[2D08-2D11] Digital technology applications (1)

[2D09] Structure optimization of multi-stage membrane separation processes using machine learning

○Keigo Matsuda¹ (1. Department of Complex Systems Science, Graduate School of Informatics, Nagoya University)

Keywords:process synthesis, simulation, machine learning

In a multistage membrane process, there are eight different configurations for a single membrane, which include selecting the membrane type, determining whether the permeate or non-permeate side is used as the product, and deciding whether purging or recycling is employed. This study utilized machine learning to optimize the process configuration with regards to energy consumption, composition, and recovery rate.