Suspended Solids Filtration In Water Recirculation Systems

Bruce L. Tetzlaff
Fisheries Research Laboratory
Southern Illinois University, Carbondale, IL

Introduction

Finfish aquaculture in recirculating systems has several unique problems as compared to extensive aquaculture in ponds or single-pass raceway systems. Among these are maintenance of dissolved oxygen, oxidation of dissolved nitrogenous wastes, and solids removal. Suspended solids can build up in recirculating systems to the level where the entire system will fail -- the biofilter will clog and suffocate the nitrifying bacteria and water flows will be disrupted. Particulate waste removal is the most problematic component of water reuse systems in aquaculture (Tetzlaff and Heidinger 1990; Lewis et al. 1981).

Three general method are used to remove suspended solids: sedimentation, centrifugal concentration, and mechanical filtration. Each of these has some inherent benefits and detriments.

Once particulates are removed from the aquaculture system, they must then be disposed of in a manner which conforms to the numerous state and federal regulations.

Particulate Waste -- The Problem

Particulate wastes in recirculating aquaculture systems come from three major sources: the fish, the feed, and heterotrophic organisms (bacteria and fungi). However, all of these result from the incomplete utilization of feed by the fish. Poor feed utilization results from: 1) poor quality of feed with a high percentage of "fines", 2) poor feeding techniques where some feed is made unavailable to the fish (washed down the standpipe), 3) overfeeding, 4) improper feed type for the fish -- some fish will chew food pellets that are too large, releasing particulates through the gill covers, and 5) a poor match between the nutrient requirements of the fish and those contained in the feed.

The aquaculturist can control the first four of these. In purchasing feed, you get what you pay for. Some low quality feeds can come with 15% "fines"; others have insufficient binding agents added so the pellets begin to fall apart as soon as they get wet. There is also some degree of "fine" production by automatic feeders. Early designs of feeders which use blowers to move feed to tanks were particularly prone to this problem. The feed pellets would grind against one another and produce a high percentage of "fines". These can contribute significantly to the suspended particle load of a system.

Poor feeding techniques, overfeeding, and pellet size are all problems which are associated with poor animal husbandry. The aquaculturist should be aware of any system flaws which allow the loss of feed before the fish can use it. Adjustments in the size of the screen covering the tank effluent may be required. Feeding less feed, but more often, prevents feed from breaking down before it is consumed by the fish. Elimination of the second problem, overfeeding, requires considerable knowledge about the animal one is rearing. Information such as optimum and maximum feed consumption as they relate to the temperature of the system, the size of the fish, and the conversion ratio of feed to fish should be known. Most of the feed given a fish beyond the amount it can use is released into the water as particulate waste. This is detrimental for the recirculating system and is also economically inefficient. The use of improper sizes of feed also results from insufficient knowledge. Many fish have pharyngeal pads which they use to move food to their stomachs. These pads are usually equipped with teeth. Fish will use these teeth to grind pellets that are too large to swallow. The resultant particulate waste can be seen as small clouds exiting their opercula. Thus, improper feed sizes should be avoided. Since there is often a several-week delay between the time feed is ordered and delivered, the aquaculturist must be able to predict the growth rate of his fish. With new systems and new animals experience is required.

Particulate matter caused by an unbalanced diet is more difficult to control. Only the salmon and trout have their nutritional requirements well known, although channel catfish nutrition is becoming better understood. For many of the other species being cultured, little is known of the basic requirements, such as optimum protein, carbohydrate, and fat levels, and which proteins can be digested. When the feed contains too high a lipid level, or lipid that are undigestible, the fish produce feces with a high level of fat. The resultant particulate waste is then less dense than water. Particulate waste removal systems that rely on the difference in density of the waste and water are thus ineffective. The result is lower food conversions and higher levels of particulate waste entering the system.

Acceptable food conversion rates for finfish range from 1.5 to 2.2. These ratios are based on dry weight of feed to wet weight of fish. Since fish are 71- 76% water, a food conversion of 2:1 is really 7.7:1 when expressed as dry weight of feed to dry weight of fish. Thus 87% of the feed given to a fish is not converted to flesh. The rest of the feed is released into the water as metabolic products and particulate waste. The greatest amount of waste is carbon dioxide and other dissolved inorganic products although settleable solids are also significant.

