Integrating artificial intelligence and conventional approaches in sugarcane bagasse biorefineries: a review towards a circular bioeconomy

Sugarcane bagasse (SCB), with global production exceeding 200 million metric tons per year, has emerged as a key lignocellulosic feedstock at the heart of the circular bioeconomy, offering a sustainable, low-carbon alternative to fossil resources through production of biofuels, biochemicals, biopolymers, and bioelectricity. This review critically examines the interplay between conventional bioprocess approaches and artificial intelligence (AI) methods in SCB-based biorefineries. It highlights that while traditional thermochemical and biochemical routes provide the physicochemical backbone of value-added processes, these conventional routes face inherent limitations in energy efficiency, operational flexibility, and environmental performance due to high pretreatment costs and biomass recalcitrance. The integration of machine learning (ML) techniques, including artificial neural networks, support vector machines, genetic algorithms, adaptive neuro-fuzzy systems, and digital twins, enables data-driven modeling, real-time process control, predictive maintenance, and multi-objective optimization across SCB pretreatment, hydrolysis, fermentation, and cogeneration units. This integration enhances resource use efficiency, product diversification, and closed-loop material flows, thereby reducing waste and advancing zero-waste circularity. The review underscores the synergistic potential of combining AI with established bioprocess knowledge to advance integrated, scalable SCB biorefineries aligned with circular bioeconomy principles. It also identifies key techno-economic, regulatory, and scalability barriers, including costly pretreatment and feedstock recalcitrance, and proposes coordinated research and policy strategies to accelerate the global deployment of sustainable, bio-based industrial systems grounded in SCB valorization.