Machine learning optimization of integrating an advanced thermal-electrochemical plant with an oxy-biogas fuel plant employing a CO2 capture process

This study addresses boosting biogas utilization through oxy-fuel combustion and integrating an innovative multigeneration system. This system allows advanced thermal-electrochemical integration for electric power, cooling, heat, and liquefied hydrogen generation. This approach reduces energy loss and incorporates a CO₂ capture unit. Hence, the integrated subsystems include an oxy-biogas combustion power plant, a supercritical-CO₂ power plant, an organic Rankine cycle, an NH₃-H₂O combined coolant and power cycle, a solid oxide electrolysis cell, and a Claude hydrogen cycle. The study presents a complete examination covering thermodynamic, sustainability, and economic perspectives and detailed parametric assessments, showing the combustion temperature as the most influential parameter. Subsequently, an optimization process is conducted, employing a multi-objective strategy utilizing machine learning techniques based on artificial neural networks and multi-objective grey wolf optimization. Considering the tri-objective scenario with the exergy efficiency, net present value, and total unit cost of products as objective functions, their optimal values are calculated at 47.75 %, 17.72 M$, and 28.13 $/GJ, respectively. Under the tri-objective optimization scenario, the total exergy destruction equals 4630 kW, with the combustion chamber as the most important contributor. Also, the sustainability index and payback period are found at 1.92 and 17.7 M$, respectively. Besides, these conditions exhibit liquefied hydrogen output of 2.9 m³/day, costing 3.37 $/GJ. This research highlights that the integration of oxy-biogas fuel combustion with the designed multigeneration system can enhance biogas utilization, achieving improved thermodynamic efficiency and economic performance while supporting the sustainable production of high-value energy products.