Summary

The study conducted deals with interdisciplinary problems in the field of fluid mechanics, organic chemistry, thermodynamics, numerical methods in applied mathematics, and informatics.

The investigated problem of “combustion of pulverised fuel in a mixture of oxygen and recycled flue gas” is per se with significant degree of complexity due to the complex nature of the processes involved.

Detailed theoretical study of coal particle combustion processes is carried out. The basic mechanisms describing the release and combustion of volatiles, coal particle ignition, coal char oxidation and gasification, the formation of pollutants, and the radiative emittance from coal particles and from combustion gases are reviewed. The factors influencing these mechanisms and the specific effects appearing due to the high content of ${\mathrm{CO}}_2$ during oxyfuel combustion are identified and analysed.

The flame stability and the related design of the PF burner, the formation of pollutants and their control, and the changes in the heat transfer are considered the main challenges for the engineers who have to design the process of PF combustion in a mixture of oxygen and recycled flue gas.

For this purpose, a CFD-based model for oxyfuel pulverised coal combustion was developed. Basic models for turbulence, turbulence-chemical reactions and interactions, coal particle pyrolysis, and coal char oxidation and gasification, as well as gas emissivity, are reviewed and analysed.

Validation against experimental data obtained for cold flows and for oxycoal flames in swirl burners showed that $k-∊$ and Reynolds stress turbulence models can correctly predict the main flow pattern; however, the peak values of the backflow axial velocity component at the burner axis and of the tangential velocity components inside the IRZ were under-predicted. Therefore, it is expected that future LES-based calculations of oxycoal swirl flames would offer improvements over $k-∊$ and RSM predictions.

Three heterogeneous reactions have been recognised to play the major role in char burnout: char ${\mathrm{O}}_2$, char ${\mathrm{CO}}_2$, and char steam. Therefore, detailed validations of char combustion submodels were performed, applying apparent, intrinsic, and Langmuir-based model approaches. The models were validated against experimental data obtained for different coal qualities upon different reactor temperatures and partial pressures of the bulk, thus demonstrating their ability to predict the reaction rate in a wide range of operating conditions. External routines for all these models were developed and integrated into the global CFD code.

Experimental and numerical investigations were carried out on oxyfuel methane combustion with the aim of assessing the importance of the chemical effects due to high ${\mathrm{CO}}_2$ concentrations on homogeneous combustion rates. Experiments have been performed in a 25 kW furnace for flameless combustion, which provides the possibilities to achieve stable combustion of methane within a wide range of oxygen concentrations in the ${\mathrm{CO}}_2/{\mathrm{O}}_2$ mixture at constant reactor temperature. Four different oxidiser mixtures (${\mathrm{CO}}_2/{\mathrm{O}}_2$ and ${\mathrm{N}}_2/{\mathrm{O}}_2$, both with 21 vol% and 18 vol% ${\mathrm{O}}_2$) have been studied by detailed in-furnace measurements for flue gas compositions and temperature. In the case of combustion in ${\mathrm{N}}_2/{\mathrm{O}}_2$ atmosphere, the CO profiles obtained for different ${\mathrm{O}}_2$ concentrations overlap, thus demonstrating that changing the ${\mathrm{O}}_2$ concentration did not affect combustion rates with keeping the temperature constant. In the case of combustion in ${\mathrm{CO}}_2/{\mathrm{O}}_2$ atmosphere, the obtained CO concentrations were much higher than those in ${\mathrm{N}}_2/{\mathrm{O}}_2$ atmosphere. In contrast to ${\mathrm{N}}_2/{\mathrm{O}}_2$, the ${\mathrm{O}}_2$ concentrations had a significant impact on the production and consumption rates of CO in oxyfuel combustion. The results demonstrated that by elimination of the influence of molar heat capacity by keeping constant reactor temperature, ${\mathrm{CO}}_2$ dissociation by keeping reactor temperature below 900 °C, and thermal radiation by achieving flameless combustion in a small-scale furnace, it can be estimated that the high ${\mathrm{CO}}_2$ leads to (1) an increase in CO production rates and (2) slower consumption rates of CO. An increase of ${\mathrm{O}}_2$ in oxyfuel led to a reduction of this impact, however, further investigations on the exact mechanism are necessary.

Numerical predictions, based on RANS models and EDC, coupled with existing global kinetic mechanisms, can provide adequate simulation of methane oxidation in ${\mathrm{O}}_2/{\mathrm{CO}}_2$ atmosphere. However, new detailed mechanisms that consider the chemical effects of ${\mathrm{CO}}_2$ on the CO reaction rate, are needed.

Further more, the performance of two different PF burner designs, single central orifice type (SCO) and single annular orifice type (SAO), under oxycoal conditions has been examined in a downfired oxycoal pilot plant.

A swirl oxycoal flame was experimentally stabilised at the burner quarl by (1) increasing ${\mathrm{O}}_2$ concentration above 34 vol% without changes to the air-firing burner design, and (2) by modifications of the burner geometry thus changing its aerodynamics.

A novel fluid dynamics-based approach for oxycoal swirl flame stabilisation, based on a thorough understanding of the underlying interactions between burner aerodynamics and reaction kinetics, is introduced. Measures for stabilisation of an oxycoal swirl flame with ${\mathrm{O}}_2$ contents equal to and lower than those in air are derived regarding compensation of the effects of higher molar heat capacity, heat consumption, and volume changes due to gasification reactions.

CFD-based burner design applying the measures for stabilisation of an oxycoal swirl flame led to the realisation of a stable flame at 23 vol% ${\mathrm{O}}_2$ concentrations for SCO burners and at 19 vol% for SAO burners. The SAO burner is scaled up from 40 kW to 100 kW based on the constant residence time-scaling approach; after that it is successfully tested in the pilot plant, thus demonstrating the viability of the scaling approach and of the measures derived.

Visual inspection and global measurements confirmed flame stabilisation at the burner quarl under oxycoal conditions with 21 vol% ${\mathrm{O}}_2$ in the ${\mathrm{O}}_2/{\mathrm{CO}}_2$ mixture (dry recycle). Furthermore, the burner was successfully tested in both oxycoal conditions using RFG (wet recycling) and in conventional air combustion. The tests demonstrated the burner’s ability to form a stable swirl flame and to provide good burnout in all three combustion modes, namely, in air, in ${\mathrm{O}}_2/{\mathrm{CO}}_2$, and in ${\mathrm{O}}_2$/RFG.

The design of an industrial-scale oxy-firing burner in the range of several MWth, however, should be based on a modification of a conventional air-firing utility-scale burner according to the measures for oxycoal swirl flame stabilisation.

Finally, in order to investigate the influence of different burner operation parameters and of the operation mode on NO_x emissions during PF oxycoal combustion, comparative experiments have been carried out with the scaled-up burner described in Section 6.4. Three combustion modes, namely, (1) air; (2) ${\mathrm{O}}_2/{\mathrm{CO}}_2$, and (3) ${\mathrm{O}}_2$/RFG have been considered. The impact of five different burner operational parameters has been studied. The main results demonstrated that:

Based on these results, it can be concluded that wet recycling is preferable in terms of NO_x emissions. However, further investigations on the influence of water vapour on NO_x mechanisms during oxyfuel coal combustion are needed.