Abstract
This research assesses the bioenergy potential of date palm branch (DPB) waste, aligning with Saudi Arabia's Vision 2030 energy and environmental goals. The study uses reaction modeling, thermokinetics, a k-nearest neighbors (KNN) machine learning approach, and techno-economic assessments. Experimental characterizations employing FTIR, SEM, and both proximate and ultimate analysis of pulverized DPB biomass reveal its lignocellulosic nature and compositional characteristics. Thermogravimetric analysis (TGA) of the material, tested between 25 and 1000 °C at heating rates of 7.5 to 60 °C per minute, revealed that the main thermal breakdown occurred from 200 to 530 °C, and was caused by the decomposition of hemicellulose and cellulose. Criado master plot analysis of the material's thermal decomposition indicated the R3 contracting cylinder model was the most suitable reaction mechanism. The Jander [D3] and Ginstling-Brounshtein [D4] diffusion models were also good fits. The kinetic analysis showed that various model-free approaches, including FWO, KAS, STK, and FR, yielded comparable activation energy values for the hemicellulose and cellulose components, with the results clustering between approximately 98.43 and 109.30 kJ/mol. The application of the KNN machine learning technique in this study yielded accurate predictions (R2~0.975) of the TGA traces following rigorous modeling that involved hyperparameter optimization and testing of the trained model on 20% unseen data. Through a global sensitivity analysis, the degradation temperature for DPB's thermal devolatilization was identified as the key parameter controlling the pyrolysis process. The techno-economic assessments of the pyrolysis operation indicate that it is a viable, financially rewarding, and environmentally friendly process, offering valuable insights for policymakers, environmental engineers, and energy professionals toward promoting sustainable waste management and a circular economy.