Thermogravimetric analysis of hazardous waste: Pet-coke, by kinetic models and Artificial neural network modeling

• Published in: Fuel (Guildford) 2021-03, Vol.287, p.119470, Article 119470
• Main Authors: S.P., Sathiya Prabhakaran, G., Swaminathan, Joshi, Viraj V.
• Format: Article
• Language: English
• Subjects: Artificial neural network; Artificial neural networks; Carbon; Coke; Combustion; Combustion indices; Deceleration; Hazardous wastes; Neural networks; Nickel; Optimization; Pet coke; Petroleum refining; Reaction mechanisms; Refineries; Refinery wastes; Solid state reaction models; Thermal degradation; Thermal degradation kinetics; Thermogravimetric analysis; Two dimensional models; Vanadium
• ISSN: 0016-2361; 1873-7153
• DOI: 10.1016/j.fuel.2020.119470

[Display omitted] •Ea (183.16 kJ/mol) and Pre-exopential factor (106–1017 s−1) were computed.•Best fit solid state reaction mechanism model identified were R3 and D2.•Performance function as MSE and transfer function as TRANSIG from NN7 was best fit ANN.•Combustion indices at variable heating rates were evaluated. Pet coke is a hazardous waste obtained from petroleum refinery on an enormous scale. Understanding pet coke’s material characteristics and its decomposition kinetics is a prerequisite for its highly stable combustion behavior. Proximate analysis revealed that high fixed carbon and low volatile matter makes pet coke a low reactive fuel. Advanced analytical techniques like FTIR, XRD, ICP-OES revealed that pet coke is predominantly inorganic, constituting graphitic carbon and hydrogenation materials like nickel & vanadium, respectively. Thermogravimetric studies were conducted, and average activation energy calculated by model-free methods was in the range of 152.70 to 206.18 kJ/mol, respectively. The study of solid state reaction mechanism by Coats-Redfern, Kennedy-Clark and Master plot methods suggested contracting volume (g(α) = (1-(1-α))1/3) and two dimensional diffusion reaction mechanism (g(α) = (1-α) ln(1-α) + α) models which falls under decelerating type reaction mechanism. Combustion indices of pet coke like CCI, DI, DB were evaluated at different rates of heating. Pet coke’s thermal degradation behavior were modelled using an artificial neural network (NN7). The research carried out permits to identify the characteristics of one of the most dominant hazardous waste, along with the futuristic potential usage and the findings provide reliable kinetic parameters for scale up, utilization, optimization, limitation for the system.