Process intensification and economic evaluation of adiponitrile production based on hybrid modeling

Adiponitrile (ADN) is a crucial chemical used to produce Nylon 66, with potential applications in civilian clothing, specialized equipment, and new energy vehicles. This paper outlines a method for producing adiponitrile from 1,3-butadiene hydrocyanation and enhances it using process simulation and machine learning (ML). The system is optimized for energy savings and assessed economically. An artificial neural network (ANN) model is developed to predict ADN product molar flow and system energy consumption. The Non-dominated Sorting Genetic Algorithm II (NSGA-II) algorithm is used to optimize two goals: increasing ADN yield and reducing system energy consumption. The prediction results show that the coefficient of determination (R²) for both system energy consumption and ADN product flow is greater than 0.989, with the Mean Absolute Percentage Error (MAPE) value less than 0.01. The optimization results indicate a 2.31 % reduction in system energy consumption and a 1.85 % increase in ADN product flow.