Piper et al. (1982) state that 0.3 kg of settleable solids are produced from each kg of feed supplied to the fish. Thus an aquaculture operation holding 25,000 pounds of fish and feeding 3% per day is producing 225 pounds of particulate waste per day. Liao and Mayo (1974) found that 70% of the total ammonia production of a system is in the particulate waste. Clarke and Phillips (1989) note that a salmon production facility will produce 40 kg of particulate nitrogen, 7-10 kg of solid phosphorus, and 250 kg of particulate carbon (organic material) per 1000 kg of fish produced. A large biological oxygen demand (BOD) and chemical oxygen demand (COD) is also produced by the particulate waste. Efficient removal of the particulate waste thus greatly reduces the oxygen and nitrification requirements of the system. The high organic content of the particulate waste permits heterotrophic bacteria populations and fungi to grow rapidly. Their growth creates high BOD in the system and the end products of their digestion release ammonia into the water.

Few studies have examined the physical properties of the particulate wastes produced in aquaculture. Generally it is agreed that most of the wastes produced by the fish have a density only slightly greater than water -- wastes resulting from uneaten feed will be slightly more dense. The size of the particulates can be highly variable, although Clarke and Phillips found that most of the waste coming from a salmon smolt facility was in the range of 50-100 micrometers. Actions such as pumping homogenize the waste, producing finer particles that are harder to settle out or filter.

Removal of Particulate Waste

Removal of particulate waste from a recirculating aquaculture system can be accomplished through settling, centrifugal separation or filtration. All of these general methods will work. The problem arises when they have to work efficiently and economically. An effective particulate removal system should: 1) concentrate the wastes so that water loss from the system is minimal, 2) remove the waste rapidly so that heterotrophs do not have time to develop and use dissolved oxygen and release ammonia, 3) not require large energy inputs which increase operating costs, and 4) not result in major pressure losses which require additional pumping or potential loss of water flow. No system totally fulfills these goals.

Settling Systems

The simplest type of particulate waste removal system is the settling basin. The velocity of the effluent water is slowed in a large tank or pond and the particulate wastes are allowed to settle to the bottom for later removal. The design of settling basins involves four inter-related factors: 1) retention time, 2) water velocity and flow distribution, 3) the density of the particulate waste, and 4) water depth (Piper et al. 1982).

Retention time is simply the time it takes to exchange the water in a container at a given flow rate. Longer retention times give the particulate waste more time to settle to the bottom. Thus, the larger the settling basin, the more waste that will be removed. The retention time needed to settle wastes depends on their settling rate which, in turn, depends on the density of the particle and the amount of water turbulence in the Basin. Depending on these factors retention times vary from 15 minutes to 2 hours (Piper et al. 1982).

Slow water velocities and laminar flow patterns enhance the settling process. Piper et al. (1982) recommend that the influent water to a settling basin should be directed through perforated diffuser plates to create laminar flow. They recommend plates with greater than 50% open area. Without diffusers, a channel effect can occur resulting in very reduced retention times for some of the wastes.

The speed which a particle settles is controlled by its mass (weight), density (weight/volume), size, and the viscosity of the fluid. The theoretical settling rate can be modeled using Stoke's Law of physics. Basically, this law states that: 1) denser particles settle faster, and 2) particles with larger diameters settle faster. Unfortunately the particulate waste produced in aquaculture is usually small, and has a density near that of water. Clarke and Phillips (1989) state that their salmon smolt operation produces particles in the range of 50-100 microns. Assuming that the density of the 100-micron particles is 1.4 (40% denser than water), these particles would settle at a rate of 0.01 cm/sec. or take 4.3 minutes to settle one inch. This calculation does not account for the effects of water movement which helps to keep particles in suspension.

Piper et al. (1982) note the importance of settling basin depth on the effectiveness of particulate waste removal. A shallow basin may allow water currents to resuspend particulates. A basin that is too deep will not allow the particles to reach the bottom during the retention time and thus be flushed out. These authors recommend a depth of 1.5 feet for a settling basin. Davis (1972) and Liao and Mayo (1972), however, recommend settling basin depths of greater than 3 feet.

