UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Determination of the effects of fish size and feed pellet size on the settling characteristics of Rainbow.. Thomson, Douglas Edward 1986

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
UBC_1987_A6_7 T46_5.pdf [ 4.27MB ]
Metadata
JSON: 1.0058685.json
JSON-LD: 1.0058685+ld.json
RDF/XML (Pretty): 1.0058685.xml
RDF/JSON: 1.0058685+rdf.json
Turtle: 1.0058685+rdf-turtle.txt
N-Triples: 1.0058685+rdf-ntriples.txt
Original Record: 1.0058685 +original-record.json
Full Text
1.0058685.txt
Citation
1.0058685.ris

Full Text

DETERMINATION OF THE EFFECTS OF FISH SIZE AND FEED PELLET SIZE ON THE SETTLING CHARACTERISTICS OF RAINBOW TROUT ( SALMO GAIRDNERI ) CULTURE CLEANING WASTES by  DOUGLAS EDWARD THOMSON B.Sc. (The University of British Columbia, 1983) A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE  in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF BIO-RESOURCE ENGINEERING  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA November, 1986 ® DOUGLAS EDWARD THOMSON, 1986  In presenting degree  this  at the  thesis  in  partial fulfilment  of  University of  British Columbia,  I agree  freely available for reference copying  of  department publication  this or of  and study.  thesis for scholarly by  this  his  or  her  purposes  Date  representatives.  BIO - RESOURCE ENGINEERING  DE-6(3/81)  APRIL 30,1987  that the  may be It  thesis for financial gain shall not  The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3  requirements  I further agree  permission.  Department of  the  for  an  advanced  Library shall make it  that permission  for extensive  granted  head  is  by the  understood be  that  allowed without  of  my  copying  or  my written  ABSTRACT  This research reports on the determination of the effects of fish size and feed pellet size on the settling characteristics of Rainbow trout (Salmo gairdneri) culture, tank cleaning wastes.  Flocculant particle settling  curves (Type II) were developed  from settling column  analysis of cleaning wastes from 11-311 gram Rainbow trout fed a moist pellet diet (Oregon Moist Pellet ®). Four feed pellet sizes were investigated: 3/32, 1/8, 5/32 and 3/16 inch.  Overall non-filterable residue removal curves and individual particle settling velocity distribution curves, derived from the Type II settling curve of each fish size and feed pellet size group, were compared. Slopes and y-intercepts of the linearized overall non-filterable residue removal curves and individual particle settling velocity distribution curves were compared using the Equality of Slope Test (S:SLTEST).  Results  of the  significant  test for a common regression  differences  equation indicated there were no  in the proportional distribution of particle sizes within the  cleaning wastes. Variations observed in the initial rates of removal within the overall non-filterable residue removal curves were considered insignificant  Settling trials were pooled in order to obtain single curves, characterizing the overall solids removal rate and the individual particle settling velocity distribution of the waste solids.  ii  TABLE OF CONTENTS  LIST OF FIGURES  iv  LIST OF TABLES  vi  ABSTRACT  ii  ACKNOWLEDGEMENT  vii  1. INTRODUCTION  1  2. LITERATURE REVIEW  7  3. THEORY  19  4. MATERIALS AND METHODS 4.1. Species 4.2. Fish Size 4.3. Feed Type and Pellet Size 4.4. Laboratory Set Up 4.4.1. Containment 4.4.2. Water Supply 4.4.3. Aeration 4.4.4. Stocking Density 4.5. Settling Column 4.6. Experimental Procedure  34 34 34 36 36 36 36 38 38 38 40  ,  5. RESULTS AND DISCUSSION  43  6. CONCLUSION  92  7. RECOMMENDATIONS FOR FUTURE WORK 7.1. Residence Time  93 93  7.2. Feed Type  94  REFERENCES  95  APPENDIX I  101  APPENDIX II  103  iii  LIST OF FIGURES  Figure 1.1: Removal of NH -N, N0 , P0 , and BOD as a function of solids removal (Muir, 1982) Figure 2.1: Settling velocities for intact Rainbow trout excreta as a function of fish size (Warrer-Hansen, 1982) Figure 2.2: Relationship between residence time and solids removal (Warrer-Hansen, 1982) Figure 3.1: Schematic of an ideal rectangular sedimentation basin, indicating the settling path of discrete particles Figure 3.2: Discrete particle velocity cumulative frequency distribution Figure 3.3: Path trajectory of particles in discrete particle settling Figure 3.4: Path trajectory of particles in flocculant particle or Type II settling Figure 3.5: Type II or flocculant particle settling curves Figure 3.6: Overall non-filterable residue removal curve Figure 3.7: Graphical representation of the percentage of non-filterable residue particulates with settling velocities greater than X Figure 3.8: Individual particle settling velocity distribution curve Figure 4.1: Schematic of 200 litre oval fibre glass tanks Figure 4.2: Schematic of settling column used Figure 5.1: Non-filterable residue removal for example data given in Table IV, as a function of residence time and column depth Figure 5.2: Type II orflocculantparticle settling curves for example data given in Table IV Figure 5.3: Overall non-filterable residue removal curve for data given in Table IV Figure 5.4: Individual particle settling velocity distribution curve for example data in Table IV Figure 5.5: Overall non-filterable residue removal curves for 11 gram Rainbow trout fed 3/32 OMP feed pellets Figure 5.6: Overall non-filterable residue removal curves for 17 gram Rainbow trout fed 3/32 OMP feed pellets Figure 5.7: Overall non-filterable residue removal curves for 21 gram Rainbow trout fed 3/32 OMP feed pellets Figure 5.8: Overall non-filterable residue removal curves for 21 gram Rainbow trout fed 1/8 OMP feed pellets Figure 5.9: Overall non-filterable residue removal curves for 41 gram Rainbow trout fed 1/8 OMP feed pellets Figure 5.10: Overall non-filterable residue removal curves for 54 gram Rainbow trout fed 1/8 OMP feed pellets Figure 5.11: Overall non-filterable residue removal curves for 54 gram Rainbow trout fed 5/32 OMP feed pellets Figure 5.12: Overall non-filterable residue removal curves for 73 gram Rainbow trout fed 5/32 OMP feed pellets Figure 5.13: Overall non-filterable residue removal curves for 93 gram Rainbow trout fed 5/32 OMP feed pellets Figure 5.14: Overall non-filterable residue removal curves for 89 gram Rainbow trout fed 3/16 OMP feed pellets Figure 5.15: Overall non-filterable residue removal curves for 311 gram Rainbow trout fed 3/16 OMP feed pellets 3  3  4  iv  5 13 15 20 23 24 24 27 29 32 33 37 39 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65  Figure 5.16: Individual particle settling velocity distribution curves for 11 gram Rainbow trout fed 3/32 OMP feed pellets 66 Figure 5.17: Individual particle settling velocity distribution curves for 17 gram Rainbow trout fed 3/32 OMP feed pellets 67 Figure 5.18: Individual particle settling velocity distribution curves for 21 gram Rainbow trout fed 3/32 OMP feed pellets 68 Figure 5.19: Individual particle settling velocity distribution curvess for 21 gram Rainbow trout fed 1/8 OMP feed pellets 69 Figure 5.20: Individual particle settling velocity distribution curves for 41 gram Rainbow trout fed 1/8 OMP feed pellets 70 Figure 5.21: Individual particle settling velocity distribution curves for 54 gram Rainbow trout fed 1/8 OMP feed pellets 71 Figure 5.22: Individual particle settling velocity distribution curves for 54 gram Rainbow trout fed 5/32 OMP feed pellets 72 Figure 5.23: Individual particle settling velocity distribution curves for 73 gram Rainbow trout fed 5/32 OMP feed pellets 73 Figure 5.24: Individual particle settling velocity distribution curves for 93 gram Rainbow trout fed 5/32 OMP feed pellets 74 Figure 5.25: Individual particle settling velocity distribution curves for 89 gram Rainbow trout fed 3/16 OMP feed pellets 75 Figure 5.26: Individual particle settling velocity distribution curves for 311 gram Rainbow trout fed 3/16 OMP feed pellets 76 Figure 5.27: Variability of the y-intercept of the linearized overall non-filterable residue removal curves as a function of the initial non-filterable residue concentration 83 Figure 5.28: Overall non-filterable residue removal curve for Rainbow trout fed an OMP diet 87 Figure 5.29: Individual particle settling velocity distribution curve for Rainbow trout fed an OMP diet 88 Figure 530: Summary of capital, operating, and maintenance costs for a sedimentation basin as a function of surface area (Underwood McLellan, 1979) 89  v  LIST OF TABLES Table I: Summary of integration calculations of flocculant particle settling curve for residence time T 28 Table II: Weight classes and corresponding feed pellet sizes used in waste trials 35 Table III: Variability of initial non-filterable residue concentrations as a function of column depth 47 Table IV: Non-filterable residue concentrations for a sample settling trial 49 Table V: SLTEST generated common regression equations for Figures 5.5 - 5.15 ....101 2  vi  ACKNOWLEDGEMENT I would like to express my sincere thanks to the following individuals for their contributions in making this thesis possible. Dr. N.R. Bulley for his supervision and guidance of my M.Sc. program and, for the valuable encouragment and patience shown throughout the production of this thesis.  I would also like to thank Prof. B. March for serving as a committee member and for making available the South Campus Fish Nutrition Lab for the waste trials. I would like to thank Bill McLean (Department of Fisheries and Oceans) for serving as a committee member and for supplying the feed used in the study.  