COMBUSTION MODELLING IN A BIOMASS-FED ...

Int. Symp. on Convective Heat and Mass Transfer in Sustainable Energy April 26 – May 1, 2009, Tunisia

COMBUSTION MODELLING IN A BIOMASS-FED FURNACE Carlo Vincenzo Iossa*, Diego Lentini, Franco Rispoli, Paolo Venturini Dipartimento di Meccanica e Aeronautica Sapienza Università di Roma, Via Eudossiona, 18, 00184, Roma, Italy (* Corresponding author: [email protected])

Recent years have witnessed a rapid growth in biomass use as a bio-fuel for industrial and residential applications. Raw biomasses often derive from agricultural waste or residues of wood processing. Furnaces must be especially designed for biomass feeding, because of its high humidity content, or the need to adopt special technologies, such as packed bed combustion. While these technologies give an environmental pay-off in terms of GHG emissions and landfill saving, the key challenge is the control of the combustion process in order to comply with current stringent standards on nitride oxide and particulate emissions. This requires testing several configurations, a task only feasible by resorting to CFD (Computational Fluid Dynamics). In order for the results to make sense, accurate and reliable prediction models are required. In this context, the present work introduces a detailed description of both heterogeneous and homogeneous combustion in the turbulent regime, with account for finite-rate chemistry and pollutant emission predictions. It is applied to the combustion of a biomass resulting from wood processing; predictions are worked out via an in-house CFD solver based on a finiteelement formulation.

MODELING OF MULTIPHASE COMBUSTION Owing to the multi-phase nature of bio-fuel combustion, the present work models it as a two-stage process. Indeed, bio-fuel remains in the solid phase until the pyrolisis temperature is attained. A packed bed combustion model (Goh et al., 2001) is used in order to model this initial stage, returning the chemical composition, temperature and flow condition of the partly-burnt gas mixture produced by the biomass pyre. In the second, purely gaseous, combustion stage, the turbulence- chemistry interaction is modelled by adopting a stretched laminar flamelets formulation (Lentini, 1994, Ionta et al., 1995), making allowance for finite-rate chemistry. This approach allows handling species characterized by relatively slow chemistry, such as those correlated with CO oxidation. The proposed model is applicable to non-premixed turbulent combustion when an appropriately defined chemical time for the processes under consideration turns out to be longer than the turbulence time scales (Bray and Peters, 1994). The termo-chemical closure model is implemented on a FEM solver for parallel computing (Corsini et al., 2005).

TEST CASE The test case considered in this study is set up after an existing furnace; the biomass fuel is composed by wood particles. The fuel forms a uniform bed, see Fig. 1, therefore mass and heat gradients in the horizontal directions can be neglected, and accordingly the fuel bed can be modeled by using a 1D model. In the computation, the fuel bed is divided into 100 identical cells (in the vertical direction).


The initial fuel and air temperatures are both set to 298.15 K, with primary air injected from under the grate, at a volumetric rate of 125 l/min (Shin and Choi 2000). Combustion in the gas phase is simulated by adopting 130240 hexahedral elements. Nodes are concentrated in regions of high gradients, which are anticipated to occur close to the inlet, where air and fuel shear develops, then in regions of high combustion intensity, and near the walls. Both the air and the fuel gas resulting from solid phase combustion enter the furnace through a plane surface, at the same velocity. At the inlet, Dirichlet boundary condition are imposed. The value of the inlet velocity, also assumed as the reference velocity, is determined as 0.12 m/s on the basis of the actual air and fuel flow rates into the furnace. The furnace radius (745 mm) is adopted as a reference length. The Reynolds number is 9484, based on the above reference values, and air viscosity. The inlet turbulence intensity is imposed equal to 8%, resulting in an eddy viscosity about 80 times as large as the molecular one. The air entering the furnace undergoes an abrupt cross-section enlargement, leading to a deceleration, partly mitigated by the density decrease due to combustion. Further, the geometry of the combustor creates a wide recirculation zone just downstream of the inlet section. The streamtraces in Fig. 2 (dark ribbons) and the vortex cores (white lines) reveal a bottom-up U vortex in the lower zone. Two vortices beside the inlet zone also appear, as a result of the shear induced at the inlet. The pressure of a large vortical region near the flow inlet is demonstrated by the conical shape of pressure isosurface indicating the core of the vortex, developing parallel to the inlet velocity edge. The maximum computed averaged temperature is 1556 K. Figures 3 and 4 show contour plots of the predicted Favre-averaged mass fraction of CO and CO2, respectively.

REFERENCES Bray, K.N.C. and Peters, N. [1994], Laminar flamelets in turbulent flames, in Turbulent Reacting Flows (Libby, P.A. and Williams, F.A., Eds.), Academic Press, London, pp. 63-113. Corsini, A., Rispoli, F., Santoriello, A. [2005], A variational multi-scale high-order finite element formulation for turbomachinery flow computations, Computetional Method in Applied Mechanics and Engineering, vol. 194/45-47, pp. 4797-4823. Goh, Y.R., Yang, Y.B., Zakaria, R., Siddall, R.G., Nasserzadeh, V., Swithenbank., J. [2001], Development of an incinerator bed model for municipal solid waste incinerator. Combustion Science and Technology; Vol. 152, pp. 37-58. Ionta, P., Lentini, D., Riccucci, G., Rispoli, F. [1995], Prediction of gas turbine combustor flow by a finite element code, Aerotecnica Missili e Spazio, January-June, pp. 55-64. Lentini, D. [1994], Assessment of the stretched laminar flamelet approach for nonpremixed turbulent combustion, Combustion Science and Technology, vol. 100, pp. 95-122. Shin, D. and Choi. S. [2000], The combustion of simulated waste particles in a fixed bed, Combustion and Flame, vol. 121, pp. 167-180.


Figure 1. Section of the furnace, and inlet layout (bottom view).

Figure 2. Dimensionless velocity magnitude contour field and streamlines.

Figure 3. CO average mass fraction contour in the middle section.


Figure 4. CO2 average mass fraction contour in the middle section.