Multiscale Modeling of Plastic Pyrolysis in a Fluidized Pyrolyzer with NN-Inspired Kinetics and Experimental Validation

The kinetic mechanism of waste plastic pyrolysis and the intricate multiphase hydrodynamics within fluidized bed pyrolyzers are of paramount importance to industrial procedures. Nevertheless, a comprehensive mathematical multiscale model that delineates the complex multiphase reaction system of HDPE pyrolysis is noticeably absent from the literature. Herein, we have formulated a novel neural network (NN)-inspired interpretable kinetics for plastic fast pyrolysis by automatically discovering the reaction pathways and kinetic parameters and later coupled it with a coarse-grained DEM-CFD model for the multiscale modeling of the plastic pyrolyzer. Microscale pyrolysis experiments in tandem microreactors were carried out to generate the data set for the neural network-inspired pyrolysis kinetics, while reactor-scale pyrolysis experiments in a fluidized bed were conducted to authenticate the multiscale model. This work provides a thorough apprehension of plastic pyrolysis at both the microscale and reactor scales, offering crucial guidance for the development and optimization of waste plastic pyrolysis processes.