Special thanks to Hugh Sparrow and Dale Larson of the Fraser Valley Trout Hatchery (Provincial Fish and Wildlife Branch) for supplying the Queens fish used in our trials.  A very special thanks are due to Jannice Wormsworth for her daily assistance in feeding and caring of the fish..  Finally I would like to' thank my family and friends for their patience, encouragement and unfailing support throughout this project  vii  1. INTRODUCTION The  effluent  discharged from a freshwater  trout farm or hatchery offers unique  problems in the areas of wastewater treatment and management  While at times  resembling that of a domestic water supply, effluents can reach strengths more frequently associated with conditions in a secondary sewage treatment plant  In the past when fish farms were small and few in number, discharged farm effluents posed no significant problems. In some cases, the contribution of solids and dissolved nutrients to oligotrophy receiving waters provided a positive benefit in the form of increased primary productivity (Samis, 1983). However, as farms have grown rapidly in size and numbers, the discharge of untreated effluents has become a source of environmental concern.  A farm producing 50-75 tonnes/year of fish has an estimated water use of 500 1/ sec which is the equivalent demand of approximately 170,000 people (Warrer-Hansen, 1979a). Bergheim and Selmer-Olsen (1978) estimated the daily loading of organics and nutrient salts from a Norwegian fish farm to be in the order of 1260 population equivalents. In Idaho, the combined waste output from 49 hatcheries and farms located along a 27 mile section of the Snake River, has been estimated to be in the order of 63-73,000 pounds of biological contaminants daily (Klontz & King, 1975; Klontz et al., 1978). Faure (1977) estimated that this would be the equivalent of a city of 1.2-1.5 million people discharging untreated sewage directly into the river on a daily basis.  Left untreated these effluents can have significant impacts on downstream receiving 1  INTRODUCTION / 2 waters. Solids discharged in the form of unconsumed feed pellets and fish faeces can settle out in slow moving streams and rivers. The accumulation of solid wastes suffocates native stream flora and fauna, while providing a rich medium for less desirable, more pollution-tolerant plant and animal species (Mayo & Liao, Bodien, 1970;  1969;  Brisbin, 1970). In turn, bacterial decomposition of the accumulated  organic solids can reduce dissolved oxygen levels in the stream (Bergheim & Silversten, 1981), resulting in fish kills under extreme conditions (Solbe, 1982).  Nutritional enrichment from the discharge of dissolved nutrients such as ammonia and inorganic phosphorus can accelerate eutrophication in low flow streams (Sumari, 1982). At higher dilutions, the added enrichment encourages the growth of noxious algae species (Cyanophyta spp.) and bacterial groups commonly referred to as "sewage fungus" (Mantle, 1982).  In the past the focus of aquaculture waste research has centered on the areas of problem surveying (Liao, 1970; Sumari, 1982; EIFAC, 1982; DFO, unpublished) and waste quantification (Liao, 1970a; Scherb & Braun, 1971; Liao & Mayo, 1972; Speece, 1973; Knbsche & Tscheu, 1974; EIFAC, 1982; Solb'e, 1982). Based on the available literature it has been possible to identify the major relationships and to develop appropriate equations  to predict the  quantity  and quality of  the  various waste  constituents in fish culture effluents.  The  quantity of these waste elements has been found to be a function of feed  quantity (Brockway, 1950; Liao, 1970a; Liao et al., 1972; Willoughby et al., 1972; Liao & Mayo, 1974; Muir, 1978; Faur'e, 1977), feed type (Solberg & Bregnballe, 1977;  INTRODUCTION / 3 Gunther et al., 1981; Butz & Vens-Cappell, 1982; Warrer-Hansen,  1982;  Stechey,  Solberg & Bregnballe, 1982;  1986), fish size (Wheaton, 1977)  and temperature  (Wheaton, 1977).  Feeding operations and cleaning activities have also been reported to influence the type and  concentration of pollutants discharged. During normal operations the level of  suspended solids in farm effluents is generally low, averaging approximately 7 mg/l (Liao, 1971). However, during cleaning operations, the concentrations of suspended solids have been shown to increase significantly. Liao (1970) reported that the level of suspended solids in the effluent increased during cleaning to an average of 96 mg/l, while Bergheim et al. (1984) reported concentrations ranging from 30 to 5800 mg/l.  Analysis of daily flows from the Summerland hatchery (British Columbia Fish and Wildlife) demonstrated that solids discharged during cleaning accounted for 20-25% of the total dishcharge of BOD  5  solids (Brisbin,  (5 day Biochemical Oxygen Demand) and suspended  1970; Underwood McLellan,  1970). In a similiar report on, U.S.  hatcheries, Liao (1970) reported cleaning flows contributed 4-52% and 5-22% of the non-filterable residue and BOD respectively to the overall waste discharged. A survey of  ten Canadian federal salmonid rearing facilities  in British Columbia, reported  cleaning flows containing 8-25% of the total waste (Underwood McLellan, 1977).  Analysis of fish culture wastes has shown a strong correlation between the removal of solids fraction and the removal of other soluble constituents. Figure 1.1 (Muir, 1982) summarizes the work reported on by Muir & Lipper (1970), Liao & Mayo (1974) and Muir (1978) on the removal of NH -N, N0 3  3  , P0 and BOD as a function of the 4  INTRODUCTION / 4 removal of solids. In other work, Willoughby et al. (1972) reported that the removal of 90% of settleable solids resulted in an overall BOD reduction of 85%. Given the correlation illustrated in Figure 1.1, removal of the suspended solids fraction should be an important consideration in the treatment of fish culture effluents.  A variety of treatment systems have been surveyed which effectively reduce suspended solids (Underwood McLellan, 1979). However, of these systems, only sedimentation or gravity solid separation has been shown to be an economically viable method of treatment (Muir, 1978; Underwood McLellan, 1979; Warrer-Hansen, 1979).  In general, the treatment methods applied to fish culture wastes have been based on the assumption that fish wastes are characteristically similiar to human wastes. As such, standard domestic-type waste treatment systems have often been prescribed. However, inadequacies  in  the  basic  information  relating  to  the  characterization  of  the  non-filterable residue from hatchery effluent wastes (Brown & Nash, 1979; Underwood McLellan, 1979) has often required that high margins of safety be incorporated into the design, significantly reducing their cost effectiveness.  The purpose of this thesis will be to adress inadequacies in the available information characterizing the physical and behavioral nature of the non-filterable residue, in trout culture cleaning wastes, within a gravity solids separation system. The thesis proposes to develop the overall non-filterable residue removal curves and individual particle size distribution curves, used in characterizing the settling behavior of a waste solid. The overall non-filterable residue removal curves and individual particle size distribution curves to be used to test the hypotheses that:l) Fish size significantly affects the  INTRODUCTION / 5  FIGURE 1.1  Removal of NH , N0 , P0 , and BOD as a function of solids(Muir, 3  3  4  1982)  NH -N (Liao 3  / /'  1— 50 Removal o f s o l i d s  ,  BOD  1 100 (%)  & Mayo, 1974)  ( M u i r , 1978)  INTRODUCTION / 6 settling behavior of the waste solids, and 2) Feed pellet size significantly affects the settling behavior of the waste solids. The results will be related to current design practices for gravity solids separation systems.  2. LITERATURE  REVIEW  A sedimentation basin or clarifier can be utilized for removing metabolic waste solids from continuous  flow  or periodic cleaning  flows discharged from fish farms or  hatcheries (Huber & Valentine, 1978; Parker & Broussard, 1977). A number of solids removal efficiencies have been reported in the literature ranging from 25% to over 90% (Underwood McLellan,  1979). However, aside  from removal efficiencies,  the  literature provides little background information on which to base the design or selection of a clarifier.  The  proper selection and sizing of a solids separation system requires a thorough  understanding of the physical characteristics of the wastes to be treated. Of primary concern is the variability in particle settling velocities within the waste solids and the factors which influence the settling characteristics.  A discrete, solid particle will accelerate in a quiescent fluid until drag (F ) reaches equilibrium with the driving force (F) or gravitational force acting on the particle. Once equilibrium is reached, the particle no longer accelerates and begins to settle at a uniform velocity. The determination of the terminal velocity of the particle can thus be obtained by equating gravitational and drag forces acting on the particle.  