CFD-aided benchmark assessment of coal devolatilization one-step models in oxy-coal combustion conditions

Abstract

Coal devolatilization, as first step in thermochemical processes involving solid fuels, has been targeted in plenty of CFD investigations: despite the variety of models that were developed to define the rate of the process and the amount of volatile matter released, not all of them are suitable for CFD simulations of large-scale coal combustors. Empirical models ensure both simplicity and feasibility, but not full reliability for all coal types and operating conditions. CFD modeling results for coal conversion during devolatilization in an entrained flow reactor and oxy-conditions are presented and compared to experimental data. A benchmark assessment of one-step models on the accuracy of the volatile yield prediction is then performed. It is shown that the heating rate strongly influences the devolatilization process and its modeling, not only in terms of rate but also in terms of yield, which differs significantly from the proximate analysis value. This indicates the need for a model able to predict the variation of coal volatile yield with temperature and heating rate, in addition to the kinetic model

Collaborative simulations and experiments for a novel yield model of coal devolatilization in oxy-coal combustion conditions

“…The model parameters, which were not tuned on the operating conditions and the coal type of interest, were considered accountable for the poor performances of the BT model. The present work is a follow-on to the results obtained and showed in [23], which emphasized the sensitivity of first-order kinetic models to the value of the ultimate volatile yield. Therefore, the modeling of the ultimate yield has been targeted and performed, using two approaches that allow the collaborative analysis of experiments and simulations: consistency analysis [24] and a Bayesian method based on Gaussian-Process Regression [25].…”

“…The solid flow rate is up to 100 g/h, while the flow rates of the carrier CO 2 stream and of the flue gases are 0.5-0.6 Nm 3 /h and 14.6-17.6 Nm 3 /h, respectively [36]. The devolatilization tests examined in this paper are the same considered in our previous work [23] where the details of the gas composition in each experiment can be found, along with a discussion on the negligible influence of oxidation and gasification reactions.…”

“…Part of the results of the CFD modeling of the IPFR can be found in [23]. Histograms of the particle temperature distributions from the inert CFD simulations are shown in Fig.…”

“…The predictions of BT model do not match the experimental values particularly at temperature lower than about 1300 K. The performances of the BT model can be related to its yield model parameters, which are not trained on the specific coal subject of investigation, as already discussed in [23]. A procedure for the updating of these parameters, based on a joint analysis of experimental data and CFD simulations, is then required.…”

“…a simple functional form that relates the ultimate volatile yield to the particle temperature, referred to here as BT model. The BT model was tested in our previous work [23] against an experimental dataset collected in a pilot-scale entrained flow reactor (EFR) and led to a non-optimal prediction of the coal yield. The model parameters, which were not tuned on the operating conditions and the coal type of interest, were considered accountable for the poor performances of the BT model.…”