Jan Erik Johnsson

Fuel nitrogen conversion in solid fuel fired systems

Abstract Understanding of the chemical and physical processes that govern formation and destruction of nitrogen oxides (NO x ) in combustion of solid fuels continues to be a challenge. Even though this area has been the subject of extensive research over the last three decades, there are still unresolved issues that may limit the potential of primary measures for NO x control. In most solid fuel fired systems oxidation of fuel-bound nitrogen constitutes the dominating source of nitrogen oxides. The present paper reviews some fundamental aspects of fuel nitrogen conversion in these systems, emphasizing mostly combustion of coal since most previous work deal with this fuel. However, also results on biomass combustion is discussed. Homogeneous and heterogeneous pathways in fuel NO formation and destruction are discussed and the effect of fuel characteristics, devolatilization conditions and combustion mode on the oxidation selectivity towards NO and N 2 is evaluated. Results indicate that even under idealized conditions, such as a laminar pulverized-fuel flame, the governing mechanisms for fuel nitrogen conversion are not completely understood. Light gases, tar, char and soot may all be important vehicles for fuel-N conversion, with their relative importance depending on fuel rank and reaction conditions. Oxygen availability and fuel-nitrogen level are major parameters determining the oxidation selectivity of fuel-N towards NO and N 2 , but also the ability of char and soot to reduce NO is potentially important. The impact of fuel/oxidizer mixing pattern on NO formation appears to be less important in solid-fuel flames than in homogeneous flames.

Formation and reduction of nitrogen oxides in fluidized-bed combustion

Abstract The subject is critically reviewed. The following topics are covered: devolatilization of nitrogen compounds from solid fuel; homogeneous gas-phase reactions; oxidation of char-nitrogen; gas-solid reactions; heterogeneous catalytic reactions; and ways in which knowledge of the chemistry and kinetics can be applied to combustor modelling and operation.

The thermal DeNOx process: Influence of partial pressures and temperature

Abstract The effect of partial pressures of the reactants in the Thermal DeNOx process has been investigated in flow reactor experiments. The experiments were performed at atmospheric pressure for temperatures ranging from 923 to 1373 K. Initial concentrations were varied for NH 3 NO ( 400 200 , 1000 500 , 2800 1400 ppm ) and O2 (0–50%). the data confirm earlier observations that in the temperature range covered, presence of O2 is required in order for NO to be reduced by NH3. As the initial O2 concentration is increased, the lower boundary for the process is shifted towards lower temperatures. The temperature range for NO reduction is widened, but the NO reduction potential decreases. At high oxygen concentrations the maximum NOx reduction is below 40%. Under these conditions, significant amounts of NO2 and N2O are formed. Two mechanisms for N2O formation in Thermal DeNOx have been identified. One is active at higher temperatures and low O2 concentrations, while the other, which presumably involves NO2 as a precursor, is dominant at lower temperatures and high O2 levels. The implications of the results for application of Thermal DeNOx in high pressure systems such as pressurized fluidized bed combustion is discussed. Comparisons of the experimental data with recent chemical kinetic models indicate that the detailed chemistry of the Thermal DeNOx system is not completely understood.

Formation and reduction of NOx in pressurized fluidized bed combustion of coal

Abstract This paper described an experimental and modelling study of NOx formation and reduction in pressurized fluidized bed combustion (PFBC) of coal. The aim was to evaluate the Thermal DeNOx process with NH3 injection and staged combustion in PFBC as measures for the reduction of NOx emissions, and to develop a mathematical model for the emission of NOx. Experiments with NH3 injection and staged combustion in a 1.6 MW PFBC test rig burning bituminous coal showed that both methods were able to reduce the emission of NOx by 50–70% under optimum conditions. Laboratory experiments were performed to elucidate the nitrogen chemistry under PFBC conditions and were used as input to the model. These experiments included an investigation of the release of nitrogen during devolatilization of coal and a kinetic study of important reactions for NOx formation and reduction catalyzed by char and bed material sampled from the test rig. The model is based on the two-phase theory of the bed and includes coal devolatilization, combustion of char and volatile matter, and NOx formation and reduction by homogeneous and heterogeneous reactions. The NOx emissions calculated with the model are in qualitative agreement with the emissions from the 1.6 MW test rig.

