Part 2: Straining and barrier methods of removal
By Ed Butts, PE, CPI
In part one of this three-part series we introduced you to separating—or removing—solids and other harmful materials from water, and the various technologies available to do it. This month, we expand upon this topic and provide an overview and detailed explanation of the first two of the four methods outlined in part one: straining and barrier removal methods.
THE STRAINING METHOD
The straining method generally involves the use of granular filtration media, such as sand, to provide a physical, often combined with a chemical, method of removing undesirable substances from water. Although the title of this series is “Solids Separation,” it is somewhat a misnomer for all the processes described can be used to remove dissolved substances, such as ferrous iron, from water as long as the appropriate pretreatment and chemical treatment is incorporated into the process.
The straining method is very flexible and applicable to treatment of water in flows down to 5 GPM for residential applications and up to several thousand GPM for municipal or industrial projects. However, it is more commonly applied to water systems with flow rates of 100 GPM or more. The barrier method is more commonly used for lower flow rates.
Here’s how I try to remember the difference between the two methods. The strainer method uses a granular media to actually “strain” the water to achieve the removal of unwanted material. The barrier method uses more of a mechanical device that provides a “barrier,” such as an inline screen, to accomplish the same task.
The straining method often uses a combination of processes to remove particles or contaminants from a water supply. In addition to the basic filtration that occurs when water is directed through a filtration media, other processes can be used to assist with removal. Several assist with the filtration process by making the particles larger and heavier, changing the electro-chemical state of the particles so the opposite charge of the filter media attracts the charge of the offending particle, or by using multiple layers (up to three or four) of progressively smaller granular filter media.
Regardless of the specific type of strainer method selected, the three most common methods of particle removal through this method are shown in Figure 1.
Figure 1a demonstrates the use of a filter cake occurring on the upper surface of the media. A filter cake occurs as the unwanted material accumulates on the upper surface of the media, resulting in a progressively tighter, but thin, layer of combined filtered material, biological action, and chemicals, if used. This method is applicable to many water filtration technologies—diatomaceous earth systems, slow sand filtration, and many adsorption type of filters.
Figure 1b is a typical straining action by using granular media as a simple physical impediment to downward particle flow.
Figure 1c indicates a combination of physical (straining) and chemical methods of filtration. This is the method most commonly used for potable and aesthetic water treatment since the average particle size, consisting of mostly very fine material, bacteria and viruses, are typically too small to effectively remove from water through strictly straining methods alone. By adding various chemicals to the flow stream, also known as coagulants, polymers, or filter aids, the unwanted material can chemically alter its electro-chemical state to physically combine with other material through a process known as flocculation. This results in a larger and more filterable particle (referred to as a floc) in order to more easily be removed through the filtration media or chemically changed to a different electrical charge so the unwanted material then “sticks” or is attracted to the opposite charge of the media. In addition to the filtering action from straining the water and furnishing coagulants to create a condition more conducive to filtration, larger and heavier particles can also be removed from water through a process known as sedimentation. During sedimentation, the water is allowed to remain reasonably quiet or stationary within a chamber, referred to as a sedimentation basin. In this basin, the quiet state of the water allows gravity to act on the larger and heavier particles in the water to settle to the bottom of the basin. This action can drastically reduce the impact upon the downstream filters by lowering the filtration burden.
In all three cases shown in Figure 1 a sand, anthracite coal, or layers of other types of granular media are used to remove undesirable material from the water. This is the primary distinction between the strainer method and the barrier method. The barrier method uses a solid physical impediment, such as a perforated screen, with a known size of round openings or slots, to remove undesirable material. The strainer method uses a single layer or a series of layers of granular filter media to accomplish the same objective. This action is shown in Figure 2 where the various physical and electro-chemical reactions that occur in a granular media are illustrated.
Particles become trapped or stick between the openings of the various grains of the media, which over time causes the flow rate through the filter to decrease and head losses through the bed to increase as more and more particles become trapped as they collect in the higher regions of the filter and travel lower and lower through the filter bed.
The filtration rate and efficiency through a barrier filter is fairly predictable up to the point where enough openings become clogged to reduce the device’s efficiency. But the filtration rate and removal efficiency through a granular media is highly dependent on the type and size of the media used; the type, size, and volume of the material removed; as well as the unit flow rate through the media. Sand in various grain sizes is the most common type of filter media used. Other types of specialty filter medias such as anthracite (coal), manganese dioxide, and manganese greensand are also used for removing specific compounds and elements, particularly from potable water.
