Copyright by David Venhuizen, One Time Rights
Copyright by David Venhuizen
David Venhuizen, P.E., waterguy@ix.netcom.com
5803 Gateshead Drive.
Austin, Texas 78745 USA
tel. 512-442-4047
fax 512-442-4057
Copyright by David Venhuizen, One Time Rights
CONTENTS
As EPA's Technology Assessment of Intermittent Sand Filters states: "Intermittent sand filtration of wastewater is not a new technology." [1] It has been reported that various forms of sand filtration have been used to purify water for centuries. [2] The modern concept of intermittent sand filtration derived from observations made in the mid-1800's at "sewage farms" in sandy soils. The drainage from these areas, onto which wastewater was intermittently dosed, was greatly purified. This led to using sand beds especially constructed for wastewater treatment. [3]
Over the last 100+ years, a vast amount of information on intermittent sand filter technology has been produced. [e.g., refs. 1,4-9] More recently, the intermittent sand filter process has been studied and used for addressing the problems of on-site and small-scale wastewater management. [e.g., refs. 10-16] These experiences have shown that, again quoting EPA's sand filter assessment: "Intermittent sand filters are ... ideally suited to rural communities, small clusters of homes, individual residences and business establishments. They can achieve advanced secondary or even tertiary levels of treatment consistently with a minimum of attention." [1]
It is precisely because sand filters operate reliably in such a trouble-free manner that this treatment device is especially well suited for on-site wastewater systems. The basic technology is very "fail-safe". Generic classification of this system is "attached growth aerobic process". The intermittent sand filter is a very lightly loaded version of that process, so mean cell residence time is very long. Also, the biology of the system is quite diverse, typically including many trophic levels of microorganisms and some macroorganisms as well. [8,9] These characteristics render the process inherently resistant to upsets.
The major failure mode is clogging of the bed. When this proceeds too far, it forms a barrier to passage of any more wastewater, so the consequence of failure is that water backs up in the system, which usually forces attention to the problem. Even when clogging has proceeded to the point where water is continuously ponded on the bed surface between doses--a condition of extreme neglect of proper maintenance--the effluent quality typically does not degrade significantly. Many studies have demonstrated that a sufficient remedy for bed clogging is scraping off the top inch or so of sand. [1,4,6,11,17] The reported intervals between scrapings have ranged from 4 to more than 54 months. [1,17] This indicates that the problem builds up slowly over a period of months, or even years, so maintenance can be conducted essentially at the operator's leisure.
This contrasts sharply with the "aerobic" systems often touted for on-site wastewater management. These systems use some variant of the activated sludge process. Treatment in those processes is affected by few trophic levels of aerobic organisms living in concentrations far in excess of those found anywhere in nature. So the process is inherently unstable, requiring constant inputs of energy and close attention to maintaining proper operating parameters in an effort to keep the process on track. Also, the consequence of any process failure is passage of poorly treated effluent in short order, since there is typically no physical barrier to passage of water through the system. Once "off track", it often takes considerable time for an activated sludge process to recover.
These contrasting characteristics dictate that the intermittent sand filter is technically far better suited than aerobic systems for use in the on-site environment. To enhance the cost efficiency of this technology--and thus promote their broader use--"advanced" design concepts need to be proliferated. These concepts provide the additional advantage of considerable reduction in effluent nitrogen concentration. This paper reviews the basis for advanced design concepts and provides insights for practical system design.
If the sand filter bed must be serviced with any frequency, the bed must be accessible. Given the low design loading rates (generally less than 3 gal/ft2/day) usually quoted for on-site system filter beds, a large bed would be required for a typical single family home. Cost efficiently covering a large bed to preclude odor problems or contact with the wastewater is problematic. Because of this, buried sand filters which must be loaded at very low rates (around 1 gal/ft2/day) are often used.
However, both the allowable hydraulic loading rate and the frequency of sand bed servicing are subject to design and operating conditions. Significant parameters include media size, loading frequency, loading method, and influent quality.
