Pia Kilpinen

Homogeneous N2O chemistry at fluidized bed combustion conditions : a kinetic modeling study

The importance of homogeneous gas-phase reactions of fuel-nitrogen species to the formation of N2O in fluidized bed combustion was studied based on detailed chemical kinetic modeling. A kinetic scheme consisting of over 250 elementary reactions was applied for an ideal plug flow reactor, and the most important reaction paths leading to N2O were identified and quantified for a number of conditions relevant for fluidized bed combustion. In addition, the effect of different operating parameters on N2O formation was investigated, and the results were compared to the existing data from laboratory and full-scale experiments. Calculations showed that if fuel-nitrogen species in form of simple cyano species (HCN) enters a fuel-lean gas phase between 1000 and 1200 K, a significant conversion to N2O is found. The N2O formation occurs principally through the reaction NCO+NO→N2O+CO where NCO originates from HCN mostly via the reaction HCN+O→NCO+H For ammonia-based fuel-nitrogen derivates (NHi) hardly any N2O was formed. Calculations indicated that N2O formation increases steeply as the temperature decreases. At high temperatures less N2O is produced because the key intermediate NCO is rapidly removed by the radicals, mostly via the reactions NCO+H→NH+CO,NCO+O→NO+CO Generally, the trends obtained in this study are in good agreement with existing data from laboratory and full scale measurements. Thus, the work indicates that the homogeneous gas-phase reactions are an important contributor to the N2O formation in fluidized bed combustion.

A reduced mechanism for nitrogen chemistry in methane combustion

A four step mechanism for combustion of methane in perfectly stirred reactors with special emphasis on formation and destruction of hydrocarbon radicals has been developed using steady state and partial equilibrium assumptions for minor species. The reduced mechanism has been extended to include the nitrogen chemistry with NO and HCN as independent reactive species. The reduced nitrogen scheme includes thermal NO and prompt NO formation, as well as the NO to HCN recycle reactions, and conversion of HCN to NO and N2. We have tested the reduced mechanism by comparing perfectly stirred reactor calculations performed with full and reduced chemistry over a wide range of stoichiometries, temperatures and residence times. The reduced model generally provides a good description of the methane oxidation process as well as formation and destruction of nitrogen oxides. However, at low temperatures or very fuel-rich conditions reduced model predictions deteriorate, partly due to neglection of the C2-chemistry and partly because the OH partial equilibrium assumption becomes less accurate.

The effects of pressure, oxygen partial pressure, and temperature on the formation of N2O, NO, and NO2 from pulverized coal

The main features of a new, pressurized, entrained-flow reactor are described and results presented of experiments investigating the formation of nitrogen oxides (N2O, NO, and NO2) from pulverized Polish coal, burned in the reactor at temperatures (T) 800–1300°C, pressures (p) 1–20, bar and oxygen partial pressures (pO2) 0.05–2.4 bar. The experimental results are compared with the results of detailed gas-phase kinetic calculations at 850°C, where HCN was used as the source of coal-nitrogen, and H2, H2O, CO and C2H4 were used to describe the gaseous products of pyrolysis and char combustion. The new reactor made it possible to control the experimental conditions with high precision. Regression equations were obtained between the dependent, y-variables (conversions of fuel-N to N2O, NO, and NgOy) and independent, x-variables (p, pO2 and T). NO formation decreased sharply with pressure, and increased, but not as strongly, with oxygen partial pressure and temperature. Total pressure and oxygen partial pressure did not affect N2O formation in the pO2 range 0.15-0.6 bar. At higher pO2 the conversion of fuel-N to N2O decreased with both total pressure and oxygen partial pressure. An increase in temperature strongly reduced N2O formation, independently of pressure and pO2. No N2O was found at or above 950°C. NO2 was formed in sufficient concentrations to find a regression model at high partial pressures (> 0.5 bar) of oxygen. Like N2O formation, the yield of NO2 decreased with temperature. But like NO, and in contrast to N2O, the formation of NO2 increased with pO2. NO was the only nitrogen oxide produced above 1000°C at 4–16 bar pressure. Under these conditions its formation obeyed a simple regression equation. Concentrations of NO, NO2 and N2O obtained in kinetic computations showed similar trends to the measured values. Calculations also showed the concentrations of O, OH and H radicals to decrease with pressure, and also that HO2 becomes the dominating radical at high pressures. These changes probably originate mostly from the three-body reaction H + O2 + M → HO2 + M, which at 850°C begins to compete with and finally dominates over the reaction H + O2 → OH + O as the pressure increases. The decrease in NO formation with increasing pressure follows as a consequence, because O and OH are key radicals in the production of NO.

