Numerical thermodynamic-economic study and machine learning-based optimization of an innovative biogas-driven integrated power plant combined with sustainable liquid CO2 and liquid H2 production-storage processes

Innovative heat recovery, CO 2 capture, and energy storage methodologies are pivotal for developing sustainable and eco-friendly solutions for the energy sector. Hence, this study proposes implementing an oxyfuel combustion process for a biogas power plant, modified by an innovative heat recovery method and a CO 2 capture-liquefaction technique. Furthermore, the design incorporates high-temperature water electrolysis to produce hydrogen, which is then introduced into a hydrogen liquefaction process utilizing a Claude cycle for adequate long-term storage. The research employs thermodynamic, exergoeconomic, and net present value assessments, accompanied by an extensive parametric study and optimization process. Hence, a machine learning algorithm is implemented using artificial neural networks combined with the NSGA-II method for multi-criteria optimization, focusing on exergy efficiency, net present value, and products' sum unit cost as objective functions. The implemented optimization reduces the optimization time to under 30 min, which is significantly more efficient than traditional heuristic techniques, which typically require several hours for similar systems. This optimization framework is highly applicable to both industrial and district energy systems. This approach enhances predictive analytics and streamlines resource management. In industrial environments, it effectively optimizes energy use and production processes by examining various operational factors, which leads to cost reductions and improved efficiency via predictive maintenance and cohesive energy strategies. The optimal outcomes reveal the mentioned objective functions' values at 47.22 %, 58.73 M$, and 33.53 $/GJ, respectively. Under these optimal conditions, liquid carbon dioxide and liquid hydrogen outputs are quantified at 4931 lit/h and 1848 lit/h, respectively. Finally, the proposed system can omit CO 2 emissions by 1.36 kg/kWh under optimal conditions, which reflects a 5.60 % better performance than the base case. Furthermore, the products’ sum unit cost decreases by 3.09 %, indicating efficient cost savings linked to the products. • Biogas-driven plant combined with liquid CO 2 and H 2 production-storage processes. • Numerical thermodynamic-economic study and machine learning-based optimization. • Using ANNs combined with the NSGA-II method for multi-criteria optimization. • Optimum exergy efficiency is found at 47.22 %. • Optimum NPV and SUCP are found at 58.73 M$ and 33.53 $/GJ, respectively.