Even though several methods are used to evaluate a filtration media for possible use, the two most common and important criteria are (1) the effective size, and (2) the uniformity coefficient.
The effective size of a granular media is defined by the size of screen opening where 90% of a sample of granular media is retained (held back) on the screen and 10% passes through the screen (referred to as D10). The larger the grain size, the faster the water or wastewater moves through the sand and the more water can thus be filtered. But if the grain size is too large, treatment efficiency will be reduced.
For wastewater, larger breakthroughs of unoxidized matter are observed due to short retention times and an instantaneous lack of oxygen within the filter bed when applying relatively large hydraulic loads to filter media with a coarse grain size, especially above 1 mm (millimeter). Often, the ideal sand for intermittent sand filters receiving domestic water or wastewater is a coarse sand with an effective size between 0.3 and 0.5 mm.
Clogging becomes a major concern when using sand with an effective size of less than 0.3 mm. Therefore it is important filters using this size of sand be lightly loaded (<1.2 GPM/ft2). Clogging is generally less a concern when using coarser sand. In a field evaluation of sand filter systems, it was found sands with too many fines have a greater chance of clogging than sand with a particle size between 0.3 and 0.5 mm, assuming other similar characteristics.
The uniformity coefficient (UC) is a numeric estimate of how sand is graded and is a dimensionless number—in other words, it has no units. The term “graded” relates to regions where the concentrations of sand particles are located by size. Sand with all the particles in two size ranges would be defined as narrowly graded sand and would have a low UC. Sand with near equal proportions in all the fractions would be defined as widely graded sand and would have a high UC value.
The UC is calculated by dividing D60 (the size of screen opening where 60% of a sample passes and 40% is retained) by D10 (the effective particle size: that size of screen opening where 10% of a sample passes and 90% is retained). The larger the UC, the less uniform the sand. For water filtration it is important the sand grains all be about the same size— that is, relatively uniform.
A UC of 4 or less is recommended for all filter media. This recommendation is intended to avoid clogging of the filter bed at higher loading rates. Sands from most natural sources are widely graded containing a variety of grain sizes, which results in a high UC. If the grain sizes vary greatly, the smaller ones will fill the spaces between the larger particles, making it easier for the filter to clog.
Two types of filters are used in current practice: gravity (or “open”) filters and pressure filters. Gravity filtration, as it is often an open system, uses only the head of water above the filter media to create the force necessary to drive the water through the media bed. The amount of head in a gravity filter varies with the application. Nonetheless, 3 to 5 feet is fairly common.
Gravity filtration can take many forms, but the most common methods are slow sand and rapid sand. The slow sand filtration method (Figure 3) uses a combination of straining through a single layer of sand with a gravel support base, which when combined with a biological action on the upper surfaces of the media removes undesirable material.
The straining action is performed at very low rates, usually between .05 and .25 GPM/ft2 or 72-360 GPD/ft2—also known as the hydraulic loading rate, which can require a large land footprint. For example, up to 2000 square feet of filter area (5% of an acre) is often needed for just 100 GPM of flow.
The biological action is performed at the very highest regions of the filter bed where a biological layer of fine material (a “schmutzdecke”) forms from the reaction between a growth that occurs with an accumulation of material in the water and biological reactions. This is called ripening of the filter. Even though both of these methods are extensively used for potable water treatment, the rapid sand method, owing to the larger filtration rate of 2-10 GPM/ft2, is much more commonly used in practice.
A pressure filter, on the other hand, uses much higher values of hydraulic head to operate, typically measured in pounds per square inch (PSI) rather than feet of head. These systems, as shown in Figure 4, consist of a closed filter vessel, generally round in shape, where the water flows through a single or series of layers of granular media under pressure.
The flow rate per area (the hydraulic loading rate) is typically much higher than with a gravity filter, averaging around 5-7 GPM/ft2 but often up to 10-15 GPM/ft2, depending on the media and the desired removal rate of the harmful material. Although the depth of filter media can vary with the application and filter media, the depth of a filter media bed is generally between 24 inches up to 48 inches for pressure filters and up to 72 inches for slow sand and rapid sand filters.
As also seen in Figure 4, the filtering process generally consists of a downward flow through the media, with backwash occurring in the opposite direction (upwards flow). Backwash of a granular media is needed to remove the entrapped particles and adequately clean and fluidize the bed. Backwash through a rapid or pressure filter is generally much higher than the hydraulic loading rate, in values typically between three up to ten times higher than the hydraulic loading rate, or 15 up to 30 GPM/ft2.
