Simulation of entrained flow gasification with advanced coal conversion submodels. Part 1: Pyrolysis

CFD modeling results for entrained flow coal gasification using advanced submodels for coal conversion are presented and compared to detailed experimental data. The focus of this investigation is on the accurate modeling of the pyrolysis process. An iterative procedure is proposed and validated to bridge the gap between detailed pyrolysis models such as CPD, FLASHCHAIN or FG-DVC and empirical models based on single-or multiple-step kinetic expressions, which are usually used in CFD. Multiple particle heating rates from the CFD solution are taken to perform detailed pyrolysis calculations and these results are used to find optimal kinetic parameters for the empirical models using an automated procedure. It is shown that the heating rate strongly influences the devolatilization process (rate and yield). CFD simulations are performed for the BYU entrained flow gasifier. Due to the high heating rate in entrained flow gasification, the volatile yield can differ significantly from the proximate analysis value. Accurate pyrolysis modeling is shown to be important to capture coal flame ignition, flame location, species distribution and outlet composition. Since the final volatile yield determines the split in carbon conversion between pyrolysis and the subsequent fast conversion in the gas phase and the heterogeneous char conversion, which is a comparatively slow process under gasification conditions, it also directly influences the overall carbon conversion. Overall, the application of the new comprehensive CFD model including the fitted kinetic rates is shown to give similar results to the full coupling of the CFD and pyrolysis software. The comparison between the simulations and the experiments shows very good agreement for three out of four coals. The fourth coal (lignite with high O/C ratio) is well outside the range for which the detailed models were developed, but reasonable agreement is still obtained. (C) 2013 Elsevier Ltd. All rights reserved.