Multistage separation process and apparatus for water treatment

Abstract

The present invention provides a membrane separation process consisting of membranes inside of pressure vessels that are structured in blocks that form multiple stages of separation. A given block may alternate in its position as part of the first stage or last stage of separation. When in the first stage it will be fed from the opposite end of the pressure vessel as when it is in the last stage. Additionally, the concentrated solution leaving the last stage of the system will be repressurized and recirculated in a closed circuit to the feed side of the last stage and only be periodically purged from the system. This combination of alternating the blocks, redirecting the feed direction, and close circuit recirculation of the concentrate from the last stage will allow for a system that can flexibly adjust the amount of concentration achieved in the final solution while also optimizing the amount of time that the membranes are exposed to the highest concentration of solution.

Description

[0001] Multistage Separation Process and Apparatus for Water Treatment

[0002] Field of Invention

[0003] The present invention generally relates to improving the recovery and prevention of precipitation fouling, in liquid filtration process. More particularly, the invention relates to method and system for preventing precipitation fouling in reverse osmosis and nanofiltration pressure -driven filtration processes and for increasing recovery by precipitation of supersaturated minerals from a concentrate stream.

[0004] Background of Invention

[0005] Constant growth in water demand for potable and industrial use, along with growth in need for smaller physical footprint, lower carbon footprint and brine management, makes membrane filtration an attractive solution, mostly for brackish water. Since traditional treatment methods are not always able to meet the requirements of environmental and drinking water regulations and quality, membrane filtration processes are becoming preferable in such applications.

[0006] Since reverse osmosis (RO) processes generate considerable amounts of reject brine, the industry has adopted numerous disposal options that usually depend on the location of the water treatment plant and type of process used. These options include: discharge to surface water or wastewater treatment plants; deep well injection, land disposal, evaporation ponds and mechanical / thermal evaporation. Management of reject brine has recently become an increasingly difficult challenge due to many factors (El -Naas., 2011).

[0007] This reject stream contains chemicals used in treatment process (e.g. HC1, H2SO4, NaOH, Antiscaling agents) which affect the environment if not well managed (Panagopoulos and Haralambous., 2020). In addition, in some cases zero liquid discharge (ZLD) is required.

[0008] The presence of ionic components such as calcium, magnesium, carbonate, sulphate, phosphate and silicate, and components such as soluble silica (silicic acid), in the feed water of an RO process can limit the achievable water recovery by forming insoluble scale that hinders membrane filtration (Sanciolo et al., 2014). Other limitations for achievement of higher recovery rates are hydraulic issues such as low flow of reject and concentration polarization which is also a significant performance limiting factor in many reverse osmosis processes. Beyond increasing transmembrane osmotic pressure and salt passage, concentration polarization exacerbates surface fouling by elevating foulant material concentrations at the membrane -solution interface. This fouling causes reductions in RO permeate flux during application; resulting in a decrease in water recovery with an associated increase in operating costs to maintain the desired flux. This constraint has to be properly addressed for sustainable RO application in water recovery. The identification of major foulants with their characterization and integration of appropriate pre -treatment options are necessary measures required for mitigating this constraint (Pandey et al., 2012).

[0009] The present invention provides advantages over previous inventions as it allows for both a flexible concentration factor through a closed-circuit recirculation of the last stage while simultaneously rotating the blocks so that the membranes in the last stage are alternated with membranes in the previous stage to reduce the effects of fouling and scaling. This also reduces the amount of excess volume required for a closed-circuit design and has the advantage that most of the system operates on the feed water, thus enabling operation on potentially descaling water and reducing the total permeate conductivity variability common to closed circuit / semi batch systems.

[0010] Pressure-driven membrane processes are defined as processes in which the feed stream is fed at a volumetric rate Qf into a membrane device (e.g., pressure vessel) equipped with membranes that divide the device space into a feed side and a permeate side and in which a pressure difference across the membranes causes the solvent (usually water) to pass from the feed space to the permeate space at a volumetric rate denoted as Qp. The remaining solution, which is now concentrated in the rejected solutes, leaves the feed space of the membrane device as a concentrate stream as a volumetric rate denoted as Qb. the fraction of feed that leaves the membrane device as permeate is referred to as the membrane recovery, %RR:

[0011] %RR = (1)

[0012] Qf

[0013] As recovery is increased the concentration of rejected solutes in the concentrate stream, Cb, is given by mass balance as:

[0014] Pressure-driven membrane processes are distinguished from each other by the types of solutes they reject or pass to the permeate side. For a given process this selectivity can be defined by the empirical solute rejection for each solute i, Ri, defined by:

[0015] Where Cf.i, Cp,i refer to the feed and permeate concentrations respectively of component i. For highly rejected species, Ri=l and then the relation defined in equation 2, between its concentration in the concentrate stream, Cb, and the recovery, %RR reduces to:

[0016] The rate at which solvent together with non-rejected solutes passes through the membrane per unit membrane area is defined as the membrane flux (denoted Jv) with units of volume / (unit area-time). During the filtration process the membranes become fouled and as a result less effective. Membrane fouling has become one of the primary impediments to their acceptance in water treatment applications. The occurrence of membrane fouling in reverse osmosis and nanofiltration processes leads to reduction in the production rate and sometimes to loss of solute rejection. There are several kinds of fouling which typically occur in such processes: colloidal fouling; organic fouling (adsorption of soluble organics on membrane surface); biofouling — formation of a biofilm which by itself or in combination with other kinds of fouling causes deterioration of membrane performance; and precipitation fouling (or scaling) due to precipitation of sparingly soluble salts and minerals.