Validation of flow simulation and gas combustion sub-models for the CFD-based prediction of NOx formation in biomass grate furnaces

While reasonably accurate in simulating gas phase combustion in biomass grate furnaces, CFD tools based on simple turbulence–chemistry interaction models and global reaction mechanisms have been shown to lack in reliability regarding the prediction of NOx formation. Coupling detailed NOx reaction kinetics with advanced turbulence–chemistry interaction models is a promising alternative, yet computationally inefficient for engineering purposes. In the present work, a model is proposed to overcome these difficulties. The model is based on the Realizable k–ϵ model for turbulence, Eddy Dissipation Concept for turbulence–chemistry interaction and the HK97 reaction mechanism. The assessment of the sub-models in terms of accuracy and computational effort was carried out on three laboratory-scale turbulent jet flames in comparison with the experimental data. Without taking NOx formation into account, the accuracy of turbulence modelling and turbulence–chemistry interaction modelling was systematically examined on Sandia Flame D and Sandia CO/H2/N2 Flame B to support the choice of the associated models. As revealed by the Large Eddy Simulations of the former flame, the shortcomings of turbulence modelling by the Reynolds averaged Navier–Stokes (RANS) approach considerably influence the prediction of the mixing-dominated combustion process. This reduced the sensitivity of the RANS results to the variations of turbulence–chemistry interaction models and combustion kinetics. Issues related to the NOx formation with a focus on fuel bound nitrogen sources were investigated on a NH3-doped syngas flame. The experimentally observed trend in NOx yield from NH3 was correctly reproduced by HK97, whereas the replacement of its combustion subset by that of a detailed reaction scheme led to a more accurate agreement, but at increased computational costs. Moreover, based on results of simulations with HK97, the main features of the local course of the NOx formation processes were identified by a detailed analysis of the interactions between the nitrogen chemistry and the underlying flow field.

PCDD/F reduction in incinerator flue gas by adding urea to RDF feedstock

The effect of urea on PCDD/F formation in a pilot incinerator was studied by incinerating urea with refuse-derived fuel (RDF) at three concentrations (0.1%, 0.5% and 1.0%, of the fuel feed). A distinct reduction in both PCDD/F and chlorophenol concentrations could be noticed when urea was introduced into the system. Partial-least-square (PLS) analysis of the data showed the importance of certain chlorophenol isomers as PCDD/F precursors, pointing to the possibility that the impact point of the urea inhibitor could be before the precursor molecules, i.e. chlorophenols, have been formed.

Influence of HCl on the homogeneous reactions of CO and NO in postcombustion conditions -- A kinetic modeling study

Abstract The influence of hydrogen chloride (HCl) on homogeneous gas-phase reactions of carbon monoxide (CO) and nitric oxide (NO) was studied in typical postcombustion conditions of industrial furnaces using detailed kinetic modeling. A well-established reaction mechanism (203 reactions) describing the oxidation of moist CO, as well as of NH 3 and HCN was extended by a recently published subset of 36 reactions for the oxidation of HCl. Validation of modeling predictions was achieved in that the effect of HCl on the CO burnout showed excellent agreement with available independent laboratory data. The modeling results led to the conclusion that the presence of HCl (100–600 ppmv) has a strong effect on the CO oxidation at low temperatures of approximately 1023 K. The effect is dependent on the H 2 O concentration and the presence of NO. Very interestingly, at high concentrations of H 2 O (7 vol %) and without any NO, HCl led to a totally unexpected acceleration of the CO burnout at residence times longer than 0.5 s. According to the reaction path analysis, CO is oxidized by OH radicals via CO + OH → CO 2 + H. The acceleration of the CO burnout is explained by the reactions HO 2 + Cl → HCl + O 2 and HO 2 + Cl → OH + ClO decreasing the concentration of the HO 2 radical and, consequently, also the rate of the reaction HO 2 + OH → H 2 O + O 2 , which competes with CO for the OH radicals. Thus, when HCl is present, more OH will be available for CO oxidation and also more H radicals are formed via the CO burnout reaction. This enhances OH formation further via the reactions H + O 2 → O + OH and H 2 O + O → OH + OH under these conditions. At lower concentrations of H 2 O (1 vol %) without any NO, and always if NO was present (150 ppmv), a deceleration of CO burnout was predicted, in agreement with available laboratory studies and most findings in practical combustors. This deceleration is explained by a decrease of the radical pool (OH). Around and above 1123 K the influence of HCl on the CO burnout was found to be very small for all conditions investigated. Furthermore, it was predicted that in the presence of ammonia, HCl extends the temperature window for NO reduction, particularly on the low temperature side.

