Evaluation of Stormwater Filtration Systems

By James Lenhart, P.E.

Learning Objectives

After reading this article you should understand:

  • Understand stormwater filtration system hydraulics,
  • Recognize media type and hydraulics, and
  • Understand stormwater filtration system structural
  • Maintenance considerations.

Online quiz for this article is no longer available.

Download the Article

Overview

As the need for effective and space-efficient stormwater treatment grows, there will be an increasing number of commercial and structural filtration systems offered to regulators. As part of their due diligence, regulators go through an evaluation process to determine if the proposed system will meet some basic criteria. If the facility meets these criteria, a pilot project is identified to evaluate the facility in the field. This paper addresses some of the stormwater filtration fundamentals that should be used to evaluate all filter systems prior to implementation of a pilot study.

To perform a successful evaluation of stormwater filtration systems, the following key elements should be considered:

 

 

 

  • System hydraulics,
  • Media hydraulics,
  • Media type,
  • Structural considerations,
  • Maintenance considerations,
  • Ongoing support, and
  • Additional considerations.

System hydraulics

This is an evaluation of the simple hydraulics of how water flows through the system. The following three steps should be part of the evaluation process:

  1. Evaluate the hydraulic grade line (HGL) at the design flow rate. Typically, a backwater calculation from the point of downstream control should be performed to ensure the system can convey the peak water quality design flow rate. This analysis should include head loss through porous media (filter head loss) and pipe entrance, exit, and barrel losses.

  2. Check scour velocities in tanks and pipes. Velocities should be evaluated with particular reference to where sediments are deposited or where high-energy flows can dislodge or scour the filtration media. For example, the velocity at which the inlet pipe discharges into the filter bay determines if there is sufficient energy dissipation to minimize scour.

  3. Ensure high flow bypassing. There is both online bypassing and offline bypassing. All filters will retain deposits of sediment, organic matter, trash, and debris. As flow through a system increases, the turbulent energy will reach a point that causes re-suspension of these materials. Therefore, it is critical to ensure that high-energy flows be kept away from the filter bed(s). Typically, areas with intense rainfall such as Type II and Type III storms require offline bypassing

Media hydraulics

This is a key factor of filtration that is poorly understood. The following steps should be part of the evaluation process:

1) Evaluate the specific flow rate (q) through the media. The specific flow rate, or flux, is in units of flow per unit area or gallons per minute per square foot (gpm/ft2). Given the specific flow rate times the surface area (A) of the filter, the total flow rate (Q) can be calculated (Q = qA). A good reference point is rapid sand filtration with rates of about 4 gpm/ft2. In general, the higher the rate, the higher the head loss. Finer media are typically more efficient for total suspended solids (TSS) removal, as well as other pollutants, but have high head loss characteristics. Coarser media can handle higher flow rates but are less efficient in TSS removal. Claims of high filtration rates with high pollutant removal capabilities and low head loss are highly suspect.

2) Compare the design specific flow rate to specific flow rates in lab and field studies. Does the specific flow rate of the proposed design match the specific flow rate associated with performance data? For example, does performance data have a specified rate of 2.0 gpm/ft2 while the design rate is 20.0 gpm/ft2? For proper filter design, this is an absolute factor. Also evaluate different model sizes of the proposed BMP. Given the same driving head, the specific flow rate should be the same. If the specific flow rates vary with model size, this should raise questions.

3) Consider the thickness and head loss of the media. Darcy's Law (q = ?hKA/L) states that flow rate (q) increases with increased head (h), surface area (A), or driving head (K) and decreases with an increased bed length (L). The thickness of the media coupled with the specific flow rate determines the amount of contact time the water has to be treated. The longer the contact time, the more effective pollutant removal will be, particularly when soluble pollutants are being removed through reactive processes. Thicker media will have higher removal rates but will increase head loss. Thicker media will also increase media costs and maintenance costs. For porous media such as perlite, TSS removal efficiency increases with thickness because there is more opportunity for particles to be captured as water follows a tortuous path though interstitial pores. For fine media such as sand, the majority of the TSS capture is at the surface, and media thickness has less influence on TSS removal. Observations and studies of sand filters show that the majority of fine solids removal occurs within the first 2 to 3 inches of bed thickness. In general, Darcy's law applies to flow through porous media. However, for many filtration systems with high conductivity and a relatively short bed, Darcy's law does not behave with the same accuracy as in a groundwater application. Another confounding factor of stormwater filters is that the conductivity is a variable. As the media loads with solids, K will slowly decline. Toward the end of the filter life, K approaches zero asymptotically.

