Design, performance assessment, and machine learning-driven optimization of a novel low-carbon urban waste-to-x polygeneration system : multi-scenario analysis of hydrogen and methane production

Abstract

Growing energy demand, waste accumulation, and greenhouse gas emissions necessitate integrated, low-carbon energy options. This study proposes a novel waste-to-x polygeneration system uniquely integrating biomass gasification with gas turbine, supercritical CO2, Kalina, organic Rankine, and steam Rankine cycles, coupled with advanced wastewater treatment, carbon capture, a proton exchange membrane (PEM) electrolysis, and methanation. The system simultaneously produces electricity, district heat, oxygen, hydrogen, and methane, advancing beyond typical waste-to-energy approaches by combining multi-vector fuel production with near-zero emissions. Under baseline operation, the system attains overall energy and exergy efficiencies of 35.0 % and 39.9 %, delivering 3510 kW net power and 1310 kW heating, and daily outputs of 131.6 kg hydrogen, 2106 kg oxygen, and 296.3 kg methane, while capturing 87 % of CO2 emissions (177.7 t/day) and treating 116.6 t/day wastewater. Exergy analysis identifies the biomass gasifier as the primary exergy destruction source (8014 kW), whereas mixers and splitters achieve the highest exergy efficiencies (>99.0 %). Employing a machine-learning-assisted multi-objective grey wolf optimizer (MOGWO), for dual fuel production scenario, enhances energy and exergy efficiencies to 49.5 % and 53.6 %, respectively; boosts hydrogen, oxygen, and methane production by 23.0 %; reduces net power by 6.9 %; and increases heating output by up to 29.1 %. Among fuel-production modes at the optimum, the hydrogen-only case achieves the highest efficiencies (49.7 % energy, 53.6 % exergy). This integrated approach offers a comprehensive and flexible option for sustainable urban resource management.