The  driving force (F), acting on the particle, is the net effect of the particle weight  acting downwards and the buoyant force of the fluid acting upward. The driving force is given by (Clark et al., 1977);  7  LITERATURE REVIEW / 8  F  where  =  ( P  s  - P )gV  ( ) n  p = density of particle s p = density of fluid g = acceleration due to gravity V = volume of particle  The drag force acting on a particle is a function of the fluid density, fluid viscosity, particle velocity and the projected area of the particle in the direction of motion. Expressed as  F_ D  where  =  C Apv n  2  s  (2.2)  = drag force = Newton's drag coefficient A = projected area of the particle in the direction of motion y  = particle velocity  p = density of fluid  UTERATURE REVIEW / 9  Equating driving and drag forces for equilibrium conditions therefore yields  Vg(  P  g  - P ) - C Apv D  2 s  (  2  J  )  Rearranging equation 2.3 for v , the particle velocity, yields s v  g  =  /2(  p  -  g  C A D  )gV  P  (2.4)  P  If the particles are assumed to be roughly spherical in shape, then V = 7rd /6 and 3  A = 7rdV4. Substituting these assumptions into equation 2.4 yields Newton's Law v_  Where  =  / 4 g ( p  - p )d  d = particle diameter  The drag coefficient (C^) is a function of the particle shape and the flow regime surrounding  the  particle. Expressed as  a  number, the  Reynolds  number (N ) R  characterizes the flow conditions surrounding the particle (Equation 2.6).  N  Where  R  =  v  d  P  (2.6)  V = relative velocity between main body of fluid and particle  LITERATURE REVIEW / 10 d = effective dimension of the particle (sphere=diameter) p = fluid density v = dynamic viscosity of the fluid  The dynamic viscosity (M) of a fluid is a measure of it's resistance to tangential or shear stress. Expressed as Newton seconds per square meter (N s/m ), values are a J  function of temperature and range from 1.781 to 0.890 for 0 and 25 ° C respectively (Tchobanoglous, 1979).  When the Reynold's number is less than 2.0, viscous forces predominate and  C  = 24/N  D  (2-7)  R  As the Reynold's number increases through the range of 2-500, a transition zone occurs in which both inertia and viscous forces are effective. The drag coefficient (C^) is therefore represented by  C  Above  D  =  1 8 . 5 / N  a Reynold's number of 500,  R  ° '  viscous  (2.8)  6  forces are not significant  and the  coefficient of drag remains constant at 0.4.  Substituting equation 2.7 into equation 2.5 generates Stokes law for small, low velocity  UTERATURE REVIEW / 11 particles. Expressed as  v  =  ( P  " P )gd  s  2 (2  .9)  18>i  where  V = terminal velocity of particle P  g  = density of particle  p = density of fluid g = acceleration due to gravity M = dynamic viscosity of the fluid d = particle diameter  The waste solids generated within a fish farm or hatchery are generally small, low velocity particulates. It should therefore be expected that Stokes Law (Equation 2.9) can be used to characterize their settling velocities.  Assuming this relationship to be correct, and the density and viscosity of the fluid to be constant, the settling velocity of the particles will thus be a direct function of the diameter and density of the particle. Any force which effectively alters either of these two factors will therefore alter the settling velocity of that, particle.  The size and density of particulate solids discharged from a fish farm or hatchery has been shown to be influenced by a variety of physical and site specific factors. These include; the species and size of fish contained (Wheaton, 1977), the type (Stechey, 1986) and pellet size of feed used (Walden & Birkbeck, 1974), the retention time of the waste within the containment unit (Warrer-Hansen, 1982) and by the amount of  LITERATURE REVIEW / 12 physical agitation and resuspension of the waste (Muir, 1978; Warrer- Hansen, 1979).  Warrer-Hansen (1979) reported settling rates for intact trout excreta to be generally high and dependent on the weight of fish (Figure 2.1). Results of settling tests on fresh excreta showed that on average, excreta from 20 gram fish settled at a rate of 3.3 cm/sec, while wastes from 5 gram fish settled at a lower rate of 1.7 cm/sec. Wastes from fish less than 5 grams were shown to settle at a correspondingly lower velocity. However, wastes which had been retained within the culture unit exhibited much lower settling velocities (Querellou et al., 1982). This was illustrated by Walden & Birkbeck (1974) who reported that wastes discharged from a hatchery exhibited two major settling fractions. Over 50% of solid particulates had settling velocities faster than 0.2 cm/sec, while 40% had settling rates faster than 0.51 cm/sec. Both of these settling velocities are significantly lower than those reported for fresh excreta (Warrer-Hansen, 1979).  It is assumed, and generally found that unless biologically or physio-chemically acted upon, waste solids are dispersed and uniformly passed through the culture system. However, waste solids often collect in areas of low velocity. The length of time a waste solid may be retained in a holding unit will thus be a function of the tank configuration, the velocity profile of the system, the degree of resuspension by fish and mechanical aeration, and by the method and schedule of cleaning.  Once a particulate is trapped in the culture unit, it is subjected to a variety of physical and chemical processes which can alter the particles physical characteristics. Bacterial decomposition, hydration and physical shearing all act to reduce the particles  UTERATURE REVIEW / 13  FIGURE 2.1  Settling velocities for intact Rainbow trout excreta as a function of fish size (Warrer-Hansen, 1982)  1 C  1 0  2.0-  - 1.0  200  600  400 Size of trout I  20 10  S>0  2.5  800  1000  ( no / Kg )  I  1  2SI  1,67  weight of trout ( g)  1  L  1.25  1.0  LITERATURE REVIEW / 14 size and density, thus reducing the settleability of the particle.  Warrer-Hansen  (1982) summarized the  results  of  settling  trials on solid wastes  discharged from 3 tank configurations; an earthen pond (Warrer-Hansen, 1979), a circular tank (Warrer-Hansen, 1979), and a tank of unspecified configuration (Muir, 1978), (Figure 2.2).  Warrer-Hansen (1979) observed that the wastes discharged from an earthen pond, exhibited poor settling characteristics. This he concluded was due to the extremely long retention times of solids within the tank, associated with the earth pond design. Low velocities, long residence times of water within the pond and infrequent cleaning all act to trap and hold waste solids for long periods of time. The waste solids are thus subjected to extensive bacterial decomposition and hydration. Frequent resuspension by fish during feeding, grading and harvesting, act to break up the solids into smaller, less settleable fractions. It is therefore not suprising that removal rates of less than 50% are observed even after 30 minutes.  The second configuration reported by Warrer-Hansen (1979), the circular tank, appears to provide conditions more conducive to rapid solids removal, as removal rates of 70% were achieved after 30 minutes. Muir (1978) reports complete solids removal (100%) after 40 minutes in wastes discharged from the unspecified tank configuration. However, assuming that conditions under which the waste solids were produced were the same (feed type, feed pellet size, fish size), the circular tank would be more desirable.  Figure 2.2 illustrates two distinct settling patterns. In both tank configurations reported  LITERATURE REVIEW / 15  FIGURE 2.2  Relationship  between  (Warrer-Hansen. 1982)  residence  time  and  solids  removal  LITERATURE REVIEW / 16 on by Warrer- Hansen (1979), the bulk of solids removal occurred within the first 15 minutes of settling, after which time there was a gradual decline in the rate of removal with time. However, the curve reported by Muir (1978), shows that the bulk settling of settleable solids (>60%) occurs between 15 and 20 minutes after settling has begun. This would indicate that the wastes solids are relatively uniform in size and density, but lighter or smaller than wastes reported in Warrer-Hansen (1979). Although the circular tank configuration does not ultimately achieve the same level of removal (100%) as Muir (1978), it does achieve comparable levels of removal (below 70%) in significantly shorter residence times.  This is desirable for two reasons: 1) rapid solids removal reduces the residence time required to achieve a desired level of removal, which reduces the size of the gravity solids separation system and consequently the capital and operating costs of the system, 2) rapid solids settling reduces the need for totally quiescent conditions within the clarifier. This effectively reduces the potential for resuspension of the waste solids which may occur as a result of occassional turbulence within the treatment system.  Walden & Birkbeck (1974) reported that feed type and pellet size also influenced the settleability of the wastes produced. Although no specific data was introduced to quantify any differences between feed types, they did provide more specific results on pellet size. It was observed that wastes produced from fish fed 3/32 inch dry pellets, achieved 48-68% solids removal within the first 5 minutes, while wastes from fish fed a finer (grower crumbles) pellet size, took longer (15 minutes) to achieve similiar removals (55%).  LITERATURE REVIEW / 17 In a recent study involving settling column analysis of a raceway cleaning effluent from an Ontario trout farm using Martins ® dry pellet feed, Stechey (1986) reports that effluents, with initial solids concentrations of 300-500 ppm, achieved 50% solids removal within 20 minutes. However, 60 minutes was required to obtain 65% solids removal.  