Synthesis and optimization of gas permeation membrane networks

We present a modeling and optimization framework for the design of gas permeation membrane networks. The modeling framework constitutes a generic superstructure of membrane units; recycle, feed, and product compressors; and vacuum pumps. Various flow patterns such as cross, countercurrent, and co-current flow can be embedded into the representation. Both conventional and novel membrane network configurations can be developed through structural optimization of the superstructure. The optimization is carried out using robust stochastic techniques in the form of simulated annealing for minimization of the total annualized network cost. Prominent industrial examples such as air separation and hydrogen recovery illustrate the potential of the design technology.