Emissions of nitrogen oxides from circulating fluidized-bed combustors: Modeling results using detailed chemistry

An emissions model for a circulating fluidized-bed combustor (CFBC) has been developed. The model consists of a one-dimensional fluid dynamics model, an 8-reaction, heterogeneous mechanism, and an elaborate, 340-reversible-reactions homogeneous mechanism with 55 species accounting for the light hydrocarbons (C 1 and C 2 ) and the nitrogen chemistry. The mechanisms are based on independent kinetic measurements made in lab-scale studies. The rate constants have not been adjusted to better fit the circulating fluidized-bed experimental measurements. For a given set of input operation and design parameters, the model estimates the fluid dynamics parameters as a function of reactor height and calculates the species concentration profiles in the reactor. The model predictions for different species provide a satisfactory description of experimental observations available in the recent literature. The model also well describes the observation that the emissions of NO x are either slightly higher or remain unchanged when biomass replaces coal as the fuel even though coal has ten times the bound nitrogen content of the biomass. When coal is fed to the reactor, a significant fraction of the NO formed is destroyed in situ. Destruction of NO in the CFBC is not significant for the case of biomass.

A kinetic study on the potential of a hybrid reaction mechanism for prediction of NOx formation in biomass grate furnaces

This paper presents the verification of a hybrid reaction mechanism (28 species, 104 reactions) by means of a kinetic study with a view to its application for the CFD-based prediction of gas phase combustion and NOx formation in biomass grate furnaces. The mechanism is based on a skeletal kinetic scheme that includes the subsets for H2, CO, NH3 and HCN oxidation derived from the detailed Kilpinen 97 reaction mechanism. To account for the CH4 breakdown two related reactions from the 4-step global mechanism for hydrocarbons oxidation by Jones and Lindstedt were adopted. The hybrid mechanism was compared to the global mechanism and validated against the detailed Kilpinen 97 mechanism. For that purpose plug flow reactor simulations at conditions relevant to biomass combustion (atmospheric pressure, 1200–1600 K) for approximations of the flue gases in a grate furnace at fuel lean and fuel rich conditions were carried out. The hybrid reaction mechanism outperformed the global one at all conditions investigated. The most striking differences obtained in predictions by the hybrid and the detailed mechanism at the residence times prior to ignition were attributed to the simplified description of the CH4 oxidation in the case of the former. The overall agreement regarding both combustion and NOx chemistry between the hybrid and the detailed mechanism was better at fuel lean conditions than at fuel rich conditions. However, also at fuel rich conditions, the agreement was improving with increasing temperature. Moreover, it was shown that an improvement in the prediction of NOx formation by the N-subset of the hybrid reaction mechanism can be achieved by replacing its C–H–O subset with that of the detailed one.

NOx and N2O Emission Formation Tendency From Multifuel CFB-Boilers: A Further Development of the Predictor

Present paper is a part of our on-going model development activities with aim to predict formation tendency of gaseous emissions from CFB combustion of different fuels, and especially, fuel-mixtures in real furnaces of various scale. The model is based on detailed description of homogeneous, catalytic, and heterogeneous chemical kinetics, and a sound but simple 1.5D representation of fluid dynamics. Temperature distribution is assumed known. With the tool, different fuels and fuel mixtures can be compared in respect to their tendency to form nitrogen oxides (NOx , N2 O). In this paper the model was tested to predict nitrogen oxide emissions from mono- or co-combustion of coal, wood, and sludge. The plants simulated were the 12MWth CFB combustor located at Chalmers Technical University (A = 2.25m2 , h = 13.6m) and the pilot scale CFB unit at the Technical University Hamburg-Harburg (d = 0.1m, h = 15m). The results gave reasonable representation of the measured emission data, and reflected correctly to the changes in the fuel characteristics and in the furnace operating conditions in most cases. An extensive laboratory fixed-bed reactor study was also performed in order to obtain input values for the kinetic constants of the catalytic reactions for the reduction and decomposition of nitrogen oxides. In literature, there is a limited data available regarding the catalytic activity of CFB solids during combustion of wood- and waste-derived fuels, especially at co-firing conditions. The kinetics for the NO reduction by CO in the temperature range of 780–910°C was determined to be of the following form (NO = 300ppm, CO = 5000ppm): −rNO = k·[NO]0.48·[CO]0.55 mol/g-s with k = 8.15·exp(−8869/T) m3/kg-s (empty reactor effect included) ork = 830·exp(−14930/T) (empty reactor effect excluded), when using a bed sample (250–355 μm) taken from the transport zone in the CTH boiler while burning a mixture of wood pellets and a pre-dried municipal sewage sludge. The role of char particle size and shape as well as the incorporation of energy balance on the char reactivity and the formation of nitrogen oxides is further illustrated by single char particle oxidation simulations.Copyright