Project Topics
Get Complete Project Material File(s) Now! »
Background of membrane distillation (MD) and state-of-the-art of coupling MD with solar energy
Membrane distillation
Membrane distillation (MD) derives from membrane contactors, which provide non-dispersive contact via microporous membranes with substantial interfacial area for gas/liquid or liquid/liquid contacting operations [29]. For separation processes like absorption, liquid-liquid extraction and distillation, membrane contactors promisingly offer a significant process intensification, as compared to conventional approaches conducted in towers, columns or mixer-settlers [30]. As a specific application of membrane contactors, by definition, MD is a process based on the evaporation of a liquid provoked by a difference in partial pressure between the two sides of a membrane that is used as a support for the liquid/vapor interface, as illustrated in Figure I.1. Water evaporation and transfer through the membrane are driven by the difference between the elevated vapor partial pressure at the membrane surface on the hot feed side, and the relatively lower vapor partial pressure created on the permeate side.
In order to generate this driving force, the temperature of the feed stream Tf is typically heated to 30°C ~ 80°C. However, the feed temperature at membrane surface Tfm, which actually induces the transmembrane vapor pressure difference, might be lower than Tf due to boundary layer effect when MD is under operation with a certain level of permeate flux. Same occasion may also exist on the permeate side, where the temperature at membrane surface Tpm is higher than that in permeate bulk Tp. Consequently, the driving force might be lowered on account of this temperature polarization phenomenon in MD [31,32], whose significance depends on the operating conditions (membrane material, feed flow, spacers, operating conditions, etc.). Similarly, the feed salt concentration at the membrane surface Cfm is higher than the salt concentration at feed bulk Cf, as shown in Figure I.1, inducing the concentration polarization [33]. However, MD has been proven not so sensitive to feed salinity [34], based on its thermal-driven process, in contrast to the widely-applied desalination technology of Reverse Osmosis (RO), which largely depends on the osmotic pressure. Description of temperature and concentration polarization equations will be introduced in the part of this chapter dedicated to heat and mass transfer.
The liquid/vapor interface is often supposed to be located close to the pore inlet. The process requires the use of microporous hydrophobic membranes, whose properties are crucial to ensure a good and sustainable quality of the produced water. Membranes should be porous enough to allow for a good productivity, but a compromise between pore size and hydrophobicity is required to avoid direct liquid transfer by convection through the pores (this phenomenon is called pore wetting [35]) and to only allow vapor passing through the pores to the other side (permeate side) after the evaporation. If no membrane wetting or other failure occurs, theoretically complete rejection of salt and other non-volatiles can be presumed [21]. For that purpose, membranes used for MD commonly have an average pore size from 100nm to 1µm [21,31], and are made of polytetrafluoroethylene (PTFE), polyvinylidene fluoride (PVDF) or polypropylene (PP). Along with the water vapor flux Jw through the pores, the heat flux Qp passes through both the solid membrane material and the pores of the membrane (Figure I.1). Heat transfer is described in more detail in Section I.2.1 of this chapter.
Historically, the concept of MD was firstly patented by Bodell in 1963 as a new approach for producing clean water via evaporation across silicone rubber membranes [36]. Four years later [37], Finley published the very first scientific paper, specifying the MD process in detail and introducing some experimental results. However, an increasing interest for MD developed in the 1980s when highly permeable membranes for MD became available in the market [21]. In the year of 1986, a “Workshop on Membrane Distillation” was held in Rome, Italy, whose major topic was to standardize the terminology for MD process [38], confirming the sprouting research interest in MD. Consequently, a steady increment of the quantity of publications on MD had been witnessed from 1990 to 2010, being described as the “emerging phase” [39]. Finally, based on this research activity, the industrial interest for MD has become significant when it was demonstrated as a process that could be complementary with reverse osmosis and/or coupled with solar energy or waste heat [22,40,27]. Since then, attempts to upscale both modules and systems have appeared, mainly with a focus for MD’s application in the huge developing sector of desalination and for some niche markets. First industrial scale modules have been made and commercialized, tested on site in different places, and some first plants based on MD have been built.
Membrane distillation applied to desalination
In terms of desalination, generally two categories of technology coexist in the global market, i.e. thermal distillation and membrane separation [9]. Within the former, Multi-Stage Flash (MSF) and Multi-Effect Distillation (MED) are the two most applied desalination technologies, which occupy about 21% and 7% of the total capacity of desalination plants according to [13] in 2017. In order to initialize the evaporation, the heating of saline water up to its boiling point at a certain vacuum pressure is needed at the first stage (or effect) in both MSF and MED [41]. Then, the latent heat of condensation is successively reused in the following stages (or effects), where operating temperature and pressure are decreased progressively. Thanks to the heat recovery, the specific thermal energy consumptions (STEC) of MSF and MED per cubic meter of water production are able to be one order of magnitude lower than the latent heat of evaporation of 667 kWh m-3 [42], being 53 ~ 78 kWh m-3 and 40 ~ 64 kWh m-3 [13], respectively.
