Numerical investigation of the combustion processes of various combustion regimes

Alganash, Blaid Sasi Abozeid (2015) Numerical investigation of the combustion processes of various combustion regimes. PhD thesis, University of Glasgow.

Abstract

This thesis concerns numerical investigations of the combustion behaviour of various combustion regimes. The simulations are based on modelling the flow of the fuels in the combustion devices. Computational fluid dynamics (CFD) modelling and analysis were used in three different works. FLUENT software, which is based on the finite volume method, is used to carry out all the simulations. Firstly, numerical simulations were carried out to investigate the turbulent non-premixed combustion of a mixture of methane (CH4) 90% and nitrogen (N2) 10%, on volume basis, inside an axis-symmetric cylindrical chamber (base case). The objective is to investigate the turbulent flow, flame propagation, temperature and species concentration and evaluate the effects of different reduced reaction mechanisms of methane and the influence of various turbulence models on them. The turbulent combustion inside the chamber occurs under a condition for which the equivalence ratio (ɸ) of 1.04 is used. Instead of using fully detailed chemical kinetics schemes and to reduce the computational costs, four global reduced chemical kinetics mechanisms are employed in the combustion model and they are named as (M-I, M-II, M-III and M-IV). The simulations, in which M-I is used, are performed by Renolds-Averaged Navier Stokes (RANS) approach with the three two-equation k-ϵ closures (standard, realizable and RNG) employed to model the turbulent flow. Concerning the chemistry-turbulence interaction, the finite-rate/eddy-dissipation model (FR/ED) is used. The first two of the above kinetics schemes are two-step reaction mechanisms and the other two are first-step and five-step reaction mechanisms, respectively. The latter one is used to assess the capability of FR/ED model for modeling such a mechanism. The influence of thermal radiation is also investigated by means of P-1 model. The standard k-ϵ model and realizable k-ϵ model are also modified and used in the course of simulations. Moreover, the reaction mechanism (M-II) is optimized to see its effects on the combustion process. The results are compared with the experimental data and gave good agreement. It is found that the best results are generally obtained using the modified standard k-ϵ model. Moreover, the simulation results using the realizable turbulence model are found to have large discrepancies compared to the experimental data. In comparison with the experimental data, the optimization of M-II (Em = 1.6x108 J/kmol) is found to have good results in terms of temperature. Increasing the dilution of the fuel by N2 is investigated. Four cases, CH4 (85, 80, 75 and 100%) on volume basis, are performed. The latter one concerns the combustion of pure methane. The results are compared with the base case and found that the base case is the best compromise to obtain the highest temperature in inside the chamber.
Secondly, an axis-symmetric combustion model based on the Euler-Lagrange approach was formulated to model the combustion of pulverized bituminous coal. Three cases with three different char oxidation models are presented. In case1 and case 2, the diffusion and kinetic/diffusion global char models are used, respectively. Whereas, to model char oxidation in case 3, the multi-surface reactions model is used. The volatiles released during the devolatilization stage, which is modelled using a single kinetic rate model, are treated as one species and its combustion is modelled using the FR/ED model. The predicted results have good agreement with the available experimental data and the best predictions are obtained from case 3. The results showed that the combustion inside the reactor was affected by the particulate size. It is found that the burnout of the particle with the diameter of 16 μm at the exit of the furnace is 100%. Whereas, the burnout of the particles with diameters of 84, 154, 222, 291 μm is approximately 86, 75, 35, 33, 29 %, respectively.
A number of simulations were carried out to find the best values of parameters suitable for predicting NOx pollutants. The chemical formation and reduction rates of NO are calculated by post-processing data obtained from the previously reacting flow simulations. This method is computationally efficient. For volatile-N is assumed that the nitrogen is released via the intermediates HCN and NH3. For char-N path way, it is assumed that all the nitrogen is released via the intermediate HCN. It is found that the assumption of the partition of volatile-N by 52% HCN, 10% NH3 and 38% NO has the best agreement with the experiment data. The influence of different operating parameters on the combustion process and NOx formation was investigated as well.
For the same operating conditions and the same particles size distribution, the combustion of pulverised biomass alone, represented by straw, was investigated followed by the investigation of its firing with coal. The former one show a promising results under such operating conditions. It is found that the temperature distribution when burning straw particles is nearly the same as that obtained from burning coal because all the saw particles are completely burned out inside the furnace when compared with the coal particles. The NOx model, in which the ratio of HCN to NH3 is suggested to be for the partitioning of volatile-N, shows that NO formation is reduced by approximately 20% for case I and 26% for case II at the exit of the furnace when compared to coal. For the latter one the results of co-firing blends of coal with 10, 20, 30 and 40% share of biomass are presented and show the influence of co-firing on the combustion process. Co-firing of straw with coal enhances the combustion behaviour and increases the burnout of coal particles compared to that of coal firing only. It is seen that the burnout of the particles with sizes 84, 154 and 222 μm is remarkably increased. On the other hand, the burnout of the other two particles (291 μm and 360 μm) does not show a great change. The share of 10% of straw shows the highest temperature.
Thirdly, Two-phase computational modelling based on the Euler–Euler was developed to investigate the heterogeneous combustion processes of carbon particles inside a newly designed combustion chamber. A transient simulation was carried out for a small amount of carbon powder situated in a cup which was located at the centre of the combustion chamber. A heat source was provided to initiate the combustion with the air supplied by three injection nozzles. The combustion simulations are performed for particle sizes with different diameters (0.5mm, 1mm, 1.5mm, 2mm, 2.5mm and 3mm). The particle of 1mm diameter is assigned to the baseline case. The results show that the combustion is sustained in the chamber, as evidenced by the flame temperature. It is shown that, up to a time of 0.55 s, the higher temperature was gained from the case of carbon particles with the diameter of 3 mm and burning the carbon particles with a diameter of 0.5 mm produces lower temperature. This may be attributed to the residence time of the carbon particles and the design of the burner. The larger particles stay longer than the smaller ones inside the chamber. This may due to the reason that the smaller particles follow the streamlines of the continuous phase and increasing the particle size leads to that the larger particles may deviate from the streamlines of the continuous phase and their slip velocity may increase resulting in enhancing convective transports of heat and species concentrations.
The influence of the chamber design was also investigated. The height of the chamber is doubled. With the same operating conditions, up to a time of approximately 0.55 s, it is found that burning carbon particles in the doubled height chamber produces higher temperature than the baseline case (particle diameter 1 mm) and after this time the opposite takes place. Most of the other cases do so.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: T Technology > T Technology (General)
• Colleges/Schools: College of Science and Engineering > School of Engineering > Systems Power and Energy
• Supervisor's Name: Paul, Dr. Manosh
• Date of Award: 2015
• Depositing User: Mr Blaid Sasi Abozeid Alganash
• Unique ID: glathesis:2015-7124
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 18 Mar 2016 09:04
• Last Modified: 13 Apr 2016 10:15
• URI: https://theses.gla.ac.uk/id/eprint/7124