Code Validation as a Reliability Problem

Thursday, September 16, 1999 - 11:20am - 12:00pm
Lind 409
Timothy Trucano (Sandia National Laboratories)
The purpose of this presentation is to accomplish three things. First, I want to discuss my view of the role that uncertainty quantification plays in code validation activities. My recent conceptualization of this is to regard code validation as a reliability problem. I will explain why I feel this way at the present time and invite criticism.

Second, I wish to summarize recent work which addresses aspects of uncertainty quantification that I feel are important for its application as a methdology relevant to code validation. In doing this, I also will also identify some technical problems that I feel are critical to achieving some success for code validation.

The narrow applications I have in mind are to shock hydrodynamics modeling, in particular Sandia's ALEGRA code development effort. Therefore, the third item I would like to briefly discuss the DOOMSDACE project at Sandia, an ASCI-supported conceptual tool development project, and its role as a testbed for some of the ideas presented in the talk.