4) Calculate the contact time. Compare the calculated contact time with the contact time used in lab or field studies. For example, if the flow rate is doubled and the thickness is reduced by half, the contact time is one fourth. This will have a direct impact on pollutant removal effectiveness. It is important to check the contact time of the design and the test data presented. Equation 1 can be used to calculate the contact time for a radial flow cartridge.

Equation 1:
eq1
Where:
t is the contact time,
R is the outer radius (feet),
ro is the inner radius (feet),
p is the porosity of the media,
Q is the flow rate(ft3/sec),
h is the height (feet).

One should also look at the flow paths through the filter. Is there a uniform pressure distribution across the media? A non-uniform pressure distribution results in differential loading of the filter and non-uniform contact time.


Photo 1: Perlite media, normally crystal white,
is occluded with fine sediments

5) Consider fouling and occlusion of the media. This will ultimately control the specific flow rate through the media. One critical aspect is the surface of the filter. In general, finer media can remove finer particles but have a much higher clogging factor. Sand can retain about 1.1 pounds of dry sediment per cubic foot before it rapidly clogs and fails (Lenhart and Calvert, 2007), whereas a coarser perlite media can retain about 6 pounds of sediment per cubic foot. This is a common tradeoff, having the benefit of finer media achieving higher performance while exhibiting a much higher clogging factor (Photo 1).

Filtration media that is effective in trapping fine solids will accumulate a thin layer of solids on its surface (smutzdecke) that begins to occlude the filter and reduce the flow rate. An accumulated layer as thin as 1 mm will control the filtration rate and reduce the specific flow rate to very small amounts (< 0.1 gpm/ft2). For example, a horizontal sand filter that builds up a smutzdecke demonstrates this. Without cleaning mechanisms to prevent surface clogging, this problem becomes endemic to all filtration systems. Following are two important aspects to consider regarding the filter surface:


Photo 2: A filter screen is used to
retain media. The total percent
open area (17 percent) equals
the screen percent open area
(43 percent) multiplied by the
fine mesh percent open area (40
percent).
  • Know and examine the percent open area of the outer wall of the filter vessel, which includes the filter body and any internal mesh (Photo 2). The lower the percent open area, the more susceptible the housing is to surface clogging and failure. If the surface of the filter is exposed to light, algal growth can rapidly clog the filter surface.
  • Understand the capability of the technology to prevent surface fouling by sediments. It is a fact that filters will eventually clog with sediments. The question is how long does it take. Clearly, if a filter has no active mechanism to remove accumulated sediment from a filter surface, its life will be less than a filter that does. Pretreatment by settling will help, but research indicates that the majority of the fouling is by fine sediments, organic matter, and bacterial growth, all of which are difficult to remove by pretreatment through rapid settling. In fact, recent research at Monash University in Melbourne, Australia, (Siriwardene, et al, 2005) indicates that it is particles of 20 µm or smaller that cause clogging of sand filters.

6) Compare a horizontal bed with vertical filters. A horizontal bed filter operates by ponding water on its surface such that the driving head causes the water to percolate through the media (Figure 1).

Figure 1: The hydraulic profile of a horizontal bed filter. The total driving head is the depth of water. The hydraulic profile of a horizontal bed filter

One characteristic of horizontal bed filters is that all collected sediments will impact the surface of the filter, thus reducing longevity, whereas vertical filters allow much of the sediment to be deposited on the floor, away from the filter. Horizontal bed filters have the benefit of a constant pressure head (equivalent to the depth of water) over the entire filter surface. Vertical filters have a reduced driving head (equivalent to half of the depth of water, see Figure 2)

Figure 2: A radial flow, gravity-driven filter cartridge. The mean pressure head is 1/2 of the water depth. A radial flow, gravity-driven filter cartridge

because of a triangular pressure distribution, unless they are operating with both a pressure on the outer surface and a suction on the inner surface to create a uniform pressure distribution across the filter media bed (Figure 3).

Figure 3: A radial flow, siphon-driven filter cartridge where a suction head on the interior and driving head on the exterior sums to a uniform pressure head equal to the depth of the water.. A radial flow, siphon-driven filter cartridge

Media type

During the past few years, a number of different filter media have been used for stormwater filtration. Media such as sand, peat, and compost have been used successfully. Recent uses of perlite, zeolite, carbon, and other "exotic" media have expanded the choices for targeting specific pollutants.

