Publications

We are concerned with transportation accidents and the subsequent fire. Progress is currently being made on a unique capability to model these very challenging events. We have identified Smoothed Particle Hydrodynamics (SPH) as a good method to employ for the impact dynamics of the fluid. SPH is capable of modeling viscous and inertial effects for these impacts for short times. We have also identified our fire code Lagrangian/Eulerian (L/E) particle capability as an excellent method for fuel transport and spray modeling. This fire code can also model the subsequent fire, including details of the heat and mass transfer necessary for thermal environment predictions. These two methods (SPH and L/E) employ disparate but complimentary length and timescales for the calculation, and are suited for coupling given adequate attention to relevant details. Length and timescale interactions are important considerations when joining the two capabilities. Coupling methodologies have been shown to be important to the model accuracy. Focusing on the transfer methods and spatial resolution, a notional impact problem is examined. The outcome helps to quantify the importance of various methods and to better understand the behavior of these modeling methods in a representative environment.

Transportation accidents and the subsequent fire present a concern. Particularly energetic accidents like an aircraft impact or a high speed highway accident can be quite violent. We would like to develop and maintain a capability at Sandia National Laboratories to model these very challenging events. We have identified Smoothed Particle Hydrodynamics (SPH) as a good method to employ for the impact dynamics of the fluid for severe impacts. SPH is capable of modeling viscous and inertial effects for these impacts for short times. We have also identified our fire code Lagrangian/Eulerian (L/E) particle capability as an adequate method for fuel transport and spray modeling. A fire code can also model the subsequent fire for a fuel impact. Surface deposition of the liquid may also be acceptably predicted with the same code. These two methods (SPH and L/E) typically employ complimentary length and timescales for the calculation, and are potentially suited for coupling given adequate attention to relevant details. Length and timescale interactions are important considerations when joining the two capabilities. Additionally, there are physical model inadequacy considerations that contribute to the accuracy of the methodology. These models and methods are presented and evaluated. Some of these concerns are detailed for a verification type scenario used to show the work in progress of this coupling capability. The importance of validation methods and their appropriate application to the genesis of this class of predictive tool are also discussed. Copyright © 2009 by ASME.

Detailed herein are the results of a validation comparison. The experiment involved a 2 meter diameter liquid pool of Jet-A fuel in a 13 m/s crosswind. The scenario included a large cylindrical blocking object just down-stream of the fire. It also included seven smaller calorimeters and extensive instrumentation. The experiments were simulated with Fuego. The model included several conduction regions to model the response of the calorimeters, the floor, and the large cylindrical blocking object. A blind comparison was used to compare the simulation predictions with the experimental data. The more upstream data compared very well with the simulation predictions. The more downstream data did not compare very well with the simulation predictions. Further investigation suggests that features omitted from the original model contributed to the discrepancies. Observations are made with respect to the scenario that are aimed at helping an analyst approach a comparable problem in a way that may help improve the potential for quantitative accuracy.

Abstract not provided.

A series of experiments has been performed in the Sandia National Laboratories FLAME facility with a 2-meter diameter JP-8 fuel pool fire. Sandia heat flux gages were employed to measure the incident flux at 8 locations outside the flame. Experiments were repeated to generate sufficient data for accurate confidence interval analysis. Additional sources of error are quantified and presented together with the data. The goal of this paper is to present these results in a way that is useful for validation of computer models that are capable of predicting heat flux from large fires. We anticipate using these data for comparison to validate models within the Advanced Simulation and Computing (ASC, formerly ASCI) codes FUEGO and SYRINX that predict fire dynamics and radiative transport through participating media. We present preliminary comparisons between existing models and experimental results.