Deep learning analysis of green ammonia synthesis: Evaluating techno-economic feasibility for sustainable production

Abstract

This research presents a comprehensive optimization of green ammonia synthesis in Morocco by leveraging advanced deep learning techniques to maximize the utilization of the country's abundant renewable energy resources. By employing deep learning models such as LSTM and LSTM_adv, we forecast ammonia production over the next decade, improving strategies for production and energy storage. Our findings confirmed the viability of sustainable ammonia production in Dakhla and highlighted its transformative potential for global sustainability goals. Under the optimal scenario of 20 % photovoltaic and 80 % wind energy, the production cost was US$575 per tNH3, delivering an energy output of 8916.64 GWh and a daily ammonia production of 2503.36 tNH3. Shifting to 100 % wind energy further enhances the potential, increasing daily ammonia production to a maximum of 3090 tNH3 while reducing the production cost to US$376 per tNH3. These results demonstrate the significant economic and environmental benefits of using renewable energy sources for ammonia production in Morocco. By leveraging a country's abundant wind and solar resources, we can contribute to a sustainable and energy-independent future.