Consider the physical properties of the media used for sediment removal. Most media remove solids by mechanical processes. The gradation of the media, irregularity of shape, porosity, and surface roughness characteristics all influence TSS removal characteristics. Finer media are more effective at removing TSS than coarse media but create higher head loss and have higher clogging factors. This trade off is a fundamental consideration. Media gradation is critical to performance, the finer the media the higher the performance but the slower the flow rate with the same amount of driving head.

Understand the chemical properties and mechanisms used to remove stormwater pollutants. Many types of pollutants such as nutrients, metals, and oil and grease are in soluble or free form and can be removed through chemical and/or biological processes. Common processes are cation exchange, precipitation, chelation, and adsorption. When claims are made for soluble pollutants, there needs to be a documented process by which these reactions take place. In addition, these reactions have limits in terms of sorption capacity and reaction kinetics. For example, media may have a sorption capacity of "X" mg/kg of media. Given the mass of the media, the total mass of pollutant that can be removed can be calculated and then compared with what is generated from the site. Reaction kinetics also cause a slowing of pollutant removal rates as media saturation increases and/or pollutant concentration decreases.

The reviewer should also consider if the media can add constituents to runoff. For example, organic media can elevate ortho-phosphorus, which leaches from the media. In some watersheds this is not a desirable media and an alternate should be selected. Other media can raise or lower pH.

Evaluate whether the properties of the media will change over time. Stormwater is a complex mixture of sediments, nutrients, organic matter, bacteria, and other pollutants. Many times, media may perform well in the short-term, but in the longterm may be compromised by biological decomposition, bacterial slimes, or simple decomposition by continuous saturation in water (Photo 3).


Photo 3: A pulp-based absorbent,
although effective for soaking up oil spills,
will rapidly decompose when used
as a stormwater filtration media.

For example, does the media decompose or dissolve when exposed to stormwater? Cellulose-based media such as treated pulp, corn cobs, or rice hulls will decompose when exposed to these elements. Does the media swell or shrink on wetting and drying cycles? Check to see if the media is free draining or submerged. Permanently submerged media can lead to anoxic conditions, causing anaerobic decomposition and release of many of the trapped pollutants.

Check media availability and cost. Many times, media are available in small production quantities or it is difficult to find media that meet all the specifications. Systems can also facilitate multiple-media options. Multiple-media systems have the versatility to fine tune media to site-specific pollutants, as well as adapt to future improvements in media effectiveness.

Structural considerations

Structural integrity is critical. Many stormwater filtration systems are designed to handle traffic loads, therefore, it is important to evaluate the structure for integrity and design life. Make sure the structures are reviewed by structural engineers to ensure expected traffic loads can be handled.

Water tightness is required by many agencies. Evaluate vertical and horizontal joints for design integrity. Vertical joints are more difficult to control because of differential settlement. Some agencies require a water tightness test prior to acceptance. All joints below the permanent pool elevation need to be watertight.

Buoyancy measures need to be considered. In areas of high groundwater, take measures to prevent system floatation.

Constructability considerations are important. Many times what appears simple on the plans can be difficult during construction. Consider construction loads, back filling, etc. Does the contractor have a track record of constructing and installing similar facilities?

Consider materials of construction. Stormwater runoff can be very corrosive. The presence of numerous nuts, bolts, differential metals contact, pivot points and hinges, and galvanized parts are all potential candidates for corrosion and ultimate failure.

Maintenance considerations

All stormwater BMP's will require maintenance at some time or time interval. Long-term maintenance costs and maintainability are very important considerations.

Evaluate availability of maintenance contracting. It is the nature of filters to occlude with captured TSS, hence maintenance is required. Acceptance of all systems should be coupled with a maintenance contract by a professional maintenance provider. Does the manufacturer stand behind the product and offer to provide maintenance?

Maintenance frequency varies from site to site. Typically, if maintenance is needed more than once a year, the project will cost more over its lifetime than if the facility had been upsized to extend the maintenance cycle. Conversely, designing for maintenance intervals greater than one year may result in higher initial costs that may never be recovered by lower lifecycle maintenance costs.

Maintenance costs are critical. If a person states a cost, ask if they will sign a contract to do that. Frequently, costs are understated because they do not include mobilization, heavy equipment rental, and mileage costs. Consult with a local maintenance provider when in doubt.

Facility access will always be needed. Even manholes are equipped for access. Stormwater filtration systems will need to be accessed for cleaning media, washing sidewalls, repairs, media installation, and facility inspection. Review plans for height restriction, ventilation, and extraction ports. Make sure the facility is also accessible by the required equipment.

Working inside can be problematic. Evaluate the complexity of the internal components and whether they pose trip hazards or make access by suction hoses difficult. Is the operator working in standing water? If you have a set of plans or photos of an existing unit, seek the opinion of people who perform the maintenance rather than rely upon the generalities of the manufacturer.