The latter category of membrane separation has flourished since the 60’s after the development of the first Reverse Osmosis (RO) membranes and of the adapted process for their use in desalination. Since 2001, the total annual volume of desalinated water produced by RO became higher than that of all thermal plants. Up to the year of 2017, RO takes up to 65% of the total capacity of all kinds of desalination plants, dominating the global market [13]. The growth of RO technology for desalination has been boosted by huge research efforts on developing RO membranes with a good salt retention and a high permeability to water at the same time, and on efficient technical equipment to recover the energy from brines [14]. RO is a barometric process, which means that the driving force is obtained by a pressure difference across a semi-permeable membrane. The applied pressure should be higher than the osmotic pressure of the feed saline water, which linearly depends on its salinity. The requirements for the RO membrane include a high permeability for water, while an extremely low permeability for dissolved salts [43]. Additionally, a physical strength of the membrane is demanded as well, in order to withstand the fairly high pressure applied on the feed side, which ranges from 55 bar to 80 bars for seawater desalination [12].
MD represents a growing interest in the field of brackish or sea water desalination and significant perspectives are rising [15], which has some characteristics and advantages of both technology categories. The intensified process (module compactness) and the applicability to small-scale systems are the major strengths over the conventional distillations of MSF and MED, added by the relatively lower operating temperature, based on its large interfacial area and feed temperature far below saturation. On the other hand, MD is a non-pressurized process and can replace or be complementary to RO because of its insensitivity to osmotic pressure, which makes possible: (i) to treat some high-salinity feed that cannot be treated by RO and (ii) to push the concentration ratio (and the water recovery rate) to a higher level, while still producing significant fluxes of pure water. In fact, MD can be even applied to the extraction of freshwater from RO brines and to the over-concentration of RO brines [22,40], as discussed in the General Introduction. Additionally, the fact that MD is a thermal process enables the coupling with waste heat and solar energy, and some alternate systems to anticipate the energy transition for fresh water production. Therefore, reduction of water squander and of brine volume without pressurization, as well as possible coupling with thermal solar energy are the major strengths of MD over RO.
Configurations of MD
In the literature, four main MD configurations have been identified and widely studied. They are categorized by the different ways of creating a low partial pressure on the permeate side, i.e. Direct Contact MD (DCMD), Air Gap MD (AGMD), Vacuum MD (VMD) and Sweeping Gas MD (SGMD) [44], as illustrated in Figure I.2. The feed side of all four configurations resembles the hot side of a heat exchanger, as the feed solution is a relatively hot stream that circulates inside one of the compartments of the MD module.
Figure I.2: Basic configurations of membrane distillation: (a) DCMD; (b) VMD; (c) AGMD; (d) SGMD
In DCMD (Figure I.2a), the cold fresh water is in direct contact with the membrane surface on the permeate side, and often in counter-current flow with the hot feed stream. The vapor migrates across the hydrophobic membrane due to the water vapor pressure gradient formed by the temperature difference between the two sides, and then condenses in the cold stream as the permeate production. In this process, the membrane has to support two liquid-vapor interfaces and the vapor only exists inside the pores. It is the simplest type of MD in both the configuration and the operation at lab-scale [39], which gives rise to its popularity. Additionally, heat recovery is possible by preheating the feed with the stream coming out of the permeate side [45]. However, the major drawback is the limited permeate flux due to the strong temperature polarization phenomenon, which decreases the transmembrane driving force created by the temperature difference between the two sides of the membrane. Moreover, the conductive heat loss is also obvious due to the direct contact of the feed solution, the membrane and the cold distillate water, lowering the thermal efficiency of the DCMD process. Besides, an initial fresh water stream is needed on the permeate side to start the operation, whose temperature has to be maintained low to keep the transmembrane driving force while the permeate flux brings in the enormous latent heat of condensation.
In VMD (Figure I.2b), the permeate side is submitted to an adjustable vacuum at a pressure lower than the vapor saturation pressure on the feed side. A vacuum pump is usually applied to remove the produced vapor and to maintain the vacuum on the permeate side, which allows the vapor to be condensed outside the module to collect the final production. Vacuum pump allows a good control of the driving force and higher permeate flux is expected in VMD process compared to other MD configurations [23], added by the advantage of negligible conductive heat loss through the membrane, and very low thermal polarization on the permeate side [46,47]. Nevertheless, extra condensation facility is needed, reducing the system compactness. Besides, electrical energy consumption is increased by the addition of vacuum pump, though it can be controlled by finding a compromise between the pumping energy and the productivity [17]. As a derivative, a combination of VMD and MED has been successfully developed and commercialized by Memsys [48], being the Vacuum Multi-Effect-Membrane-Distillation (V-MEMD) module.