Once the particulate waste has settled out, it must be concentrated and removed. If the waste is not removed rapidly, nutrients can be released through leaching and bacterial action. These will increase the ammonia load on the biofilter, provide nutrients for heterotrophic bacteria, and raise the BOD and COD in the system. Sludge is generally removed from settling basins either by de-watering the basin and scraping, or by vacuum pumping. De-watering the system requires that a method be installed to by-pass the settling basin during cleaning, and that the basin be drainable. Some settling basins have mechanical sludge scrapers installed in the bottoms to facilitate cleaning.

Based on the above information, it is obvious that settling basins are not efficient particulate waste removal systems for large in-door facilities. A system running 500 gallons/minute would require a 1 5,000-gallon settling basin. Piper et al. (1982) discuss the use of two 30 ft X 100 ft basins, 4 feet deep, to handle a 600-gallon/minute influent. In this system, 85% of the wastes settle out.

In a small aquaculture facility settling basins may be more realistic. A settling basin could be incorporated below the rearing tanks and thus reduce the space requirements for the basin.

Several modifications of the basic settling basin are commercially available. They incorporate a series of parallel plates or tubes to create areas of no turbulence, allowing for more rapid settling of particulate waste. These units require less settling time, and thus permit a much smaller unit. These plate or slant-tube settlers are still controlled by the physics of the particulate waste -- the particles must be sufficiently dense to settle.

The benefits of settling systems are: 1) low cost, 2) ease of operation -- they operate on gravity and produce no loss in pressure, 3) low energy requirement, and 4) low water requirements when properly designed. The detriments of these systems include: 1) large size which precludes their use in in-door facilities, 2) relative inefficiency -- small or low-density particles will not settle out, and 3) the continued contact of the waste with the culture water which allows ammonia and other nutrients to leach into the water.

Centrifugal Systems

Particulate waste concentrators which operate on centrifugal force have been used in manufacturing for many years, but have only been used for the treatment of particulate waste from aquaculture for about 10 years. Centrifugal concentrators are also known as swirl separators, hydrocyclones, and vortex concentrators. They operate on the principle of spinning particulate-laden water so that the denser particulate waste is forced to the wall of a cone by centrifugal force. Influent water enters a closed cone tangentially, causing a swirling motion in the cone. Particulates are forced to the wall of the cone where they settle and are removed with a continuous effluent of water. This waste effluent can compose as much as 5% of the total system flow. Scott and Allard (1983) restricted this water loss to less than 3.8% of total flow. The cleaned water exits the cone through an effluent pipe in the top of the unit.

The effectiveness of these units depends on: 1) the amount of centrifugal force placed on the particles, and 2) the density of the particles in relation to that of water. The centrifugal force in the cone is dependent on the velocity of the entering water. Pumping is usually required to obtain sufficient water velocity to force the particulates to the cone wall.

The same factors that affect the ability of a particle to settle in a settling basin affect the ability of a centrifugal concentrator to remove it, namely; the mass, the diameter, and the density of the particle. Thus those particles with densities similar to that of water are not removed. In addition, the pumping of the waste water to produce sufficient velocity homogenizes it, producing particles with smaller diameters.

The efficiency of these units has been evaluated by Scott and Allard (1983, 1984) and Enqvist and Larrson (1986?). Scott and Allard (1983) found that this type of particulate removal system would remove 56% of the net dry solids circulating in their system. In 1984, these authors further evaluated centrifugal concentrators in a recirculated trout culture system. They found that 90% of the particulate wastes produced by the trout were larger than 77 micrometers in diameter and that the concentrator removed 70% of these. With smaller particle sizes, the concentrator was less efficient, removing only 10% of the waste entering it. Enqvist and Larrson (1986?) examined a centrifugal concentrator in a single-pass salmon hatchery. They found that a centrifugal concentrator removed 66% of the total solids, 33% of the BOD, 18% of the total nitrogen, and 40% of the total phosphorus.

