This page is Membrane Filtration

Overview of Membrane Filtration

There are several classes of treatment processes that constitute membrane filtration for the purposes of Long-Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR) compliance. These processes include: microfiltration (MF), ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO). In addition, cartridge filtration devices that meet the criteria for a membrane filtration process as defined under the rule would also be eligible for Cryptosporidium removal credit as membrane filtration (40 CFR 141.719(b)(1)). For the purposes of this guidance manual, these devices are termed membrane cartridge filtration (MCF).

Each of these technologies utilizes a membrane barrier that allows the passage of water but removes contaminants. The membrane media is generally manufactured as flat sheets or as hollow fibers and then configured into membrane modules. The most common membrane module configurations are hollow-fiber (consisting of hollow-fiber membrane material), spiral- wound (consisting of flat sheet membrane material wrapped around a central collection tube), and cartridges (consisting of flat sheet membrane material that is often pleated to increase the surface area).

Although the spiral-wound and cartridge configurations are also termed as “elements” and “cartridges,” respectively, under the LT2ESWTR the term “module” – defined as the smallest component of a membrane unit in which a specific membrane surface area is housed in a device with a filtrate outlet structure – is used to refer to all of the various membrane module configurations for simplicity of nomenclature.

In addition to the various module configurations, there are a number of different types of membrane materials, hydraulic modes of operation, and operational driving forces (i.e., pressure or vacuum) that can vary among the different classes of membrane filtration (i.e., MF, UF, NF, RO, and MCF). Each of these characteristics of membrane filtration systems may be considered tools that a manufacturer may utilize to meet the particular treatment objectives for a given application.

Basic Principles of Membrane Filtration

A membrane filtration process is defined by two basic criteria

  1. The filtration system must be a pressure- or vacuum-driven process and remove particulate matter larger than 1 μm using an, engineered barrier, primarily via a size exclusion mechanism.

  2. The process must have a measurable removal efficiency of a target organism that can be verified through the application of a direct integrity test.


The ability of each of type of membrane filtration system to remove various drinking water pathogens of interest on the basis of size is illustrated in Figure 2.1. The figure shows the approximate size range of viruses, bacteria, Cryptosporidium oocysts, and Giardia cysts, as well as the ability of MF, UF, NF, RO, and MCF, respectively, to remove each of these pathogens on the basis of size exclusion. Overlap between the range covered by a membrane filtration process with a given pathogen size range indicates the ability of that process to remove the pathogen. Note that the molecular weights listed do not correspond precisely to the indicated pathogen size range, but are rough generalizations depicted as a result of the fact that NF, RO, and some UF processes are rated according to a “molecular weight cutoff” on the basis of their ability to remove dissolved phase constituents.

Although each of the classes of membrane filtration functions as a filter for various sizes of particulate matter, the basic principles of operation vary between MF/UF, NF/RO, and MCF systems. Each of these types of systems is described in the following sections.

Microfiltration and Ultrafiltration

MF and UF are the two processes that are most often associated with the term “membrane filtration.” MF and UF are characterized by their ability to remove suspended or colloidal particles via a sieving mechanism based on the size of the membrane pores relative to that of the particulate matter. However, all membranes have a distribution of pore sizes, and this distribution will vary according to the membrane material and manufacturing process. When a pore size is stated, it can be presented as either nominal (i.e., the average pore size), or absolute (i.e., the maximum pore size) in terms of microns (μm). MF membranes are generally considered to have a pore size range of 0.1 – 0.2 μm (nominally 0.1 μm), although there are exceptions, as MF membranes with pores sizes of up to 10 μm are available. For UF, pore sizes generally range from 0.01 – 0.05 μm (nominally 0.01μm) or less, decreasing to an extent at which the concept of a discernable “pore” becomes inappropriate, a point at which some discrete macromolecules can be retained by the membrane material. In terms of a pore size, the lower cutoff for a UF membrane is approximately 0.005 μm.

