Assessing the Accuracy of Vulnerability Models by Comparison with Vulnerability Experiments
Abstract
Over the past three decades the business of vulnerability analysis has been conducted more as art than as science. For a number of reasons, there has been recent movement away from art and towards science. One reason is internal to the Ballistic Research Laboratory. If a substantial work force of scientists and engineers spends several decades doing vulnerability analysis it would be both surprising and disappointing if improved clarity and on the foundations of the subject were not achieved. A second reason is the declining cachet associated with computer models. Several decades ago it was virtually the case that if a quantitative result issued from a computer model, it was sufficiently warranted as correct. The crude extrapolations which then passed as vulnerability models benefited from that halo effect. Today, people are understandably more skeptical about models. Questions are asked and answers expected about accuracy and validation. The pressure of such questioning has forced vulnerability analysts into seeking greater scientific respectability for the fruits of their labors. A third cause of the scientific turn is the recent explosion of computer technology. Vulnerability phenomena are complex; realistic simulation was out of the question with the tools available to early vulnerability analysts.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1989
- Accession Number
- ADA210871
Entities
People
- Michael W. Starks
Organizations
- Ballistic Research Laboratory