Modelling hydrogen deblending from natural gas using Pd-based dense metallic membranes with an empirical polarization model

Low carbon hydrogen can be produced using excess of electric power and further stored into the existing natural gas grid. Deblending processes are required at the end-users side. However, conventional technologies such as adsorption methods can be complex, especially at the small scales, making the investment not convenient. Membrane technology is gaining interest. In scientific literature, Pd-based dense metallic membranes were already identified as a potential solution to separate high purity hydrogen. They can also be coupled with electrochemical hydrogen pump and TSA. However, Pd-based membrane potentiality had not yet been deeply investigated. In this article, a new empirical correlation for concentration polarization losses is found. Thanks to its simplicity and flexibility, it is used to show how an appropriate design of the separation modules can lead to outperform benchmark performance. Different scenarios, with blending at 5 %, 10 % and 20 % and at low (8 bar), mid (40 bar) and high (66 bar) grid pressure are studied to separate 25 kg/day of pure hydrogen. Separation costs obtained with 10 % hydrogen blending are below 4.44 <euro>/kg even at low pressure, outperforming benchmark estimations using PSA, that resulted 7.64 <euro>/kg. Polarization correlation is recommended to further improve module design in H2 deblending.