Because some UF membranes have the ability to retain larger organic macromolecules, they have been historically characterized by a molecular weight cutoff (MWCO) rather than by a particular pore size. The concept of the MWCO (expressed in Daltons – a unit of mass) is a measure of the removal characteristic of a membrane in terms of atomic weight (or mass) rather than size. Thus, UF membranes with a specified MWCO are presumed to act as a barrier to compounds or molecules with a molecular weight exceeding the MWCO. Because such organic macromolecules are morphologically difficult to define and are typically found in solution rather than as suspended solids, it may be convenient in conceptual terms to use a MWCO rather than a particular pore size to define UF membranes when discussed in reference to these types of compounds. Typical MWCO levels for UF membranes range from 10,000 to 500,000 Daltons, with most membranes used for water treatment at approximately 100,000 MWCO. However, UF membranes remove particulate contaminants via a size exclusion mechanism and not on the basis of weight or mass; thus, UF membranes used for drinking water treatment are also characterized according to pore size with respect to microbial and particulate removal capabilities.

Nanofiltration and Reverse Osmosis

NF and RO constitute the class of membrane processes that is most often used in applications that require the removal of dissolved contaminants, as in the case of softening or desalination. The typical range of MWCO levels is less than 100 Daltons for RO membranes, and between 200 and 1,000 Daltons for NF membranes. While NF and RO are sometimes referred to as “filters” of dissolved solids, NF and RO utilize semi-permeable membranes that do not have definable pores. NF and RO processes achieve removal of dissolved contaminants through the process of reverse osmosis, as described below.

NF/RO membranes are designed to remove dissolved solids through the process of reverse osmosis. Osmosis is the natural flow of a solvent, such as water, through a semi- permeable membrane (acting as a barrier to dissolved solids) from a less concentrated solution to a more concentrated solution. This flow will continue until the chemical potentials (or concentrations, for practical purposes) on both sides of the membrane are equal. The amount of pressure that must be applied to the more concentrated solution to stop this flow of water is called the osmotic pressure. An approximate rule of thumb for the osmotic pressure of fresh or brackish water is approximately 1 psi for every 100 mg/L difference in total dissolved solids (TDS) concentration on opposite sides of the membrane.

Reverse osmosis, as illustrated in Figure 2.2, is the reversal of the natural osmotic process, accomplished by applying pressure in excess of the osmotic pressure to the more concentrated solution. This pressure forces the water through the membrane against the natural osmotic gradient, thereby increasingly concentrating the water on one side (i.e., the feed) of the membrane and increasing the volume of water with a lower concentration of dissolved solids on the opposite side (i.e., the filtrate or permeate). The required operating pressure varies depending on the TDS of the feed water (i.e., osmotic potential), as well as on membrane properties and temperature, and can range from less than 100 psi for some NF applications to more than 1,000 psi for seawater desalting using RO.


Both NF and RO are pressure-driven separation processes that utilize semi-permeable membrane barriers. NF differs from RO only in terms of its lower removal efficiencies for dissolved substances, particularly for monovalent ions. This results in unique applications of NF, such as the removal of hardness ions at lower pressures than would be possible using RO. Consequently, NF is often called “membrane softening.” The differences between NF and RO are irrelevant with respect to the removal of particulate matter.

Because semi-permeable NF and RO membranes are not porous, they have the ability to screen microorganisms and particulate matter in the feed water; however, they are not necessarily absolute barriers. NF and RO membranes are specifically designed for the removal of TDS and not particulate matter, and thus the elimination of all small seal leaks that have only a minor impact on the salt rejection characteristics is not the primary focus of the manufacturing process. Consequently, NF and RO spiral-wound elements are not intended to be sterilizing filters and some passage of particulate matter may occur despite the absence of pores in the membrane, which can be attributed to slight manufacturing imperfections (Meltzer 1997).