[0017] Approaches to control membrane fouling involve hydrodynamic and chemical methods, periodic backwashing, chemical cleaning, changing operating conditions, and reducing the operating flux. Another solution for controlling membrane fouling proposes changing the flow direction in order to reduce concentration polarization and fouling in general (“Ultrafiltration Membranes and Applications”, Breslau, B. R. at al, Polymer Science and Technology, Plenum Press, Vol. 13; “Flux Enhancement Using Flow Reversal in Ultrafiltration”, Hargrove, S C and Ilias, S., Sep. Sci. Technol., 34 (6&7), 1319).

[0018] A flow reversal process and device are described in U.S. Pat. No. 5,690,829 (to Lauer), which particularly relates to the cleaning of the membrane from dirt particles. Another possible solution for reducing membrane fouling is described in U.S. Pat. No. 5,888,401 (to Nguyen), which suggests periodically increasing the permeate pressure next to the membrane by partially closing a valve on the permeate side, which results in reductions in the permeate flow rate. This last method reduces the rate of overall permeate recovery which is a disadvantage. Other possible solutions for reducing membrane fouling are described in US 9,649,598, US 8,137,539, and US 8,632,682.

[0019] As will be apparent to those skilled in the art, an efficient solution for preventing precipitation fouling has also implications for biofouling since the stagnant layers and surfaces of scale layers can allow biofilms to attach and develop with less shear forces to remove them.

[0020] In many processes sparingly soluble salts can limit the recovery of desalination processes as their concentration increases in the brine as more product water is pulled out of the feed flow (Stover., 2016). Different techniques have been used to cope with this problem (Section 9.4 in Water Treatment Membrane Processes, Mallevialle, I, Odendaal, P, Wiesner, M. eds., McGraw-Hill, 1996). Chemical softening has been proposed to precipitate sparingly soluble salts most of which are salts of alkali earth metals (Ca, Sr, Mg, Ba). The problem with this approach is that it requires stoichiometric amounts of chemicals to precipitate all of the metal ions of sparingly soluble salts, which is often costly. For example, brackish water containing 100 mg / L of calcium, 30 mg / L of magnesium, and 150 mg / L of carbonate alkalinity as bicarbonate will require 91 g / m3of hydrated lime and 135 g / m3of soda ash to completely remove the calcium. At 80 $ / ton for hydrated lime and 180 $ / ton for soda ash this would involve a chemical cost of 1.8 cents / m3. In addition, the sludges formed in lime softening are often voluminous and hard to remove. This can be prevented by using advanced precipitation processes that combine precipitation softening with microfiltration, also known as Membrane Assisted Crystallization (MAC) or filtering through a filter cake of calcium carbonate seeds which is also known as Compact Accelerated Precipitation Softening (CAPS). Another method would be fluidized bed reactor (pelleting).

[0021] Alternatively, the pH can be reduced by adding acid and removing the carbonic acid formed by air stripping. This type of treatment eliminates the carbonate scales problems, but the problems associated with sulfate scales will still remain. Alternatively, the pH can be increased by adding NaOH or other base for elimination of silica scales problems, however most water does have some sort of carbonate system which might then form a scale since saturation will be high at high pH.

[0022] Consequently, the most common approach today is to use antiscalants which allow operation at various values of super-saturation. However, even with the most advanced antiscalants used today, there are limits on the super-saturation ratios (for example -2.60-3.0 for calcium sulfate, a Langelier Saturation Index (LSI — log 10 of supersaturation ratio) of 2.8 for calcium carbonate, and supersaturation ratio of 2.0 for silica. This means that recoveries are often limited to 75-90%.

[0023] In view of the aforementioned problems there is a need for filtration processes capable of effectively and efficiently controlling and preventing / minimizing precipitation fouling in pressure- driven membrane desalination processes, and of improving the recoveries of such processes.

[0024] It is an object of the present invention to provide a high-flux filtration process for controlling and preventing precipitation fouling in pressure-driven membrane processes.

[0025] It is another object of the present invention to provide a method and system for efficiently controlling flow reversal in a membrane filtration process for preventing precipitation fouling in pressure-driven membrane processes.

[0026] It is still another object of the present invention to provide a method and system for preventing precipitation fouling in a membrane filtration process wherein the composition of the solution next to the active membrane surface exceeds the effective saturation limit of sparingly soluble salts.