The mode and length of time between cleanings, and the method of solids handling has also been shown to influence settleability. In a report prepared by Hydroscience Inc.(1977) on wastewater treatment and control in Idaho hatcheries, it was observed that cleaning operations employing vacuums or pumps for handling wastes, produced wastes of poor settleability. It was further observed that if pumps were to be used, then diaphragm type pumps should be employed as they minimized solids break-up.  Given the wide range of factors which influence the physical nature of the wastes and their subsequent settleability, it is not surprising that there is wide variation in reported clarifier design and treatment efficiency. A number of overflow rates and residence times have been reported in the literature as providing treatment for fish culture wastes. Liao & Mayo (1974) report that an overflow rate of 49,000 1/mVday, with a retention time of 15-30  minutes was adequate for treatment Walden &  Birkbeck (1974) reported that a retention time of 8-15 minutes at an upflow rate of 6.8-9.2 mVmVday (converted from gal/ftVday) was sufficient to reduce suspended solids and BOD by 37-52% and 16% respectively. Sparrow (1981) reported that a 12 minute retention time in a 0.46m deep settling basin was sufficient to remove 90% of settleable solids. Bergheim and Selmer-Olsen (1978) suggested a medium retention period of 20 minutes would be sufficient to achieve satisfactory settling. Petit (1978)  LITERATURE REVIEW / 18 reported suspended solids removal of 92% after 30 minutes. Muir (1978) reported 66% of solids settled within 5 minutes while 90% settled within 15 minutes.  3. THEORY A clarifier or sedimentation basin is designed for a particular waste by selecting a suitable settling velocity (V ) such that all particles with settling velocities greater than o V  q  will be removed. The rate at which clarified water is produced can then be  expressed as (Tchobanoglous, 1979) Q = AV  where  (3.1)  Q  Q = flow rate (mVday) A = surface area of clarifier (m ) 2  V  o  = settling velocity of waste particle (m/day)  Rearranging equation 3.1 for V  q  (Equation 3.2) 3  V  q  = Q/A  = overflow  rate  (m /m  2 /day)  (32)  shows that the overflow rate or surface loading rate, is equivalent to the settling velocity (Figure 3.1).  From the geometry of Figure 3.1, it can be seen that if the area of the triangle, having legs H and L, represents 100% removal of particles, then the removal or particles having settling velocities less than V ( V ) will be in the ratio h/H. Given o s the depth a particle will settle is equal to the product of the settling velocity of that particle, and the retention time (t ), then the ratio at which particles with settling  19  THEORY / 20  FIGURE 3.1  Schematic of an ideal rectangular sedimentation basin, indicating the settling path of discrete particles  Surface area A-bL  Inlet  Inflow«Q  Out flow«Q o u t l e t weir  THEORY velocities less than V  will be removed will be  o  ,  V  t  V  ^=-L--^ _-i_  (3.3)  =  V  H  The  efficiency  of a  o  t  sedimentation  o  o  V  basin is a  measure of the removal  solid particulates at a given overflow rate V . From the discussion and Figure  3.1,  for a  / 21  of suspended observation of  clarification  rate of V , it has been shown that those particles o having settling velocities greater than V will be completely removed. The fraction of q  particles  removed  fraction  of  particle  with  removed efficiency percentage  from  particles a  the  with  suspension  settling  settling velocity  would  be  of the  equal  to  sedimentation  thus  velocities  less  the  will  than  ratio  be  less  than  V , it has o  of  1-C  o  .where  C  V . However,  for  o  been  shown  V /V . Therefore, when s o  basin for removing  represents the  o  each  that the  size  fraction  considering the  particles within this category, the  removal of these particles is given by (Clark et al, 1977)  f ° V  (3.4)  s  dc V  0 The of V  o  overall removal of suspended solids in a clarifier with a designed o  is thus determined  using equation 3.5. Expressed as a total percent removal  V Total  %  Removal  =  (  1 -  C  o  )  C  o  =  G s  +  o' where  overflow rate  V  • d  (3.5)  C  fraction of particles having settling velocities less than  V  o  THEORY / 22 Equation 3.5 is generally solved by integration of the area between 0 and a selected overflow rate (V ) on the particulates settling velocity distribution curve (Figure 3.2).  Figure 3.2 can be derived in two ways; 1) by selective sieve analysis and hydrometer tests in combination with Newton's Law (Equation 2.5), or 2) using settling column analysis of the waste. For the purpose of continuity of discussion, the application of method 2 will be described during the discussion of Type II settling and the derivation of the Type II or flocculant particle settling curve.  The overflow rate (V ) is subsequently chosen such that the desired percentage of suspended solids removal can be achieved. However, this design procedure is only effective if the waste solids conform to Type I or discrete particle settling only.  In Type I settling, each particle behaves as an individual entity and settles at a uniform rate determined by equations 2.1-2.9. The particle is said to follow a linear path, illustrated in Figure 3.3.  Chesness et al. (1975) observed that fish culture wastes exhibited two forms of settling behavior. In effluents of low suspended solids concentrations, like those produced during normal hatchery operations, Type I settling predominates. However, during cleaning and feeding  operations,  suspended  solids  concentrations  can increase  to  levels where  particulates begin to collide and coalesce, assuming the settling velocity of the new, larger or heavier particle. The particles exhibiting Type II or flocculant settling will thus follow a curvlinear path as illustrated in Figure 3.4.  THEORY / 23  FIGURE 3.2  Discrete particle velocity cumulative frequency distribution  THEORY / 24  THEORY / 25 The degree to which coalescence or agglomeration may occur will be due in part to the opportunity for contact between particles. This in turn is related to a number of factors, including; the suspended solids concentration, the range of particle sizes present in the waste, the distance the particles are to settle, and the velocity gradient within the system.  In Type I settling, the overall solids removal efficiency is a function of the overflow rate only and independent of the depth of the clarifier. However, in Type II settling, where particle interaction will be a function of depth and time, retention time and clarifier depth must also be considered when determining solids removal efficiency. Given the high level of particle interaction in Type II settling, the necessary design parameters can no longer be set on the basis of a simple mathematical relationship as in Type I settling. Overall suspended solids removal efficiency  can presently be  determined by empirical methods only. This is accomplished using a settling column equal  in height  to the  clarifier depth, and of at least 15 cm in diameter  (Tchobanoglous, 1979).  The settling column is used to approximate conditions within a clarifier in order to produce a graphical representation of the rate of solids removal as a function of time and depth. Effluent samples collected from the settling column are analyzed for non-filterable residue and subtracted from the initial non-filterable residue concentration in order to determine the level of removal of non-filterable residue over time. Expressed as a percentage of the initial concentration, the level of non-filterable residue removed, is plotted as a function of residence time and sample port depth. With the inclusion of iso-concentration curves, representing points of equal levels of  THEORY / 26 removal, the Type II orflocculantparticle settling curve is obtained (Figure 3.5).  The overall level of solids removal, within the column, at a specific residence time (T), is then determined by integrating the percent solids removal over the height of the column using equation 3.6. Overall solids removal  where  A  h n  (  R  %^ -  +  &h<W  n  +  2__ . ., , , . . ( A hl , + A hz0 + . . . ,A h n  +i)  " ^ V V W  2  A  + R n  l W  h  A  2 v  )  '  = average solids removal concentration within a section of  ~~2  the column depth h.  As an example to illustrate the use of equation 3.6, the overall non-filterable residue removal concentration will be determined from Figure 3.5 for a residence time of T . 2  For the example shown in Figure 3.5, equation 3.6 may be rewritten as  A  (90+80)  hJT ~~2  where  + Ah h  2  5  (80+70)  ~2  + Ah (70+60) h 2 3  5  hj = height of the seuling column  + Ah^(6(H50)  h  5  ~2  ( 3 ? )  THEORY / 27  FIGURE 3.5  Type  II or flocculant particle settling  curves (Note: Percentage  removal curves (R ) are interpolations of the solids fraction removed n as a function of residence time and column depth)  Time  (minutes)  THEORY  / 28  Results of the computations are summarized in Table I TABLE I  Summary  of integration  a residence time T h  R +R ,, n n+l  0.10 x (90+80) 2  =  8.5  0.10 x (80+70) 2  =  7.5  0.13 x (70+60) 2  =  8.45  0.67 x (60+50) 2  =  36.85  5  1.00 this  particle removal for  2  P e r c e n t Removal  h  Continuing  of flocculant  =  n  x  calculations  Total  integration  % Removal = 6 1 . 