The main benefit of centrifugal concentrators is their compact size. This makes them adaptable to in-door culture facilities. However, there are numerous weaknesses in this system. Centrifugal concentrators require high water velocity to produce the force to separate the particles. This necessitates pumping, and thus homogenizing the waste. It also requires that pumps produce a higher pressure than is necessary to just move the water. This is energy inefficient. These systems also require a continuous waste effluent of up to 5% of total system flow. Scott and Allard (1984) their concentrators to reduce this loss. The concentrator waste effluents were directed through 100-foot coils to reduce flow and pressure. However, they still had continuous water loss and thus needed a freshwater make-up supply. This fresh water would have to be purchased or pumped from well and then potentially heated. A 3% make-up on a 500 gallon/minute system totals 21,600 gallons/day, This increases the costs production. It is possible to retreat the effluent of the concentrators to reduce this water loss. As an example, the effluent from a centrifugal concentrator would be 15 gallons/minute from a system running 500 gallons/minute (3%). The 15 gallons/minute could be directed to a settling basin for further concentration of the sludge. At this flow rate, a settling basin of only 900 gallons would provide a 1-hour retention time.

Mechanical Filtration

Mechanical filters can be divided into two subgroups; media filters and screen filters. Both operate by acting as a physical barrier to the particulate waste. The main advantage of mechanical filters over the previous two systems is that filters do not rely on the density of the particle -- floating objects are filtered equally as well as heavy objects. Filtration is only limited by the size of the particle. Unlike sedimentation basins and centrifugal separators, mechanical filters require a cleaning cycle which utilizes energy.

Media Filters

Media filters are typically composed of a container holding a fixed volume of the filtering media. Sand is the most common filtering media, although very small systems (aquaria) may use diatomaceous earth. Manufactured media consisting of very small beads is also being used. Media filters can be pressurized or open-flowing. Open-flow filters are very limited in the amount of water flow they can handle. The maximum for sand filters is about 1 gallon/minute for every square foot. This discussion is therefore limited to pressurized systems.

Pressure sand filters consist of a closed chamber holding the filter sand. Raw water is pumped into a diffuser and filters through the sand. Cleaned water exits through the bottom of the filter. These filters are very effective at removing particulate waste. They are capable of filtering particles as small as 15 micrometers.

There are, however, numerous problems with pressure sand filters. They have high costs involved with backwashing, and create significant pressure losses. As particulate matter is removed from the culture water, it accumulates on top of, and in, the sand media. When this happens, water pressure on the upstream side of the filter increases and water flow decreases on the downstream side. The SIUC Fisheries Research Laboratory had one pressure sand filter explode from high pressure when using a extremely fine-grain medium. As the filter clogs, flows to the system can be restricted significantly and even totally stop.

The only means to prevent clogging is frequent backwashing of the medium. Backwashing involves reversing the direction of the water flow and providing adequate pressure and volume to suspend the sand while flushing the particulate material to waste. A pressure sand filter typically requires approximately 20-40 volumes of water to backflush -- a 60-gallon filter thus requires 1,200 gallons of water.

A second common problem encountered with these filters is adhesion of the media by heterotrophic bacteria. Growth of these bacterial populations within the filter create viscous masses of sand, bacteria, and particulate waste that are not removed by normal backwashing. Water flow is inhibited by these masses, reducing the efficiency of the filter. Chlorination, combined with mechanical agitation, is required to eliminate the bacterial masses. The filter must be taken off line during this procedure.

These problems make pressure sand filters one of the most effective, but least efficient methods for removing particulate waste from a recirculated system. They require high capacity pumps both to push the raw water through the filter and to backwash the filter, and require large amounts of water for backwashing.

Screen Filters

For this discussion, screen filtration includes any mechanism where particulate matter is trapped on a thin porous membrane. The filter membrane may be composed of woven wire or other material, a perforated plate, or a set of parallel wires with fixed distances between them. Screen filters can have a stationary filter plate or a moving membrane. The moving membrane filters often take the form of a rotating drum or plate. Screen filtration systems vary from a simple plate or bag placed in a tank to complex automatic systems costing in excess of $50,000.

Three factors govern the effectiveness of all screen filtration systems: 1) the efficiency of particulate removal, 2) the rate of clogging, and 3) the ease and efficiency of cleaning. Traditionally screen filter systems used in industrial applications have not been successful when applied to aquaculture using water reuse, The efficiency of particulate removal is governed by the mesh size of the screen, the amount of open area compared to the area screen (% open area), and whether the system forces soft particulate material through the mesh. Microscreens are commercially available with extremely small mesh sizes.