Membrane Materials, Modules, and Systems

There are a number of different types of membrane materials, modules, and associated systems that are utilized by the various classes of membrane filtration. While several different types of membrane modules may be employed for any single membrane filtration technology, each class of membrane technology is typically associated with only one type of membrane module in water treatment applications. In general, MF and UF use hollow-fiber membranes, and NF and RO use spiral-wound membranes.

Membrane Materials

The membrane material refers to the substance from which the membrane itself is made. Normally, the membrane material is manufactured from a synthetic polymer, although other forms, including ceramic and metallic “membranes,” may be available. Currently, almost all membranes manufactured for drinking water production are made of polymeric material, since they are significantly less expensive than membranes constructed of other materials.

The material properties of the membrane may significantly impact the design and operation of the filtration system. For example, membranes constructed of polymers that react with oxidants commonly used in drinking water treatment should not be used with chlorinated feed water. Mechanical strength is another consideration, since a membrane with greater strength can withstand larger transmembrane pressure (TMP) levels allowing for greater operational flexibility and the use of higher pressures with pressure-based direct integrity testing. Similarly, a membrane with bi-directional strength may allow cleaning operations or integrity testing to be performed from either the feed or the filtrate side of the membrane. Material properties influence the exclusion characteristic of a membrane as well. A membrane with a particular surface charge may achieve enhanced removal of particulate or microbial contaminants of the opposite surface charge due to electrostatic attraction. In addition, a membrane can be characterized as being hydrophilic (i.e., water attracting) or hydrophobic (i.e., water repelling). These terms describe the ease with which membranes can be wetted, as well as the propensity of the material to resist fouling to some degree.

MF and UF membranes may be constructed from a wide variety of materials, including cellulose acetate (CA), polyvinylidene fluoride (PVDF), polyacrylonitrile (PAN), polypropylene (PP), polysulfone (PS), polyethersulfone (PES), or other polymers. Each of these materials has

different properties with respect to surface charge, degree of hydrophobicity, pH and oxidant tolerance, strength, and flexibility.

NF and RO membranes are generally manufactured from cellulose acetate or polyamide materials (and their respective derivatives), and there are various advantages and disadvantages associated with each. While cellulose membranes are susceptible to biodegradation and must be operated within a relatively narrow pH range of about 4 to 8, they do have some resistance to continuous low-level oxidant exposure. In general, for example, chlorine doses of 0.5 mg/L or less may control biodegradation as well as biological fouling without damaging the membrane. Polyamide (PA) membranes, by contrast, can be used under a wide range of pH conditions and are not subject to biodegradation. Although PA membranes have very limited tolerance for the presence of strong oxidants, they are compatible with weaker oxidants such as chloramines. PA membranes require significantly less pressure to operate and have become the predominant material used for NF and RO applications.

A characteristic that influences the performance of all membranes is the trans-wall symmetry, a quality that describes the level of uniformity throughout the cross-section of the membrane. There are three types of construction that are commonly used in the production of membranes: symmetric, asymmetric (including both skinned and graded density variations), and composite. Cross-sectional diagrams of membranes with different trans-wall symmetry are shown in Figure 2.3. Symmetric membranes are constructed of a single (i.e., homogeneous) material, while composite membranes use different (i.e., heterogeneous) materials. Asymmetric membranes may be either homogeneous or heterogeneous.

In a symmetric membrane, the membrane is uniform in density or pore structure throughout the cross-section, while in an asymmetric membrane there is a change in the density of the membrane material across the cross sectional area. Some asymmetric membranes have a graded construction, in which the porous structure gradually decreases in density from the feed to the filtrate side of the membrane. In other asymmetric membranes, there may be a distinct transition between the dense filtration layer (i.e., the skin) and the support structure. The denser skinned layer is exposed to the feed water and acts as the primary filtration barrier, while the thicker and more porous understructure serves primarily as mechanical support. Some hollow- fibers may be manufactured as single- or double- skinned membranes, with the double skin providing filtration at both the outer and inner walls of the fibers. Like the asymmetric skinned membranes, composite membranes also have a thin, dense layer that serves as the filtration barrier. However, in composite membranes the skin is a different material than the porous substructure onto which it is cast. This surface layer is designed to be thin so as to limit the resistance of the membrane to the flow of water, which passes more freely through the porous substructure. NF and RO membrane construction is typically either asymmetric or composite, while most MF and UF membranes are either symmetric or asymmetric.

