Optimization of OSN Membrane Cascades for Separating Organic Mixtures

Given the recent advances in organic solvent nanofiltration (OSN) membranes for the separation of organic mixtures, this paper investigates the design of OSN membrane cascades for hydrocarbon mixtures. A superstructure optimization approach based on mixed-integer nonlinear programming (MINLP) is considered to determine the most promising cascade configurations. The superstructure integrates OSN membrane modules, mixers, and the classical solution-diffusion transport models for the description of the permeate flux and the rejection coefficients at each membrane stage. The goal of the optimization is to determine the cascade design and operating conditions maximizing product purity, for a given composition of the inlet stream. A binary system of heptane and hexadecane is taken as an example to mimic the light-key and heavy-key in the multicomponent environment of a representative petrochemical/hydrocarbon feed stream while simplifying the analysis. This case study demonstrates the effectiveness of the proposed approach of superstructure optimization for a lab-scale experimental setup and provides a proof-of-concept for crude oil refining using membrane cascades.