The major limitation with screen filtration has been the rapid growth of bacteria and fungi on the screen and ineffective backwashing techniques to remove that growth. This is particularly true in warm-water aquaculture systems. Moving screen systems may also have a tendency to grind soft particulate matter until some of it will pass through the mesh.

Of the many proprietary screen filtration systems on the market, the Triangelfilter by the Swedish company, Hydrotech, is gaining popularity among fish culturists. The design of the Triangelfilter is based on a slanted filter plate. The slanted plate allows the input water to help in cleaning the filter screen. Wastewater flowing over the top of the plate pushes some of the sludge down the slanted screen to a sludge collection trough. When the filter screen begins to clog, a spray arm washes the screen. Water loss from the Triangelfilter is designed to be 0.2-0.5% of input flow. An additional 1000 gallons per day is used for backwash, although much of this actually enters the system. Thus for a 500 gallon/minute system, the total water use would be 2,400-4,600 gallons. This would require an input of 1.2-3.2 gallons/minute of new water. The manufacturers recommend a post-Triangelfilter settling basin to further concentrate the sludge. Use of this would further reduce the water requirements. The Triangelfilter has been evaluated in salmon hatcheries in Sweden by Enqvist and Larrson (1988?), and Canada by Goldberg et al. (1988). In these applications the filter removed 80-100% of the total dissolved solids, 27-70% of the total nitrogen, 67-79% of the total phosphorus, and 75-82% of the BOD. Variations in the efficiency of the filter were attributed to the size of fish being reared and the speed with which the particulates were being removed. Smaller fish produced finer waste which reduced efficiency of the filter -- self cleaning circular tanks reduced sludge breakdown and increased the efficiency.

I have not found an evaluation of this filter in a warm-water fish culture operation using water recirculation. The rapid growth of bacteria in warm water may cause bio-fouling of the filter screen similar to that seen on other screen filtration systems. If this occurs, an expensive retrofit for steam cleaning would be required.

Screen filtration systems vary considerably on the amount of energy input required to operate them. Rotating drums and disks have the greatest power consumption. Some systems require high pressure pumps or steam generators for backflushing. These add to the production costs. Another consideration is the head loss in the filter. All screen filters should be capable of operating on gravity. However, the physical dimensions of some systems results in some head loss. The Triangelfilter screen has a 16-inch head loss at maximum flow rate. This amount of head loss makes this system marginal when considering air lift pumping.

Waste Disposal

Depending on the food conversion ratio, a fish culture operation will produce approximately 0.5-1.5 tons of particulate waste (dry weight) for each ton of fish produced. However, when it leaves a particulate filter this material is mixed with a lot of water. Goldberg et al. (1988) measured 348 mg/L of total suspended solids in the sludge effluent of their Triangelfilter. The dry solids level was thus only 0.0035%. Using a post-filter settling basin as a sludge concentrator, the percent dry matter can be increased to 8-10%. The fish culturist must thus dispose of 0.5-1.5 tons of waste mixed in 5-15 tons (1,200-3,600 gallons) of water for each ton of fish. Based on the data of Clarke and Phillips (1989) this mixture will contain 210 pounds of nitrogen, 30-40 pounds of phosphorus and 500 pounds of particulate carbon. A commercial aquaculture operation thus creates a large quantity of nutrient rich manure.

This waste can be disposed of by field application, sending it to a municipal sewage treatment facility, drying and burning, or discharging it to a lagoon. Because of the high costs of vacuum de-watering and the regulations governing incineration of sludge, this option is unlikely.

The other methods of disposal of this waste are regulated by one or more government agencies, including the U.S. Army Corps of Engineers, the U.S. EPA, the state EPA, the Department of Conservation, and the Department of Agriculture. If a fish culture operation produces more than 20,000 pounds of aquatic animals per year, or discharges wastewater for more than 30 days per year, a National Pollution Discharge Elimination System (NPEDS) Permit is required. Although a federal program, NPEDS permit application are obtainable from the Illinois EPA. Basically, applications requesting permits to discharge waste into any "navigable water", including streams, rivers, lakes, natural ponds, and wetlands will be denied.

Field application of this waste will fall under two guidelines, both administered by the Illinois EPA. If a vegetative filter system in employed, the information in the "Design Criteria Regarding Runoff Field Application Systems" guidelines must be followed. Since a iart~culate filtrat~on system is a type of wastewater treatment system, the guidelines of the "Land Application of Sludge Permit" and the "Design Criteria for Field Application of Livestock Waste" must be followed. These rules basically are the same that a swine producer must follow.