Membrane Modules

Membrane filtration media is usually manufactured as flat sheet stock or as hollow fibers and then configured into one of several different types of membrane modules.

Module construction typically involves potting or sealing the membrane material into a corresponding assembly, which may incorporate an integral containment structure, such as with hollow-fiber modules. These types of modules are designed for long-term use over the course of a number of years. Spiral-wound modules are also manufactured for long-term use, although the design of membrane filtration systems that utilize spiral-wound modules requires that the modules be encased in a separate pressure vessel that is independent of the module itself. Alternatively, a module may be configured as a disposable cartridge with a useful life that is typically measured in weeks or months rather than years. Membrane cartridges may either be inserted into pressure vessels that are separate from the module (as with spiral-wound modules) or manufactured within a casing that serves as an integral pressure vessel. Each of these three types of modules, along with some other less common module designs, is discussed in the following subsections.

Hollow-Fiber Modules

Most hollow-fiber modules used in drinking water treatment applications are manufactured to accommodate porous MF or UF membranes and designed to filter particulate matter. As the name suggests, these modules are comprised of hollow-fiber membranes, which are long and very narrow tubes that may be constructed of any of the various membrane materials described in section 2.3.1. The fibers may be bundled in one of several different arrangements. In one common configuration used by many manufacturers, the fibers are bundled together longitudinally, potted in a resin on both ends, and encased in a pressure vessel that is included as a part of the hollow-fiber module. These modules are typically mounted vertically, although horizontal mounting may also be utilized. One alternate configuration is similar to spiral-wound modules in that both are inserted into pressure vessels that are independent of the module itself. These modules (and the associated pressure vessels) are mounted horizontally. Another configuration in which the bundled hollow fibers are mounted vertically and submerged in a basin does not utilize a pressure vessel. A typical commercially available hollow-fiber module may consist of several hundred to over 10,000 fibers. Although specific dimensions vary by manufacturer, approximate ranges for hollow-fiber construction are as follows:

  • Outside diameter: 0.5 – 2.0 mm

  • Inside diameter: 0.3 – 1.0 mm

  • Fiber wall thickness: 0.1 – 0.6 mm

  • Fiber length: 1 – 2 meters

A cross section of a symmetric hollow-fiber is shown in Figure 2.4.

Hollow-fiber membrane modules may operate in either an “inside-out” or “outside-in” mode. In inside-out mode, the feed water enters the fiber lumen (i.e., center or bore of the fiber) and is filtered radially through the fiber wall. The filtrate is then collected from outside of the fiber. During outside-in operation, the feed water passes from outside the fiber through the fiber wall to the inside, where the filtrate is collected in the lumen. Although inside-out mode utilizes a well-defined feed flow path that is advantageous when operating under a crossflow hydraulic configuration (see section 2.5), the membrane is somewhat more subject to plugging as a result of the potential for the lumen to become clogged. The outside-in mode utilizes a less well- defined flow feed flow path, but increases the available membrane surface area for filtration per fiber and avoids potential problems with clogging of the lumen bore.

