Independent parallel pyrolysis kinetics of model components in sewage sludge analyzed by BPM neural network.

Abstract

Analyzing the kinetic behavior of sewage sludge pyrolysis is essential for the design of efficient reactors to produce biofuel and syngas. To understand the complex pyrolysis process of sewage sludge, we pyrolyzed six model components (i.e., cellulose, hemicellulose, lignin, protein, soluble sugars, and lipid) using a thermogravimetric analyzer. The effects of the heating rate on the pyrolysis process were examined at four different heating rates (5, 15, 25, and 50 °C/min). As temperature increased, the derivative thermogravimetric peaks shifted to higher temperature zones. The temperature ranges of the maximum mass loss rate for cellulose, hemicellulose, lignin, protein, soluble sugars, and lipid were within 326.1-368.0 °C, 288.7-315.5 °C, 375.1-429.4 °C, 291.9-308.0 °C, 251.0-314.1 °C, and 410.8-454.1 °C, respectively. The apparent activation energies of the model components were obtained using non-isothermal kinetic analysis methods (Flynn-Wall-Ozawa and Kissinger-Akahira-Sunose). In addition, a back-propagation artificial neural network with a momentum algorithm (BPM) was developed to predict the relationship between the pyrolysis experiment and the activation value. The best BPM model (BPM5) for predicting the cellulose pyrolysis was identified.