Generating buoyancy-driven convection in membrane distillation

Membrane distillation (MD) is a thermally-driven desalination process that can treat hypersaline brines. Considerable MD literature has focused on mitigating temperature and concentration polarization. This literature largely neglects that temperature and concentration polarization increase the feed density near the membrane. With gravity properly oriented, this increase in density could trigger buoyancy-driven convection and increase permeate production. Convection could also be strengthened by heating the feed channel wall opposite the membrane. To investigate that possibility, we perform a series of experiments using a plate-and-frame direct contact MD system with an active membrane area of 300 cm and a feed channel wall heated using a resistive heater. The experiments measure the average transmembrane permeate flux for two gravitational orientations, feed Reynolds numbers between 128 and 1128, and wall heat fluxes up to 12 kW/m. The results confirm that with gravity properly oriented, wall-heating can trigger buoyancy-driven convection for a wide range of feed Reynolds numbers, and increase permeate production between roughly 20 and 130 %. We estimate, however, that at high Reynolds numbers (), more than 70 % of the wall heat is carried out of the MD system by the feed flow, without contributing to permeate production. This suggests the need for longer membranes and heat recovery steps in any future practical implementation.

Miles Mabry, Federico Municchi, Yiming Liu, Jingbo Wang, Tzahi Cath, et al.. Generating buoyancy-driven convection in membrane distillation. Journal of Membrane Science, 2024, 709, pp.123043. ⟨10.1016/j.memsci.2024.123043⟩. ⟨hal-05042235⟩

Journal: Journal of Membrane Science

Date de publication: 01-09-2024

Auteurs:
  • Miles Mabry
  • Federico Municchi
  • Yiming Liu
  • Jingbo Wang
  • Tzahi Cath
  • Craig Turchi
  • Michael Heeley
  • Eric M.V. Hoek
  • David Jassby
  • Denis Martinand
  • Nils Tilton

Digital object identifier (doi): http://dx.doi.org/10.1016/j.memsci.2024.123043


x >