The latter appears to be especially significant. Some references state that, to preclude premature clogging, it is expected that organic loading rates should be limited to less than 0.005 lb. BOD5/ft2/day. [1] It was stated in early studies of the intermittent sand filter that "... the volume of sewage that can be purified ... is dependent upon the amount of organic matter present in the wastewater rather than the volume of wastewater in which this organic material is held." [1] However, EPA observes: "Allowable [hydraulic] loading rates increase with the degree of pretreatment. A strict relationship establishing an organic loading rate ... has not yet been clearly defined in the literature." [17]
Many investigations have demonstrated that, with proper system design, hydraulic loading rates much higher than those generally assumed can be accommodated without requiring frequent maintenance. This would allow a filter bed sized for a home to be cost efficiently contained within a buried tank, allowing it to operate like an open access sand filter while precluding odor and public health problems.
The original concept of intermittent sand filter operation was to utilize rather fine media--in the range of 0.2 to 0.4 mm effective size--and to load the entire daily dose at one time, allowing it to drain and rest until the following day. [4,6,17] In studies conducted at the University of Florida in the 1940's and 50's, it was noted that, all other factors remaining equal, splitting the daily load into two doses increased removal efficiencies and allowed sands of a given size to be loaded more heavily. [6]
This finding led to further investigations of more frequent loadings. The conclusion of these studies was that, even without improving the influent quality before application to the sand filter, more frequent loadings can allow higher loading rates to be readily accommodated with no degradation in treatment and without premature clogging of the filter. In fact, for filters employing larger media, treatment actually improved with more frequent loadings even when higher hydraulic loading rates were applied.
A filter with an effective sand size of 1.04 mm demonstrated a 96% BOD5 removal efficiency when loaded at 13.8 gal/ft2/day and dosed hourly, while at loading rates ranging from 4.02 to 9.76 gal/ft2/day, removal efficiency was generally in the range of 70-80% when the filter was only loaded twice per day. During the period when the filter was dosed hourly, average organic loading rate was 0.0119 lb./ft2/day, and it was above 0.005 lb/ft2/day at all loading rates except 4.02 gal/ft2/day. Two other filters, with sands of 0.44 mm and 0.46 mm effective size, demonstrated BOD5 removal efficiencies ranging from 80% to 93% when loaded twice per day at rates from 2.87 to 7.46 gal/ft2/day. With 4 loadings per day at a rate of 6.89 gal/ft2/day, removal efficiencies were elevated to 95-97%. Here again, when hydraulic loading rates were at or above 6.89 gal/ft2/day, organic loading rates were generally above the 0.005 lb./ft2/day "limit". [5]
Scherer and Mitchell [14] studied a "stratified" sand filter, consisting of three sand layers with effective sizes of 0.52 mm, 0.27 mm, and 0.17 mm, from top to bottom, respectively. The filter was loaded with septic tank effluent at a rate of 6.3 gal/ft2/day. Mean influent BOD5 was 101 mg/l--implying an average organic loading rate of 0.0053 lb./ft2/day--and mean influent TSS was 59 mg/l. The effluent averaged <1 mg/l of both BOD5 and TSS, representing 99+% removal. Clogging was not reported to be a problem over the year that the project was operated. Apparently the coarser top layer "pre-strains" the wastewater, protecting the very fine lower layers from clogging at higher loading rates.
Mitchell [10] demonstrated the ability of sand filters to accommodate elevated loading rates when the influent is septic tank effluent which has been upgraded by passing it through an anaerobic upflow filter. Four sand filters, each containing media of 0.52 mm effective size, were loaded at rates of 1.5, 3, 9 and 15 gal/ft2/day. Loading frequency was not reported. Filter influent averaged 85 mg/l of BOD5, and the effluent from all four filters was below 5 mg/l throughout the 6-month study period, which implies over 94% removal. At 15 gal/ft2/day, the average organic loading rate was 0.0106 lb./ft2/day. While influent levels were not reported, TSS concentrations in the effluents were less than 3 mg/l throughout the study period. Mitchell reported that there was no evidence of clogging in any of the filters. He offered the opinion that such a system could definitely operate maintenance-free at 8 gal/ft2/day and could probably do so at rates in excess of 15 gal/ft2/day. [18]
Swanson and Dix [15] studied a recirculating filter treating household septic tank effluent which utilized bottom ash as the filter media. The "sand" filter was constructed over a rock filter, which also served as a reservoir for the mixture of septic tank effluent and "sand" filter effluent. Several operating periods were reported, with forward flow loading rates ranging from 3.0 to 8.2 gal/ft2/day and actual total hydraulic loading rates (forward flow plus recirculation flow) ranging from 13.8 to 44.8 gal/ft2/day. Except for the trials conducted at a hydraulic loading rate of 3.0 gal/ft2/day, organic loading rates were all in excess of 0.005 lb./ft2/day. Over the 8-month study period, effluent BOD5 averaged 3.4-12.1 mg/l, representing average removal efficiencies of 91-98%. Though not statistically significant, it is interesting to note that the lowest efficiency was achieved at the lowest forward flow and total hydraulic loading rate, and the highest efficiency was experienced at the highest forward flow and second-highest total hydraulic loading rate. For TSS, the overall average effluent strength was 7.7 mg/l, representing a 90% removal efficiency. The range was from 84% to 93%, with the highest efficiencies again being observed at the highest loading rates.