Filter manufacturers use a statistical method for rating a filter’s capability to remove particles or organisms, called a log removal rating or value (LRV). In order to determine the validity of a manufacturer’s claim, the designer must understand how LRVs work and how they are reported.
LRVs are reported using a percentage removal number based on an initial challenge. This number is commonly reported using a multiple of “9s” (example: 99.9%). Each “9” reported indicates the challenge level as well as the filter’s removal capability. For example, a filter that reports four “9s” (99.99%) corresponds to a “4-log” removal and indicates the filter was challenged with at least 10,000 particles or organisms for every milliliter (mL) of water that was tested, and 0 particles or organisms were detected downstream. A mL of water is about the same size as a sugar cube.
An understanding of industry standards for LRVs is crucial in determining if a filter is qualified for the intended use. A company may report a filter with 99.99% removal of bacteria, and to an uninformed user this value may look very impressive. But when compared to the industry standard for bacteria removal of ≥99.9999%, an informed user would realize the filter falls way short of the intended mark. In fact, when tested at >1,000,000 particles or organisms/mL required to achieve a 99.9999% rating, a 99.99% removal rating indicates that 100 particles/mL water still got through the filter.
Industry standards are often cited as follows. Protozoan cysts: ≥99.9% removal (“3-log”). Bacteria: ≥99.9999% removal (“6-log”). Viruses: ≥99.99% removal (“4-log”)
When a designer is considering the straining method for filtering water, it is vital to fully consider the physical and electrical state of the material to be removed and the water supply itself since almost no granular media can remove unwanted material with a physical size of 10-20 microns and smaller, such as bacteria or viruses, without some type of chemical addition such as a coagulant, polymer, or filter aid. This may mean the water supply must undergo various laboratory and other tests to determine the best combination of media and chemicals to use.
Often, these tests comprise a “jar test” where various chemicals are introduced to the water under a controlled setting to determine the best coagulant or polymer for a specific water. Media selection is also conducted based on the filtration rate desired plus the size and chemical state of the undesirable material. In virtually all cases, a reduced scale test of a full scale operation, known as a pilot test, is warranted to determine the best selection of a filter type, media, and chemical for a specific application and water.
THE BARRIER METHOD
So to review, the strainer method of solids removal uses granular media, such as sand, to remove undesirable material through a combination of straining the material through the pore spaces that exist between the grains and an electrochemical reaction between the filter media and the particles.
The barrier method, conversely, removes unwanted material through simply inserting a perforated screen or similar stationary or moving (travelling) device into the flow stream. These screening devices are available in numerous forms and styles—wye, inline, suction screens, and “tee” configurations as well as “passive” (well screens) and “non-passive” (travelling or self-cleaning) configurations. In screening applications for water wells, a wedge, or “vee” slot screen is often used to prevent large sand and gravel grains entering into the well.
A typical well screen design is shown in Figure 5. The degree of rejection is based on the gap, or “slot size” between the wires. Proper selection and development of a well screen is critical since it is vital that larger granular material outside of the well, sand and gravel, be used to create an exterior filter and stabilizing pack. If the slots are too small, much of the formation will remain outside of the well during development and result in an inefficient well with lower capacity than desired. A filter pack will not form, resulting in the formation breaking down and entering the well during pumping. This determination is made through a “sieve analysis.”
A sieve analysis examines the various grain sizes of the material within the formation so a proper well screen selection can be made. Like Goldilocks, the filter pack should be “just right” with not too many fines and not too much of the larger material. Although the criteria varies with the type of formation, the grain sizes, and personal preference, the most common size range to screen against (retainage) is 40%-50% of the average size of the material in the aquifer, assuming use of a natural gravel pack.
An artificial gravel pack uses a different value of formation retainage, depending on the material used in the filter pack. Although this value can vary between aquifer material and driller preference, for the most part, the entrance velocity of a well screen is typically designed for around .10 feet per second (FPS). Although, in some cases, values as high as .50 FPS have still resulted in a successful design. In all cases, selection of a well screen should be performed by individuals knowledgeable and trained in the art and science of well design.
