Optimizing green urea production: Integration of process simulation, artificial intelligence, and sustainable technologies

Abstract

The production of fertilizers, such as urea, faces significant environmental challenges due to greenhouse gas emissions associated with conventional processes. This study addresses these problems by integrating innovative technologies such as electrolysis to generate green hydrogen, carbon capture, green ammonia and the use of renewable energies. In addition, an approach combining artificial neural networks (ANN) and multi-objective optimization techniques is proposed to design efficient and sustainable green urea production processes. Using simulations in Aspen Plus and sensitivity analysis, databases were generated to train ANN models with optimized hyperparameters, achieving accurate representations of the systems. These models were used in optimized schemes using deterministic, metaheuristic and Bayesian methods. The results highlight the ability of ANN to predict with high accuracy variables such as production, costs and energy consumption, with coefficients of determination above 0.99. The optimal solutions balanced sustainability and costs, showing that deterministic methods are more robust and faster, while metaheuristic and Bayesian techniques achieved higher yields, albeit with higher costs. This approach not only demonstrates the technical and economic feasibility of green urea, but also establishes an innovative framework for sustainability in chemical processes using artificial intelligence and advanced optimization.