CFD Simulations of coal gasification in a fluidized bed

Abstract

In the world there is a trend towards new forms of energy, generated mainly by the progressively increment in demand for non-renewable resources, which are becoming more expensive and scarce. This has motivated countries and companies to seek new alternatives and technologies related to the exploitation and use of other energy sources. An alternative is coal, whose global reserves constitute about 65% of fossil fuel reserves in the world (Shafiee and Topal, 2009). In this work, ANSYS-Fluent 14.0 is used for simulating coal gasification in a pilot-scale fluidized bed whose characteristics are detailed in the experimental work published by Ocampo et al. (2002). Heterogeneous reactions representing gasification and combustion have been programmed in C ++ and coupled to the software resolution algorithm. A molecular formula has been derived for the volatile material present in the coal particles. It has also been proposed an appropriate stoichiometry for the reaction that represents the release of pyrolysis gas. In addition, homogeneous reactions are also considered in the simulations performed. The bed reactor used by Ocampo et al. (2002) has a side feeding of coal. The height of the unit is 2 m. The bed has initially a height of 1m and is composed by sand and limestone. The simulations were performed on a 2D system with a structured mesh of 4000 cells. Multiphase Euler-Euler approach was used in order to solve the unsteady system. The particles of coal and limestone are spherical and uniform in size. Gidaspow drag model was adopted. Additionally, Gunn correlation was selected for modeling heat transfer between the gas phase and granular phases, as suggested in ANSYS-Fluent (2011) for granular systems. In this paper we present the results of the gas composition at the reactor outlet for different operating conditions and we compare them with experimental results reported by Ocampo et al. (2002). Furthermore the results are compared with those obtained by Armstrong et al. (2011) which have also been obtained by CFD simulation. A stationary bed temperature has also been reported throughout this process, in concordance with the experimental procedure. The results are in good agreement with experimental data and a significant improvement over previous simulations (which use the same technique) has been achieved. It constitutes an important validation of the results obtained in this work.