Because of the high nutrient content of aquaculture sludge, Piper et al. (1982) suggests that the waste could be sold to fertilizer manufacturers. This may entail the use of expensive de-watering systems and thus may not be economical. Moving the waste also requires an Illinois EPA "Special Hauling Permit" for each vehicle used to transport the sludge.

Construction of a lagoon or other waste treatment facility requires another Illinois EPA permit, "Construction and/or Operating Permit, Division of Water Pollution Control." Treatment works which may, or may not, have discharge may need this permit.

The rules governing discharges from aquaculture in Illinois are still being refined. Currently, the major concerns are the release of non-native species and the high levels of phosphate and solids.

The Department of Lands, Forest and Water Resource Services in British Columbia have developed rather strict, but straight-forward rules for aquaculture operation. Their rules are based on the weight of fish currently being reared and are defined as pounds of pollutant/100 pounds of fish/day.

With the new state and federal incentives that are available, construction or restoration of wetlands to process the waste may be a disposal method worth evaluating.

Discussion

Particulate waste removal a major, if not the major, problem with recirculating aquaculture systems. The high nutrient load and BOD associated with particulate wastes have caused more biofilter systems to fail than any other cause.

The main difficulty in efficiently and economically removing this waste is the similarity in density of the waste and water. This similarity prevents systems based on gravity or centrifugal force from removing much of the waste.

Of the current technology available, screen filtration seems to offer the best removal of particulate wastes. However, because most of the technology was developed for salmonid production, the problem with bio-fouling of the filter screens by rapidly-growing bacteria has not been addressed. Current methods for filter cleaning are expensive, either using considerable power or water.

Designs for recirculating fish culture systems should incorporate criteria for the efficient transport and removal of particulate waste, including: self-cleaning tank designs, gravity flow to the particulate filter, rapid or frequent cleaning of the filter, and rapid isolation of the filter sludge from the culture water. These criteria minimize the breakdown of the waste into particles that are difficult to remove, and isolate them from the culture water so that nutrients cannot leach from the sludge, and place additional burdens on the biofilter and oxygenation system.

Sources Cited

Clarke, R., and M. Phillips. 1989. Environmental impacts of salmon aquaculture. AAC Bulletin 89(4):24-31.

Goldberg, H., J. Korn, F. Berry, R. Carswell, A. Ismond, and A. Martin. 1988. A case study: salmon hatchery effluent treatment in Sechelt, B.C. Proceeding of the Aquaculture International Congress and Exposition, Sept. 6-9, 1988. Vancouver, B.C., Canada.

Davis, J.T. 1977. Design of water reuse facilities for warm water fish culture. Ph.D. Thesis, Texas A&M University. 109pp.

Enqvist, M., and P-O. Larrson. (1988?). Cleaning effects on effluents from fish farms by Triangelfilter and swirlseparator. Report submitted to Hydrotech, Vellinge, Sweden. 5pp.

Lewis, W.M., R.C. Heidinger, and B.L. Tetzlaff. 1981 Tank culture of striped bass: production manual. Fisheries Research Laboratory, Southern Illinois University, Carbondale. 115pp.

Liao, P.B., and R.D. Mayo. 1974. Intensified fish culture combining water recirculation with pollution abatement. Aquaculture 3:61-85.

Piper, R.G., l.B. McElwain, L.E. Orme, J.P. McCraren, L.G. Fowler, and J.R. Leonard. 1982. Fish Hatchery Management. U.S. Department of Interior, Fish and Wildlife Service, Washington, D.C

Scott, K., and L. Allard. 1983. High-flowrate recirculation system incorporating a hydrocyclone prefilter for rearing fish. Progressive-Fish Culturist 45:148-153.

Scott, K., and L. Allard. 1984. A four-tank water recirculation system with a hydrocyclone prefilter and a single water reconditioning unit. Progressive-Fish Culturist 46:254-261.

Tetzlaff, B.L., and R. C. Heidinger. 1990. Basic principles of biofiltration and system design. Illinois Aquaculture Resource/Research Center. Southern Illinois University, Carbondale. 22pp.