Uwe Schnell

Comparison of Different Global Reaction Mechanisms for MILD Combustion of Natural Gas

Emissions of nitrogen oxides from fossil fuel combustion are a major environmental problem because they have been shown to contribute to the formation of acid rain and photochemical smog. MILD (Moderate and Intensive Low oxygen Dilution) combustion is a promising technology to decrease pollutant emissions and to improve combustion efficiency. A combination of air preheating and fuel dilution with combustion products of low oxygen concentration are the main features of this technique. In MILD combustion mode, preheated air and fuel are gradually mixed with large amounts of recirculated exhaust gas. The objective of the present work is to investigate a global reaction mechanism for natural gas combustion to predict the observed nitrogen oxide and carbon monoxide levels in MILD combustion mode. For this purpose, the comparison of several global reaction mechanisms during the combustion process is paramount. The interaction between turbulence and chemistry is modeled by an advanced Eddy Dissipation Concept (EDC) in order to compare some global reaction mechanisms. For validation purposes, this model is applied to a non-premixed turbulent jet flame of a pilot methane-air mixture (Sandia flame D), and to the semi-industrial scale experiments at the IFRF Research Station. The modeling results are discussed and compared with measurements.

A Kinetic Model for the Prediction of NO Emissions from staged Combustion of Pulverized Coal

Emissions of nitrogen oxides from coal combustion are a major environmental problem because they have been shown to contribute to the formation of acid rain and photochemical smog. Air staging and the application of low-NOx burners are effective in reducing the In-furnace-formed NOx. Fuel staging, or reburning, is another effective method to reduce NOx emissions in the combustion chamber. For a successful application of these processes on industrial scale, the governing parameters must be evaluated very carefully. Computer modeling is an efficient tool for acquiring a better knowledge of the optimum process parameters. In order to predict NO emissions from furnaces operated with advanced combustion technologies, a knowledge of the fare of coal nitrogen during the combustion process is paramount. An advanced NOx model for staged combustion of pulverized coal is presented. It is closely connected to a coal combustion model that includes primary pyrolysis and secondary reactions of tars formed during primary pyrolysis. The present NOx model takes into account the different pathways of coal nitrogen release during primary and secondary pyrolysis, char, and soot combustion. The subsequent conversions of nitrogen-containing species comprise formation of NO from fuel nitrogen and air nitrogen (thermal NO) as well as reduction of NO by hydrocarbon and NHi radicals, char, and soot. The interaction between turbulence and chemistry is modeled by an advanced eddy dissipation concept (EDC). The NOx model is used to predict NO profiles that are compared to measurements obtained from combustion tests carried out at a bench-scale entrained-flow reactor. Comparisons are made for air-staged and fuel-staged combustion of pulverized coal using methane and coal as reburn fuel.

Detailed modeling of hybrid reburn/SNCR processes for NOx reduction in coal-fired furnaces

Abstract Mechanism reduction has made the detailed kinetic modeling of combustion problems much easier; it also offers potential improvement of modeling accuracy and flexibility in comparison to global mechanisms. The present work applies mechanism reduction in conjunction with the CHEMKIN library and develops an automatic reduction program code. Regarding the hybrid re-burn/selective non-catalytic reduction (SNCR) (“advanced re-burning”) conditions in coal-fired furnaces and based on a full mechanism “GADM98,” a skeletal mechanism with 39 species, 105 reactions, and further a 10-step/14-species reduced mechanism were established. The reduced mechanism was implemented into a 3D-combustion computational fluid dynamics (CFD) code. The eddy-dissipation-concept model was used to describe the influence of turbulence on the combustion chemistry. A large number of simulations for reburning and hybrid reburn/SNCR processes in a coal-fired reactor were executed; the predicted results were compared with experimental measurements. The reduced mechanism and the comprehensive modeling give quite satisfactory results over a wide range of mole ratios for β = [NH 3 ]/[NO] and air/fuel equivalence ratios λ 2 in the reburn zone. From the modeling results, it was found that adding ammonia premixed with reburn fuel (CH 4 ) effects no further reduction of NO x or even impairs the reduction efficiency compared to pure reburning, and in contrast, staged addition of ammonia downstream of the CH 4 injection in the reburn zone provokes a significant further reduction of NO x over a wide range of parameters. According to the predictions, NO x -reduction rates of 50–60% and of 70–80% can be achieved through pure reburning and hybrid reburn/SNCR approaches, respectively, at λ 2 = 0.95 and β = 1.5. Concerning the computational procedure, essential measures were taken to optimize convergence and computing time. The computing time with the present reduced mechanism is ∼2.5 times that with the traditional global mechanism for the same iteration number. Tabulation of the rate constants reduced the computing time of the reaction kinetics by ∼50%.

