Design of the Ethylbenzene production process using machine learning

Ethylbenzene (EB) is a raw material used to produce the styrene monomer, and the EB market size has been increasing. Recently, EB has been produced by a production process using zeolite catalysts. However, the fuel consumption in this process is an issue because the energy load in the separation process is high. The objective of this study was to design an EB production process with low energy consumption per unit production. In the proposed method, the process structure is first optimized, followed by optimization of the design variables. In the conventional process design, the design variables are optimized by calculations based on knowledge of the chemical engineering and repeated simulations, but it is labor and time intensive. The proposed method uses machine learning to efficiently optimize the design variables. By using machine learning, we achieved the design of a process with a low fuel energy intensity while satisfying multiple reaction conditions.