Numerical study of the characteristics of CNG, LPG and Hydrogen turbulent premixed flames
thesisposted on 2015-11-24, 16:32 authored by Mohamed A. Abdel-Raheem
Numerical simulations have proven itself as a significant and powerful tool for accurate prediction of turbulent premixed flames in practical engineering devices. The work presented in this thesis concerns the development of simulation techniques for premixed turbulent combustion of three different fuels, namely, CNG, LPG and Hydrogen air mixtures. The numerical results are validated against published experimental data from the newly built Sydney combustion chamber. In this work a newly developed Large Eddy Simulation (LES) CFD model is applied to the new Sydney combustion chamber of size 50 x 50 x 250 mm (0.625 litre volume). Turbulence is generated in the chamber by introducing series of baffle plates and a solid square obstacle at various axial locations. These baffles can be added or removed from the chamber to adapt various experimental configurations for studies. This is essential to understand the flame behaviour and the structure. The LES numerical simulations are conducted using the Smagorinsky eddy viscosity model with standard dynamic procedures for sub-grid scale turbulence. Combustion is modelled by using a newly developed dynamic flame surface density (DFSD) model based on the flamelet assumption. Various numerical tests are carried out to establish the confidence in the LES based combustion modelling technique. A detailed analysis has been carried out to determine the regimes of combustion at different stages of flame propagation inside the chamber. The predictions using the DFSD combustion model are evaluated and validated against experimental measurements for various flow configurations. In addition, the in-house code capability is extended by implementing the Lewis number effects. The LES predictions are identified to be in a very good agreement with the experimental measurements for cases with high turbulence levels. However, some disagreement were observed with the quasi-laminar case. In addition a data analysis for experimental data, regarding the overpressure, flame position and the flame speed is carried out for the high and low turbulence cases. Moreover, an image processing procedure is used to extract the flame rate of stretch from both the experimental and numerical flame images that are used as a further method to validate the numerical results. For the grids under investigation, it is concluded that the employed grid is independent of the filter width and grid resolution. The applicability of the DFSD model using grid-independent results for turbulent premixed propagating flames was examined by validating the generated pressure and other flame characteristics, such as flame position and speed against experimental data. This study concludes that the predictions using DFSD model provide reasonably good results. It is found that LES predictions were slightly improved in predicting overpressure, flame position and speed by incorporating the Lewis number effect in the model. Also, the investigation demonstrates the effects of placing multiple obstacles at various locations in the path of the turbulent propagating premixed flames. It is concluded that the pressure generated in any individual configuration is directly proportional to the number of baffles plates. The flame position and speed are clearly dependent on the number of obstacles used and their blockage ratio. The flame stretch extracted from both the experimental and numerical images shows that hydrogen has the highest stretch values over CNG and LPG. Finally, the regime of combustion identified for the three fuels in the present combustion chamber is found to lie within the thin reaction zone. This finding supports the use of the laminar flamelet modelling concept that has been in use for the modelling of turbulent premixed flames in practical applications.
- Mechanical, Electrical and Manufacturing Engineering
Publisher© M.A. Abdel-Raheem
Publisher statementThis work is made available according to the conditions of the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) licence. Full details of this licence are available at: https://creativecommons.org/licenses/by-nc-nd/4.0/
NotesA Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.