[0027] An additional object of the invention is to provide a method and system for achieving high recoveries in a membrane filtration process operating with high local super-saturations.

[0028] Other objects and advantages of the invention will become apparent as the description proceeds.

[0029] The Invention

[0030] In membrane separation systems, spiral wound reverse osmosis (RO) or nanofiltration (NF) membrane modules are connected in series inside pressure vessels (usually up to 8 membrane modules per pressure vessel). The pressure vessels can be assembled in blocks. A pressurized aqueous stream to be treated flows through the pressure vessels, to produce permeate and concentrate streams. The invention provides a multistage flow design, managing the flow of concentrate streams produced by individual blocks installed in RO and NF system, before the system eventually discharges the concentrate.

[0031] To this end, the blocks are divided into two (or more) groups, corresponding to different flow stages denoted Si (i=l, 2, ...). The number of blocks designated to each distinct flow stage Si is denoted Bi. The number of blocks in any stage may be equal to or greater than 1. In some nonlimiting embodiments, the number of blocks in a stage may be comprised between 1 -50, 1-20, 2- 15, 2-12. For example, in a non-limiting case where i=2, and the blocks are divided into two stages Si and S2, the numbers of blocks in stages Si and S2 may be 2<Bi<12 and B2 =1. In its most general form, the invention is based on directing the aqueous stream from a feed supply unit to the blocks designated to the first stage Si and circulating the concentrate stream produced by the blocks operating at stage Si through one or more blocks designated to the consecutive stage Si+i, with or without reversal of flow direction between stages and occasional or regular block rotation between stages.

[0032] The preferred mode of operation includes reversal of flow direction between stages Si and Si+i. The invention relates to a multistage membrane separation process for water treatment using membrane modules connected in series inside pressure vessels assembled in blocks, comprising: dividing the blocks into two or more stages Si (i =1, 2, ... p); passing a pressurized aqueous feed stream from a feed supply unit through the blocks designated to stage Si, wherein the aqueous feed stream flows in a first direction through each of the Bi blocks operating at stage Si, to produce Bi permeate streams and Bi concentrate streams, wherein usually 2<Bi<10; joining the Bi concentrate streams and circulating the joined concentrate stream of stage Si through one or more of the B2 blocks belonging to stage S2, wherein typically 1<B2<3, wherein the concentrate stream flows at a direction opposite to said first direction; when i>2, circulating a joined concentrate stream of stage SM through one or more blocks designated to successive stage Si, wherein the concentrate stream flows in stage Si in a direction opposite to the flow direction of SM; discharging the concentrate stream circulated at Stage Si (i>2); and rotating blocks between stages (e.g., at a frequency determined by the induction time).

[0033] The basic design of a membrane separation process and system according to the invention is described conceptually in reference to Figures 1 to 5, showing an example with four blocks of pressure vessels, denoted by capital letters A, B, C, and D. Usually, all blocks, apart from one, are designated to operate at Stage Si (say, blocks B, C and D) whereas the remaining block (namely, block A) is set apart for operation in Stage S2. That is, in the example illustrated in Figures 1 to 5, i =2, Bi=3 and 62= 1. The blocks are supplied with an aqueous stream from a feed water supply unit, pressurized by pump Pl to produce a feed flow. The design shown in Figures 1-5 is adapted to allow reversal of flow direction between the stages. Optionally, bypass lines can be added for further adaptability of the system, for example if flow reversal is not intended between stages.

[0034] The pressurized feed solution consisting of the water to be treated, for example, wastewater, brackish water or sea water, is split into subsidiary streams which are directed (via a manifold) to the individual blocks, or in fact, most of the blocks: only blocks operating at Stage Si receive their incoming stream from the original feed solution. The direction of flow in blocks operating at Stage Si is defined herein as the regular, or normal, direction, shown by three parallel arrows in Figures 1 to 5, pointing from left to right. The permeate streams produced by blocks B, C and D operating at flow Stage Si are collected in a conventional manner (not shown), whereas the outgoing concentrate streams produced by blocks B, C and D are joined to form a combined concentrate stream, which is guided to flow and circulate through block A in a direction opposite to the normal flow direction, as shown by the arrow pointing from right to left in Figures 1 - 5.

[0035] The system comprises a concentrate flow line which diverges to a concentrate circulation loop and a concentrate discharge line provided with a pump P2 (and optionally aided by pump P3 to improve hydraulics performance). The flows through the concentrate circulation loop and the concentrate discharge line are regulated by valves V circulation and V discharge, respectively (these valves are also denoted herein AVF and P / FV, respectively).

