Green ammonia production using biogas-solar-wind hybrid systems: Optimization and deep learning methods for techno-economic evaluation

The decentralized and completely electrified green ammonia (GA) generation represents a promising strategy for achieving energy transition targets in various energy-intensive sectors. Although many techno-economic studies have concentrated on solar and wind sources for GA generation, there is limited research on integrating biopower to mitigate their intermittency challenges. This study integrates biogas, solar, and wind energy sources to supply the necessary power for generating 100 tons of GA per day. This techno-economic study uses biogas electrical generators (BGEG), solar photovoltaic (SPV) arrays, and wind turbines (WT) to meet the power needs. Moreover, the analysis used natural gas generators (NGG) to compare BGEG's advantages over conventional power sources. The HOMER Pro software optimizes the entire system for four scenarios, calculating the major performance and economic parameters. To overcome the complex optimization process, based on the capacities and cost parameters of major components, and fuel properties, the study evaluated the predictive capabilities of machine learning (ML) and deep learning (DL) models. For an optimum system with BGEG, SPV, and WT, the excess energy (EE), levelized cost of H2 (LCOH), and levelized cost of NH3 (LCOA) are 12 %, 3.03 $/kg, and 740 $/ton, respectively. The integration of 2.7 % of biopower reduces LCOA by 4.3–17.6 %, as compared to the systems without BGEG. The proposed DL and ML models showed R2 values exceeding 0.99 and 0.98, respectively. Although this study used weather and cost data from Oman, results indicate that the utilization of biogas from organic waste is an attractive strategy for reducing LCOA.