Global minimization of power consumptions for multicomponent gas membrane cascades

As an emerging separation technology, membrane has been applied to many applications, among which many require the separation of a multicomponent feed into desired products. In this work, we developed an MINLP formulation that, for a given multi-component feed, identifies the minimum power consumption and the corresponding optimal membrane cascade configuration. This MINLP formulation can simultaneously guarantee global optimality and accuracy for multicomponent membrane cascades. In this formulation, we introduced a new approximation model for multicomponent membrane module, which is much simpler than the model obtained by conventional collocation method. We also introduce several auxiliary variables which help us solve cascades with 6 components and up to 5 stages. We solved the resulting MINLPs to within 5% of global optimality. Finally we apply the algorithm to a H2 separation case and identified optimal membrane cascade configurations with 2, 3, and 4 compressors.