Numerical investigation on the heat transfer of plastic waste pyrolysis in a rotary furnace.

• A CFD-DEM model simulated heat transfer between plastic and sand in a rotary furnace. • Loading solid heat carriers in the kiln can promote the heat transfer for pyrolysis. • The volatile residence time in the kiln increased with the loading of heat carriers. • Increasing the solid heat carrier loading could reduce the wax yield from pyrolysis. • Loading more heat carriers favored the production of pyrolysis oil and gas. The increasing plastic wastes have become a serious concern to the environment. Plastic wastes could be efficiently converted to fuel via pyrolysis. In the process of plastic pyrolysis, the key process parameters will affect the production of the final product, the distribution of pyrolysis gas, pyrolysis oil, and pyrolytic wax. Based on previous experiments, the discrete element model and computational fluid model were established in this study to compare the influence of heat carrier loading and furnace rolling speed on the pyrolysis of waste plastics in a rotary furnace. The simulation results indicate that when the granular heat carriers were loaded, as the heat carrier loading increased, the translational velocity and angular velocity of the waste plastic particles increased. Compared with the heat carrier-free case, the heating rate of the waste plastic particles is greatly increased when the heat carrier was loaded. The increase in the heat carrier loading also increased the residence time of the pyrolysis volatiles in the rotary furnace. With the acceleration of the rolling speed of the furnace, the translational velocity and angular velocity of plastic particles all increased, but the improvement of heat transfer efficiency is not obvious. This is consistent with the insignificant change in the pyrolysis product in the previous experimental results. The CFD-DEM simulation (excluded the pyrolysis reaction model) provided detailed information on particle movement, heat transfer and volatile residence time which could be used as an effective tool to interpret the experimental observation and optimize the process parameters. [ABSTRACT FROM AUTHOR]