Optimisation of a wet FGD pilot plant using fine limestone and organic acids

Abstract The effects of adding an organic acid or using a limestone with a fine particle size distribution (PSD) have been examined in a wet flue gas desulphurisation (FGD) pilot plant. Optimisation of the plant with respect to the degree of desulphurisation and the residual limestone content of the gypsum has been the aim of the work. In contrast to earlier investigations with organic acids, all essential process parameters (i.e. gas phase concentration profiles of SO 2 , slurry pH profiles, and residual limestone in the gypsum) were considered. Slurry concentrations of adipic acid in the range of 0– 7 mM were employed. The overall degree of desulphurisation in the plant increased from 83% at 0 mM to 90% at 3 mM and the residual limestone level was reduced from 4.6 to 1.4 wt % . Increasing the slurry concentration of adipic acid above 3 mM gave only a slightly higher degree of desulphurisation. The wet FGD model of Kiil et al. (Ind. Eng. Chem. Res., 37 (1998) 2792) was extended to include buffer systems and verified against experimental data. Subsequently, the model was used as a tool to identify the optimal organic acid dissociation constants (as pK a values) and concentration levels at different operating conditions. At a holding tank pH of 5.5 and a temperature of 50°C, simulations with Bryozo limestone and a monoprotic buffer suggested that the optimum pK a value is between 4.5–5.5 and 5.5–6.5 with respect to the degree of desulphurisation and the residual limestone level, respectively. Adipic acid has pK a values close to these ranges ( pK 1 =4.40 and pK 2 =5.41 at 50°C). Changing limestone type (in the absence of organic acids) to one with a lower average particle size (i.e. from 20 to 4 μm ) increased the overall measured degree of desulphurisation from 83 to 87% and reduced the residual limestone level from 4.6 to 1.3 wt % . Increasing the holding tank pH level from 5.5 to 5.8 affected the degree of desulphurisation and the residual limestone level only slightly. At holding tank pH levels between 5.88 and 5.90, a high degree of desulphurisation was observed, but the residual limestone content in the gypsum increased to somewhere between 19 and 30 wt % , making this pH range unsuitable for use in a full-scale plant. The investigations have shown that both the addition of organic acids and the use of a limestone with a fine PSD can be used to optimise wet FGD plants.

Simulation studies of the influence of HCl absorption on the performance of a wet flue gas desulphurisation pilot plant

Abstract The mathematical model of Kiil et al. (Ind. Eng. Chem. Res. 37 (1998) 2792) for a wet flue gas desulphurisation (FGD) pilot plant was extended to include the simultaneous absorption of HCl. In contrast to earlier models for wet FGD plants, the inclusion of population balance equations for the limestone particles enabled a quantitative description of the influence of HCl absorption on essential process parameters such as the degree of desulphurisation and the residual limestone level of the gypsum produced. Simulations showed that the presence of 100 ppmv HCl in the flue gas reduced the degree of desulphurisation from 85 to 84% and increased the residual limestone level of the gypsum from 2.1 to 2.4 wt %. It was found that these undesired effects from HCl absorption could be counteracted by adding adipic acid to the slurry in a concentration of about 1 mM . The influence of holding tank pH and the inlet flue gas concentration of SO2 on the degree of desulphurisation and the residual limestone level was found to be almost the same irrespective of HCl was present ( 100 ppmv ) in the flue gas or not. The results presented are of importance in the analysis of the performance of wet FGD plants installed at power plants firing coals of varying Cl contents.

KINETIC MODELING OF FUEL-NITROGEN CONVERSION IN ONE-DIMENSIONAL, PULVERIZED-COAL FLAMES

Abstract A detailed reaction mechanism for converting fuel nitrogen to nitric oxide and molecular nitrogen in gas flames is used to model nitrogen chemistry in fuel-rich coal-dust/oxidizer flat flames. In the devolatilization zone a simplified pyrolysis model is used and a hydrocarbon oxidation scheme supplement the nitrogen reactions. The predicted distributions of nitrogenous species (HCN, NH3, NO and N2) are compared to time-resolved experimental data obtained for two stoichiometrics and coal-types. The model accounts for the radical removal on particle surfaces, and various heterogeneous reactions for reduction of NO are considered. In the devolatilization flame zone the coal-N conversion mechanisms appear to be augmented by physical and chemical processes occurring in the vicmity of the devolatilizing particles. The current analysis supported by other investigations indicate that processes such as formation of fuel-rich volatile clouds and heterogeneous reduction of NO on surfaces of coal-particles a...