Experimental and Numerical Investigation of NOx Formation and its Basic Interdependences on Pulverized Coal Flame Characteristics

ABSTRACT The formation of nitrogen oxide in pulverized coal combustion is analysed by experiments in a 500 kW test rig and by numerical predictions. This study extracts some of the dozens of parameters that were investigated during the trials. It turned out that the huge number of process variables that have an impact on NOx formation can be reduced to a smaller figure by an appropriate data analysis. Major parameters presented in this contribution are the fineness of the coal dust and the amount of swirl. Special emphasis is put on the fact that the level of NOx concentration must necessarily be monitored together with the degree of bumout of the coal dust

Simulation of detailed chemistry in a turbulent combustor flow

This paper describes the inclusion of detailed chemical reaction mechanisms in the framework of a turbulent flame simulation. Calculations are based on a finite-volume solution procedure including submodels for turbulent flow, radiative heat transfer, combustion of fuel, and pollutant formation. The interaction of chemical reactions and turbulence is modeled using the eddy dissipation concept (EDC), which has been extended to include detailed chemical reaction mechanisms. For the oxidation of methane, a detailed C 1 / C 2 mechanism is compared with a skeletal mechanism, which is also used to calculate the formation of nitrogen oxide. The basic idea of incorporating the reaction mechanism into the EDC is described. The numerical effort of the resulting coupled partial differential equation system involved investigations of adequate reduction methods. The proposed model is applied to a 400-KW turbulent diffusion methane flame in a cylindrical furnace of which experimental results are available for a detailed evaluation of the proposed method. Predictions are performed with full and skeletal mechanisms. The measured trends in temperature and species concentrations of CH 4 , O 2 CO, CO 2 and NO are reproduced adequately by the predicted profiles. Steady-state conditions have been introduced for many of the radical concentrations. In this way, the numerical effort can be lowered considerably without affecting the results compared to calculation without steady state conditions. Postprocessor calculation of the NO chemistry shows some differences compared with the coupled solution. However, considering the uncertainties and simplifications included in a turbulent flame calculation. the postprocessor calculation shows reasonable agreement. Investigations on the mechanisms of NO formation reveal that the calculated thermal NO cannot account for the experimentally observed NO, and prompt NO makes a significant contribution to the NO emission of this flame.

Comparison of coherent anti-Stokes Raman-scattering thermometry with thermocouple measurements and model predictions in both natural-gas and coal-dust flames

Mobile coherent anti-Stokes Raman-scattering equipment was applied for single-shot temperature measurements in a pilot-scale furnace with a thermal power of 300 kW, fueled with either natural gas or coal dust. Average temperatures deduced from N(2) coherent anti-Stokes Raman-scattering spectra were compared with thermocouple readings for identical flame conditions. There were evident differences between the results of both techniques, mainly in the case of the natural-gas flame. For the coal-dust flame, a strong influence of an incoherent and a coherent background, which led to remarkable changes in the spectral shape of the N(2)Q-branch spectra, was observed. Therefore an algorithm had to be developed to correct the coal-dust flame spectra before evaluation. The measured temperature profiles at two different planes in the furnace were compared with model calculations.

Numerical modelling of solid fuel combustion processes using advanced CFD-based simulation tools

Computational modelling of combustion processes has been the subject of coninuous research at the Institute of Process Engineering and Power Plant Technology (IVD) over the last two decades. To this end, finite-volume-based computer codes have been developed. In the present paper, some fundamental ideas and approaches of the applied mathematical models and the numerical methods are described, followed by some examples of typical applications of the procedures with special emphasis on the validation of simulation results. These examples show that the application of combustion simulation codes has been extended to comprise a wide range of several different areas ranging from huge bituminous coal-fired utility boilers for electricity production to decentralised small-scale furnaces and tile stove heating inserts for domestic heating.

Some remarks on the PISO and SIMPLE algorithms for steady turbulent flow problems

Abstract Properties of the PISO and SIMPLE algorithms of solution of momentum equations are examined by comparison of the computational effort required for reaching the same convergence criterion in two test problems. In the present paper swirling and non-swirling flows are considered in axisymmetric geometry. In the computations both, interative and time marching versions of these methods are considered. Optimal values of numerical parameters, on which effectivity of both algorithms depends for various differential grids, were found and the results for swirling flow problem were verified on the basis of available experimental data.

A model of the combustion of a porous carbon particle in oxygen

Abstract A model of the combustion of a porous carbon particle in oxygen is developed. The model considers heat and mass transfer in both the gas phase above the particle’s surface and inside the porous particle. The conditions for the model having a solution have been determined by solving the diffusion equation in the gas phase around the particle. Two regimes of particle combustion have been analyzed: 1) the regime when carbon reacts with oxygen in the all volume of the particle, and 2) the regime with carbon reacting with oxygen in a layer at the particle’s exterior. Both carbon monoxide and carbon dioxide can be formed in the first regime, but carbon monoxide can only be formed in the second regime. The conditions for each regime of particle combustion are determined.

Modelling potassium release and the effect of potassium chloride on deposition mechanisms for coal and biomass-fired boilers

A model, predicting the release of potassium compounds and its effect on the deposition of superheaters, has been recently developed. The model has been implemented into the three-dimensional CFD program AIOLOS. Concerning the release of potassium compound, the vaporization of the potassium and its reactions in the gas phase are considered. The influence of turbulence on chemistry is considered by using the eddy dissipation concept. Two simulations, one for a coal with high content of chlorine and the other for a coal with low content of chlorine were performed on a small-scale entrained flow reactor. The modelling results are discussed and compared with measurements. Furthermore, the effect of the released potassium chloride on the deposition mechanisms has also been considered. In order to predict the deposition rate, two major deposition mechanisms i.e. condensation and inertial impaction are considered in the model. The sticking probability is modelled based on the melting behaviour of the species involved in the deposition process. Deposit formation in a 0.5 MW semi-industrial pulverized-fuel combustion test facility is predicted considering different operating conditions. Deposition rates on the deposition probe from two kinds of biomass are compared and discussed.