Check the weight of the media. How is the media being extracted? How much would a media "vessel" that is full of sediment and has a high water content weigh? Is it practical that it be removed or lifted? As a rule of thumb, use a minimum of 70 lbs/ft3 for a lightweight media such as perlite (weighs about 5 lbs/ft3 when fresh and dry). Some media such as sand can result in cartridges weighing in excess of 300 pounds.

In addition, once media becomes clogged with sediments, it can become firmly lodged in the filter body. Typically spent media needs to be sucked, shaken, or dug out of the filter body, as it rarely falls from the filter body under its own weight.

Standing water is costly and expensive to remove. What is the volume of water if the system is drained down and what is the volume of water if the filters are fouled? Does the system cause standing water in the upstream pipe network?

Product support

Does the manufacturer warranty the product? Typically there is a one year warranty on the structure and components.

Does the manufacturer provide support to the owner? Filtration systems are more effective than simple settling or vortexing devices and require media replacement. It is important that the manufacturer supply on-going long-term support to ensure proper operation.

Other considerations

Check references and speak with other agencies where the facilities have been installed, then use the scope of the information to establish credibility. Remember, because of the variable nature of stormwater runoff, all types of facilities — including ponds, swales, filters, settling devices, and others — will have examples of poor performance, but overall the assessment should be positive. Checking manufacturers' claims is critical. A large number of reports do not necessarily imply a system is well tested or verified.

Integrate all of the considerations above into an overall assessment of how a proposed system matches performance data from prototypes. This is a critical review. If the design flow rates, media thickness, etc. do not match the studies, upsizing the facility to the test values is warranted.

Make sure the data presented are consistent with pollutant concentrations and characteristics typically associated with stormwater runoff. Common claims that illustrate inconsistencies include the following:

  • performing a study with sand or grit and equating the percent removal to the removal of silts and clays in the field;
  • performing a study with very high concentrations of oil in water to assess oil and grease (O&G) removal and then using these data to claim high percentage removals on stormwater runoff, which has much lower concentrations of O&G (Does the data presented reflect the reality of what is typically found in the field?); and
  • performing studies conducted at a fraction of the design flow rate (If a filter is designed to a high specific flow rate, it should be tested at that rate).

Evaluate lab data versus field data. Lab data are collected under controlled conditions and can provide a lot of insight on filter behavior. Studies successfully executed at the laboratory level include removal of TSS relative to different flow rates and hydraulic behavior.

Properties that cannot be evaluated in the laboratory include fouling characteristics, maintainability, and pollutant removal characteristics with complex hydrology and water chemistry. Laboratory data provides rapid data collection and good insight to filter performance but should not be used as a sole method to judge performance.

Programs for filtration system verification

New Jersey and Washington have both implemented statewide programs to evaluate filtration technologies (and other technologies as well) for the purpose of ensuring that the technology can meet the water quality objectives. The programs, known as the TARP and TAPE processes, are well documented and available on the Internet. The TARP process has been largely administrated by the New Jersey Corporation for Advanced Technology (NJCAT).

In short, these processes outline monitoring protocols, treatment goals, reporting standards, and other criteria that all new BMP's must be subjected to for final verification and approvals.

Conclusion

These factors are meant to serve as guidelines for the preliminary review of new products. If this review meets the satisfaction of the reviewer(s), the next consideration should be pilot facilities that lead to system acceptance. If a submittal does not appear to meet the criteria listed above, the design engineer or reviewer needs to seek clarification or redesign prior to setting up a pilot program.


James Lenhart, P.E., chief technology officer for CONTECH Stormwater Solutions, is one of the founders of Stormwater Management, Inc., and led the design efforts of the StormFilter and StormScreen. He currently directs CONTECH Stormwater Solutions' Research and Development, Product Evaluation, and Regulatory Services departments. Lenhart is an active member of the Water Environment Research Foundation and serves as a board member of the New Jersey Corporation for Advanced Technology.


REFERENCES

  • Siriwardene, N.R., A. Deletic, and T.D. Fletcher, 2006, "Preliminary studies of the development of a clogging prediction method for stormwater infiltration systems," Proc., 4th International Conference on Water Sensitive Urban Design,. Melbourne, Australia, April, V1.211-V1.218.
  • Lenhart, James H., P.E., and Paula P. Calvert, 2007, "Mass loading and mass load design of stormwater filtration systems," Environmental Water Resources Institute, Proceedings, April 2007.