Physicochemical properties and pyrolysis behavior of petcoke with artificial neural network modeling

Abstract

Petcoke is a byproduct of heavy crude oil refining at an enormous scale, and insights into the physicochemical properties, thermal degradation behavior, decomposition kinetics, and thermodynamic analysis of petcoke pyrolysis are crucial for the efficient design of pyrolysis reactor systems. In this study, proximate and ultimate analyses were performed, and petcoke was found to have a high fuel ratio with high carbon, high sulfur, and considerably low ash content, implying that it is a less reactive fuel. Advanced analytical techniques such as SEM, BET, FTIR, and petrography indicated that petcoke has a considerably low pore volume and predominantly inorganic graphitic carbon. Furthermore, thermogravimetric analysis was examined at four different heating rates of 10, 20, 30, and 40 K/min and petcoke exhibited a low pyrolysis performance, which was confirmed by the devolatilization index and pyrolytic parameters. The activation energies and frequency factors estimated by three model-free methods (DAEM, FWO, Friedman) were in the range of 221.6–235.9 kJ/mol and 1011–1012 s−1, respectively. In addition, thermodynamic analyses were examined and the thermal degradation behavior of petcoke pyrolysis was modeled using an artificial neural network; the NN-2-12-12-2 model with Tansig-Tansig transfer functions was found to be the best fit. This study enabled the identification of the fundamental characteristics of petcoke fuel, and these results provide useful information regarding the design, utilization, optimization, and limitations of petcoke pyrolysis systems.