A comparison of the RANS and LES turbulence models in the simulation of emissions from flares to the Environment

Abstract

Purpose: This study is an attempt to simulate wake-stabilized flares in the petroleum and gas industry using mathematical equations governing the flow, turbulence and combustion in flames as encoded in the ANSYS-HYSYS simulation software package. The work compares the RANS and LES turbulence models in conjunction with the partially premixed combustion model. The model uses the mixture fraction approach in order to predict the flame appearance and the thermochemical properties of the wake stabilized cross flow flame. Design/ Methodology/ Approach: The research strategy involves using the RANS (Reynolds Averaged Navier Stokes) and the LES (Large Eddy Simulation) mathematical Models to simulate and to study the physical structure and the thermochemical properties of natural gas flares in the presence of crosswind. The wind-tunnel geometry was built and meshed using the ICEM software, the calculations were carried out in the ANSYS-Fluent CFD software and the computational data generated was processed and analyzed using the Tecplot CFD post-processing software package. The results of the simulation were then validated against the experimental work of Huang and Wang. Findings: The findings demonstrated that the LES turbulence model outperformed the RSM turbulence model in terms of predicting temperature trends and pollutant species. However, the peak temperatures at the analyzed measurement locations were predicted by both models accurately. The LES model also improved CO2 concentration predictions. In general, the LES turbulence model predicts more accurately than the RANS model, but the RANS model still provides a respectably accurate forecast of the thermo-chemical characteristics of the flame, making it a viable substitute for the more expensive LES. Practical Implications: The practical implications of the work is that simulation can be used in place of experiments to save money on some of the more expensive experimental projects since the results from the simulation agrees fairly well with the experimental data.