Computational fluid dynamics (CFD) assisted analysis of profiled membranes performance in reverse electrodialysis

Abstract

Implementation of reverse electrodialysis (RED) is economically limited by the relatively high ion-exchange membranes price. Additionally, the shadow effect of non-conductive spacers reduces the membrane area available for counter-ion transport and increases the stack electric resistance. A promising alternative could be utilization of profiled membranes, since the reliefs formed on their surface keeps the membranes separated and provides channels for solutions flow. Herein, we have simulated, through computational fluid dynamics (CFD) tools, fluid behavior in channels formed by various profiled membranes. The highest net power density values were obtained for corrugations shape and arrangement in a form of chevrons due to the increase of the available membrane area and an excellent balance between enhancement of mass transfer and the increase of the pressure drop in the channel. When properly designed, corrugated membranes may offer a better performance even compared to the case of conductive spacers. The proposed membrane corrugation design in not limited to the RED application, and could be also extended to other electromembrane processes, such as electrodialysis and Donnan dialysis, in which high ionic mass transport rates are desirable at as low as possible energy costs.

Pressure drop at low Reynolds numbers in woven-spacer-filled channels for membrane processes: CFD prediction and experimental validation

“…Only few works have been devoted to the CFD simulation in the low Reynolds numbers range typical for RED. Beyond the works of our research group [19,[39][40][41][42], the only example is the work of Pawlowski et al [43].…”

A hierarchical model for novel schemes of electrodialysis desalination

“…However, in this case they have limited prediction capabilities, providing accurate results only under some conditions [47]. On the contrary, distributed parameters models [49][50][51][52][53][54][55][56] are more accurate, but at the cost of a larger implementation effort. In summary, with respect to the theoretical models, 1-D semi-empirical models are preferable as process simulators thanks to their features of versatility, robustness and effectiveness.…”

“…In particular, the majority makes use of either empirical information [42][43][44]48,52,55,57] (e.g. limiting current density) or 3-D computational fluid dynamics (CFD) simulations [50,51,54,56].…”

“…For a NaCl salt solution, the equivalent conductivity can be estimated by the correlation of Islam et al [62], as described in Appendix C. The shadow factor is generally a function of the geometrical characteristics of the spacer. Therefore, it is usually calculated as a function of the channel porosity [52,54], open area [51], or both [63,64]. The porosity represents the fraction of channel volume occupied by the liquid, while the open area represents the free fraction of membrane area projected in the direction perpendicular to membranes.…”

The role of membranes in the use of natural salinity gradients for reverse electrodialysis

“…However, the electric resistance due to concentration polarization phenomena was found to be significantly higher in stacks with those profiled membranes in comparison to that measured in stacks with spacers [22,23]. The fluid mixing in the channels formed by relatively simple structures at the membrane surface was less efficient and dead zones were formed due to sub-optimal corrugation designs, as observed by the use of ink-inject techniques [23] and CFD simulations [28,29].…”