The barrier method includes screening devices in various configurations, sizes, and screen types and materials. The screen bodies can be built from cast or ductile iron, steel, PVC, or bronze. Screening elements are usually constructed from bronze or stainless steel with round (perforated) openings or slots of a uniform size. A typical screen filter is generally rated by the filter inlet/outlet sizes and/or the mesh size. The inlet and outlet sizes simply refer to the piping connections on both inlet and outlet sides, usually in NPT or ANSI flange sizes. The particle removal ability is rated through the mesh size of the internal screening element itself.
For reference, the relationship between inches, microns, and mesh was illustrated in last month’s column (May, page 37, Figure 4). As well as the mesh size, the screen’s open area and micron rating are additional terms used to describe the opening size and filtration capability of a particular screen. The open area of a screen refers to the pore area or the sum of the combined areas of the holes or slots in the screen through which the water passes. Filtration open area is usually expressed as a percentage of the effective total area. For example, a screen may be said to have an open area of 10 square inches, or conversely, may be a screen with 50 square inches of total area with 20% of open area.
To be effective, a screen must contain an adequate percentage of open area to permit the effective transfer of flow along with providing resistance against a specific size of particle across the openings of the screen, with minimal head loss while also avoiding an excessive percentage of open area where the holes or slots are too close together or too large, causing the structural integrity of the screen to be compromised, which could result in collapse of the screen element. The open area of a screening element varies with the type of opening, the diameter of the screen, and the specific design. For the most part, values of open area between 10% and 50% are fairly common in practice.
Depending on the micron filtration degree, weave-wire screens often contain 30% to 50% of open area while wedgewire screens (typical well screens) consist of open areas between 5% up to 40%. The micron rating of a screen represents the smallest nominal size of particle (in microns) the filter will remove. Manufacturers may use one or both terms when rating their devices, so it is important to understand the relationships between the terms. Generally speaking, the relationships in Table 1 between mesh size and microns apply.
The barrier method of solids removal is usually recommended when the amount of contaminants in the water is light to moderate and the contaminants consist of settleable and non-settleable solids of adequate physical size. A critical factor to consider for a screening filter is that the screen element will require periodic cleaning and maintenance.
The typical operation and backwash of a screen is shown in Figure 6a and Figure 6b. Water is normally flowing from right to left with the water passing through an inline screening element. Undesirable particles are trapped and impinged onto the surface and in the holes of the screen. As more and more particles become trapped on and in the screen, a filter cake, similar to that in a granular media filter, begins to develop. As the area of the screen becomes more clogged, a pressure decrease and reduction in flow is realized across the screen. The loss in pressure across the screen is also referred to as the pressure drop. Eventually, as the reduction in area across the screen increases, the pressure drop also increases proportional to the square of the velocity through the screen.
With a pressure screen, backwash or removal/cleaning of the screen often occurs when the pressure drop is between 5- 10 PSI. Since this mathematical relationship applies to most fixed screening elements, this also means that doubling the open area of a screen increases the time interval between backwash cycles at the same pressure drop by a factor of four (4). For example, a screen with 20 square inches of open area will typically function four times longer between backwash or cleaning cycles than the same type of perforation design with 10 square inches of open area (both set at a 5 PSI pressure drop to initiate backwash). This is a primary reason why a screen’s open area is so critical to the proper selection of a barrier-method device and why a larger screen is sometimes desirable for a given flow rate.
Depth filters rely on a torturous path to capture particles within the matrix or “depth” of the filter. Basically, particles are caught within the depth pores of the filter as they come into contact with obstructions within the filter element. There is rarely a uniform or defined pore structure in a depth filter and in many configurations, such as fibrous filters, there are no pores at all. Even though there may not be defined pores, depth filters can still be performance rated based on challenge testing.
In these tests, the filter is challenged with a pre-set quantity of a defined size of particles or organisms. When finished, this type of testing renders the filter unusable and is referred to as destructive testing. Manufacturers perform these tests on a representative sample of each filter batch. Since every filter cannot be realistically tested and verified individually, a “nominal” rating is generally associated with depth filters as a “typical” value of removal.
Depth filters can be produced using several methods. The most common types are fiber filters and sintered filters. Fiber filters are either spun or woven into a cloth or felt. A common example of a fiber filter is shown in Figure 7. These inline sediment and carbon filters are used extensively in residential applications as “point of entry” (POE) or “point of use” (POU) devices to filter and remove unwanted sediments or contaminants or to correct taste and odor issues with water as it enters the residence or appliance. Depth filters are commonly rated in such nominal filtration values as 5 microns, 20 microns, 50 microns, and 100 microns, and in flows between 1 GPM up to 30 GPM, depending on the size and type of the unit. Sintered filters such as ceramic, metal, or porous plastic filters are formed by fusing particles together under heat and pressure. The spaces between the particles create the flow path or pores of the filter. Inline ceramic filters, often used for drip irrigation applications, are an example of sintered filters.