3  for each  unit  of  time,  thus  provides  a  graphical  representation of the level of solids removal as a function of residence time (Figure 3.6).  The solids  clarifier  removal  by selecting  curve  illustrated in Figure  the residence  time  required  3.6 is then  to achieve  used  in sizing the  the desired  %  solids  removal.  As  previously discussed, the settling column may also  be used to derive the individual  particle settling velocity distribution curve.  The equal  isc—concentration curves on the Type II settling curve represent lines or points of levels  of non-filterable  residue  maximum  trajectories  of particle  flocculant  suspension.  For example,  removal.  settling  paths  in Figure  However, for a  they  specific  also  represent the  concentration  3.5, 60% of the particles  will  in a have  THEORY / 29  FIGURE 3.6  100  Overall non-filterable residue removal curve  _  O UJ  a: UJ  •ZD  Q  CO UJ UJ _J CD <  OH  UJ  < UJ  > o  TIME [minutes]  THEORY / 30 velocities greater than h /T 2  at the time the particles reach a depth of h . Simplified,  2  2  the curve shows that for a particle to be removed from the column, it's settling velocity must be greater than the depth to be traversed (h ) divided by the allowed 2  time or residence time (T ). Therefore for the same example, the curve shows that 2  ( V ) and will therefore o be removed. However, within our discussion of discrete particle settling, it was shown 50% of the particles have settling velocities greater than h /T 5  that particles with settling velocities less than V  Q  2  ( V ) will be removed in the ratio of  V / V . The overall level of removal as a function of settling velocities is therefore so obtained by integration of the Type II settling curve using equation 3.5. Where 1 C  represents the percentage of solids with settling velocities greater than V  q  q  (h^/T^).  The second term within equation 3.5 therefore represents the fraction or percentage of particles with settling velocities V less than  which are removed.  The total fraction of particles which are removed with settling velocities greater than, equal too and less than h /T ( V ) is ^ t o o  (  1  Where  -  c  o  )  +  V  W  +  _ V W  +  ••• i < V i - V < > h  18  1 - C = fraction of particles removed at the maximum depth (h^) of the settling column given a residence time T :  0  h^ = total depth of settling column or clarifier h = n  depth within the column at vertical midpoint between particle  trajectory curves R and R , n n-1  In Figure 3.5 the fraction of particles removed at a residence time of T is expressed 2  THEORY / 31 as  50 + h (60-50) + M 7 0 - 6 0 ) + h (80-70) + hi (90-80) — — —± — h h h h 4  5  ?  5  5  (3.9)  5  However, this calculation has already been indirectly solved in the derivation of the overall non-filterable residue removal curve. The overall non-filterable residue removal curve is a graphical representation of the level of solids removal as a function of residence time. Restated, this curve is a representation of the percentage of particles with settling velocities greater than h^/T . Figure 3.6 may therefore be rearranged as the percentage of non-filterable residue with settling velocities greater than h / T t o (Figure 3.7).  The individual particle settling velocity distribution curve can therefore be obtained from Figure 3.7 by plotting the residual of 1 - (the fraction of non-filterable residue with settling velocities greater than h/T ), to obtain the fraction of non-filterable t o residue with settling velocities less than h / T (Figure 3.8).  THEORY / 32  PARTICLE SETTLING VELOCITY  FIGURE  3.8  Individual  p a r t i c l e settling v e l o c i t y  distribution  PARTICLE SETTLING VELOCITY  curve  4. MATERIALS AND METHODS  4.1. SPECIES Rainbow trout (Salmo gairdneri) were selected for use in this study as representative of the family of salmonids presently raised in British Columbia. The selection of the species was based primarily on two factors; 1) the entire life cycle from egg to market, occurs in fresh water, 2) Rainbow trout are grown extensively in the Lower Mainland for both the commercial and sports fishery markets. The species was therefore readily available at all times and in the sizes required. Rainbow trout were obtained from two sources; Fraser Valley Trout Hatchery (Ministry of Environment, Fish and Wildlife Branch) and Spring Valley Trout Farm.  4.2. FISH SIZE Uniform populations were chosen on the basis of weight Weight classes were selected from the feed manufacturers feeding guide to provide waste trials for the lower, middle and upper range of fish sizes recommended for each pellet size used. Table II summarizes the pellet sizes used, and the corresponding weight classes. The use of three weight classes for each pellet size used was chosen to provide a means of comparing the effects of weight on waste solid settling rates, independent of pellet size. Overlap of weight classes amongst pellet sizes, provided a means of comparing the effects of pellet size, on waste settling rates, independent of weight The weight sizes used in the study ranged from 11 to 311 grams.  34  MATERIALS AND METHODS / 35  TABLE II  Weight classes and corresponding feed pellet sizes used in waste trials  G r o u p //  Standard n Deviat ion  11.0  1.4  200  3/32  4.5  2  17.0  1.2  140  3/32  6 . 0  3  21.0  1.3  83  3/32  4.5  4  21.0  1.3  83  1/8  4.5  5  41.0  3.3  104  1/8  7.5  6  54.0  3.1  150  1/8  4.5  7  54.0  3. 9  150  5/32  4.5  8  72.0  2.9  100  5/32  7.5  9  93. 1  4.1  100  5/32  4.5  10  89 .0  2.5  89  3/16  4.5  23 . 7  25  3/16  5.0  Oregon  311.0  Moist  Pellet:  Moore-Clarke  Washington n  Feed Pellet Water Size(inches) * (°C)  1  11  *  Fish Size (grams)  Population  size  i n holding  tanks  Feed  Mills,  Laconner  MATERIALS AND METHODS / 36 4.3. FEED TYPE AND PELLET SIZE Oregon Moist Pellet ®, produced by Moore-Clark, was used in the feed-waste trials. The fish in each weight class were fed the feed pellet size, at a rate recommended by the manufacturers feeding guide. The feed was distributed on a daily basis over a 15 hour period in 8 equal portions by means of an automated type feeding system. The pellet sizes used in the feed-waste trials, as summarized in Table II, were the 3/32, 1/8, 5/32, and 3/16 inch pellet  4.4. LABORATORY SET UP  4.4.1.  Containment  Each test population was held in 200 litre (50 gallon) oval fibre glass tanks (Figure 4.1)  which were obtained  from the  Department of Fisheries  and Oceans West  Vancouver Marine Laboratory. Each tank was equipped with a 10 cm (4 inch) central standpipe drain which regulated the water level within the tank and reduced the discharge of settled waste solids.  4.4.2. Water Supply Each tank was provided with dechlorinated water through an over-head pipe. Overall flow rates were controlled at the main supply level by means of a constant head tower. Individual tank flow rates were maintained using a 1.3 cm (1/2 inch) ball valve. Water temperature ranged from 4 ° C to 7 ° C during the course of the study.  Flow rates for each test population were based on two factors; 1) the minimal flow  MATERIALS AND METHODS / 37  FIGURE 4.1  Schematic of 200 litre fibre glass tanks  MATERIALS AND METHODS / 38 rate required to prevent scouring of the tank bottom and loss of solids, and 2) the minimal flow rate required to maintain suitable water quality within the holding tanks.  4.4.3. Aeration  Aeration was provided at the main constant head tower, and at each 200 litre tank using a venturi aspirator attached to the 1.3 cm (1/2  inch) valve. Aeration was  maximized at the air/water interface of the water jet stream by adjusting the angle of the water stream from the aspirator to 60  0  (from the horizontal)(Chesness et al,  1973).  4.4.4. Stocking Density  Stocking densities were selected on the basis of; 1) the number of fish available within the size range desired, and 2) the number of fish in a particular size class required to produce a measurable quantity of waste solids.  4.5. SETTLING COLUMN  A plexiglass column, 230 cm in height by 15 cm in diameter was used in the settling rate trials. The column was equipped with 7 sampling portes, located, starting at 22 cm from the bottom of the column and then continuing up the column at 30.5 cm intervals (Figure 4.2 ).  Pressurized air was supplied to the base of the column to provide mixing.  MATERIALS AND METHODS / 39  FIGURE 4.2  Schematic of settling column used in settling trials  T  Dimensions i n centimeters (cm)  30.5  30.5 S e t t l i n g column i n s i d e diameteT" 15.0 cm 30.5  T 30.5 217  3 -+« 30.5  sampling port  2  30.5  22.0  230  MATERIALS AND METHODS / 40 4.6. EXPERIMENTAL PROCEDURE  Weight classes were acclimatized to the feed and pellet sizes for a period of three days prior to waste collection. This was carried out to insure that the wastes produced were of the pellets being fed, and to insure that the fish were feeding normally.  The tanks were thoroughly cleaned of any accumulated solids prior to waste solids collection. The solid wastes were then allowed to accumulate over a 24 hour period. During this 24 hour period, the fish were fed according to the program outlined in Section 4.3. At the end of the 24 hour period, the settled solids were collected from the tank and placed in the settling column. Collection of the settled waste solids from the bottom of the holding tanks was carried out using a modified vacuum siphon tube. A modified 2x2x2 cm (3/4x3/4x3/4 inch) T-fitting, attached to the end of the siphon hose produced a configuration similiar to the floor wand of a vacuum cleaner . This increased the functional area of the siphon and reduced turbulence. The large diameter of the tubing used in the siphon, and the low vacuum pressures used also prevented the fragmentation of the fecal pellets and unconsumed feed pellets.  