[0036] A key feature of the invention is that during the treatment, blocks are periodically transitioned between stages. That is, the block operating in Stage S2 (e.g., block A), is exchanged after a while with one of the blocks operating in Stage Si, say, block B, with block A being simultaneously shifted to operate in Stage Si, and so forth, in a cyclic manner (i.e., B«->C, then C«->D, then D«->A). The redesignation of the two blocks replacing each other is achieved with the aid of an array of valves. A useful array of valves is shown in Figure 6 and explained in a tabular manner below. Briefly, each block A, B, C and D is equipped with two pairs of valves; a first pair which regulates a flow of a stream supplied by the feed solution in the normal direction across the block, and a second pair of valves, which regulates the flow of the joined concentrate in the opposite direction across the block. So, based on the state of these valves, each block can be disconnected from the supply of the feed solution and connected to receive the concentrate as a feed solution, transitioning between Stage Si and Stage S2, and vice versa.

[0037] In another aspect, the invention relates to a process described herein wherein the rotation of the blocks between stages is performed through the opening and closing of an array of valves. In the case of a two stage system, said rotation can be performed by disconnecting at least one block of stage Si from the supply of the feed solution in the normal direction and connecting it to receive the concentrate as a feed solution in the opposite direction and disconnecting at least one block of stage S2 from receiving concentrate as a feed solution in the opposite direction and connecting it to the supply of the feed solution in the normal direction. The process can also comprise disconnecting at least one block of any stage.

[0038] The timing of the transition of blocks between stages is determined based on considerations described below.

[0039] The treatment process involves cycles consisting of distinct steps corresponding to Figures 1 to 5:

[0040] In the first step, which is shown in Figure 1, V circulation is open, and V discharge is closed. The feed flow (denoted Qi) is pressurized in pump Pl and passed in the normal direction through blocks B, C and D which perform at Stage Si. The rejected concentrate streams from blocks B, C and D are joined and the joined concentrate stream, denoted QfC, is directed by pump P2 to flow in the opposite direction through a circulation loop which includes Block A, which was initially set apart for working at Stage S2. Qfc is mixed and circulated with the concentrate produced by block A operating in Stage S2. Because V discharge is closed, then all the concentrate produced by the system is circulated through Block A, and no concentrate is withdrawn from the system, such that the total system discharge (denoted Qb) is null. The circulation of the concentrate in this mode (with Qb=0) is continued for a duration determined based on one or more selected properties indicative of the composition of the stream QfC, as explained in more detail below. For example, because solution conductivity increases with increasing solute concentration, conductivity can serve as a good indicator for the composition of the circulated stream QfC. When the measurable / estimated result associated with the selected property falls outside a predetermined range or is approaching the limit of that range, then the system switches to operate under the conditions of the second step.

[0041] In the second step, which is shown in Figure 2, V circulation is only partially open, and V discharge is at least partially open so that the total concentrate stream rejected by block A operating at Stage S2 is divided between the circulation line (Qc) and the discharge line (Qb). Thus, the concentrate is partially withdrawn from the system (i.e., Qb>0). The extent to which V circulation is open is adjusted to keep the measurable / estimated result associated with the selected property of Qcwithin the predetermined range.

[0042] In the third step, which is shown in Figure 3, V circulation is closed, and V discharge is open. Therefore, no concentrate flow is circulated, namely, Qcequals 0, and Qb>0.

[0043] In the fourth step, which is shown in Figures 4-5, the transitioning of blocks between stages is done, by determining the state of the relevant pairs of valves to a) disconnect block B from the supply of the aqueous feed solution and connect Block B to the concentrate line, such that blocks A and B both temporarily serve in Stage S2 with their concentrate being discharged (Qc=0, Qb>0), as shown in Figure 4, and b) disconnect block A from the concentrate line and connect block A to the supply of the aqueous feed solution, thereby reaching the flow configuration shown in Figure 5, which corresponds to the first step, with the difference that Blocks A and B were exchanged for one another, that is, blocks A, C and D operate at Stage Si, and block B operates at Stage S2. It should be understood, however, that the fourth step does not necessarily follows steps 1-3 in a sequential manner. Induction time is the predominant factor in determining when rotation between blocks should be performed, as explained in more detail below, and block rotation takes priority accordingly.

[0044] Another aspect of the invention is a multiple stage reverse osmosis (RO) or nano filtration (NF) system comprising: A feed water supply unit operated by a first pump and connected by a feed line and a feed manifold to the feed ports of pressure vessels assembled in blocks, said pressure vessels housing membrane modules; a permeate recovery line connected to permeate ports of the pressure vessels; a concentrate flow line connected to concentrate ports of said pressure vessels, wherein the concentrate flow diverges into a concentrate circulation loop equipped with a circulation valve (V circulation) and a concentrate discharge line equipped with a discharge valve (V discharge), operated by a second pump and optionally a third pump; an array of valves associated with the blocks of pressure vessels; a control unit determining the designation of blocks to flow stages, by connecting / disconnecting the pressure vessels to and from the feed line and the concentrate flow line by adjusting the states of the valves associated with the pressure vessels; optionally, a conductivity meter positioned along the concentrate circulation line, wherein the control unit is programmed to regulate the flow of the concentrate stream in the concentrate circulation loop and concentrate circulation line based on indication received from said conductivity meter; optionally, an energy recovery device (ERD) before P / FV between Qb and Qc(ERD could be Isobaric type (Pressure exchange), Work exchange or Turbo Charge types as energy converted from brine (concentrate) stream to feed stream of one of the previous stages (see Figure 7).