Both the inside-out and outside-in operating modes for hollow-fiber modules utilizing pressure vessels are illustrated in Figure 2.5. When a hollow-fiber module is operated in an inside-out mode, the pressurized feed water may enter the fiber lumen at either end of the module, while the filtrate exits through a filtrate port located at the center or end of the module. In outside-in mode, the feed water typically enters the module through a inlet port located in the center and is filtered into the fiber lumen, where the filtrate collects prior to exiting through a port at one end of the module. Most hollow-fiber systems operate in “dead-end” or direct filtration mode (see section 2.5) and are periodically backwashed to remove the accumulated solids. Note that the submerged hollow-fiber membranes operate in outside-in mode, but do not utilize the pressure vessels (and the associated inlet ports) that are illustrated in Figure 2.5.

Spiral-Wound Modules

Spiral-wound modules were developed as an efficient configuration for the use of semi- permeable membranes to remove dissolved solids, and thus are most often associated with NF/RO processes. The basic unit of a spiral-wound module is a sandwich arrangement of flat membrane sheets called a “leaf” wound around a central perforated tube. One leaf consists of two membrane sheets placed back to back and separated by a fabric spacer called a permeate carrier. The layers of the leaf are glued along three edges, while the unglued edge is sealed around the perforated central tube. A single spiral-wound module 8 inches in diameter may contain up to approximately 20 leaves, each separated by a layer of plastic mesh called a spacer that serves as the feed water channel.

Feed water enters the spacer channels at the end of the spiral-wound element in a path parallel to the central tube. As the feed water flows across the membrane surface through the spacers, a portion permeates through either of the two surrounding membrane layers and into the permeate carrier, leaving behind any dissolved and particulate contaminants that are rejected by the semi-permeable membrane. The filtered water in the permeate carrier travels spirally inward around the element toward the central collector tube, while the water in the feed spacer that does not permeate through the membrane layer continues to flow across the membrane surface, becoming increasingly concentrated in rejected contaminants. This concentrate stream exits the element parallel to the central tube through the opposite end from which the feed water entered. A diagram of a spiral-wound element is shown in Figure 2.6.

Spiral-wound membranes for drinking water treatment are commercially available in a variety of sizes. Modules that are either 4 or 8 inches in diameter and either 40 or 60 inches long are most common, although other sizes may be used. Some bench- and pilot-scale applications utilize modules that are 2.5 inches in diameter, while modules up to 16 inches in diameter or more may be used in full-scale facilities.

Membrane Cartridges

Under the LT2ESWTR, cartridge filters that meet the criteria would be eligible to receive Cryptosporidium removal credit as a membrane filtration process.
In this case, the cartridge filter element would constitute a membrane module for the purposes of the rule. The ability of these modules to be subjected to direct integrity testing in the field during the course of normal operation, a feature that has not been widely utilized in association with cartridge filters in municipal water treatment applications, is a critical aspect of these systems that distinguishes what is considered to be MCF under the LT2ESWTR.

Membrane cartridge filters are manufactured by placing flat sheet membrane media between a feed and filtrate support layer and pleating the assembly to increase the membrane surface area within the cartridge. The pleat pack assembly is then placed around a center core with a corresponding outer cage and subsequently sealed, via adhesive or thermal means, into its cartridge configuration. End adapters, typically designed with a double o-ring sealing mechanism, are attached to the filter to provide a positive seal with the filter housing. A representative diagram of membrane cartridge filter is shown in Figure 2.7.

Most membrane cartridge filters are manufactured as disposable components that are inserted into a housing. Once the filter fouls to the point at which the maximum TMP is reached, the cartridge is replaced. Because the cartridges are designed to be disposable, and thus relatively inexpensive to replace, cartridge filtration systems have not historically utilized backwashing or chemical cleaning. However, some systems that feature these processes have recently been introduced. Cartridge filters are available in various sizes and pore sizes, although the device would have to be capable of filtering particulate matter larger than 1 μm to comply with the definition of a membrane filter under the LT2ESWTR.