To minimize the shear forces inherent in transfering the collected waste solids from the collection buckets to the settling column, the column was first partially filled to a depth of 180-200 cm with clarified effluent from the collection buckets. The settled solids were then added to the column. Rinse water from the collection buckets was used to bring the total column depth to 217-220 cm.  Mixing of the effluent within the column was accomplished using the compressed air system, located at the base of the column. The release of compressed air at the  MATERIALS AND METHODS / 41 bottom of the column produced conditions similiar to an air lift pump. The use of the air lift pump system for mixing did not appear to effect the particle sizes. Although care was taken, during the collection and manipulation of the waste solids, to minimize particle fragmentation, it must be assumed that in a hatchery or farm situation, fragmentation of fecal pellets and unconsumed feed pellets will occur. This is unavoidable as the solid particulates are subjected to shear forces during mechanical aeration and as the waste effluent cascades through a series of raceways or ponds.  Two sets of 100 ml samples were taken from each sampling port during mixing as a test for uniformity, and to provide a baseline concentration from which further samples would be compared. Mixing was terminated immediately after the last mixing sample was taken.  After termination of aeration, 100 ml samples were then collected from each sample port at 1, 3, 6, 15 and 30 minute intervals. A final 200 ml sample was taken from each port at 60 minutes.  Each sample taken from the column was then filtered through preweighed 5.5 cm Whatman ® glass fibre filters (934-AH). The filters and non-filterable residue were then dryed at 105 ° C for 24 hours.  After 24 hours the filters and residue were removed from the drying oven and placed in a desiccator to cool, after which time they were weighed and the amount of non-filterable residue determined.  MATERIALS AND METHODS / 42 The two sets of samples collected during mixing were pooled and an average taken in order to provide a baseline  or starting concentration from which the level of  non-filterable residue removed would be obtained. The level of non-filterable residue removed in  each sample  was  calculated as  the  difference  between this initial  non-filterable residue and the amount of non-filterable residue present on each of the sample filter papers. The level of non-filterable residue removed, expressed as a fraction or percentage of the baseline or starting concentration was then plotted as a function of residence time and sample port depth to provide the framework for the Flocculate particle or Type II settling curve illustrated in Figure 3.5. The Type II settling curves were then used in the derivation of the overall non-filterable residue removal curves and the individual particle settling  velocity distribution curves as  outlined in Chapter 3.  The  sampling procedure used in this study differed from the standard methods  reported in the literature, in that no makeup water was added to the column after a sample was taken. This modification required that a variable column height be used when calculating settling velocities from the Type II settling curve.  5. RESULTS AND DISCUSSION Non-filterable residue removal curves and individual particle settling velocity distribution curves are used to characterize a waste solids settling behavior. These curves were selected for use in this study as a means of evaluating and comparing the effects of Fish size and feed pellet size on the settleability of the waste solids produced. Although the study deals primarily with waste produced during the feeding of a moist pellet diet, similiar trials were carried out using a dry pellet feed (Purina Trout Chow ®), and will be discussed where appropriate.  In producing the non-filterable residue removal curves and individual particle settling velocity distribution curves to be presented in this study, a number of assumptions were made. Waste solids produced from the feeding trials were allowed to accumulate in the holding tanks over a 24 hour period prior to collection and analysis. During this time, the waste solids, composed of faeces and unconsumed feed pellets, were subjected to hydration and physical shearing. These two processes, combined with periodic resuspension of the  solid particulates, resulted in the breakup of solid  particulates and the loss of some solids. Due to the arrangement of the tank discharge, and the large volumes of water used, it was not possible to determine the quantity or sizes of the waste solids lost However, it was assumed that under actual operating conditions these same solids would be lost through resuspension. They would therefore form part of the average daily waste loading of the farm. As this study was concerned primarily with the treatment of the significantly higher strength cleaning flows (Liao, 1970; Bergheim et al, 1984), we were interested in only that fraction which was retained within the culture unit The waste solids lost through resuspension were therefore ignored.  43  RESULTS AND DISCUSSION / 44 Initially the study proposed to evaluate, in addition to fish size and feed pellet size, the effect of feed type on the physical characteristics of the waste produced. Two feeds were selected which represented alternate milling processes; 1) a frozen feed (Oregon Moist Pellet ®), produced by Moore-Clark and 2) a dry pellet (Purina Trout Chow ®) produced by Ralston Purina.  In feeding trials using the Oregon Moist Pellet ® feed, sufficient solids were retained within the tanks to produce the desired solids removal and particle settling velocity distribution curves. However, in trials using the dry pellet feed, of the three pellet sizes used, only the # 4 size produced any measurable quantity of waste solids. In trials using the # 3 and # 5 pellet sizes, most of the solids were lost Those solids which were retained, only after severely restricting water flow through the tanks, tended to form a gelatinous mat which adhered to the tank bottom and discharge screens.  It was also observed that in trials using the # 4 pellet, a large quantity of yellowish grit was found to collect in the tank. This may have been undigested corn particulates or part of the vitamin premix.  Due to this inability to obtain consistent measurable quantities of waste solids, it was decided to terminate further studies using the dry pellet feed.  Once the accumulated waste solids had been collected, they were placed in the settling column for analysis. The column simulates a portion of a clarifier or sedimentation basin. Sampling ports located along the side of the column allow for samples of the  RESULTS AND DISCUSSION / 45 effluent to be withdrawn for analysis.  The settling column can be used to derive the non-filterable residue removal and individual particle settling velocity distribution curves used in this study provided two conditions are met; 1) the waste solid particulates are uniformly distributed within the column, and 2) quiescent conditions exist in the column during sampling.  Waste solids  were kept in suspension  within the column using compressed air.  Compressed air was released at the bottom of the column and allowed to rise through the column, creating conditions similar to an air lift pump. The bubbles produced were kept large to reduce air flotation of the waste particulates. In this process, small bubbles adhere to  the  solid particles in suspension,  producing lift  The solids  subsequently float to the surface where they collect as a foam or scum. However, even with the large bubbles used some flotation of small solid particulates was observed. Once aeration was terminated , small low velocity waste solid particulates continued to rise, forming a scum at the surface of the column. No attempt was made to quantify the amount of solids removed by this process, as agitation of the layer resulted in it's breakup and settlement It is assumed however that in review of the high rate of settleable solids removal observed within the first 5-8 minutes (Figure 5.5 -  5.15), the air flotation removal of these low density particulates does not  contribute significantly to the overall rate of removal during this period. The inclusion of these particulates only slightly reduces the slope of the overall non-filterable residue removal curve during the first 5-8 minutes of settling.  As a test for uniformity of waste solids within the column, two sets of samples were  RESULTS AND DISCUSSION / 46 collected from each sampling port during mixing of the waste effluent From this set of samples it was observed that there was a high degree of variability in the concentration of suspended solids within the column (Table III). In general, suspended solids concentrations increased from the top of the column, down to the bottom. A general decline in the overall level of suspended solids within the column over time was also observed. Both of these observations indicated that some settling of solids was occurring during mixing. However, increasing the level of aeration to over come this settling, resulted in loss of the waste water out of the top of the column. It was decided therefore that some settling of solids was unavoidable.  