Predictive Modeling and Analysis of Bio Briquettes from Challenging Agricultural Residues

The escalating demand for sustainable energy sources has prompted extensive research into bio briquettes derived from agricultural waste, particularly problematic agro residues that pose environmental and disposal challenges. This study focuses on the production, characterization, and performance prediction of bio briquettes manufactured from rice husk, sugarcane bagasse, and corn cob wastes known for their high silica content, low biodegradability, and contribution to open burning pollution. The experimental work involved blending these materials in varying proportions (e.g., 50:30:20, 40:40:20) with cassava starch as a binder (5–10% by weight), followed by densification using a hydraulic press at 10–15 MPa. Characterization encompassed proximate analysis (moisture content: 4.2–7.5%, volatile matter: 60–75%, ash: 10–25%, fixed carbon: 15–25%), ultimate analysis (C: 35–45%, H: 4–6%, O: 40–50%, N: 0.5–1.5%, S: <0.5%), physical properties (bulk density: 450–650 kg/m3, compressive strength: 2.5–4.5 MPa), and combustion performance (calorific value: 14–18 MJ/kg, burning rate: 0.8–1.2 g/min). For performance prediction, an Artificial Neural Network (ANN) model was developed using inputs from proximate and ultimate analyses, achieving an R2 of 0.92 for calorific value prediction and 0.89 for compressive strength. Novelty lies in integrating machine learning for real-time prediction of briquette quality from waste composition, enabling optimization without extensive testing. Results indicate that a 40:40:20 blend yields optimal properties, with calorific values comparable to traditional wood fuels (16.5 MJ/kg) and reduced emissions (CO: 150–250 ppm, PM: 50–100 mg/m3). The study concludes that these bio briquettes offer a viable, eco-friendly alternative to fossil fuels, addressing waste management in agro-intensive regions while promoting circular economy principles.