Biomass slow pyrolysis inside a screw reactor was studied using three-dimensional (3D) computational fluid dynamics (CFD) modeling. A multi-component (cellulose, hemicellulose and lignin with detailed elemental composition), multi-step kinetic model was incorporated into the CFD model. Modeling results were compared with experimental results with respect to the steady-state axial temperature distribution in the reactor (mean deviation, MD, of 12.6 K at the operating temperature of 598 K), the product yields (MDs of 7.7 wt%, 6.2 wt% and 1.7 wt% for the mass yields of char, liquid product, and gas product, respectively) and the elemental composition of char (MD of 1.6 wt%, 0.9 wt% and 2.5 wt% for the mass fractions of C, H and O, respectively). The model was then used to predict the energy density (higher heating value, HHV) and the energy yield of char, and these modeling results were compared against experimental data (MD of about 1.0 MI kg(-1) for the HHV and of about 9.4 wt% for the energy yield). Simulation results indicate that the model can well predict biomass pyrolysis and can predict the quality (e.g. HHV) and the energy yield in the resulting char. (C) 2019 Elsevier Ltd. All rights reserved.