2024 AIChE Annual Meeting

(44g) Accurate Prediction of Pilot-Scale OptipermTM Olefins Membranes with Multiphysics Simulation

Authors

DeJaco, R. - Presenter, National Institute of Standards and Technology
Burghardt, B., University of Pennsylvania
Parrish, C., Compact Membrane Systems
Ensuring local conservation of mass, momentum, and energy in numerical simulations of membrane modules can provide highly accurate cost estimates and accelerate development of next-generation module configurations. However, process simulation packages only have built-in capabilities for membrane technologies that are mature and popular. To address these limitations for our OptipermTM Olefins Membrane, a spiral-wound module containing an active layer that facilitates transport of olefins over paraffins, we developed a mathematical model and implemented a numerical method in customized software. This model is critical to ensure that process designs and economics can be validated as part of the technology development process.

Building on top of previous work,1 our approach accounts for variation in composition, pressure, velocity, temperature, and permeance along the flow path, as well as an additional variable to be described following internal acceptance of Intellectual Property disclosures. We use a finite-difference method to discretize the differential equations, and a mesh-refinement analysis indicates the scheme achieves first-order convergence. After using orthogonal distance regression to infer parameters from laboratory data, numerical simulations closely predict stage cuts and purities of pilot plant data, with average absolute differences less than 5 %. Writing custom software in a compiled language sets the stage for extension to optimal design and process control without the need to make an approximation of an approximation (i.e., a surrogate model).

[1] DeJaco RF, Loprete K, Pennisi K, Majumdar S, Siepmann JI, Daoutidis P, Murnen H, Tsapatsis M. Modeling and simulation of gas separations with spiral-wound membranes. AIChE J. 2020; 66:e16274. https://doi.org/10.1002/aic.16274.