Experimental and Numerical Studies on Two-Stage Combustion of Biomass

Abstract

In this thesis, two-stage combustion of biomass was experimentally/numerically investigated in a multifuel reactor. The following emissions issues have been the main focus of the work:

1- NOx and N2O

2- Unburnt species (CO and CxHy)

3- Corrosion related emissions

The study had a focus on two-stage combustion in order to reduce pollutant emissions (primarily NOx emissions). It is well known that pollutant emissions are very dependent on the process conditions such as temperature, reactant concentrations and residence times. On the other hand, emissions are also dependent on the fuel properties (moisture content, volatiles, alkali content, etc.). A detailed study of the important parameters with suitable biomass fuels in order to optimize the various process conditions was performed.

Different experimental studies were carried out on biomass fuels in order to study the effect of fuel properties and combustion parameters on pollutant emissions. Process conditions typical for biomass combustion processes were studied. Advanced experimental equipment was used in these studies. The experiments showed the effects of staged air combustion, compared to non-staged combustion, on the emission levels clearly. A NOx reduction of up to 85% was reached with staged air combustion using demolition wood as fuel. An optimum primary excess air ratio of 0.8−0.95 was found as a minimizing parameter for the NOx emissions for staged air combustion. Air staging had, however, a negative effect on N2O emissions. Even though the trends showed a very small reduction in the NOx level as temperature increased for non-staged combustion, the effect of temperature was not significant for NOx and CxHy, neither in staged air combustion or non-staged combustion, while it had a great influence on the N2O and CO emissions, with decreasing levels with increasing temperature. Furthermore, flue gas recirculation (FGR) was used in combination with staged combustion to obtain an enhanced NOx reduction.

The fate of the main corrosive compounds, in particular chlorine, was determined in an experimental campaign using fuel mixtures. The corrosion risk associated with three fuel mixtures was quite different. Grot (Norwegian term used for tree’s tops and branches) was found to be a poor corrosion-reduction additive and could not serve as an alternative fuel for co-firing with straw. Peat was found to reduce the corrosive compounds only at high peat additions (50 wt%). Sewage sludge was the best alternative for corrosion reduction as 10 wt% addition almost eliminated chlorine from the fly ash.

Numerical studies were also performed to estimate the emission level in the flue gas using a comprehensive mechanism in a configuration which simulated two-stage combustion of biomass. Furthermore, a reduction of the comprehensive chemical mechanism was performed since the mechanism is still complex and needs very long computational time and powerful hardware resources. The selected detailed mechanism in this study contains 81 species and 703 elementary reactions. Necessity analysis was used to determine which species and reactions that are of less importance for the predictability of the final result and, hence, can be discarded. For validation, numerical results using the derived reduced mechanism were compared with the results obtained with the original detailed mechanism. The reduced mechanism contains 35 species and 198 reactions, corresponding to 72% reduction in the number of reactions and, therefore, improving the computational time considerably. Yet the model based on the reduced mechanism predicts correctly concentrations of NOx and CO that are essentially identical to those of the complete mechanism in the range of reaction conditions of interest. The modeling conditions are selected in a way to mimic values in the different ranges of temperature, excess air ratio and residence time, since these variables are the main affecting parameters on NOx emission.