The Fire Dynamics Simulator (FDS) code recently developed by the National Institute of Standards and Technology (NIST) is particularly well-suited for use by fire protection engineers for studying fire behavior. It makes use of Large Eddy Simulation (LES) techniques to directly calculate the large-scale fluid motions characteristic of buoyant turbulent diffusion flames. However, the underlying model needs further development and validation against experiment in the areas of soot formation/oxidation and radiation before it can be used to calculate flame heat transfer and predict the burning of solid or liquid fuels. WPI, Factory Mutual Research, and NIST have undertaken a project to make FDS capable of calculating the flame heat transfer taking place in fires of hazardous scale. The temperatures predicted by the FDS code were generally too high on the fuel side and too low on the oxidant side when compared to experimental data from small-scale laminar diffusion flames. For this reason, FDS was reformulated to explicitly solve the conservation of energy equation in terms of total (chemical plus sensible) enthalpy. This allowed a temperature correction to be applied by removing enthalpy from the fuel side and adding it to the oxidant side. This reformulation also has advantages when using probability density function (PDF) techniques in larger turbulent flames because the radiatively-induced nonadiabaticity is tracked locally with each fluid parcel. The divergence of the velocity field, required to obtain the flow-induced perturbation pressure, is calculated from an expression derived from the continuity equation. A new approach to soot modeling in diffusion flames was developed and added to the FDS code. The soot model postulated as part of this work differs from others because it is intended for engineering calculations of soot formation and oxidation in an arbitrary hydrocarbon fuel. Previous models contain several fuel-specific constants that generally can only be determined by calibration experiments in laminar flames. The laminar smoke point height, an empirical measure of a fuel?s sooting propensity, is used in the present model to characterize fuel-specific soot chemistry. Two separate mechanisms of soot growth are considered. The first is attributed to surface growth reactions and is dependent on the available surface area of the soot aerosol. The second is attributed to homogeneous gas-phase reactions and is independent of the available soot surface area. Soot oxidation is treated empirically in a global (fuel-independent) manner. The local soot concentration calculated by the model drives the rate of radiant emission. Calibration against detailed soot volume fraction and temperature profiles in laminar axisymmetric flames was performed. This calibration showed that the general approach postulated here is viable, yet additional work is required to enhance and simplify the model. The essential mathematics for modeling larger turbulent flames have also been developed and incorporated into the FDS code. An assumed-beta PDF is used to approximate the effect of unresolved subgrid-scale fluctuations on the grid-scale soot formation/oxidation rate. The intensity of subgrid-scale fluctuations is quantified using the principle of scale similarity. The modified FDS code was used to calculate the evolution of soot in buoyant turbulent diffusion flames. This exercise indicated that the subgrid-scale fluctuations are quantitatively important in LES of turbulent buoyant diffusion flames, although no comparison of prediction and experiment was performed for the turbulent case.
Identifer | oai:union.ndltd.org:wpi.edu/oai:digitalcommons.wpi.edu:etd-theses-1103 |
Date | 15 January 2002 |
Creators | Lautenberger, Christopher W. |
Contributors | David A. Lucht, Department Head, Nicholas A. Dembsey, Advisor, , Jonathan R. Barnett, John L. de Ris |
Publisher | Digital WPI |
Source Sets | Worcester Polytechnic Institute |
Detected Language | English |
Type | text |
Format | application/pdf |
Source | Masters Theses (All Theses, All Years) |
Page generated in 0.0018 seconds