Impact of module design on heat transfer in membrane distillation

Nusselt correlations originally developed for estimating heat transfer rates in heat exchangers poorly describe heat transfer in membrane distillation (MD) processes. In this work, we assess the impact of module design in bench-scale experiments, simplified treatment of heat transfer rates in MD models, and the effect of permeate flux on temperature polarization as sources of error in Nusselt correlation estimates of heat transfer rates. To test these effects, we systematically vary membrane structure, module sizes, temperatures, and Reynolds numbers to generate a large dataset (n = 240) of MD experiments. We apply this dataset to estimate the heat transfer rate for each unique membrane/module combination and compare our predictions to the classical Sieder-Tate Nusselt correlation (Nu(s-t)). Our results show that heat transfer rates in small modules can be up to five times higher than predicted by Nu(s-t). The heat transfer rate decreases with increasing module size, with heat transfer in large modules adequately described by the Sieder-Tate correlation. We demonstrate that this high heat transfer rate in small modules is a result of an entrance effect, which increases fluid mixing over the membrane area. These results validate the use of Nu correlation in large membrane modules while highlighting issues with their application in small scale systems. This work also emphasizes the importance of bench-scale module design in materials evaluation and process characterization. Finally, it highlights the need for direct measurement techniques that better characterize interfacial processes in membrane systems with modeled driving forces.