Numerical modelling of spectral and turbulence effects on gas-phase radiative heat transfer in fires

Radiative heat transfer plays a significant role in combustion systems, yet its reliable prediction remains as one of the most persistent challenges in computational fire dynamics. Conventional gray approach introduces strong approximations in representing the spectral aspect of the phenomenon. The dependency to prescribed path length further limits predictive capability, especially in scenarios with multiple radiative length scales. This research aims to address the spectral and turbulence effects of gas phase radiative heat transfer in fires. To handle the spectral aspect, several global non-gray spectral models were constructed and validated against line-by-line benchmarks, in a series of canonical decoupled test cases. This effort led to development of rank correlated full spectrum k-distribution (RC-FSK) with multiplication-approximation approach, which offered the best compromise between predictive accuracy and computational efficiency. Consequently, RC-FSK was selected to be implemented in the Fire Dynamics Simulator (FDS). To account for the turbulence-radiation interaction (TRI), a two-zone subgrid temperature model was introduced and implemented into FDS. The TRI model performance was assessed using the University of Maryland line-burner case, where the TRI model yielded reduced grid sensitivity in under-resolved simulations and improved prediction of the radiative fraction of the heat release rate. The integrated RC-FSK and TRI models were further evaluated using the methanol pool fire test case. The analysis comprised of decoupled (radiation only) and coupled CFD simulations. In the decoupled calculations, RC-FSK showed closer agreement with line-by-line Photon Monte Carlo (PMC-LBL) benchmarks compared to the default gray model in FDS. In the coupled CFD simulations, the RC-FSK combined with the TRI model predicted higher radiative emission, absorption, and source terms than the gray approach, highlighting the importance of advanced nongray and TRI treatments in predictive fire modeling.