The aim of the current thesis is to model the various fluid-particle interactions in an 150g/h bubbling fluidised bed reactor. Mass, momentum and heat transfer from the bubbling bed of the reactor to the discrete biomass particles are modelled and analysed. The Eulerian-Eulerian approach is used to model the bubbling behaviour of the sand, which is treated as a continuum. The biomass particle motion inside the reactor is computed by integrating the equations of motion using drag laws, dependent on the local volume fraction of each phase. Reaction kinetics are also incorporated in the computational code according to the literature, using a two stage global model which takes into account the intra-particle secondary reactions due to the catalytic effect of char, resulting on secondary vapour cracking.

Co-combustion of biomass with coal has been investigated in a 0.15 m diameter and 2.3 m high fluidised bed combustor under various fluidisation and operating conditions. Biomass materials investigated were chicken waste, rice husk, palm kernel shells and fibres, refuse derived fuel and wood wastes. These were selected because they are produced in large quantities particularly in the Far East. The carbon combustion efficiency was profoundly influenced by the operating and fluidising parameters in the decreased following order: fuel properties (particle size and density), coal mass fraction, fluidising velocity, excess air and bed temperature. The smaller particle size and lower particle density of the fuels (i.e. coal/chicken waste, coal/rice husk and coal/wood powder), the higher carbon combustion efficiency obtained in the range of 86-90%, 83-88%, 87-92%, respectively. The carbon combustion efficiency increases in the range of 3% to 20% as the coal fraction increased from 0% to 70%, under various fluidisation and operating conditions. Also, the carbon combustion efficiency increases with increasing excess air from 30- 50% in the range of 5 - 12 % at 50% coal mass fraction in the biomass mixture. However, further increased of excess air to 70% will reduced the carbon combustion efficiency. Relatively, increasing fluidising velocity contributed to a greater particle elutriation rate than the carbon to CO conversion rate and hence increased the unburned carbon. Furthermore, the bed temperature had insignificant influence of carbon combustion efficiency among the biomass fuels. Depending upon excess air ranges, fluctuations of CO emissions between 200 - 1500 ppm were observed when coal added to almost all biomass mixtures. In ash analyses, the percentages of unburned carbon were found to have increased in the range 3 to 30% of the ash content with the increases of coal fraction in the coal! biomass mixture. Furthermore, no fouling, ash deposition and bed agglomeration was observed during the combustion runs for all tests due to lower operating bed temperature applied. Lastly, a simple model was developed to predict the amount of combustion in the freeboard. This study demonstrated the capability of co-firing biomass with coal and also demonstrated the capability to be burnt efficiently in existing coal-fired boilers with minimum modification.

We study in this thesis the influence of prescribed flows on three distinct types of flame, namely flame balls, premixed flames and triple flames. The interaction between flame propagation and fluid flow is examined using asymptotic analyses and numerical simulations of thermo-diffusive models. We consider first the effect of a flow of hot inert gas, either a source or a sink, located at the origin of flame balls. It is shown that the flow gives rise to new kinds of flame balls characterised by having nonzero burning speeds, which we refer to as generalised flame ball. The structure and stability of these flames are found to strongly depend on the direction and rate of the flow. When the flow is a source, there exist two stationary solutions, small stable flames and large unstable flames including the classical Zeldovich flame balls. When the flow is a sink, on the other hand, only large unstable flames are predicted except for sufficiently large values of the Lewis number and large negative values of the flow rate, where three flame balls exist, the medium one being stable.

The importance of turbulent premixed combustion comes from the fact that it is present in many engineering applications such as spark ignition engines and gas turbines. It is also present in explosion incidents where premixed mixture is burnt rapidly. Interaction of the flame with highly turbulent flows causes flame acceleration. Unsteady turbulent premixed combustion is a complex engineering phenomenon that is still not well understood. The nature of the unsteady turbulent premixed flame is so unstable, much wrinkled and too fast. A wide range of turbulence scales are formed in the flow region around the flame, which produce a variety of premixed combustion regimes during the course of flame propagation. Studies on premixed combustion have been a need for long time. Despite of its importance as a physical reference, combustion experiment is an expensive tool in both research and industry. Advances in premixed combustion predictive tools together with growing computational power have made computer simulation of flow and combustion in practical systems an effective alternative to experimental tests.