A demonstration project conducted by the Town of Washington on Washington Island, Wisconsin, monitored the performance of 5 recirculating sand filter systems under "field" conditions. Four of the systems served residences, and one served the island's grocery store. The circumstances of each system varied, and the usage of some of the systems varied with the seasons. In general, however, removal rates for BOD5 and TSS were in excess of 95% at hydraulic loading rates ranging from 2.5 to over 10 gal/ft2/day. Septic tank effluent strengths were typically quite high, especially at the store (where the relatively low 2.5 gal/ft2/day hydraulic flow rate was observed), so average organic loading rates were well above 0.005 lb/ft2/day in all systems. [19]
These experiences introduce the recirculation concept. Originally, this feature was employed to minimize odors when dosing open sand filters in the on-site environment. [11,12] However, it was found to also improve operating efficiency of the filter. This is almost certainly due in part to the more uniform loading schedule usually experienced in recirculating systems. The foregoing discussions have indicated that better treatment efficiency can be expected if the sand filter is loaded more uniformly throughout the diurnal cycle. A ubiquitous characteristic of small-scale systems--especially those serving a single family home--is extreme variability in the diurnal flow profile. This suggests that recirculation for the purpose of enhancing temporal stability of sand filter loadings is a desirable strategy.
Improved operation is also due to dilution of effluent strength imparted by the recirculation flow. For example, in the Swanson and Dix system reviewed above, average organic loading rate was generally somewhat in excess of 0.005 lb/ft2/day. However, with recirculation ratios ranging from 3.4:1 to 7:1, the organic strength of the applied influent was greatly reduced. [15]
One of the sand filter systems in the Washington Island project provided a graphic example of the benefits of recirculation in an on-site system. The filter was operated in a single-pass mode for almost a year. During this period, BOD5 removal averaged only 78.6% (60.2 mg/l average strength), TSS removal averaged 80.9%, and nitrification was inconsistent. Within two weeks after a recirculation system was installed, high degrees of both BOD5/TSS removal and nitrification were achieved, despite this system being quite heavily loaded (>10 gal/ft2/day) for the four months following startup of the recirculation system. The system operated with this improved efficiency until the end of the observation period over a year later. BOD5 removal averaged 94% (12.4 mg/l average strength), even though there was a flaw in the recirculation system which caused the recirculation ratio to vary widely. It is expected that a combination of decreased organic strength of applied influent, more uniform loading cycle, and more frequent dosing of the filter bed produced this improved performance. [19]
This discussion highlights a widespread misconception about sand filter technology. Many regard single-pass intermittent sand filters and recirculating sand filters as different technologies, with different design parameters being relevant to each. However, both types are intermittently loaded and the operating principles are the same for each. It is clear from the above that recirculation is nothing more than a device to control timing of filter loadings and to manipulate the applied organic strength.
The Washington Island project also confirmed that little treatment efficiency would be lost by employing coarser media in recirculating sand filters. Three of the systems employed a very coarse gravel media in the size range of 1/4"-3/8" (6-9.5 mm), yet still provided superior performance. [19] This has been observed in other efforts as well, and larger media is being routinely used in many applications. [20] Good results have been obtained, however, with a range of media sizes, from about 1 mm to over 6 mm. Slightly lower effluent BOD5 and TSS levels and better removal of indicator bacteria--presumed to imply better removal of pathogens--are generally obtained with smaller media.
All of these considerations indicate that sand filter systems can be designed to support forward flow hydraulic loading rates in the range of 6-10 gal/ft2/day. Filters loaded this heavily should employ larger media, should be dosed frequently with the total daily load spread uniformly throughout the diurnal cycle, and should employ recirculation or some other means to reduce organic strength of the influent.
