The demand for renewable energy solutions has increased research in biomass conversion methods, particularly pyrolysis. Among various biomass feedstocks, coconut shell is one of the most promising feedstocks for biochar production due to its high carbon content and thermal stability. This work presents the intermediate pyrolysis of coconut shell waste. It introduces a new hybrid optimization approach that combines Response Surface Methodology (RSM), Adaptive neuro-fuzzy inference system approach, particle swarm optimization algorithm (PSO-ANFIS), Adaptive neuro-fuzzy inference system approach, and Genetic Algorithm (GA-ANFIS) to improve biochar yield prediction and process efficiency. The performance of both PSO-ANFIS and GA-ANFIS surpassed RSM through superior predictions and error reduction exceeding 15%, so they proved more effective for optimizing hybrid ANFIS models. The results obtained indicate that the optimal biochar yield (47.2 wt.%) was attained at a moderate temperature of 348.6 °C, lower heating rate of 6 °C/min, residence time of 5 min, larger particle size of 4.5 mm, and nitrogen flow rate of 10 mL/min. Fourier-transform infrared spectroscopy (FTIR) analysis confirmed Biochar’s absorption and catalysis potential ability by detecting hydroxyl, carbonyl, and ether functional groups, attributing the presence of alcohol, esters, and conjugated acids halides. A scanning electron microscope (SEM) showed that Biochar possessed a porous shape with interconnected channels, which improved its ability to adsorb materials used in energy storage and environmental applications. X-ray Diffraction (XRD) analysis showed abundant graphitic and amorphous carbon structures, reinforcing the findings about sequestration. The analytical methods confirm that Biochar demonstrates three main usages: energy storage capability alongside functionality in soil improvement and industrial processes, thereby addressing sustainable waste-driven energy transformation. This research enhances the efficiency of biomass pyrolysis through a hybrid model approach, which provides recommendations for real-time operational improvements in industrial use.
Onokwai et al. (Wed,) studied this question.