- Author
-
Rouson, D. W.
- Title
- Fire, Smoke and Particles: FAST and DNS. BFRL Fire Research Seminar. VHS Video.
- Coporate
- Exponent Failure Analysis Association, Menlo Park, CA
- Report
-
Video
August 2, 1999
- Keywords
-
smoke
|
computer programs
- Identifiers
- FAST
- Abstract
- NOTE: NO AUDIO AVAILABLE The engineer tasked with modeling multiphase and multi-component flows faces a paucity of accurate software tools. The older, industry-standard codes approximate a broad class of flows with a small set of empirical formulas. The newer, state-of-the-art tools are built largely on statistical theories that are difficult to validate experimentally. At one end of the spectrum, a nave user of FAST will choose from a small set of heat release rate curves to model all fires. At the other end, the designers of solid particulate transport devices benefit from a 78-year history of extensive theoretical development. Yet the fundamental parameters in these theories defy direct measurement. Results from two studies addressing these issues will be presented. The first involved a full-scale building burn and fire safety demonstration in the wake of the nation's second-worst domestic fire. Temperature, carbon monoxide, and oxygen levels measured during the test will be compared with the predictions of FAST. The successes and difficulties encountered will be discussed. The second study involved the direct numerical simulation (DNS) of the turbulent transport of airborne solids, an important phenomenon in problems ranging from dust explosions to pollution dispersion. The results presented will focus on a set of second-order, particle-Lagrangian statistics that form the underpinnings of most particle dispersion and turbulence modification theories. A new class of functions will be proposed for modeling the particle-Lagrangian fluid velocity autocorrelation tensor. Autocorrelations calculated by DNS will also be used to validate a novel laboratory technique. And a structural deficiency in the state-of-the-art turbulence modification models will be demonstrated.