are based upon the questionable assumption that the dominant defect or flaw can be identified before the beginning of the main body of the analysis. This concept is fundamental to most (but not all) treatments of probabilistic failure prediction and accept/reject optimization. This oversimplification and associated logical tangles are obviated by a more comprehensive approach to defect characterization and probabilistic failure prediction, in which it is assumed in the pertinent stochastic models that the various significant types of defects occur in all possible numbers. In this paper, we present a version of such an approach that involves a single type of defect and a promising approximation methodology.