Types of Membrane Filtration Systems

In drinking water treatment applications, each of the four traditional types of pressure- driven membrane processes (i.e., MF, UF, NF, and RO) is generally associated with a single type of membrane filtration system that is designed around a specific type of module. MF and UF systems typically utilize hollow-fiber modules, while NF and RO systems typically utilize spiral- wound modules. An overview of each of the two types of systems that utilize these respective modules is provided in the following subsections. Because the concept of a MCF system as defined under the LT2ESWTR is a new concept introduced with the rule, a standard type of MCF system has not yet been developed.

Hollow-Fiber (MF/UF) Systems

In drinking water treatment applications, each of the four traditional types of pressure- driven membrane processes (i.e., MF, UF, NF, and RO) is generally associated with a single type of membrane filtration system that is designed around a specific type of module. MF and UF systems typically utilize hollow-fiber modules, while NF and RO systems typically utilize spiral- wound modules. An overview of each of the two types of systems that utilize these respective modules is provided in the following subsections. Because the concept of a MCF system as defined under the LT2ESWTR is a new concept introduced with the rule, a standard type of MCF system has not yet been developed.

With few exceptions, most MF/UF processes utilize systems designed around hollow- fiber modules. Hollow-fiber membrane filtration systems are designed and constructed in one or more discrete water production units, also called racks, trains, or skids. A unit consists of a number of membrane modules that share feed and filtrate valving, and each respective unit can usually be isolated from the rest of the system for testing, cleaning, or repair. A typical hollow- fiber system is composed of a number of identical units that combine to produce the total filtrate flow.

Most of the currently available hollow-fiber membrane systems are proprietary, such that a single supplier will manufacture the entire filtration system, including the membranes, piping, appurtenances, control system, and other features. The manufacturer also determines the hydraulic configuration and designs the associated operational sub-processes – such as backwashing, chemical cleaning, and integrity testing – that are specific to its particular system. As a result, there are significant differences in the proprietary hollow-fiber membrane systems produced by the various manufacturers, and the membranes and other components are not interchangeable.

Although each manufacturer’s system is distinct, all of the hollow-fiber membrane systems fall into one of two categories – pressure-driven or vacuum-driven – according to the driving force for operation. In a pressure-driven system, pressurized feed water is piped directly to the membrane unit, where it enters the module and is filtered through the membrane. Typical operating pressures range from 3 to 40 psi. Most applications require designated feed pumps to generate the required operating pressure, although there are some water treatment plants that take advantage of favorable hydraulic conditions to operate a MF or UF system via gravity flow.

A schematic of a typical pressure-driven hollow-fiber membrane filtration system is shown in Figure 2.8. In the example shown, the system is operated in a “dead-end” hydraulic configuration (see section 2.5) and uses a liquid backwash.

While all hollow-fiber systems employ pressure as a fundamental driving force, a vacuum-driven system is distinguished by its utilization of negative pressure and, consequently, its significantly different design and configuration. Unlike pressure-driven systems, in which each membrane module incorporates a pressure vessel, vacuum-driven systems utilize hollow- fiber modules that are “submerged” or “immersed” in an open tank or basin. While the ends are fixed, the lengths of the hollow-fibers are exposed to the feed water in the basin.

Because the feed water is contained in an open basin, the outside of the fibers cannot be pressurized above the static head in the tank. Therefore, a vacuum of approximately -3 to -12 psi is induced at the inside of the fibers via pump suction. The water in the tank is drawn through the fiber walls, where it is filtered into the lumen. By design, vacuum-driven membrane filtration systems cannot be operated via gravity nor in an inside-out mode. However, a favorable hydraulic gradient might enable the use of a gravity-based siphon to generate the suction required to drive the filtration process in a vacuum-driven system. In some cases with a substantial hydraulic gradient, the large amount of available head could be used to generate the power for suction pumps via on-site turbines.

A representative schematic of a vacuum-driven system is shown in Figure 2.9. In the example shown, the membrane process may be designed with either continuous (Option A) or intermittent discharge (Option B) of concentrated waste.