It is also important to dose the filter bed as uniformly as practical over the entire surface. This is especially so when employing larger media. If water is applied over only a small area of the filter, it would not spread over other parts of the filter area until significant clogging of the area receiving the flow had occurred. Then adjacent areas would receive all the flow until they too clogged, and so on, until this "progressive failure" resulted in the whole bed becoming clogged. Uniform distribution can be accomplished with either a very dense network of perforated headers or a spray distribution system. The latter is typically more practical and cost efficient to construct.
Another one of the sand filters on Washington Island demonstrated the importance of uniform distribution. Due to a design flaw, the spray heads partially clogged, so that influent ran onto the filter surface only in a small area around each of the six heads rather than being thrown over the full surface. This situation persisted for several months, during which system performance--which had been excellent for many months prior to this--degraded badly. As soon as the spray system was repaired, BOD5 and TSS removal efficiency drastically improved and the filter also began nitrifying again, even though by this time the condition of the bed was highly compromised. This experience is also a graphic example of how resilient the sand filter technology is. [19]
Another basic design parameter is media depth. Design criteria generally recommend a depth of 24-36 inches. [1,17] It has been recognized, however, that most of the purification is affected within the top 9-12 inches of media. [1,4,5,9,17] It is asserted that a greater depth of media tends to produce more consistent effluent quality. [1,17] The reasons given are that deeper beds are not as severely affected by rainfall and that they permit the removal of more media before replacement becomes necessary. It is clear that a filter bed contained in a buried tank and for which very, very infrequent scraping is expected to be required would be immune to these concerns. Therefore, it is expected that perhaps even less than 24 inches of sand depth would be sufficient in these sand filters.
It turns out that design features which would enhance sand filter
treatment efficiency in regard to BOD5 and TSS removal and allow
it to be more heavily loaded also generally promote a high degree
of nitrification, which is the conversion of ammonium-nitrogen
in the wastewater to the nitrate form. This is, of course, a
prerequisite for eliminating nitrogen by denitrification.
A number of works have explored the use of modified recirculating
sand filter concepts to effect removal of nitrogen from wastewater.
This is accomplished by first nitrifying the wastewater by passage
through the sand filter, then recirculating the nitrified effluent
back through the anaerobic "front end" of the system,
in which denitrification takes place. It had been observed that
some degree of denitrification occurs in conventional recirculating
sand filters when sand filter effluent mixes with septic tank
effluent in the dosing tank. [21] That concept is illustrated
in Figure 1.
Modifications of that process aim to provide a better
anoxic environment to increase the degree of denitrification.
Piluk [21] and Sandy [22] reported on studies employing "batch" concepts in which a horizontal flow rock bed underlying the sand filter bed served as an anoxic reactor. Septic tank effluent was routed through this rock bed, where it mixed with effluent flowing out the bottom of the sand filter. Organic carbon in the septic tank effluent provided the energy and nitrate in the sand filter effluent provided the substrate for denitrifying bacteria. Reductions in total nitrogen (TN) on the order of 70% were observed in these studies. However, this system configuration results in the overall system effluent being a mixture of denitrified sand filter effluent and septic tank effluent which had traversed only the rock bed, somewhat defeating the purpose of investing in the high quality treatment capabilities of a sand filter system. Hydraulic residence time (HRT) in the rock beds of these systems was quite high, which the investigators expected would be required to achieve high rates of denitrification. [21,22]
Mitchell, however, demonstrated 70-75% total nitrogen removal
using a system with an anaerobic upflow filter (vertical flow
rock bed anoxic reactor) and sand filter in sequence, with relatively
low HRT's in the upflow filter. [23] This was a primary basis
for the system concept, illustrated in Figure 2,
which was investigated
in the Washington Island project. Results from that project indicated
that 60-90% reduction in TN concentration could be consistently
obtained. Average influent TKN (organic plus ammonium nitrogen)
ranged from 37.9 mg/l in one of the residential systems to 130
mg/l in the grocery store system. Removal percentage typically
increased with higher influent nitrogen content. Effluent TN
concentrations generally less than 15 mg/l were observed in all
of the Washington Island systems when proper operating conditions
were maintained. [19]
In the Washington Island systems, when sand filter effluent was recirculated back into the septic tank rather than directly into the upflow filter, the vast majority of recirculated nitrate was denitrified in the septic tank. This occurred even though recirculation flowed only through the second chambers of two-chamber septic tanks in which the HRT's were fairly low. It was found that excessive clogging of the upflow filters was a potential maintenance problem, and that organic strength reductions effected by the upflow filter were less than expected. These factors urged elimination of the upflow filter, resulting in the modified conventional recirculating sand filter concept, illustrated in Figure 3. [Note: Use your browser's Back button to return to this page.]