For the purpose of this study, a mean of the twelve samples taken during mixing, as shown in Table  III,  was  used  as  the  initial  solids  concentration or baseline  concentration. However, using a mean value in the percent removal calculations to represent a column with varying suspended solids concentrations provided some error in the values obtained. At the top of the column, actual solids concentrations will be higher than the mean used, therefore the percentage of solids removal calculated from the samples taken from this region will be overstated. However, the opposite is true for the  lower portion of the  column. The use  of  a mean suspended solids  concentration will underestimate the actual percentage of solids removed.  Given that the non-filterable residue removal curves reflect the rate of removal within the entire column, small localized variations will tend to be buffered out To minimize the possible spread between the mean concentration used in our calculations and the actual initial solids concentration, sampling for the settling trials was begun immediately after the last mixing samples were taken.  RESULTS A N D  T A B L E III  DISCUSSION  / 47  Variability of initial non-filterable residue concentrations as a function of column depth  Trial Sample  Port*  1 2 3 4 5 6  Initial *  I  Trial  Sample 1 (grams)  Sample 2 (grams)  Sample I (grams)  Sample 2 (grams)  0.0631 0.0689 0.0583 0.0540 0.0550 0.0503  0.0582 0.0598 0.0474 0.0472 0.0455 0.0432  0.0629 0.0720 0.0665 0 . 065 1 0.065 1 0 . 0624  0 .0754 0.0654 0.0604 0.0620 0.0614 0 .0626  X = 0 .0542 S = 0 .0078 542 Concentration  R e f e r to location  II  F i g u r e 4.2 f o r  mg/l  settling  Note: N o n - f i l t e r a b l e r e s i d u e i n T r i a l sample volume o f 100 ml.  column  I and T r i a l  X = 0 .065 1 S = 0 .0045 651 sample  port  I I i s based on  mg/:  RESULTS AND DISCUSSION / 48 The second requirement in running settling trials in the column is that quiescent conditions be maintained within the column. The settling column was located in the fish culture lab, which due to the high volumes of water contained in the holding tanks, maintained an ambient temperature equal to the incoming water. The effluent placed in the column was therefore free of any external temperature gradients, and the convection currents which are produced when a gradient exists.  Some minor upwelling was observed at the start of the settling trials as a result of the air lift pump mixing system used. However, this was generally of short duration. The column was observed to be quiescent within 5 to 7 minutes after termination of aerated mixing.  Given that the waste solid particulates were uniformly distributed within the column, and that the column was quiescent, the settling column can be used to approximate conditions within a clarifier or settling basin. Samples collected from the column can thus be used to derive the Type II orflocculantparticle settling curves for the waste. The flocculant particle settling curve can in turn be used to obtain the overall non-filterable  residue  removal curve and the  individual  particle  settling  velocity  distribution curve for the waste.  As previously discussed, the two sets of samples collected during mixing were used to determine the baseline initial non-filterable residue concentration within the settling column (Table III). Samples collected after termination of mixing were then measured for non-filterable residue and compared to the initial concentration to determine the level  of  non-filterable  residue  removed  over  time  (Table IV). The level  of  RESULTS AND DISCUSSION / 49  TABLE IV  imple #  1  Sample P o r t Location *  Non-Filterable Residue (grams)  Z  Removal  Residence Time ( m i n )  0.0399 0 .0432 0.0381 0.0349 0.0410 0.0282  26.4 20.3 29 .7 35 .6 24.4 48.0  1 :00  1  1 2 3 4 5 6  2 2 2 2 2 2  1 2 3 4 5 6  0.0297 0.0305 0.0296 0.0262 0.0285 0.0209  45 .2 43.7 45 .4 51.7 47.4 61.4  3:00  3 3 3 3 3 3  1 2 3 4 5 6  0.0243 0.0242 0.0207 0.0168 0.0167 0.0121  55 .2 55 .3 61.8 69.0 69.2 77 . 7  6 :00  4 4 4 4 4 4  1 2 3 4 5 6  0.0143 0.0142 0.0135 0.0119 0.0108 0.0085  73.6 73.8 75 .1 78.0 80. 1 84.3  15 :00  5 5 5 5 5 5  1 2 3 4 5 6  0.0100 0.0114 0.0103 0.0094 0.0075 0.0062  81.5 79 .0 81.0 82 .7 86 .2 88.7  30 :00  6 6 6 6 6 6  1 2 3 4 5 6  0.0071 0.0074 0 .0077 0.0079 0.0070 0.0069  86 .9 86 .3 85 .8 85 .4 87 .1 87.3  60:00  1 1 1 1  *  Non-filterable residue concentrations for a sample settling trial  Refer to Figure column sampling  4.2 f o r l o c a t i o n port  of  settling  RESULTS AND DISCUSSION / 50 non-filterable residue (X ), expressed as a percentage, was then plotted as a function o of column depth and residence time in order to provide the frame work for the Type II settling curves (Figure 5.1). With the inclusion of iso-concentration lines, representing points of equal levels of non-filterable residue removal, the Type II or flocculant particle settling curve was obtained (Figure 5.2). The Type II settling curve (Figure 5.2) thus provided a graphical representation of the level  of solids removal over time at various depths within the settling column.  However, to be of use as a design tool, it was necessary to know what the overall level of solids removal was for the entire column. This was obtained from the Type II curve by integration of the various iso-concentration curves using equation 3.6 as outlined in Chapter 3. Plotting of the resulting overall non-filterable residue removal concentrations for each unit of time resulted in the overall non-filterable residue removal curve (Figure 5.3).  The individual particle settling velocity distribution curve was in turn obtained by plotting the residual of 1-X , where X is the level of overall non-filterable residue o o removed at a residence time T (expressed as a fraction), against the particle settling o velocity as outlined in Chapter 3. Figure 5.4 illustrates the resulting individual particle settling velocity distribution curve from the example data provided in Table IV. Overall non-filterable residue removal curves (Figure 5.5 - 5.15) and individual particle settling velocity distribution curves (Figure 5.16 - 5.26), were subsequently derived for each of the test groups summarized in Table II.  RESULTS AND  FIGURE 5.1  Percentage  of non-filterable  residue  removed  DISCUSSION / 51  as a  function of  residence time and column depth for data given in Table III  61.4 77. + + + |48.0  84.3 +  i.7 +  87.3  47.4 69.  80.1 +  86.2 +  87.1  78.0  82.7,  85.4  81.C  83.8  79. C  86.3  6*-|  5  "4 +  ,  *  + 124.4  +  +  33 H  CU w a 5=  s S3  o  4  i  +  5^1.7 _IS9.0  +  [35. &  cj  z •J  H H W cn  61.8  3*4 69.7  73.8 60. 3  .  45.2  55.2  H - 4 - 4 — 26.4  1 3  6  -  15  —4-  -4-  30  60  RESIDENCE TIME  *  Refer  to  Figure  86.9  81.5  73.6 - +  4.2  for  sample  (minutes)  port  location  and  depth  RESULTS AND DISCUSSION / 52  FIGURE 5.2  Type II or flocculant particle settling curves for example data in Table IV  26.4  13  6  15  30 RESIDENCE TIME ( m i n u t e s )  *  Refer  to Figure  4.2  f o r sample  port  location  and depth  RESULTS AND DISCUSSION / 53  FIGURE 5.3  Overall non-filterable residue removal curve for example data in Table IV  100-1  O UJ  • A"  aoH  . -A"  cc  Q  00 UJ  cc  /  UJ _J  m <  cc  UJ  /  60 A  40  O 20 A  <  cc  UJ  > O  10  20  30  40  ~r~ 50  TIME [minutes]  60  70  80  RESULTS AND DISCUSSION / 54  FIGURE 5.4  Individual particle settling velocity distribution curve for example data in Table IV  x  z <  1".  X  o.a-\  A/  A  /  /  /  / A  1  2  3  4  PARTICLE SETTLING VELOCITY [ c m / s e c ]  RESULTS AND DISCUSSION / 55  FIGURE 5.5  Overall. non-filterableresidue removal curves for 11 gram Rainbow trout fed 3/32 OMP feed pellets (NoterLegend indicates initial non-filterable residue concentrations)  >  o  00 Q  10 1  20 1  30 1  40 1  50 1  TIME [minutes]  60 1  70 r  80  RESULTS AND DISCUSSION / 56  FIGURE 5.6  Overall non-filterable residue removal curves for 17 gram Rainbow trout fed 3/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  100-1  v 0  i  1  10  20  1  1  1  1  1  30  40  50  60  70  TIME [minutes]  RESULTS AND DISCUSSION / 57  FIGURE 5.7  Overall non-filterable residue removal curves for 21 gram Rainbow trout fed 3/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  100-,  Legend O 1007 m g / l  -T—  40  TIME [minutes]  A  676 m g / l  X  179 m g / l  50  s  126 91  s  60  25 70  RESULTS AND DISCUSSION / 58  FIGURE 5.8  Overall non-filterable residue removal curves for 21 gram Rainbow trout fed 1/8 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  •A O  UJ OH UJ ZD  Q  CO UJ  OH  60-  CD <  OH 40-  Legend 20-  < OH UJ >  o  10  20  30  40  TIME [minutes]  O  903 m g / l  s  A  794 m g / l  s  X  584mo^l  s  50  = 135 =  60  9  .  