Spiral-Wound (NF/RO) Systems

Virtually all NF and RO membrane processes applied for potable water treatment in the United States utilize systems designed for spiral-wound membrane modules. Although some MF and UF membranes may also be manufactured as spiral-wound modules, these are seldom used in municipal drinking water applications. Consequently, the discussion in this section is focused on NF/RO spiral-wound systems.

In a spiral-wound membrane filtration system, the spiral-wound modules are contained in a pressure vessel that is independent of the module itself. Typically, a single pressure vessel houses six or seven modules, although vessels that accommodate other numbers of modules can be custom manufactured. The modules are arranged in series in the pressure vessel such that the concentrate from each preceding element represents the feed water for the next. A brine seal around the outside of the feed end of each element separates the feed water from the concentrate and prevents the feed water from bypassing the element. Although the recovery for a single NF/RO module is typically less than 15 percent, the cumulative recovery associated with a six- module pressure vessel may be 50 percent or more. A diagram of a typical pressure vessel containing spiral-wound modules is shown in Figure 2.10.

A group of pressure vessels operating in parallel collectively represent a single stage of treatment in a NF/RO spiral-wound system. The total system recovery is increased by incorporating multiple stages of treatment in series, such that the combined concentrate (or reject) from the first stage becomes the feed for the second stage. In some cases in which higher recovery is an objective, a third stage may also be used. This configuration is sometimes referred to as “concentrate staging.” Because some fraction of the feed to the first stage has been collected as filtrate (or permeate), the feed flow to the second stage will be reduced by that fraction. As a result, the number of total pressure vessels (and hence the number of modules) in the second stage is also typically reduced by approximately that same fraction. Similar flow, module, and pressure vessel reductions are propagated through all successive stages, as well. Although the potential system recovery is a function of the feed water quality, as a rough approximation, a two-stage design may allow recoveries up to 75 percent, while the addition of a third stage can potentially achieve recoveries up to 90 percent.

Although concentrate staging is most often used in drinking water applications, another arrangement called “permeate staging” may also be employed. In this configuration the filtrate (or permeate) from a stage (rather than the concentrate) becomes the feed water for the subsequent stage. While this arrangement is more commonly employed in ultra-pure water applications (typically in industry), it may also be used for drinking water treatment when the source water salinity is very high, such as with seawater desalination. In these cases, the product water must pass through multiple stages to remove a sufficient amount of salinity to make the water potable quality.

The combination of two or more stages in series is called an array, which is identified by the ratio of pressure vessels in the sequential stages. An array may be defined by the ratio of either the actual number or relative number of pressure vessels in each stage. For example, a 32:16:8 array expressed as the actual number of pressure vessels may be alternatively called a 4:2:1 array in relative terms. Two-stage arrays, such as 2:1 and 3:2 (relative), are most common in drinking water treatment, although the specific array required for a particular application is dictated in part by the feed water quality and targeted overall system recovery. Figure 2.11 illustrates the configuration of a typical 2:1 (relative) array, showing both plan and end- perspective views.

As with hollow-fiber systems, spiral-wound membrane systems are designed and constructed in discrete units that share common valving and which can be isolated as a group for testing, cleaning, or repair. For spiral-wound systems these uniform units are typically called trains, or alternatively racks or skids. NF and RO treatment processes consist of one or more trains that are typically sized to accommodate a feed flow of up to about 5 MGD per train. A schematic of a typical NF/RO system is shown in Figure 2.12.

Unlike hollow-fiber systems, spiral-wound membrane filtration systems are not manufactured as proprietary equipment. With the exception of the membrane modules, spiral- wound systems are generally custom-designed by an engineer or an original equipment manufacturer (OEM) to suit a particular application. Although the membrane modules are proprietary, standard-sized spiral-wound NF/RO modules share the same basic construction, and thus membranes from one manufacturer are typically interchangeable with those from others.

Close Bitnami banner