Turning waste into energy through a solar-powered multi-generation system with novel machine learning-based life cycle optimization

This study presents an innovative solar-powered multi-generation system aiming at converting waste into diverse forms of energy, including dimethyl ether (DME), hydrogen, power, and heat. Concurrently, a systematic and computationally efficient optimization framework is developed to unlock the maximum potential of this complex waste-to-energy system. The system integrates plasma gasification, DME synthesis, combining heat and power generation, solar-driven electrolysis and desalination. Life cycle assessment and techno-economic assessment have been implemented for system comprehensive optimization which is formulated as a large-scale nonlinear program (NLP) model. Based on rigorous process simulation results, a machine learning-based framework is proposed to accelerate optimization. Using medical waste treatment as a case study, the solution of the NLP problem reveals optimal levelized costs per kWh energy range from $0.1064 to $0.1304, with total life cycle carbon emissions ranging from 0.2748 to 0.5083 kg CO₂-eq/kWh energy. The findings demonstrate the proposed system's environmental sustainability and economic viability.