Flame-acoustic wave interactions have been studied widely in the combustion community; however the whole physicochemical mechanism is still not clear. The present research aims to further investigate the unfinished areas and focus on the acoustic excitation effect on flow and flame characteristics. An experimental approach has been adopted, and the results have been analysed and discussed. PIV diagnostic system and high speed schlieren visualisation system have been applied and built together with a signal synchronisation system which enables the phase-locked observation. The acoustic field and the interactions between acoustic waves and flames in an acoustically excited tube were studied in detail. The acoustic wave induced by external excitation and its effect on both cold jet fuel flow and jet flames have been investigated. Results indicate that the flame behaviour is affected mostly by the variation of the excitation frequency. The infrasound (Frequency < 20 Hz) was observed to have less influence on air motion. The natural flame flicker was suppressed by the forcing infrasound because of the excitation effect on the fuel jet. In the case of harmonic frequencies at which a standing wave is formed, the ambient air and flame were less affected by the excitation in the velocity node area. In contrast, the cold fuel jet, flame and ambient air experienced large velocity variations in the anti-node area. At non-resonant frequencies, which are between the 1st and 2nd harmonic frequencies (65 Hz to 220 Hz), the flame pattern and luminosity were very different in each excitation case. The frequency analysis of flames has shown that flame/acoustic coupling behaviour results in a complex nonlinear coupling effect. The excitation frequency may be coupled with the sub-harmonic frequencies and the flame flickering frequency. These frequencies were then found to couple with each other and create complex nonlinear frequency couplings. It can be seen that the surrounding air, cold jet fuel flow and flames are strongly affected by the excitation frequencies, phase angles and the nozzle location relative to the tube, and the coupling effect creates complex flame dynamics.

Mathematical modelling of the turbulent combustion process is becoming increasingly applied in calculations to assist in the design and analysis of practical combustion devices for efficiency-improvement and emission reduction. The current requirement to accurately predict pollutant emissions in many applications has increased the need for linking turbulent flow calculations and finite- rate chemistry effects in a rigorous way. Several methodologies are available for modelling such interactions, including the transported probability density function (PDF) approach and the conditional moment closure (CMC) method. Although in the early stages of its development, the CMC method has been shown to be a promising technique for predicting a wide range of practical problems. These include both premixed and non-premixed combustion, relatively slow chemistry effects, and ignition and extinction phenomena. This study concerns the CMC approach, and addresses the application of a number of models to a wide range of flows displaying varied compositions and geometries, including hydrogen and methane, and rim-stable and lifted jets. The impact of the choice of chemistry mechanism is considered for all the flows, and a higherorder CMC chemistry closure is investigated for the hydrogen flames. Analysis is made as to the ability of a parabolic CMC model to predict such flows, and the performance of the sub-model interactions is also reported on. The method of coupling the turbulent mixing field and the chemical kine tics is also investigated, and the effects of Reynolds stress and k - E turbulence closures upon subsequent CMC calculations are compared in all the flows considered. Overall, the results shown and conclusions drawn are very promising with respect to the possible future development of CMC. Requirements essential for this step forward of CMC methodologies for use in modelling practical geometries are specified, and an outline for the continuation of these studies is presented.

No description available.

