#### Date of Award

12-2016

#### Document Type

Dissertation

#### Degree Name

Doctor of Philosophy (PhD)

#### Legacy Department

Mechanical Engineering

#### Committee Member

Dr. Richard S. Miller, Committee Chair

#### Committee Member

Dr. Donald E. Beasley

#### Committee Member

Dr. Richard S. Figliola

#### Committee Member

Dr. Xiangchun Xuan

#### Abstract

Direct Numerical Simulation (DNS) data for high pressure *H _{2}/O_{2}* and

*H*flames using the compressible flow formulation, detailed kinetics, a real fluid equation of state, and generalized diffusion are analyzed. The DNS is filtered over a range of filter widths to provide exact terms in the Large Eddy Simulation (LES) governing equations, including unclosed terms. The filtered heat flux vector is extensively compared with the heat flux vector calculated as a function of the filtered primitive variables (i.e. the exact LES term is compared with its form available within an actual LES). The difference between these forms defines the subgrid heat flux vector. The analyses are done both globally across the entire flame, as well as by conditionally averaging over specific regions of the flame; including regions of large subgrid kinetic energy, subgrid scalar dissipation, subgrid temperature variance, flame temperature, etc. In this work, both the subgrid heat flux vector and its divergence are found to be substantially larger in reacting flows in comparison with mixing due to the associated larger temperature gradients. However, the divergence of the subgrid heat flux vector tends to be significantly smaller than other unclosed terms in the energy equation with decreasing significance with increasing Reynolds number. Then a reduced (29 step, 10 species)

_{2}/Air*Kerosene/Air*mechanism including a semi-global soot formation/oxidation model associated with an optically thin medium radiative heat flux model has been added to the same code to investigate soot formation/oxidation processes in a temporarily developing hydrocarbon flame operating at both atmospheric and elevated pressures for both a real gas law (RGL) and the ideal gas law (IGL) equations of state (EOS). Btoh 3D the RGL and the IGL EOS predictions of the soot formation/oxidation processes good agreement with the limited literature of atmospheric pressure flames [45, 46, 96] has been achieved. High values of the soot volume fraction have been shown to be independent from high temperature flame regions by occupying the flame volumes whose temperature varies from 1300 K to 1800 K. Additionally, the soot number density has been shown to be highly dependent on the temperature, while the soot volume fraction is dominated by local flow characteristics which is also in good agreement with Ref. [96]. Lignell et al. [46, 45] have reported two distinct behaviors of soot mass fraction: I- the slow soot nucleation process has caused the soot mass fraction to be widely scattered in the flame, and II - turbulent transportation has carried the soot to the fuel rich region emphasizing the importance of the turbulence transportation in sooting flames. Similar behavior has also been observed in the current work. The soot generation rate has been shown to have a similar trend with soot mass fraction, while Lignell et al. [45, 46] have observed a high dependency on flame temperature for the soot generation rate. Furthermore, a slight difference has been observed between the RGL and the IGL EOS model predictions of soot quantities in atmospheric pressure flames. This implies employing the IGL EOS might be reasonable to eliminate the complexity of mathematical models of real gas effects. In addition to the atmospheric pressure flames, inter-mediate (i.e. 5 and 10 atm) pressure flames are also investigated by artificially increasing the Planck mean absorption coefficient of the optically thin medium model (Lignell et al. [45] have reported due to the short simulation time, and small soot load in comparison to the domain size, radiative heat transfer is found to be insignificant). In adiabatic 5 and 10 atm flames slight differences for soot volume fraction, (fv), and the scalar dissipation rate are observed between the RGL and the IGL EOS models which are eliminated by the addition of the radiative heat loss. However, at 35 atm, the IGL EOS model has been shown to extremely over-predict not only the flame temperature but also the soot quantities by 25% to 100% in comparison with the RGL EOS model predictions. In elevated pressure flames, a similar trend has been observed for the soot volume fractions and the soot number density (N), while high values of the soot mass fraction, (

*Y*), exhibit a less scattered profile in the mixture fraction coordinate which is limited in the range of φ = 0.4 to 0.9. The soot generation rate has been observed to a have smaller standard deviation than its means in elevated pressure flames, while it has been observed to be larger than its means in atmospheric and inter-mediate pressure flames. As the time evolves in the RGL EOS elevated pressure flame by increasing the intensity of the flow more scattered trends have been detected in both flame characteristics and soot properties indicating soot properties are highly affected by the local flow characteristics. After testing the validity of the current model with past literature, and revealing the importance of real gas effects on the soot formation/oxidation process, 2D DNS have been conducted for all cases to investigate pressure effects on the process in a much deeper manner. It has been known for decades that in hydrocarbon flames soot production has been increased by increased ambient pressure. Such a behavior has been noted in the current work by investigating flames of 1, 5, 10 and 35 atm with the RGL and the IGL EOS models. These predictions show the effects of pressure on the soot production/oxidation processes. For the first three pressures the IGL EOS model has not deviated from the RGL EOS model significantly. However, for the flames of 35 atm the differences becomes highly significant. The unity Lewis (Le) number assumption on the soot formation/oxidation process has been studied in 3DDNSof atmospheric pressure flames for both theRGLand the IGL EOS models. The comparison of non-unity and unity Le number adiabatic atmospheric pressure flames has been done since it has a crucial importance to help to better understand the influence of the unity Le number assumption on these flames. In order to verify the acceptability of the unity Le assumption in hydrocarbon flames, similar trends in flame structure and soot properties should be obtained. Testing the validity of the unity Le assumption for the adiabatic atmospheric pressure flame predictions of the RGL and the IGL EOS models has revealed that coupling the EOS models with the unity Le number assumption under-predicts the flame temperature and soot properties. The known effect of the unity Le number on the enthalpy has been observed in these atmospheric pressure flames. Ignoring non-unity Le number effects has been shown to under-predict the soot quantities by at least an order of magnitude.

_{s}#### Recommended Citation

Korucu, Ayse, "Analysis of High Pressure H2/O2, H2/AIR AND KEROSENE/AIR Reacting Shear Flows" (2016). *All Dissertations*. 1837.

http://tigerprints.clemson.edu/all_dissertations/1837