Biochar Production in Torrefaction Process from Bamboo Biomass to Predict Highest Heating Value

Abstract

This research investigates the effects of torrefaction on the higher heating value (HHV) of bamboo biomass. The process involves heating biomass at temperatures ranging from 200 to 400 °C under oxygen-limited or inert atmospheric conditions. Agricultural bamboo residues were sun-dried (average temperature 35.27 °C), ground, and sieved through mesh sizes of 20, 60, and 100. A custom-designed torrefaction system was developed, consisting of a horizontal tubular furnace, a vapor condensation unit, and a biomass feed chamber. The process was simulated and optimized using Aspen Plus V12.1 and Design Expert V13.0 to evaluate the HHV in megajoules per kilogram (MJ/kg). Proximate analysis of the bamboo biomass yielded values of 60.45% volatile matter, 6.73% moisture, 32.05% fixed carbon, and 1.84% ash. Under optimal experimental conditions—60 mesh particle size, 25 mL/min nitrogen flow rate, 340 °C temperature, and 90 minutes duration—the maximum HHV obtained was 19.3945 MJ/kg, with a mean prediction error of only 3.143%. Further optimization showed that the ideal conditions were a particle size of 33.6813 mesh, nitrogen flow rate of 28.9080 mL/min, temperature of 339.225 °C, and residence time of 52.2639 minutes, yielding an HHV of 19.4937 MJ/kg. Machine learning models were also applied to predict HHV based on 279 experimental data sets. Kernel Ridge Regression (KRR) demonstrated superior performance, achieving a prediction accuracy of 87.64% during training and 81.16% in testing, outperforming Support Vector Machine (SVM), which showed 60.60% and 57.85% accuracy, respectively. Therefore, KRR is identified as a more effective model for predicting the HHV of torrefied bamboo biomass.