Flow and mass transfer in spacer-filled channels for reverse electrodialysis: a CFD parametrical study

Influence of the boundary conditions on heat and mass transfer in spacer-filled channels

“…It can be observed that Nu (2) and Sh (2) are always lower than Nu (1) and Sh (1) , and that the difference increases with Re. Note also that the difference between Sh (1) and Sh (2) (Sc=600) is larger than that between Nu (1) and Nu (2) (Pr=4). …”

“…Membrane processes offer examples of heat or mass transfer in plane channels bearing mixing promoters of more or less tortuous geometry, aimed at reducing temperature or concentration polarization phenomena thus improving the process performance [1,2]. Promoters usually consist of polymeric filaments and play also the role of spacers, keeping a fixed distance between the opposite channel walls.…”

“…Obvious modifications apply in the case of mass transfer. Dimensionless numbers Nu (1) , Nu (2) or Sh (1) , Sh (2) can be obtained from the above definitions by choosing a suitable length scale. Nu (1) and Nu (2) coincide when a uniform wall temperature is imposed, while they differ when a uniform wall heat flux or a wall thermal resistance are imposed.…”

“…Figure 5 shows the behavior of the averages Nu (1) and Nu (2) as functions of R for one-side heat transfer, =45° and P/H=4, Re42 (graph a) or P/H=2, Re126 (graph b). The more reliable of the two averages, i.e., Nu (2) , decreases monotonically with R in both cases; the highest values of Nu (2) are attained for R=0 (isothermal wall conditions) and the lowest for R. Note that this behavior is opposite to that reported in the Appendix for the case of a spacerless channel, where Nu increases monotonically with R and thus is larger for an imposed uniform wall heat flux than for an imposed uniform wall temperature.…”

“…The left column is for R=0 (uniform wall temperature); the central column is for R=0.75 (mixed, or 3 rd type boundary condition); finally, the right column is for R (uniform wall heat flux). Figure 6 shows that the choice of the thermal boundary conditions at the wall affects the local distributions of thermal quantities even more than it affects average quantities such as Nu (2) . Therefore, in studies discussing the wall distributions of thermal quantities, and even more so in studies comparing experimental and predicted results for the same quantities, a realistic choice of the thermal boundary conditions appears to be crucial.…”