Ammonia Co-Firing in a Staged-Air Pulverized-Coal Boiler: Reactor-Network Prediction of Carbon Reduction and Nitrogen-Oxide Emissions

Abstract

As power systems pursue deep decarbonization while retaining dispatchable capacity, retrofitting the existing coal-fired boiler fleet with carbon-free fuels such as ammonia is a pragmatic option, but practical implementation is constrained by the coupled behavior of combustion stability, char burnout, and NO formation. In this work, a Chemkin-based chemical reactor network (CRN) is developed for a 350 MW pulverized-coal boiler with staged air, combining perfectly stirred reactors for the main furnace zones with detailed gas-phase kinetics and heterogeneous NO-char/soot reactions. The CRN is validated against staged drop-tube-furnace measurements for pure coal and 20% ammonia cofiring (dry 6%O2), predicting outlet O2, CO2 and CO within ∼10% and capturing the measured NO rise. With ammonia introduced in the flame zone, simulations for 0–80% blending (20% step) show near-linear reductions in CO2 and CO (−21%/–73% at 20%; −84%/–98% at 80%), while char consumption is delayed and the NO/burnout peaks shift downstream. NO varies nonlinearly, increasing from 20% to 60% but decreasing at 80%. Reaction analysis attributes this trend to competition for H/OH radicals: at high ammonia fractions, NH3 dehydrogenation depletes H/OH, promotes N2 formation via NHi, and suppresses NO production. At 20% cofiring, injection location is critical: postflame injection minimizes CO2/CO (−33%/–80%) but increases NO (∼+88%), whereas burnout-zone injection causes a much larger NO penalty (about +571%). Among the 20% cofiring cases, flame-zone injection yields the lowest predicted NO emission. The outlet NH3 values are conditional on the ideal-mixing assumptions of the CRN and are not interpreted as quantitative predictions of ammonia slip under practical high-ratio operation. The results provide a chemical-kinetic interpretation of the effects of blending ratio and injection location within the modeled conditions. The results relate the predicted outlet species concentrations to nitrogen-conversion pathways under the prescribed network and operating conditions.