NOX control in combustion of alternative fuels

Abstract

Renewable fuels such as biomass and municipal waste have received growing interest, to reduce the share of coal in heat and power production. However, the emissions of nitrogen oxides, i.e. NOx (NO and NO2), from combustion of these fuels continue to be an environmental concern, as NOx leads to formation of photochemical smog and acid rain. A commonly used technology to reduce NOx from stationary facilities is the selective non-catalytic reduction (SNCR) process. In this process, a chem-ical reducing agent, e.g. NH3 or urea (CO(NH2)2), is added to the post-combustion region, to reduce NO to N2. An additional challenge in combustion of biomass and waste fuels arises due to the content of inorganic matter (e.g. alkali and heavy metals) and chlorine in the fuels. Vaporized alkali salts released during combustion can generate sticky and corrosive ashes, which may deposit in the post-combustion region, and potentially cause shutdown of the power plant. Addition of sulfur containing compounds is known to decrease these issues, by converting alkali chlorides to less corrosive alkali sulfates. For this reason, ammonium sulfate ((NH4)2SO4) has been suggested as an additive in the SNCR process, which upon decomposition yields NH3 for reduction of NO, and SO2/SO3 for sulfating metals. NOx emissions from mobile sources, i.e. marine and automotive diesel engines, are typically reduced by employing the selective catalytic reduction (SCR) process, using urea as additive. However, for-mation of undesirable byproducts during urea decomposition is known to cause deposition in the exhaust system, and may lead to shutdown of the SCR operation. It has therefore been of interest to develop computational tools for the description of urea decomposition and byproduct formation. The goal of these tools are to gain an improved understanding of the process, and to predict optimal op-erating conditions that reduces the risk of deposition in urea-SCR systems. Previous models for urea decomposition are mainly limited to low heating rate conditions (2-20 °C/min). Such low heating rates are not representative of practical urea-SCR systems, where the urea-water-solution (UWS) droplets of <100 μm may be heated by up to 105 °C/min. The high heating rate is expected to have a significant influence on the process. Therefore, it has been of particular interest in this work to de-velop models that reliably extends to higher heating rates. The SNCR performance of ammonium sulfate (AS) has been investigated experimentally in a flow reactor setup under well-controlled conditions. Additionally, a detailed kinetic reaction model for the SNCR chemistry was developed, containing subsets for H/N/O, H2/O2, H2S/O2, alkali/sulfur/chlorine and N/S interactions, along with a decomposition reaction for ammonium sulfate. As an important prerequisite for studying the SNCR process with ammonium sulfate, the effects of sulfur oxides on NO reduction was investigated. For SNCR experiments using NH3 in absence and presence of SO2, it was found that SO2 has a negligible effect on the NO reduction, in line with numerical results of the model. SNCR experiments using ammonium sulfate was demonstrated to achieve up to 95 % NO reduction, in a temperature interval of 1025-1075 °C at AS/NO molar ratios of 1.1-2.3. The model accurately predicted the operating temperature window for efficient NO reduction, but slightly un-derestimated the NO reduction at temperatures above the optimum. Increasing the residence time by a factor of 40 caused a downwards shift in the operating temperature window by approximately 150 oC, which was well captured by the model. Finally, it was desired to investigate the effect of KCl on NO reduction, as well as the KCl sulfation potential when using ammonium sulfate. Addition of 390 ppm KCl was found to have a promoting effect on NO reduction, by shifting the operating tempera-ture window by 100 °C towards lower temperatures, while attaining a similar degree of NO reduction to that without addition of KCl. However, this promoting effect was not captured by the model, and more work is required to solve this issue, and to determine if the effect is caused by gas phase or condensed phase interactions of KCl with the SNCR chemistry. Finally, the results implied that 60 % of the added KCl was converted by ammonium sulfate at temperatures between 800-1025 °C. The urea decomposition process was investigated experimentally, by thermogravimetric analysis (TGA), Fourier-transformed Infrared Spectroscopy (FTIR), and Raman spectroscopy. Based on the results of these methods, it was indicated that urea mainly decomposes directly from condensed phase, whereas evaporation of urea is negligible. A kinetic reaction scheme including 14 reactions for urea decomposition and formation of biuret (H2N-CO-NH-CO-NH2), cyanuric acid (HOCN)3, and ammelide (C3H4N4O2), was developed. The kinetic parameters in the reaction scheme were derived from fitting to TGA experiments under low to medium heating rate conditions (5-500 °C/min). The model yielded a good description of the urea, biuret, and cyanuric acid TGA curves for all investi-gated heating rates. Moreover, the results implied that an increasing heating rate causes a shift in production of cyanuric acid to ammelide, while reducing the overall deposit formation. Additionally, isothermal TGA experiments of urea extending over 96 hours demonstrated that 280 °C is insufficient to attain fully decompose urea, whereas 380 °C yielded complete decomposition within 55 hours. The model yielded a precise description of the isothermal characteristics of urea at 280 °C, but slightly overestimated the decomposition rate at 380 °C. A model for UWS droplet evaporation and decomposition in urea-SCR systems was developed, in-cluding the reaction scheme for byproduct formation as described above. The water evaporation model compared favorably with experimental evaporation rates of 0.82-0.92 mm UWS droplets at ambient temperatures of 200-400 °C. The decomposition rates of urea particles were generally un-derestimated, most pronounced at higher temperatures. The model discrepancies were mainly at-tributed to the experimentally observed micro-explosions of UWS droplets, which is not captured by the present model. The model was further used to predict the evaporation and decomposition behavior of practical UWS droplets of 7-70 μm at 200-400 °C. The numerical results indicate that it is favorable for the conversion of UWS droplets to minimize the droplet size, while operating at high ambient temperature. However, the residence time of the droplets in the exhaust gas before reaching the cata-lyst is typically too short to attain high conversion of urea. Therefore, the results indicate that consid-erable amounts of urea and byproducts will reach the catalyst in practical urea-SCR systems.