A study of benzene formation in a laminar flow reactor

The formation of benzene in a series of fuel-rich premixed reactant systems with a constant carbon-to-oxygen ratio of approximately 0.6 is investigated experimentally in a laminar flow reactor at temperatures between 1073 and 1823 K and at atmospheric pressure. The experimental data are compared to model results using a chemical kinetic mechanism based on the work of Pope and Miller. Modifications to their mechanism include changes in the reaction subsets of C 2 H 2 , C 3 H 4 , and 1,3-C 4 H 6 . The experimental data show that benzene formation may exhibit two distinct peaks as a function of the reaction temperature. A high-temperature peak is observed between 1500 and 1600 K, and it appears with a similar magnitude of concentration for all sets of reactants. A low-temperature peak is observed between 1200 and 1300 K for reactant sets CH 4 /C 2 H 2 , CH 4 /C 3 H 4 , and CH 4 /1,3-C 4 H 6 . The low-temperature peak is comparable in magnitude for the CH 4 /C 2 H 2 and CH 4 /1,3-C 4 H 6 mixtures, while it is 5 times larger in the CH 4 /C 3 H 4 system. In general, there is good agreement between modeling and experimental results. However, some improvements are needed, in particular in the acetylene chemistry and the initiation kinetics for 1,3-butadiene.

Influence of mixing on the SNCR process

Abstract An experimental and theoretical investigation of mixing in the SNCR process was performed. The experiments were carried out in a bench scale reactor using the flue gas from a natural gas burner as the main gas and injection of a jet of NH 3 mixed with carrier gas in crossflow. The results show a dependency on the carrier gas flow at temperatures above the optimum temperature for NO reduction. No dependency on the variation of the O 2 concentration in the carrier gas from 0 to 21 vol% was observed. It was found that an increasing momentum ratio of the jet to the main gas improves the NO reduction up to a limiting value of the momentum ratio of approximately 20. Above this value no further improvement was observed. Chemical kinetic modelling of the initiating reactions involving NH 3 showed that the reaction with OH radicals is the primary initiating reaction. It was also shown that process performance is influenced by the O 2 concentration in the flue gas. The experimental results were used to verify the droplet diffusion model proposed by Ostberg and Dam-Johansen (1995, Chem. Engng Sci. 50 , 2061–2067), using an empirical kinetic model valid for 4 vol% O 2 in the reacting gas.

Modelling of NOx emissions from pressurized fluidized bed combustion—a parameter study

Abstract Simulations with a mathematical model of a pressurized bubbling fluidized-bed combustor (PFBC) combined with a kinetic model for NO formation and reduction are reported. The kinetic model for NO formation and reduction considers NO and NH3 as the fixed nitrogen species, and includes homogeneous reactions and heterogeneous reactions catalyzed by bed material and char. Simulations of the influence of operating conditions: air staging, load, temperature, fuel particle size, bed particle size and mass of bed material on the NO emission is presented and compared to results from the literature. In general, the trends predicted by the model are in good agreement with the experimental observations. A rate of production analysis for the nitrogenous species is used to identify the important reactions for formation and reduction of NO. According to the kinetic model, the reduction of NO by CO catalyzed by bed material is very important, especially at low temperatures where the CO concentration in the bed is high. The sum of the reduction of NO by char and by CO catalyzed by char increases with increasing temperature, but is lower than usually attributed to these reactions. In the temperature range 973–1273 K, 20–30% of the fuel-N in the form of NH3 is oxidized catalytically to N2 over bed material and so this reaction is important for a low conversion of fuel-N to NO. The importance of the homogeneous oxidation of NH3 to NO and reduction of NO by NH3 increases with increasing temperature. The sensitivity of the simulated NO emission with respect to hydrodynamic and combustion parameters in the model is investigated and the mechanisms by which the parameters influence the emission of NO is explained. The analysis shows that the most important hydrodynamic parameters are the minimum fluidization velocity, the bubble size, the bubble rise velocity and the gas interchange coefficient between bubble and dense phase. The most important combustion parameters are the rates of CO and CH4 combustion and the CO (CO + CO 2 ) ratio from char combustion.