Work with this concept in Anne Arundel County, Maryland, reported by Piluk and Peters [24] also confirmed that, given proper system design, little loss of nitrogen removal efficiency would result from elimination of the upflow filter. Results from monitoring of 3 systems yielded an average of 64% TN reduction, with effluent TN concentration averaging 20 mg/l. [24] In each of these systems, recirculation flow is routed through the second chamber of a two-chamber septic tank. More recent monitoring of 4 systems in which recirculation flow was routed into a single-chamber septic tank yielded an average TN reduction of 60.6% and an average effluent TN concentration of 21.1 mg/l. [25]
SYSTEM DESIGN FEATURES FOR OPTIMUM PERFORMANCE
Applying the results from these studies, along with basic wastewater engineering principles, to the system concept in Figure 3 yields design features expected to enhance denitrification performance and minimize maintenance liabilities. A single chamber septic tank with a long HRT should be used, and recirculation flow would be routed into that chamber. By recirculating into the chamber which receives raw wastewater, where high organic loads would assure anoxic conditions and provide an abundant carbon source, and by providing a relatively long contact time, denitrification potential will be maximized.
However, recirculating into the primary septic tank chamber imposes a high total flow rate through that chamber, creating the potential for excessive solids carryover. This can be prevented by installing an effluent filter at the outlet of this chamber. These types of filters have proven to be effective at reducing BOD5 and TSS concentrations in septic tank effluent. [26] Thus, not only would excessive washout of solids be prevented, the effluent filter would even replace to some extent the function which was fulfilled by the upflow filter in regard to reduction of BOD5 and TSS levels. While these effluent filters can also become clogged, available information indicates that they need to be cleaned rather infrequently. [26] Even if it turns out that they must be washed off fairly often, this is a simple and easy maintenance procedure relative to backflushing an upflow filter when it becomes clogged.
Since the effluent filter would cause more solids to be retained in the primary septic tank chamber, sludge may build up faster, potentially increasing the frequency at which the tank must be pumped. This is blunted by using a tank with a large sludge storage volume. This is "automatically" provided at no additional cost because the septic tank is already designed with a very long HRT to enhance denitrification potential, as just detailed.
As shown in Figure 3, recirculation flow can either be pumped or the system can be configured to allow gravity flow. Most on-site recirculating sand filter systems employ various schemes to split the sand filter effluent stream by gravity, routing part of it to the recirculation loop and part of it to the disposal field, as shown in the top drawing of Figure 3. As usually practiced, however, this method imposes a major design compromise.
Gravity flow concepts can only provide a recirculation flow when the sand filter is dosed, since there is no effluent flow otherwise. If the dosing pump is controlled by float switches, no recirculation flow would occur except during periods when the system is receiving forward flow. Flow into an on-site system is typically dominated by morning and evening flow peaks, resulting in episodic loading of the sand filter. Experience indicates that this is detrimental to general system performance. Also, if recirculation flow only occurs in response to forward flow, there would be less recirculation flow when forward flow is low, and this would result in a lower denitrification potential. Further, if there is no forward flow for a period of time--e.g., if the residents are on vacation--there would be no flow at all through the sand filter. This would impose a "restart" period of lesser performance on the system when forward flow resumes.
System designers attempt to get around these problems by controlling the sand filter dosing pump with a timer. A large dosing sump provides for flow equalization. The level in this sump builds up during periods of high forward flow and is drawn down during periods of little or no forward flow. Using this scheme requires that the average daily forward flow be accurately estimated and that the dosing system be set up to match that rate.
If the daily flow rate is underestimated, the sump would fill up and trip the alarm. The pump and alarm control systems can be designed to circumvent these "false alarms", but the additional electrical circuitry is probably as great an operational liability as is the additional pump in the pumped recirculation system.
If the flow is overestimated, there would be periods when the sump was empty, again resulting in less uniform loading of the sand filter. It can be demonstrated that, when actual flow rate falls significantly below the design value, there would be long periods with no doses through the night and, depending on the timing of forward flows, perhaps through the afternoon as well. Since statutory design flow rates are often somewhat higher than actual flows, this problem would probably be ubiquitous. Also, regardless of how accurately the dosing schedule was matched to the average forward flow rate during periods of "normal" operation, this scheme would still not keep the sand filter active during extended periods of no forward flow.