[0045] The present invention provides a membrane separation process which separates an original solution into a volume with higher concentration of soluble and / or insoluble material and a volume of lower concentration of soluble and / or insoluble material. This process consists of 1 to 9 spiral wound reverse osmosis (RO) or nanofiltration (NF) membranes inside of pressure vessels that are structured in blocks that form multiple stages of separation. A given block may alternate in its position as part of the first stage or last stage of separation. When in the first stage it will be fed from the opposite end of the pressure vessel as when it is in the last stage. Additionally, the concentrated solution leaving the last stage of the system may be repressurized (with the aid of the ERD device) and recirculated in a closed circuit to the feed side of the last stage and only be periodically purged from the system. The number of blocks, stages and passes can vary depending on recovery rate, TDS (Total Dissolved Solids), saturated ion or mineral, system size, flow and water quality. Stages can be separated by pumps for required pressure adjustments. A "Stage" is defined as a configuration where multiple RO units are connected in series. In this setup, the brine concentrate from the first unit becomes the feedwater for the second unit, and so on. Each additional stage increases the desalination efficiency and water recovery but also requires more energy and infrastructure.

[0046] The present invention, through the recirculation of the concentrated solution from the last stage will allow for the concentration factor to be adjusted while maintaining appropriate crossflow velocities within the membranes. Additionally, the alternation of the blocks and redirecting the feed direction will allow membranes to alternate in their position from being exposed to the highest concentrations to the lowest concentrations. These adjustments may be used to maximize the amount of the concentration factor, while minimizing the effects of fouling and scaling on the membrane surface.

[0047] Some percent of reject (brine / concentrate) circulation during operation is known and acceptable for hydraulic adjustments however, in conventional systems outside of the scope of this invention, circulation is normally from brine of last reverse osmosis stage to first stage feed (regardless of number of stages) Figure 8.

[0048] Reject circulation is designed for brackish water desalination and water reclamation and is operated by alternating between filtration and drain phases. During this phase of circulation, the flow rate of water entering at the inlet of membrane elements equals the sum of the feed water flow and the recycled flow while the permeate flow equals the feed water flow. As most of the salt remains in the closed loop, the salinity of the water in the loop increases with time, and therefore the transmembrane pressure (TMP) increases to maintain constant water flux through the membrane. (Dhrubajit et al., 2024)

[0049] The present invention provides advantages over previous inventions as it allows for both a flexible concentration factor through a closed-circuit recirculation of the last stage while simultaneously rotating the blocks so that the membranes in the last stage are alternated with membranes in any one of the previous stages to reduce the effects of fouling and scaling. This also reduces the amount of excess volume required for a closed-circuit design, utilizes the benefits the advantages that most of the system operates on the feed water- enabling operation on potentially descaling water and reducing the total permeate conductivity variability common to closed circuit / semi batch systems. The time required for a supersaturated aqueous solution to start precipitating sparingly soluble salts and / or minerals therefrom (in the absence of seed crystals or other appropriate pre-existing crystal growth surface) is known as the “induction time” (T). Most preferably, the process according to the present invention comprises the estimation of the induction time associated with the composition of the water to be treated and the nature of the membrane surface, to permit the flow reversal to be carried out at the desired periodicity (as mentioned in reference to step 4).

[0050] According to one preferred embodiment, the induction time is estimated experimentally, for example, by carrying out a preliminary experiment, in which the water stream corresponding to the final concentrate composition is passed along a small membrane unit under flow conditions generating the concentration polarization that would be expected in a full-scale unit, without reversing the flow direction. The period of time measured until sparingly soluble salts from said water stream precipitate onto the membrane surface as reflected in a drop in membrane water permeability, or until a drop in the bulk concentration of the sparingly soluble salt is detected, is subsequently used to schedule the flow reversal accordingly. The occurrence of the precipitation of sparingly soluble salts onto the membrane surface may be conveniently tracked by measuring changes to the membrane water permeability, Lp, as measured by the flux, Jv, divided by the net pressure driving force (NDP) as defined by:

[0051] Where AP and An are the transmembrane pressure and the osmotic pressure difference between feed and permeate solutions respectively.

[0052] Alternatively, the induction time is estimated using the following equation: wherein:

[0053] A and B are constants related to the salt, the hydrodynamics and the membrane surface, since the nucleation that must be prevented is the nucleation at the membrane surface. Constant A includes the effect of the surface energy at the nucleating surface and molar volume of the salt. B includes the frequency factor for the nucleation rate. Values of A and B have already been determined, for example, for calcium sulfate and polyamide low pressure RO membranes (e.g., D. Hasson at al, “Induction times induced in an RO system by antiscalants delaying CaSO4 precipitation”, Desalination, v.157 (2003), p 193; Alimi, F. et al, Desalination V 157, pp 9-16 (2003); He, S. et al, J. Colloid and Interface Science, V 162 pp. 297-303 (1994)). A and B may also be readily obtained experimentally.