LARGER SOLIDS SEPARATIONIn addition to the straining and barrier methods of solids removal, a final example of removal of unwanted material in a flow stream is shown in Figure 8. Commonly known as a sand or centrifugal separator, it uses centrifugal force with gravity to separate and remove sand and other particles up to 98% in sizes of 74 microns (200 mesh) and larger from a water flow. Although these units may occasionally remove some material slightly smaller than 74 microns, the performance rating is usually based on that value since the gravity of smaller particles vary greatly.
As shown in the figure, as water enters the side inlet of the separator, the water is accelerated to create a spin around a central chamber. This resulting velocity physically separates heavier and larger particles, such as sand and most larger debris, from the water through centrifugal force. The water is then discharged out through the top of the unit with the settled material allowed to settle down to the bottom of the separator through the specific gravity (weight) of the material. Over time enough accumulation of this material is generally gained to initiate a “blowoff” sequence. This typically consists of an automatic or manual valve on the lower end of the separator (purge) to provide cleaning of the solids chamber. A blowoff sequence is generally initiated through an automatic timer for automatic operation or a written schedule with a manual function.
Sand separators are quite effective for solids removal in the proper application and design. First, the separator must be designed for the appropriate flow range of the device. For example, a 1-inch unit may have a stated flow range of 5-15 GPM—and it is critical the operation remain within that range for effective performance. Unlike most filters, a separator must be designed to function within this given flow range. It cannot be allowed to work below the stated minimum flow rate; otherwise, the needed centrifugal force and removal of solids will not occur.
Second, the high range of flow must also be observed since exceeding the maximum flow rate will result in excessive wear and pressure drop of the unit, resulting in premature failure due to sand erosion.
Next, a sand separator must be designed for and function under a sustained and constant pressure drop of around 5-10 PSI. Any attempt to lessen or increase the pressure drop appreciably above or below these thresholds will negate the effectiveness of the separator’s removal.
Finally, a sand separator must be recognized for what it is and is not. It is not a filter or method of removing fine sediment, silt, or particles from water. Remember the most effective range of a separator is to remove up to 98% of 74 microns (and larger) material from a water supply. Although separators have been known to remove much larger sand and gravel from flow streams, they are essentially intended for sand and similarly sized particles between 74-100 microns. Gravels and large, abrasive materials have been known to cut through the upper centrifugal chamber in a short time, resulting in early failure of the device. For total effectiveness, all manufacturers’ guidelines for design and installation must be adhered to and the flow rate ranges and pressure loss through the unit followed.
The barrier method can be very effective at removing larger material (>75 microns or 200 mesh), such as silt and sand, from most freshwater supplies at reasonable levels of the total particle volumes, flow rates, filter runs (time between backwashes or cleaning of the screen), and at a minimum of pressure drop (5-7 PSI). The most efficient barrier method uses a screening device of the proper mesh and inlet/outlet size to handle the flow rate and degree of particle removal desired. The high end of the flow rate and removal efficiency through most barrier types of filtration is generally limited to 500 GPM or less and particle sizes of 50 microns and larger.
Anything above that requires consideration of a granular media filter, possibly with chemical addition or a system with automatic backwash of the screen to avoid rapid clogging of the screen. Although the application of a barrier-method device is not as critical as that with the strainer method, it is still important to consider all aspects of the intended application and select a screen of the proper mesh, configuration, and size.
This will wrap up part two of this series on solids separation methods and techniques. Next month we will conclude this series with a discussion on the final two methods: adsorption and chemical alteration techniques.
Until then, work safe and smart.
McDowell-Boyer, L., J. Hunt, and N. Sitar. 1986. Particle transport through porous media. Water Resources Research v. 22, no. 13, pp. 1901-1921.
“Introduction to Filtration.” AWWA Training Manual.
Sand Separator and Filtration Data, Lakos Website and Literature, 2016.
Brown, Dennis. “Understanding Terminology in Regards to Filter Ratings.” 2012. Aquamira Technologies.
Ed Butts, PE, CPI, is the chief engineer at 4B Engineering & Consulting, Salem, Oregon. He has more than 35 years experience in the water well business, specializing in engineering and business management. He can be reached at email@example.com.