Automated Machine Learning-Assisted Low-Carbon Operation Optimization for Catalytic Cracking Units: A Multi-Objective Optimization Perspective

This study proposes a low-carbon operational optimization framework for fluid catalytic cracking (FCC) units and conducts a comprehensive evaluation of their economic performance, energy utilization, and environmental impacts. First, a low-carbon energy-saving operational optimization framework was developed to enhance the economic benefits of FCC units while exploring their carbon reduction potential. Subsequently, a surrogate-assisted model based on an automated machine learning (AutoML) framework was constructed to reduce computational costs. A carbon tax scenario was then introduced, converting CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions into carbon tax values. This approach can improve the computational efficiency of the multi-objective optimization problem. The experimental results demonstrate that optimizing operational conditions increased unit-time profits by 0.5%, reduced CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions by 1088.92 kgCO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>eq/h, and decreased wind energy and steam energy consumption by 627.67 m<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup>/h and 324.28 m<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup>/h, respectively. The scientific optimization of operational parameters enhances production efficiency while mitigating environmental impacts, providing valuable guidance for energy conservation and emissions reduction in industrial processes and supporting sustainable development.