[0054] S is the ratio between the maximum actual activity product of a given sparingly salt in the solution next to the feed side of the membrane and the thermodynamic solubility product, or, for an undissociated mineral (e.g. silica), it is the ratio of the actual maximum mineral concentration to the saturation concentration of that mineral for the given composition. It will increase when the product water recovery is increased, also increasing the concentration polarization. This ratio can be calculated from commercial desalination process simulation programs available in the open market (such as IMSdesign from Hydranautics, Inc or WAVE from Dupont Inc. etc.), or independently as described in D. Hasson at al, “Inception of CaSCE scaling on RO membranes at various water recovery levels”, Desalination, 139, 73-81 (2001). S may be also calculated upon measuring data related to the concentration level of salts / mineral in the water stream, and data related to the permeate and concentrate flows. This data can be obtained utilizing appropriately placed flow meters and sensing devices within the system, in order to provide corresponding signals (e.g., conductivity) for calculating S.

[0055] Having obtained and / or experimentally determined the values of A, B and S, for the relevant salt and system, one may easily estimate the induction time, T. The block rotation, achieving flow reversal, will then be set to take place at some time less than the T, for example at 0.1 to 0.9 the value of the induction time. In the event that there is more than one sparingly soluble salt, the salt with the shortest induction time will be the one used for determining the period for flow reversal. Typically, induction times are above 10 minutes, e.g., from 45 to 180 minutes.

[0056] Alternatively, the induction time may be determined during system operation via appropriately situated sensors for monitoring permeation rate (flow meter or weigh scale) or scaling ion concentration (e.g. using calcium sensor, or conductivity sensor) as a function of time, to see when these indicators change as a result of precipitation in the system.

[0057] The invention is now described in reference to Figure 6 - 7, explaining the role of the valves used in the multistage process and system of the invention.

[0058] In first step, feed flow (Qf) is inlet flow. This feed is pressurized in pump Pl while valves AFI, AFO, BRI, BRO, CRI, CRO and DRI, DRO are closed and valves ARI, ARO, BFI, BFO, CFI, CFO and DFI, DFO are open allowing flow direction from circulation valve AVF. Brine-to-drain valve P / FV is closed, allowing feed flow (Qf) to pass from left to right in Blocks B, C and D which perform as 1ststage and reject flow (Qfc) from these Blocks to flow from right to left in Block A as Block A is 2ndstage (or n stage in more complex systems). Pump P3 is optional for better hydraulics performances. QfCis mixed with circulated concentrate from 2ndstage and regulated by valves P / FV and / or AVF which are controlled by conductivity of QfC. As QfCconductivity is below threshold as calculated from equation 6 or by prior experiments. Qb is total system discharge.

[0059] In step 2, Qcconcentrated flow is circulated back to stage 2 feed. If the conductivity of QfCincreased above design minimum valve P / FV is opened. The opening of valve AVF is controlled to stabilize QfCconductivity; the use of pump P3 is optional in this step. Qb is then greater than 0 and stage 2 concentrate discharges to the drain.

[0060] In step 3 while valve AVF is fully closed, valve P / FV is open and Qcequals 0, in this scenario, no concentrate flow is circulated. P3 pump is optional to operate to stabilize hydraulics.

[0061] In step 4, Block A switches position with block B via a sequence of switching of the AFI, AFO, BFI, BFO, ARI, ARO, BRI, BRO valves to prevent dead end and water hammer effects. While conductivity of QfCreaches threshold as calculated from equation 6 or by prior experiments, valves BFI, BFO, ARO and ARI are closed and valves BRO, BRI, AFI and AFO are open. Then Blocks A, C and D operates as stage 1 and Block B operates as stage 2. An energy recovery device (ERD) can optionally be added before P / FV between Qb and Qc.

[0062] Pumps can be of centrifugal type with flow-pressures specs as calculated for feed water flow rate, recovery rate and osmotic pressure and will be based on design calculations.

[0063] Pneumatic ball valves designed for pressures and diameter can be installed for flow control and transition. Membranes include spiral wound membranes and can be installed in GRP or stainless- steel pressure vessels.

[0064] Valves used are normally of pneumatical or electrical type. Pumps should be capable of pressure variety and equipped with VFD (Variable Frequency Drive). Brine valves normally match the changes in flows.

[0065] Pressure vessels and membranes of rotating stages (normally first and last stages) should be capable of hydraulics involved both as first and last stages (flows, pressures, fluxes).

[0066] Examples:

[0067] A full experimental system was designed, built and named RP03 (referred as FR-RC System) and experiments conducted upon the new design. The system can be seen in Figure 9.