The pumped recirculation scheme illustrated in Figure 3 solves all of these problems. Given the high reliability and low cost of the small submersible pump required to implement a pumped recirculation system, this scheme would not impose significantly greater operational liabilities. The effluent pump control system--usually "on" and "off" float switches--is arranged so that a "permanently" ponded depth remains below the "off" level. This provides a sump for the recirculation pump. Because recirculated flow loops through the system and returns to the effluent tank, this sump is replenished even when there is no forward flow through the system.
The recirculation pump is controlled by a timer. To minimize disturbance in the septic tank and to distribute the total daily recirculation flow more evenly throughout the diurnal cycle, the pump delivers a series of short flow "pulses". Controlling this flow with a timer dictates that total daily recirculation flow is constant, so that recirculation ratio will vary with the actual daily forward flow rate. The sand filter dosing pump is controlled by float switches to deliver a fixed volume per dose. Since low forward flow decreases the total flow rate through the system, fewer doses would occur. In practice, this does not compromise performance because the timing of recirculation doses can be set to assure a minimum sand filter dosing frequency in the absence of forward flow.
Even this liability can be circumvented, however, by adding a simple component to the gravity recirculation system, optimizing system performance while eliminating a pump. An effluent bypass valve, as illustrated in Figure 4, is placed in the effluent line from the side of the sand filter bed which flows to the disposal system. This valve is set so that, if the dosing sump level drops below a preset depth, the valve opens and this flow returns to the dosing tank. Thus, when forward flow is low, there would be two recirculation loops. One is the normal recirculation flow from the "recirculation side" of the sand filter through the septic tank. The other is from the "effluent side" of the sand filter directly into the dosing tank through the bypass valve. Therefore, the dosing sump can never run out of water, and no sand filter doses would be skipped.
Using this scheme, the sand filter is loaded with the same total volume of wastewater on the same schedule every day, regardless of forward flow rate. Only the organic strength of the wastewater will vary with more or less forward flow. Volume of the dosing sump would be a large fraction of the daily design flow rate, so it is expected that organic strength would not vary widely through the diurnal cycle. The sand filter would operate in a "steady state" mode, maximizing its efficiency.
Using this concept, the system's design recirculation ratio is set by the ratio of flow onto the two sides of the sand filter bed. Based upon results of the Washington Island project and other studies of this technology, it is concluded that a 2:1 ratio on the design flow rate is a good compromise. Using this ratio, total hydraulic load--the sum of forward flow plus recirculation flow--onto the sand filter is moderate, and good nitrogen removal performance is expected.
This recirculation ratio is implemented by using twice the number of spray heads on the "recirculation side" as on the "effluent side" of the filter. If all heads are built the same and system piping is arranged to assure minimal head loss difference in feed lines to each side, flow out of each head will be about the same. Therefore, 2/3 of every dose will flow to the recirculation loop and 1/3 will flow to the effluent tank--a 2:1 recirculation ratio.
The split filter bed separates sand filter effluent into two distinct flow streams. It was quickly realized that, since it just flows back through the septic tank, the quality of flow out of the "recirculation side" of the sand filter can be allowed to degrade slightly. A 30 mg/l BOD5 concentration would serve just as well as a 10 mg/l level, as long as sufficient nitrification was achieved to provide the level of denitrification desired. Therefore, rather than using a filter bed surface area ratio of 2:1 and applying the same hydraulic loading rate to each side, the "recirculation side" area can be reduced and the "effluent side" enlarged, resulting in a lower loading rate onto the "effluent side". This will quite likely enhance quality of the "effluent side" drainage and slightly degrade quality in the recirculation loop.
Because the sand filter system consistently delivers a low turbidity effluent, it is practical to use a subsurface drip irrigation field for disposal. Where site constraints dictate that high quality pretreatment--including reduction of nitrogen content--is necessary to render on-site wastewater management an environmentally sound proposition, it would usually also be necessary to employ a very shallow, lightly loaded disposal field in any case. [27] So sand filter treatment and drip irrigation reuse/disposal is a natural marriage of alternative on-site management methods. Where freezing of drip lines is a concern, a modified low-pressure-dosed system which operates as a "pseudo-drip" irrigation system can be used. These strategies would allow on-site systems to be used on very restrictive sites without compromising public health or environmental values.
