Poultry manure combustion parameters to produce bioenergy: A thermogravimetric analysis by isoconventional models and machine learning

Abstract

Poultry manure has immense potential to generate bioenergy to supply some of producers’ electrical consumption. However, if desired, the combustion of this biomass requires the investigation of important thermal parameters for its efficient and automated industrial-scale application. This study determined combustion parameters such as ignition and burnout temperatures, activation energy, and combustibility indices by thermogravimetric analysis under two approaches: isoconventional models with a set of differential and integral equations and previously trained and validated machine learning models. Poultry manure combustion ignition and burnout temperatures were 212.28 and 443.94, 214.30 and 452.70, and 223.31 and 464.09 °C for 10, 25 and, 20 °C min−1 heating rates, respectively. The combustion index calculated from these temperatures was of 4.78, 6.13, and 4.22 × 10−7. Activation energy was calculated based on isoconventional models, whose values were 99.11 (Friedman), 117.78 (FWO), and 107.50 kJ mol−1 (KAS). Isoconventional models adequately represented mass degradation as a function of temperature. Moreover, after training and testing, our machine learning models (neural networks, decision trees, and random forests) satisfactorily described the combustion of poultry manure, enabling the optimization of parameters such as ignition and burnout temperatures. Poultry manure combustion showed interesting thermal characteristics that can be optimized by machine learning.

Keywords

• Biomass
• Bioenergy
• Burnout and ignition temperatures
• Mathematical modeling
• Poultry manure
• Machine learning