In order to reduce the conductive heat loss through the membrane, AGMD differentiates from DCMD by applying a stagnant air gap on the permeate side (Figure I.2c). The permeated vapor has to pass through both the membrane and the air gap, and finally condense on a cold surface located inside the module. Consequently, the conductive thermal resistance in AGMD is enhanced and the heat loss is therefore much alleviated compared with DCMD. However, the trade-offs here are a more complex module design and a simultaneously strengthened transfer resistance through the membrane [49], which results in a comparatively lower permeate flux. Another advantage of AGMD lies in the convenient heat recovery by simultaneously condensing the permeate vapor and preheating the feed, when using the feed source to cool the cold condenser surface. Some innovative design based on AGMD has emerged in recent years, such as Permeate Gap MD (PGMD) and Material Gap MD (MGMD), altering the composition of the air gap. The former directly replaces the air in the gap with stagnant permeate [50,51], yielding higher permeate flux but more heat loss compared to conventional AGMD. The latter diversifies into using either non-conductive material [52] (such as sand or even vacuum in Vacuum Gap MD [50]) or conductive material [51] (such as metal mesh) to fill the gap.
Table of contents :
General introduction
Thesis objectives
Thesis organization
I. Background of membrane distillation (MD) and state-of-the-art of coupling MD with solar energy
I.1. Membrane distillation
I.1.1. Membrane distillation applied to desalination
I.1.2. Configurations of MD
I.1.3. Configurations of MD modules
I.2. Transfer mechanisms of MD
I.2.1. Heat transfer
I.2.2. Mass transfer
I.2.3. Profiles along the flow direction
I.3. State of the art: MD driven by solar energy
I.3.1. Solar thermal collectors
I.3.2. Coupling solar thermal collectors with MD
I.3.3. Direct integration of solar thermal collectors and MD
I.3.4. Observations of the literature
I.4. Conclusions (in English)
I.4. Conclusions (en français)
II. Direct integration of a vacuum membrane distillation module within a solar collector for small-scale units adapted to seawater desalination in remote places: Design, modeling & evaluation of a flat-plate equipment
II.1. Introduction (in English)
II.1. Introduction (en français)
II.2. Modeling of an integrated VMD-solar module
II.2.1. Solar radiation modeling
II.2.2. Simultaneous modeling of heat and mass transfer for an integrated VMD-solar module
II.3. Configuration for a dynamic recycling batch system
II.3.1. Temperature-based control strategy for the recycling batch system
II.3.2. Dynamics for the recycling batch system
II.3.3. Solution procedure
II.3.4. Performance indicators
II.4. Results and discussions
II.4.1. Consistency of the models: solar radiation and VMD
II.4.2. General set of parameters for a daily varying operation
II.4.3. Performance under temperature-controlled batch regime
II.4.4. Improved performance under continuous MD operation
II.5. Conclusions (in English)
II.5. Conclusions (en français)
III. Comparative study of flat-plate DCMD and VMD modules with integrated direct solar heating (DCMD-FPC and VMD-FPC)
III.1. Introduction (in English)
III.1. Introduction (en français)
III.2. Module & system description and modeling
III.2.1. DCMD module configuration
III.2.2. Description of mass and heat transfer in MD modules
III.2.3. Description of the dynamic system
III.2.4. Pumping energy consumption
III.2.5. Model coupling and resolution procedure
III.3. Results and discussion
III.3.1. Parameter settings and daily operation
III.3.2. Influence of parameters
III.3.3. Discussions on a high potential: Heat recovery & solar concentration
III.4. Conclusions (in English)
III.4. Conclusions (en français)
IV. Optimization and design of a novel integrated vacuum membrane distillation – solar flat plate collector module with heat recovery strategy through heat pumps
IV.1. Introduction (in English)
IV.1. Introduction (en français)
IV.2. Process description: coupled solar collector – VMD
IV.3. Design configuration and modeling of VMD-FPC with integrated heat pump
IV.3.1. Coupled solar flat-plate vacuum membrane distillation collector
IV.3.2. Theoretical study of heat recovery from condensation by heat pump
IV.3.3. Modeling structure, recirculation and system dynamics
IV.4. Performance assessment and analysis
IV.4.1. Decision variables, design parameters and main performance indicators
IV.4.2. Sensitivity analysis via Delta Moment-Independent (DMI) indicator
IV.4.3. Fast multi-objective optimization on design and operating conditions
IV.5. Results and discussions
IV.5.1. Sensitivity variation due to heat recovery from permeate condensation
IV.5.2. Importance of heat recovery level
IV.5.3. Global optimization and performance improvement using heat pump
IV.5.4. Benchmark optimization of VMD-FPC at fixed heat recovery levels
IV.5.5. Pareto-based study of decision variables and key indications on design
IV.6. Conclusions (in English)
IV.6. Conclusions (en français)
V. Practical recommendations on the design of a small MD-FPC system for autonomous and decentralized seawater desalination in remotes areas
V.1. Introduction (in English)
V.1. Introduction (en français)
V.2. Choice of integrated MD – direct solar heating module and optimal design of recirculation system
V.2.1. VMD-FPC module materials
V.2.2. Operating conditions
V.2.3. Collector positions and module dimensions
V.3. Seasonal performance of the VMD–FPC based desalination system
V.4. Dynamic behaviors of the integrated desalination unit
V.4.1. Representative daily variations in summer (August 1st)
V.4.2. Representative daily variations in winter (February 1st)
V.5. Conclusions (in English)
V.5. Conclusions (en français)
General conclusions and perspectives