Being able to execute the concept shown in Figure 4 in a cost efficient manner is a key to proliferating this strategy. Growing out of the Washington Island project, a "standardized" system is being produced by Crest Precast, Inc., of La Crescent, Minnesota, utilizing a tank system which efficiently implements this concept. This manufacturer will also install the pumps, internal piping, pump control devices, and electrical boxes at the factory. The system design is illustrated in Figure 5, showing the septic/dosing/effluent tank, and Figure 6, showing the sand filter tank.
The concept can also be implemented using standard precast septic tanks with minimal modifications. A paper detailing how this has been done using the standard products of tank manufacturers around Austin, Texas, has been produced by the author and is available on the internet. [28] Although these companies have not chosen to do so to date, the pumps, piping, controls and electrical boxes could easily be factory-installed in these systems as well.
With the system being "packaged" in this manner, site installation is greatly simplified. The contractor needs only to prepare the excavations, hook up the house drain and effluent line, plumb the sand filter and main tanks together, install the filter media, and backfill the tanks. Experience indicates that this is quite a bit less expensive than site-modifying "stock" tanks and installing the internal parts in the field.
The "advanced" sand filter systems detailed in this paper, along with drip or "pseudo-drip" irrigation reuse/disposal systems, provide enhanced pretreatment and improved disposal methods. Many site conditions demand that on-site systems utilize such methods. These include shallow groundwater, which drove the investigations in Anne Arundel County, and fractured bedrock, especially in aquifer recharge zones, which urged development of these systems on Washington Island. The use of this method is encouraged wherever lack of sufficient soil resources creates a concern about threats to the environment and public health from on-site wastewater systems.
Sand filter/drip irrigation systems route system effluent to a highly efficient irrigation system, providing a high degree of beneficial reuse as well as an environmentally benign wastewater disposal system. In more arid climates, this strategy can greatly decrease dependence upon potable water systems to serve irrigation demands. [29] In many of these areas, such as Central Texas, water resources constraints dictate that wastewater reuse will be increasingly beneficial to the regional water economy as time goes by. So the use of these systems are also encouraged anywhere that beneficial reuse of treated wastewater is good public policy.
Despite more than 100 years of operating history and more than two decades of progressive development for use in the on-site environment, sand filters continue to be under-employed as a solution to these types of problems. Dispersing knowledge about "advanced" design concepts and establishing them as a standard method for on-site/small-scale wastewater management is the frontier of this technology.
1. D. L. Anderson, R. L. Siegrist & R. J. Otis, Technology Assessment of Intermittent Sand Filters, U. S. EPA, Office of Research and Development, Municipal Environmental Research Laboratory, Cincinnati, 1985.
2. A. Montiel, B. Welte & J. M. Barbier, "Improvement of Slow Sand Filtration", Slow Sand Filtration, N. J. D. Graham, ed., Ellis Horwood Limited, Chichester, England, 1988, pp. 47-59.
3. R. L. Dymond, "Design Considerations for Use of On-Site Sand Filters for Wastewater Treatment", M.S. Thesis, Institute for Research on Land and Water Resources, Pennsylvania State University, 1981.
4. H. W. Clark & S. Gage, A Review of Twenty-One Years Experience Upon the Purification of Sewage at the Lawrence Experimental Station, Fortieth Annual Report of the State Board of Health of Massachusetts, Public Document No. 34, 1909.
5. T. deS. Furman, W. T. Calaway & G. R. Grantham, "Intermittent Sand Filters--Multiple Loadings", Sewage and Industrial Wastes, Vol. 27, No. 3, March 1955, pp. 261-276.
6. G. R. Grantham, D. L. Emerson & A. K. Henry, "Intermittent Sand Filter Studies", Sewage Works Journal, Vol. 21, No. 6, November 1949, pp. 1002-1015.
7. T. deS. Furman, "Sewage Plant Design Criteria for the Semitropics", Sewage and Industrial Wastes, Vol. 26, No. 6, June 1954, pp. 745-758.
8. W. T. Calaway, "Intermittent Sand Filters and Their Biology", Sewage and Industrial Wastes, Vol. 29, No. 1, January 1957, pp. 1-5.
9. W. T. Calaway, W. R. Carroll & S. K. Long, "Heterotrophic Bacteria Encountered in Intermittent Sand Filtration of Sewage", Sewage and Industrial Wastes, Vol. 24, No. 5, May 1952, pp. 642-653.
