OCR Text |
Show In a previous work (Kilpinen et al, 1998) the effect of relevant operating and process parameters was investigated on N O x emission prediction from a G T C C burning an air-blown biomass gasification gas. For these simulations a simple plug flow model was adopted; thus the mixing between the staged air and the fuel gas was assumed occurring perfectly and instantaneously. Long time ago authors underlined how, in order to obtain representative predictions of many reactive systems, the time delay in gas mixing needs to be properly considered: Danckwerts (1957) and Zwietering. The effect of mixing delay on N O x reduction by reburning technique has been recently illustrated by Kilpinen et al. (1992) and by Alzueta et al. (1998). In this study the effect of modeling of the mixing between air and fuel gas have been investigated in a G T C C . The design and operational conditions has been treated in a similar way as by Kilpinen et al. (1998). T w o different ways of modeling the mixing delay have been tested. The aim of this paper is to compare the prediction obtained by the instantaneous mixing model with those obtained by the delayed mixing models. The attention will focus the affect of the variables characterizing the mixing delay in the models on the predictions of nitrogen emission. The effects of the operative variables of the G T C C predicted by the different models will be investigated. BACKGROUND In a previous work (Kilpinen et al., 1998), a 1- dimensional model was used for studying the nitrogen chemistry behavior during combustion in conditions typical for a Gas Turbine Combustion Chamber (GTCC). Such conditions were deduced from the specifications of several commercial gas turbines, from the work of Perry and Green (1997), and from the paper of Chomiak et al. (1989). The reference fuel was an air-blown biomass gasification gas. The gas composition assumed for the base case is reported in Table 1 (Leppalahti et al., 1997). To be noted is that the fuel nitrogen is completely bonded in the ammonia, so that the nitrogen conversion is here intended as ammonia oxidation. In an I G C C the fuel-nitrogen is often converted to mainly ammonia in the gasifier. During the S C O process ammonia is converted to mainly N 2 but also N O and N 0 2 ; H C N and N 20 may be also formed. Investigations were made on the effect of the several nitrogen compounds in the fuel gas. In the synthetic fuel gas adopted for the calculations such compounds will be referred to as "syngas nitrogen dopants". Air was always the oxidizing agent. The air-staging configuration in the combustion chamber was adopted. The total residence time in the G T C C was set to 15 ms, equally shared by each stage. Table 1. Composition of the Syngas for the Base Case Species CO H, CK, Nj 02 H20 C02 NH3 Mole-% 13 12 1 50.9 0 10 13 01 Mass-% 14.09 0.94 0.62 55.17 0 6.97 22.14 0.07 Three simple plug flow reactors (PFRs) were respectively the simulators of the three stages composing the G T C C in some 110 cases. Therefore the staged air was mixing instantaneously with the evolving gas along the combustion chamber, where isobaric and isothermal conditions were imposed. The following conditions defined the base case: 10 bar pressure, 1000 degC temperature, 3 stages with an air-to-fuel ratio of respectively 0.8, 1.0 and 2.0. Detailed nitrogen chemistry was implemented in the simulator. The reaction mechanism was taken from Kilpinen (1997). It is originally based on the mechanism developed by Glarborg et al. (1993a), (1995), and by Miller and Glarborg (1996). Some changes were also made in order to allow simulations at high pressure. Since the validation of these last modifications is not complete yet, some reference cases at atmospheric pressure were investigated in our studies here. The calculations of the homogeneous gas phase chemical kinetics were performed via the code S E N K I N (Lutz et al., 1993) of the commercial package C H E M K I N - H (Kee et al., 1993) by means of the mechanism as well as the postprocessors developed within this work. The results always showed that the nitrogen chemistry was more active along the first stage, in sub-stoichiometric conditions. There, the nitrogen compounds were converted to N 2 and N O . A minor formation of N O was predicted along the second stage, in stoichiometric conditions. The N O itself was partially reduced to N2. The third stage, in over-stoichiometric conditions, deled just with conversions NO/N02 . In the cases that will refer in this paper, the N 20 compound was never predicted in considerable amount. The variables that mainly affected the predictions were pressure and temperature. At lower pressure and higher temperature the oxidation of N H 3 to N O was enhanced along the firsts two stages, as well as the reduction of the above mentioned N O to N 2 along the second stage. Considering the pressure range from 1 to 10 bar and the temperature range from 1000 to 1300 degC, the best fixed nitrogen decrease, i.e. conversion of N H 3 to N2, at the gas turbine outlet was calculated at higher pressure and lower temperature. Quite similar fixed nitrogen reduction was achieved when the syngas dopant was either N H 3 , N O or N 0 2 . A worst result was predicted when H C N dopant and a remarkably higher nitrogen reduction was calculated in case of N2 0. Intermediate results were obtained when two of the above mentioned species were simultaneously in the syngas. |