[0068] The system consists of one, two or three stages while each pressure vessel (PV) is built of two equal PVs in series with four membranes each. The overall eight membranes in each of the two PVs in series will be labeled and referred to as a "Block". The flexible pilot system can operate in different modes of operation. For the trials presented in this study, only FR-RC mode was operated and as such the operation of the blocks are described below.

[0069] During the experiment, four blocks are on 1ststage and two blocks will operate as 2ndstage. The rotation between 1ststage and 2ndstage operation will be controlled by valves ARO, AFI, ARI, AFO, BRO, BFI, BRI, BFO, CRO, CFI, CRI, CFO, DRO, DFI, DRI, EFO, ERO, EFI, ERI, EFO, FRO, FFI, FRI, FFO.

[0070] Circulation will be controlled by valves AVF and P / FV which will range from 0% to 100% circulation. During 100% circulation operation, valve P / FV will be fully closed. During the 0% circulation ("Dump" step), AVF will be fully closed and P3 will shut down (see Figure 10 - Basic system configuration). When operating in FRRC mode, there will be no circulation to 1ststage feed as CiV is closed.

[0071] The membrane elements used are 4" in diameter type BW30-4040 membranes.

[0072] Literature such as Stover (2016), mentions that closed loop and concentrate circulation systems are designed with shorter membrane arrays, typically consisting of three or four elements, compared to traditional RO systems that use six to eight elements. This design helps in providing a more balanced flux distribution along the membrane array, reducing the flux through the first element and, consequently, decreasing the risk of fouling.

[0073] In the following experiments and invention, though closed circuit is indeed applied along with phases of reject circulation, reject is circulated into same stage feed, in this case, two stage system while reject from 2ndstage is circulated back to 2ndstage (and not 1ststage) feed in range of O% to 100%. Another major difference is the block rotation which applied between first and last stages as well and long pressure vessels with six to eight membranes maintain the possibility of operation as conventional reverse osmosis system.

[0074] Experiments were performed in-situ with Rotec pilot system.

[0075] Feed water profile (as estimated by water quality in the area as well as direct measurements) are presented in Table 2.

[0076] Table 2-Feed water profile

[0077] Conventional operation

[0078] For better evaluation of the new suggested technology, the control experiment was operated in 4- 2 configuration as shown in Figure 11 - Experiment system design.

[0079] For conventional operation Experiment:

[0080] Feed water: 7 m3 / h

[0081] Overall recovery rate: 77%

[0082] Projection performed showed operation with the given feed water as stable at 85% recovery rate. However, due to hydraulic and chemistry limitations, 85% recovery rate operation could not be performed at a commercial level.

[0083] To deal with the hydraulic issues, circulation could be added (CiV OPEN).

[0084] As circulation was applied, it removed the hydraulic issues; however, saturation indexes showed severe precipitation over 1stand 2ndstages. This system could not operate for longer time periods. Evaluation of FR-RC operational mode

[0085] System configured with reject circulation from 2ndstage to 2ndstage feed was operated in following step:

[0086] Steps

[0087] Step 1 : Full circulation

[0088] Stage 1 : Blocks A, B and C

[0089] Stage 2: Block D

[0090] Circulation: 100% for 10 minutes

[0091] Step 2: Partial circulation

[0092] Stage 1 : Blocks A, B and C

[0093] Stage 2: Block D

[0094] Circulation: 50% for 40 minutes

[0095] Step 3: Dump

[0096] Stage 1 : Blocks A, B and C

[0097] Stage 2: Block D

[0098] Circulation: 0% for 10 minutes

[0099] Step 4: Rotation

[0100] Valves: CFI and CFO closed

[0101] Valves: CRI and CRO opened

[0102] Valves: DFI and DFO opened

[0103] Valves: DRI and DRO closed

[0104] Stage 1 : Blocks A, B and D

[0105] Stage 2: Block C

[0106] Circulation: 0%

[0107] The above steps repeat as blocks rotate [A, C and D as stage 1 and B as stage 2], [B, C and D as stage 1 and A as stage 2] and so on (Figure 12- Step by Step FRRC with block rotation).

[0108] Experiment:

[0109] Feed: 7 m3 / h

[0110] Recovery rate 1# phase: 62% Recovery rate 2# phase: 60%

[0111] Overall recovery rate: 85% to 100% along steps 1 and 4

[0112] Goal:

[0113] Stable operation

[0114] Stable Differential Pressure (DP)

[0115] Overall recovery rate will be calculated and documented during the study.

[0116] Results

[0117] Operation in FRRC mode showed stable operation with slight conductivity decrease which could be explained by variation of feed water profile and present normal phenomenon. Pressures throughout the system remained stable as the recovery rate was stable at 91% (Figure 13- Recovery rate, pressures and conductivity over time (h)), and flows remained stable during the operation as shown in Figure 14 (Feed flow 1st and 2nd stages over time (h)).