10. D. T. Mitchell, "Sand Filtration of Septic Tank Effluent", Small Flows, Fall 1986, U.S. EPA National Small Flows Clearinghouse, pp. 2-3.
11. M. H. Hines & R. E. Favreau, "Recirculating Sand Filters: An Alternative to Traditional Sewage Absorption Systems", Home Sewage Disposal, Proc. of the National Home Sewage Disposal Symposium, ASAE, Pub. Proc-175, 1975, pp. 130-136.
12. M. G. Teske, "Enhanced Treatment for Surface Discharge", Individual Onsite Wastewater Systems, Proc. of the Sixth National Conference, 1978, Ann Arbor Science Publishers, Inc., Ann Arbor, 1979, pp. 137-145.
13. D. G. Perley, "Use of Open and Buried Intermittent Sand Filters as a Low-Cost Cluster Treatment Alternative", On-Site Wastewater Treatment, Proc. of the Fourth National Symposium on Individual and Small Community Sewage Systems, ASAE, Pub. 07-85, 1985, pp. 169-173.
14. B. P. Scherer & D. T. Mitchell, "Individual Household Surface Disposal of Treated Wastewater without Chlorination", On-Site Sewage Treatment, Proc. of the Third National Symposium on Individual and Small Community Sewage Treatment, ASAE, Pub. 1-82, 1982, pp. 207-214.
15. S. W. Swanson & S. P. Dix, "Onsite Batch Recirculation Bottom Ash Filter Performance", On-Site Wastewater Treatment, Proc. of the Fifth National Symposium on Individual and Small Community Sewage Systems, ASAE, Pub. 10-87, 1987, pp. 132-141.
16. D. K. Sauer & W. C. Boyle, "Intermittent Sand Filtration and Disinfection of Small Wastewater Flows", Home Sewage Treatment, Proc. of the Second National Home Sewage Treatment Symposium, ASAE, Pub. 5-77, 1977, pp. 164-174.
17. U.S. Environmental Protection Agency, "Design Manual--Onsite Wastewater Treatment and Disposal Systems", EPA 625/1-80-012, 1980.
18. D. T. Mitchell, University of Arkansas, Department of Civil Engineering, personal communication, April 1987.
19. David Venhuizen, "Demonstration Systems Performance Analysis--Final Report", Town of Washington (Wisconsin) Wastewater Management Facility Plan, 1994.
20. Steve Fishel, Tennessee Water Pollution Control Division, personal communications, 1994, 1995, 1996. Robert Morriss, P.E., personal communications, 1995, 1996. Observation of Mayo Peninsula sand filter plant, 1995.
21. R. J. Piluk, "Field Study of an Innovative On-Site Wastewater Disposal System for Nitrogen Reduction", M. S. Thesis, Graduate School of the University of Maryland, 1988.
22. A. T. Sandy III, "Nitrogen Removal Using a Batch Recirculating Bottom Ash Filter", Problem Report submitted in partial fulfillment of requirements for M.S., College of Engineering, West Virginia University, 1987.
23. D. T. Mitchell, personal communication, March 1991.
24. R. J. Piluk & E. C. Peters, "Small Recirculating Sand Filters for Individual Homes", draft copy of paper submitted for presentation at Seventh National Symposium on Individual and Small Community Sewage Systems, 1994.
25. Richard Piluk, personal communication, August 1996, transmitting data from Sept. 1995 to June 1996 for systems being monitored under EPA's National Onsite Demonstration Project.
26. Test results furnished by Zabel Wastewater Filter Systems and Orenco Systems, Inc., and unpublished study of effluent filter effectiveness conducted at Tennessee Tech University, 1996.
27. David Venhuizen, "Soil Treatment Mechanisms," Wisconsin Department of Industry, Labor & Human Relations, 1995.
28. David Venhuizen, "A Minnesota Regulator's Guide to the Venhuizen Standard Denitrifying Sand Filter Wastewater Reclamation System", 1996. Available on internet here.
29. Analysis performed by David Venhuizen as part of water conservation study for Barton Springs/Edwards Aquifer Conservation District indicated that, among customers using significant fraction of demand for irrigation, total water use during May-September peak irrigation season would be reduced by 40-70% if wastewater were reused for irrigation.
Return to the SepticTankInfo.com Home
Page.
