Detailed Modeling of Turbulent Fire Systems

Abstract

Fire spans a large range of temporal and spatial scales, and involves multiphysical processes. Turbulence, homogeneous and heterogeneous chemical reactions, radiative heat transfer, and multiphase transport can simultaneously be significant in fire. To maintain reasonable computational cost, reduced-order models are often required to describe these processes in order to predict the dynamics of fire. Such reduced-order models often require extension or re-formulation whenever a new aspect of the underlying physics demands attention. In this talk, our recent effort on developing high-fidelity models for the fire environment is introduced. In particular, we are concerned with the scenario of fire extinguishment where turbulence-droplet-radiation interactions are important. The solutions obtained from the high-fidelity models are leveraged to provide benchmark data for the reduced-order models or to provide closure information. Two modeling aspects are detailed in this talk, including the development of a comprehensive Monte Carlo ray tracing solver that accounts for radiative interactions between gas, soot, wall and water droplets, as well as the development of a reduced-order soot model. Finally, the numerical aspects of the underlying flow solver (OpenFOAM), as well as the effort in enabling large-scale direct numerical simulation on the multicore computing architectures (i.e., Intel Knights Landing) are briefly discussed.

XINYU ZHAO

University of Connecticut