Designing filtration devices using volume-, flow-, and storage-based methods
BY JOANNA B. OGINTZ, P.E.
Across the country, many regulatory agencies are establishing criteria for sizing stormwater filtration devices to meet their water quality standards.
There are a number of ways to size these devices, and these sizing methodologies can be tailored to meet the local requirements.
The most common sizing methodologies are volumebased, flow-based, and storage-based. Volume-based sizing lends itself well to the design of horizontal bed filtration technologies such as sand filters. However, with the development of the manufactured Best Management Practice (BMP) market, new radialflow filtration cartridges are gaining prominence as viable alternatives. Although these devices historically have been flowbased technologies, new research is demonstrating that these systems also can be designed using volume-based, or even hybrid, storage-based methods.
Volume-based sizing Sizing methodologies were based originally on the idea of treating the first flush” of stormwater runoff. Much research has demonstrated that in many cases most pollutants in stormwater are carried off a site during the first flush of a storm. This is true particularly for small, highly impervious catchments and highintensity storms. In addition, the typical frequency distributions for rainfall depth show that the majority of the average annual runoff volume is from higher-frequency, small storms.
For example, according to the Georgia Stormwater Management Manual, hydrologic studies have shown that small, frequent storms account for the majority of rainfall events that generate runoff (AMEC Earth and Environmental, et al, 2001).
The state’s design storm of 1.2 inches was determined to generate runoff from the 85th percentile storm event, or the storm event that is greater than 85 percent of the storms that occur within an average year.” This first flush phenomenon led to development of volumebased sizing methodologies. These were first used to design treatment facilities, such as ponds or sand filters, where the facility could be sized to capture and treat a certain volume of water. In this methodology, a depth of runoff is used to calculate a water quality volume (WQv) that must be treated. This runoff depth over the site – for example, 1.0 inch per impervious acre – is specified typically by the local regulatory agency and, in general, is meant to equal the runoff depth that yields the first flush.
Alternatively, some jurisdictions may specify a depth of rainfall that must be routed over the site to create an equivalent depth of runoff. The WQv is calculated by multiplying the depth of rainfall or runoff by the site area, and then by a weighted coefficient to represent the land use of the site. The WQv to be treated is governed by the total amount of rainfall or runoff. A typical volume-based system design also must meet the local jurisdiction’s requirements for drain-down time, such as less than 40 hours, or between 5 and 24 hours.
To design a filtration BMP, such as a sand filter, to meet a volume-based design requirement, the structure generally would be designed to capture the WQv. The WQv then would be treated gradually as the system drains down through the filter bed. Thus, the system consists of two components – a storage component followed by a filtration component. Some regulatory agencies may require capture of 100 percent of the WQv, while others may allow some credit for treatment through the filtration component as the system is filling. For example, some jurisdictions require capture of only 75 percent of the WQv, which helps to reduce the overall size of the unit.
One advantage of a volume-based sizing methodology is that it ensures treatment of the WQv exactly according to a regulatory requirement. Also, the volumebased design can extend the life of the filtration component by allowing for settling of larger sediment particles in the storage component. This reduces the total mass that must be removed by the filtration component.
An advantage of radial-flow filtration cartridges as compared with horizontal filter beds is that the storage and filtration components can be connected hydraulically, allowing the filter to control flow through the unit. For sand filters, this typically would not be recommended because there is no outlet control for the filter bed, and it may result in high flows through the filter bed, possibly resulting in loss of accumulated sediments. However, for radialflow filter cartridges, the cartridge can be designed to control flow through the media, taking advantage of the full driving head available. A disadvantage, however, is that a volume-based design typically results in large treatment facilities to guarantee capture and treatment of the required volume, which may result in the loss of available land space for parking or other amenities, or increased costs because more land area may be required to install the system.
Flow-based sizing Many jurisdictions have found that volume-based sizing methodologies derived to capture the first flush do not sufficiently take into account the variability of pollutant transport with runoff intensity. Therefore, some have moved toward using a flow-based sizing method instead. With a flow-based methodology, a design storm event is used to calculate a design discharge flow rate, which is expressed as a volume per time.
Typically, the design storm event is specified by the local regulatory agency as some depth of rainfall over a certain period of time, for example 1 inch in 24 hours, or as some intensity, such as 0.5 inch per hour, with a specified rainfall distribution. The rainfall distribution may vary with time or may be a straight-line relationship.
Commonly, the Soil Conservation Service (SCS) rainfall distributions are accepted.
Depending on how the storm is specified, a number of hydrologic models can be used to calculate the discharge flow rate from the site given the design storm.
Some models, such as the Rational Method, simply result in a single flow rate value for the discharge flow rate. Others, such as TR-55 or the Santa Barbara Urban Hydrograph, result in a peaked hydrograph, where the flow rate of the site increases as the rainfall increases, and drops off as the storm subsides. For a peaked hydrograph, the discharge flow rate is considered to be the peak flow rate of the hydrograph.
The discharge flow rate is influenced primarily by the distribution of the rainfall and the design storm event or intensity. Flow-based sizing methodologies are used frequently to design treatment structures such as flow-through swales.
A flow-based design methodology has the advantage of taking into account the distribution of the rainfall and runoff. However, it has the disadvantage of potentially oversizing a treatment unit to treat a flow rate that results from a high, infrequent peak storm with a short duration of intense rainfall.
Storage-based sizing Filtration BMPs also can be sized using a hybrid design methodology that optimizes the size of the unit by incorporating routing calculations into the design of the system. Similar to the volume-based sizing method, the filtration BMP must consist of two components – a storage component and a filtration component.
These components can be configured either in-series (storage followed by filtration) or in-parallel, a configuration also known as a surge tank.
The standard storage approach is based on the in-series configuration.
Using a distributed rainfall hyetograph to generate an influent hydrograph, the inflow to the treatment facility is routed through the storage component and then directed to filtration.
This routing allows for taking some credit for filtration that occurs while the storage facility is filling, which helps to reduce the overall size and cost of the treatment facility. Figure 1 (page 36) provides an example of how routing can attenuate the peak flow off a site. For all practical purposes, the level of treatment remains the same, but a system designed using this sizing methodology provides a more cost-effective solution. In addition, long-term maintenance costs may be reduced.
The surge tank approach is based on an in-parallel configuration. When a runoff hydrograph results in an extremely high peak flow rate, a surge tank can be used to trim the peak” and reduce the size of the filtration component. The minimum size of the filtration component can be determined by local requirements, such as required filter bed surface area, or massloading requirements. During a storm, the runoff flows directly into the filtration component until the hydraulic capacity is reached. If the inflow exceeds the filter capacity, the peak of the hydrograph is routed to the surge tank, which is aligned in-parallel to the filter. The size of the surge tank is determined by calculating the volume of water remaining in the peak of the hydrograph greater than the capacity of the filtration system. Once the storm subsides and the water level in the filter begins to drop, the water stored in the surge tank is released gradually into the BMP until the tank is empty. This ensures full treatment of the design storm.
In summary, there are many ways to design filtration BMPs to meet local regulatory requirements. These different methodologies can be used to meet specific water quality challenges facing local regulatory jurisdictions while offering cost-effective and space-efficient designs.
Joanna Ogintz, P.E., is regulatory engineer in the Research and Development department at Stormwater Management, Inc., in Portland, Ore. She can be reached at 800-548-4667 or via e-mail at firstname.lastname@example.org.