[0118] During the experiment, DP remained stable (Figure 15 - Differential Pressure over time (h)) and the overall recovery rate was 91%,

[0119] As shown herein, the novel approach of the present invention, using full and semi recirculation of the reject stream to the second stage and rotation between blocks of the first and second stages, enabled flux adjustments along the RO membranes and operation with a relatively low feed flow without major flux and brine flow concerns. The rotation of stages facilitated the introduction of undersaturated feed water to membranes operating in the later stages under high saturation conditions, as brine from these stages was recirculated via the second stage's feed. Throughout the experiments, there was no increase in differential pressure and system feed pressure, which indicates stable operation. This operational method, combined with induction time calculations, allowed for sustained operation at higher recoveries compared to conventional two stage RO system.

[0120] The pilot system used (i.e. RP03) allowed block rotation, semi as well as full circulation of reject stream and operation of two stage RO in conventional and FR-RO (flow reversal - reverse osmosis) was flexible and allowed comparison of different technologies over same system configuration.

Claims

Claims1) A multistage membrane separation process for water treatment using membrane modules connected in series inside pressure vessels assembled in blocks, comprising: dividing the blocks into two or more stages Si (i =1 , 2, ... ), wherein the number of blocks assigned to stage Si are labeled Bi; passing a pressurized aqueous feed stream from a feed supply unit through the Bi blocks designated to stage Si, wherein the aqueous feed stream flows in a first direction through each of the Bi blocks operating at stage Si, to produce Bi permeate streams and Bi concentrate streams, joining the Bi concentrate streams and circulating the joined concentrate stream of stage S i through one or more blocks belonging to stage S2, wherein the concentrate stream flows at a direction opposite or identical to said first direction; and when i>2, circulating a joined concentrate stream of at least one of stages SI-SM through one or more blocks designated to stage Si, wherein the concentrate stream flows at stage Si in a direction opposite to or identical to that of Si-i; discharging the concentrate stream circulated at stage Si (i>2); and rotating blocks between stages.2) The process of claim 1 wherein the concentrate stream flows at stage S2 in a direction opposite to that of Si.3) A process according to claim 1 or 2, wherein i=2, and the blocks are divided to two stages Si and S2, the numbers of blocks in stages Si and S2 are 2<Bi<10 and B2 =1.4) The process of any one of the previous claims, comprising monitoring at least one physical parameter of at least one concentrate stream using at least one sensor.5) The process of claim 4, wherein said parameter comprises the conductivity of said stream.6) The process of any one of the previous claims comprising monitoring the permeation rate and / or scaling ion concentration.7) The process of any one of the previous claims wherein said block rotation is performed after a suitable induction time.8) The process of any one of the previous claims, wherein a proportion of the concentrated stream of stage Si is recirculated and adjoined to said joined concentrate stream of at least one of stages Si-Si-i.9) The process of any one of the previous claims, wherein said concentrated stream of stage Si is initially entirely recirculated and adj oined to said j oined concentrate stream of at least one of stages Si-Si-i.10) The process of claim 9 wherein at least part of said concentrated stream of stage Si is discharged once at least one monitored physical parameter has reached a predetermined threshold.11) The process of any one of claims 8 wherein the proportion of Si recirculation and discharge is calculated based on the conductivity monitoring of at least one concentrate stream.12) The process of any one of the previous claims wherein said rotating blocks between stages is performed through the opening and closing of an array of valves.13) The process of claim 12 comprising disconnecting at least one block of stage Si from the supply of the feed solution in the normal direction and connecting it to receive the concentrate as a feed solution in the opposite direction and disconnecting at least one block of stage S2 from receiving concentrate as a feed solution in the opposite direction and connecting it to the supply of the feed solution in the normal direction.14) The process of claim 12 comprising disconnecting at least one block of any stage.15) A multiple stage reverse osmosis (RO) or nano filtration (NF) system comprising:a feed water supply unit operated by a first pump and connected by a feed line and a feed manifold to the feed ports of pressure vessels assembled in blocks, said pressure vessels housing membrane modules; a permeate recovery line connected to permeate ports of the pressure vessels; a concentrate flow line connected to concentrate ports of said pressure vessels, wherein the concentrate flow diverges into a concentrate circulation loop equipped with a circulation valve (V circulation) and a concentrate discharge line equipped with a discharge valve (V discharge), operated by a second pump and optionally a third pump; an array of valves associated with the pressure vessels; a control unit determining the designation of blocks to flow stages, by connecting / disconnecting the pressure vessels to and from the feed line and the concentrate flow line by adjusting the states of the valves associated with the pressure vessels; optionally, a conductivity meter positioned along the concentrate circulation line, wherein the control unit is programmed to regulate the flow of the concentrate stream in the concentrate circulation loop and concentrate circulation line based on indications received from said conductivity meter; optionally, an energy recovery device (ERD) installed upstream to V circulation (AVF), wherein the ERD is preferably selected from isobaric type (pressure exchange) ERD, work exchange type ERD or turbo charge type ERD.

Images