The aims of this thesis are to apply the Large Eddy Simulation (LES) and beta Probability Density Function (β- PDF) for the simulation of turbulent non-premixed reacting flow, in particularly for the predictions of soot and NO production, and to investigate the radiative heat transfer during combustion process applying Discrete Ordinates Method (DOM). LES seeks the solution by separating the flow field into large-scale eddies, which carry the majority of the energy and are resolved directly, and small-scale eddies, which have been modelled via Smagorinsky model with constant Cs (Smagorinsky model constant) as well as its dynamic calibration. This separation has been made by applying a filtering approach to the governing equations describing the turbulent reacting flow. Firstly, LES technique is applied to investigate the turbulent flow, temperature and species concentrations during the combustion process within an axi-symmetric model cylindrical combustion chamber. Gaseous propane (C3H8) and preheated air of 773K are injected into this cylindrical combustion chamber. The non-premixed combustion process is modelled through the conserved scalar approach with the laminar flamelet model. A detailed chemical mechanism is taken into account to generate the flamelet. The turbulent combustion inside the chamber takes place under a fuel-rich condition for which the overall equivalence ratio of 1.6 is used, the same condition was used by Nishida and Mukohara [1] in their experiment. Secondly, the soot formation in the same flame is investigated by using the LES technique. In this thesis, the soot formation is included through the balance equations for soot mass fraction and soot particle number density with finite rate kinetic source terms to account for soot inception/nucleation, surface growth, agglomeration and oxidation. Thirdly, the NO formation in the flame is studied by applying the LES. The formation of NO is modelled via the extended Zeldovich (thermal) reaction mechanism. A transport equation for NO mass fraction is coupled with the flow and composition fields. Finaly, the radiative heat transfer in the flame is investigated. Both the luminous and non-luminous radiations are modelled through the Radiative Transfer Equation (RTE). The RTE is solved using the Discrete Ordinates Method (DOM/Sn) combining with the LES of the flow, temperature, combustion species and soot formation. The computed results are compared with the available experimental results and the level of agreement between measurements and computations is quite good.

With the depletion of hydrocarbon fuels, the hydrogen-hydrocarbon combustion system provides a good solution for the transition period from a hydrocarbon-based energy sys- tem to a hydrogen-based energy system because of its desirable combustion characteristics and the low level of modification to current combustion systems. Though extensive re- search has been carried out to investigate the combustion process of hydrogen-hydrocarbon fuels, no experiments have been reported to study the Particulate Matter (PM) formations in hydrogen-hydrocarbon combustion systems. To measure the PM concentrations in a laminar diffusion flame, a new optical diagnostic technique, called Cone-Beam Tomographic Three Colour Spectrometry (CBT-TCS) has been developed to measure the spatially distributed temperature, soot diameter and soot volume fraction. This technique is based on the principle of three colour pyrometry, but uses a more rigorous light scattering model to calculate the soot diameter and soot volume fraction. The cone beam tomography technique has also been used to reconstruct the 3D property fields from the 2D flame images. The detailed theoretical principles, the exper- imental setup, the optical considerations, the reconstruction algorithm and the sensitivity analysis are all introduced. The CBT-TCS technique has been successfully applied to several laminar diffusion flames to study the PM formation. The temperature and soot volume fracction profiles measured by CBT-TCS for a ethylene laminar diffusion flame are consistent with the data reported by Snelling et al. [77]. The helium-ethylene-air flame tests show that adding helium reduces the PM formation (due to the dilution effect). The hydrogen-ethylene-air flame tests show that adding hydrogen is more effective in reducing the PM formation due to the combined effect of dilution and direct chemical reaction. A PM sampling system has also been de- veloped to verify the PM size distributions measured by CBT-TCS. The comparison results show that the CBT-TCS tends to overestimate the particle size. Several optical engine experiments have also been undertaken to investigate the effect of adding hydrogen on the PM emissions from a Gasoline Direct Injection (GDI) engine. The hydrogen-ethylene engine tests show that adding hydrogen can reduce the PM emissions without sacrificing the power output. The hydrogen-base fuel (65% isooctane and 35% toluene) tests show that adding hydrogen can improve the combustion stability and reduce the PM emissions, especially at low load. Adding 5% stoichiometric of hydrogen can reduce the total PM number concentration by 90% for a stoichiometric mixture and 97% for richer mixture at low load. At high load, adding 10% stoichiometric of hydrogen can also reduce the total PM number concentration by 85% for richer mixture but has little effect upon the stoichiometric mixture.