NOx formation and selective non-catalytic reduction (SNCR) in a fluidized bed combustor of biomass

Abstract

Caledonian Paper (CaPa) is a major paper mill, located in Ayr, Scotland. For its steam supply, it previously relied on the use of a Circulating Fluidized Bed Combustor (CFBC) of 58 MWth, burning coal, wood bark and wastewater treatment sludge. It currently uses a bubbling fluidized bed combustor (BFBC) of 102 MWth to generate steam at 99 bar, superheated to 465 degrees C. The boiler is followed by steam turbines and a 15 kg/s steam circuit into the mill. Whereas previously coal, wood bark and wastewater treatment sludge were used as fuel, currently only plantation wood (mainly spruce), demolition wood, wood bark and sludge are used. Since these biosolids contain nitrogen, fuel NO is formed at the combustion temperature of 850-900 degrees C. NOx emissions (NO + NO2) vary on average between 300 and 600 mg/Nm(3) (dry gas). The current emission standard is 350 mg/Nm(3) but will be reduced in the future to a maximum of 233 mg/Nm(3) for stand-alone biomass combustors of capacity between 50 and 300 MWth according to the EU LCP standards. NOx abatement is therefore necessary. In the present paper we firstly review the NOx formation mechanisms, proving that for applications of fluidized bed combustion, fuel NOx is the main consideration, and the contribution of thermal NOx to the emissions insignificant. We then assess the deNO(x) techniques presented in the literature, with an updated review and special focus upon the techniques that are applicable at CaPa. From these techniques, Selective Non-catalytic Reduction (SNCR) using ammonia or urea emerges as the most appropriate NOx abatement solution. Although SNCR deNO(x) is a selective reduction, the reactions of NOx reduction by NH3 in the presence of oxygen, and the oxidation of NH3 proceed competitively. Both reactions were therefore studied in a lab-scale reactor and the results were transformed into design equations starting from the respective reaction kinetics. An overall deNO(x) yield can then be predicted for any operating temperature and NH3/NOx ratio. We then present data from large-scale SNCR-experiments at the CFBC of CaPa and compare results with the lab-scale model predictions, leading to recommendations for design and operation. Finally the economic impact is assessed of implementing SNCR-technology when applying an NH3 SNCR or urea SNCR to the CFBC at CaPa. (C) 2010 Elsevier Ltd. All rights reserved.