Conjugate Heat And Mass Transfer In Channels Of Membrane-based Liquid Desiccant Air Dehumidification

People's lives are seriously influenced by high humidity, especially in hot and humidclimates. Air dehumidification is very significant on indoor air quality and socialdevelopments. There are many air dehumidification methods. Among them, due to thecoherent virtues of high efficiency, no liquid water condensation, low-grade energy use, andenergy storage by solution, liquid desiccant air dehumidification has gained much progressand engineering applications. Packed columns, in which the desiccant and the air streamsexchange heat and moisture in a direct contact, are the common equipments for such atechnology. However, since the liquid desiccant and the process air are in a direct contact,small liquid desiccant droplets may be carried over by the process air to indoor environment,which are rather harmful for occupant health, building structure and furniture.Recently, semi-permeable membranes have been used in the liquid desiccant airdehumidification to overcome the problem of liquid solution droplets cross-over. The biggestbenefit of this method is that the traditional dehumidifier and regenerator, i.e., the packedcolumns, are replaced by membrane modules. In such modules, the membranes are used toseparate the desiccant solution from the process air. The membranes can prevent the solutionfrom crossing over to the process air, while selectively permitting the transport of heat andmoisture between the solution stream and the air stream. In the membrane-formed channelsused for liquid desiccant air deumidification, the fundamental data such as the Nusselt andSherwood numbers under the uniform temperature or heat flux boundary conditions are notsuitable. It is because that the boundary conditions on the membrane surfaces are neitheruniform temperature (concentration) nor uniform heat flux (mass flux) conditions. Rather,they are naturally formed by coupling between the desiccant and the air streams through themembranes. In order to disclose the conjugate heat and mass transfer mechanisms in theprocess of the liquid desiccant air dehumidification in the membrane modules, followingnumerical work is performed:(1) Parallel-plates membrane channels. When the dehumidifiers are in parallel-plates, themomentum, energy and mass conservation equations governing the air and the solution flowsare solved together with the relevant boundary conditions. The data such as friction factor,Nusselt and Sherwood numbers are then obtained and validated experimentally. The resultsfound that for the air stream, the local Nusselt number (NuC,a) and Sherwood number (ShC,a)obtained under the naturally formed boundary condition are between the corresponding values under the uniform temperature condition (T) and heat flux boundary condition (H).(2) Counter flow hollow fiber membrane channels. When the dehumidifiers are incounter flow hollow fibers, the momentum, energy and mass conservation equationsgoverning the air and the solution flows are solved together with the relevant boundaryconditions. The data such as friction factor, Nusselt and Sherwood numbers are then obtainedand validated experimentally. The results found that for the air stream, the local Nusseltnumber (NuC,a) and Sherwood number (ShC,a) obtained under the naturally formed boundarycondition are between the corresponding values under the uniform temperature condition (T)and heat flux boundary condition (H).(3) Cross-flow hollow fiber membrane bundles, neglecting the interactions betweenfibers. For the cross-flow hollow fiber membrane module, due to the complexity of thestructure, a representative free surface model is used to establish laminar and turbulent models.The governing equations are solved together with the boundary conditions. The fundamentaldata are then calculated and validated experimentally. The results are that, for the air stream,when the fraction factor of the module (φ) is less than0.4, the average Nusselt number (Nuave,a)under the naturally formed boundary condition are between the average Nusselt number underuniform temperature condition (Nuave,T) and the average Nusselt number under uniform heatflux condition (Nuave,H). However, when φ is larger than0.4, both the average Nusselt number(Nuave,a) and the average Sherwood number (Shave,a) are higher than Nuave,H.(4) Cross-flow hollow fiber membrane bundles, considering the interactions betweenfibers. The free surface model can provide certain guidance, however it cannot accuratelypredict the hollow fiber membrane-based liquid desiccant air dehumidification because thefiber-to-fiber interactions are not considered. In this chapter, two representative periodic unitcells in in-line and staggered packed fiber bundles are selected to establish the laminar andturbulent models. The fundamental data are obtained and validated experimentally. Generallythe air side Nuave,ais between Nuave,Tand Nuave,H. For in-line arrangement, when Reais lessthan400and φ is less than0.16, Nuave,ais higher than Nuave,H. For staggered arrangement, theair side Nuave,ais somewhat less than Nuave,T.Above work provides the fundamental data of Nusselt and Sherwood numbers formembrane channels under the conjugate heat and mass transfer boundary conditions. Thesebasic data is the tools for future component structural optimization and system design.