A  8  5  70  RESULTS AND DISCUSSION / 59  •FIGURE 5.9  Overall non-filterable residue removal curves for 41 gram Rainbow trout fed 1/8 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  IOO-I  O UJ  80H  OH Q 00  UJ  OH  60 H  < OH y 4o-i  Legend  20 H < OH  UJ  > O  Op0  10  20  30  O 1755 m g / l  s  186  A 1148 m g / l  s  132  X  422 m g / l  s  39  •  329 m g / l  s  60  40  TIME [minutes]  50  60  -1  70  RESULTS AND DISCUSSION / 60  FIGURE 5.10  Overall non-filterable residue removal curves for 54 gram Rainbow trout fed 1/8 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  lOO-i  TIME [minutes]  RESULTS AND DISCUSSION / 61  FIGURE 5.11  Overall non-filterable residue removal curves for 54 gram Rainbow trout fed 5/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  100'  O  Z£  UJ  80-  cn UJ =>  Q  (/) UJ  cn  60-  UJ  _i  CD <  cn  40-  Legend O 792 mg/l  O  20-  <  cn  >  o  0 Q0  10  -~r— 20  30  A  744 mg/l  X  595 mg/l  •  262 mg/l  40  TIME [minutes]  50  13  g  126  g  120  s  83  s  60  70  RESULTS AND DISCUSSION / 62  FIGURE 5.12  Overall non-filterable residue removal curves for 73 gram Rainbow trout fed 5/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  lOO-i  80-  60-  40-  Legend  20-  0 Q0  O 792 m g / l  S =  105  A  693 m g / l  S =  127  X  657 m g / l  S  •  543 m g / l  S = 50  =148  i  10  20  30  40  TIME [minutes]  50  60  70  RESULTS AND DISCUSSION / 63  FIGURE 5.13  Overall non-filterableresidue removal curves for 93 gram Rainbow trout fed 5/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  100-1  O UJ  80H  ct: UJ  ZD Q  CO UJ CU UJ _l  CD < CU  Legend -z.  O  594 m g / l  s  150  o  A  503 m g / l  s  198  X  397 m g / l  s  397  •  332 m g / l  S  65  <  cn  UJ  > o  0 Q  0  1-  10  20  30  40  TIME [minutes]  50  60  70  RESULTS AND DISCUSSION / 64  FIGURE 5.14  Overall non-filterable residue removal curves for 89 gram Rainbowtrout fed .3/16 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  100-i  80H  60  H  40 H  Legend O 1254 mg/l  20 H  0 Q0  10  20  30  S = 257  A  889 mg/l  S - 77  X  286 mg/l  S - 48  40  TIME [minutes]  50  60  70  RESULTS AND DISCUSSION / 65  FIGURE 5.15  Overall non-filterable residue removal curves for 311 gram Rainbow trout fed 3/16 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  TIME [minutes]  RESULTS AND DISCUSSION / 66  FIGURE 5.16  Individual particle settling velocity distribution curves for 11 gram Rainbow trout fed 3/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  Legend  1  2  3  O  651 mg/l  A  542 mg/l  4  PARTICLE SETTLING VELOCITY [ c m / s e c ]  s  =  4  5  s = 78  RESULTS AND DISCUSSION / 67  FIGURE 5.17  Individual particle settling velocity distribution curves for 17 gram . Rainbow trout fed 3/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  x  i-i  < X to CO  o < Lu  0-j  0  1  1  1  1  1  1  2  3  4  5  PARTICLE SETTLING VELOCITY [ c m / s e c ]  RESULTS AND DISCUSSION / 68  FIGURE 5.18  Individual particle settling velocity distribution curves for 21 gram Rainbow trout fed 3/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  z <  X — I  (/) to  0.8-  o o  _J LU  >  0.6-  to X  0.4-  to  Legend  Q _I  o to  L_  0.2-  O Z  o o < 1  1~  2 2  3  O  1007 m g / l  s  =  126  A  676 m g / l  s  =  91  X  179 m g / l  s  =  25  4  PARTICLE SETTLING VELOCITY [ c m / s e c ]  RESULTS AND DISCUSSION / 69  FIGURE 5.19  Individual particle settling velocity distribution curves for 21 gram Rainbow trout fed 1/8 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  RESULTS AND DISCUSSION / 70  FIGURE 5.20  Individual particle settling velocity distribution curves for 41 gram Rainbow trout fed 1/8 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  Legend  1  2  3  O  1755 m g / l  s  186  A  1148 m g / l  s  132  X  422 m g / l  s  39  •  329 m g / l  s  60  4  PARTICLE SETTLING VELOCITY [ c m / s e c ]  RESULTS AND DISCUSSION / 71  FIGURE 5.21  Individual particle settling velocity distribution curves for 54 gram Rainbow trout fed 1/8 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  x < X  I—  OO CO UJ -J  >-  0.8-  o  _l  UJ  > o  0.6-  ui OO  0.4  Q CO  Legend  _i  o to  0.2  z o  — (  o <  1  2  3  O  1945 m g / l  s  278  A  1308 m g / l  s  216  X  616 m g / l  s  66  i  4  PARTICLE SETTLING VELOCITY [ c m / s e c ]  5  RESULTS AND DISCUSSION / 72  FIGURE 5.22  Individual  particle settling velocity distribution  curves for 54 gram  Rainbow trout fed 5/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  Legend O  792 m g / l  s  13  A  744- m g / l  s  126  X  595 m g / l  s  120  •  262 m g / l  s  83  "T"  1  2  3  4  PARTICLE SETTLING VELOCITY [ c m / s e c ]  n  5  RESULTS AND DISCUSSION / 73  FIGURE 5.23  Individual particle settling velocity distribution curves for 73 gram Rainbow trout fed 5/32 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  RESULTS AND DISCUSSION / 74  FIGURE 5.24  Individual particle settling velocity distribution curves for 93 gram Rainbow trout fed 5/32 OMP feed pellets (Note: Legend indicates initial non-filterableresidue concentrations)  Legend 594 mg/l  s  150  A 503 mg/l  s  198  X 397 mg/l  s  397  •  s  65  O  332 mg/l  1 2 3 4 PARTICLE SETTLING VELOCITY [ c m / s e c ]  RESULTS AND DISCUSSION / 75  FIGURE 5.25  Individual particle settling velocity distribution curves for 89 gram Rainbow trout fed 3/16 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  Legend mg/l  S  257  A  889 mg/l  s  77  X  286 mg/l  s  48  O  1  2  3  1254  4  PARTICLE SETTLING VELOCITY [ c m / s e c ]  RESULTS AND DISCUSSION / 76  FIGURE 5.26  Individual particle settling velocity distribution curves for 311 gram Rainbow trout fed 3/16 OMP feed pellets (Note: Legend indicates initial non-filterable residue concentrations)  RESULTS AND DISCUSSION / 77 Comparison of the settling curves within a feed pellet size and between feed pellet sizes should therefore identify any differences in the settling behavior of the wastes, related to fish size and feed pellet size respectively. non-filterable  residue  removal curve  is  not  ideally  However, the form of the suited  to  analysis.  Statistical  procedures for handling complex non-linear functions are often complicated and more difficult than for handling linear relationships. In some situations, however, it may be possible to transform the x, and/or y variables in such a way that the resulting function is close to being linear. A linear regression model can then be formulated in terms of the transformed variables, and the appropriate analysis can be based on the transformed data (Bhattacharyya & Johnson, 1977).  The data from the settling trials was therefore transformed in an effort to obtain an approximate linear function. Two transformations were tested; LogX,Y and X.X/Y, a linearization of the hyperbolic relationship  Y  Where  =  ( AX  ) /( B +  X )  Y  percentage of non-filterable residue removed  X  residence time (minutes)  A.B  (5.1)  constants  Equation 5.1 is then linearized using the relationship  Z = b  o  + b X  1  (5.2)  RESULTS AND DISCUSSION / 78 Where  Z b o b. 1  X/Y  (5.2.1)  B/A  (5.2.2)  1/A  (5.2.3)  Both transformations produced linear regressions with high i Determination) values, of 0.87 - 0.99 and 0.9996 -  2  (Sample Coefficient of  1.0000 respectively. However, the  use of the LogX,Y transformation to fit the severe curve in the data is only valid at short residence times. The relationship obtained in this transformation assumes that as the residence time increases, so will the level of solids removal. Removal rates of greater than 100% are thus possible at long residence times. This is clearly impossible. However,  in  the  hyperbolic  relationship,  as  the  residence  time  increases  the  non-filterable residue removal curve increases to an asymptote or maximum level which is less than 100%. This more accurately represents the situation occurring in the level of removal of non-filterable residue as a function of time. The hyperbolic function and it's linearized transformation was subsequently selected for this study.  It is important to note that in cases where the range in Y values is small, the r  2  can be misleading. When Y is small, the transformation X/Y is similiar to dividing X by a constant (k). When X/k is plotted as a function of X it will naturally produce a perfect fit linear regression. Therefore the transformation should only be used when the variation in Y values is large. Within the data obtained in the settling trials, the level of removal (Y) varies from 24 - 93% during the 60 minute retention time (X). This should be sufficiently  large to permit the use of the X.X/Y transformation.  Appendix I summarizes the hyperbolic and the corresponding linear regression equations for each of the overall non-filterable residue removal curves illustrated in Figures 5.5  RESULTS AND DISCUSSION / 79 - 5.15.  Differences in the overall non-filterable residue removal curves could now be identified by comparing the linear regression equations for each X.X/Y transformed data group.  Regression equations may differ for two reasons; 1) they have different slopes, that is they are not parallel , or 2) if. they are parallel, they may differ in level; that is their intercepts are not equal . Two hypothesis that must be considered and tested are therefore: 1) Slopes of linear functions are the same, slopes are parallel 2) Intercepts of linear functions are the same, intercepts are equal  One computer program suitable for performing these two tests is the Equality of Slope Test program S:SLTEST. SLTEST is a self-contained program, available through the public file S: SLTEST. A copy of the program description may be obtained through the computing, center.  Simply, SLTEST computes the linear regression equation for each set of data and tests whether the differences  in the regression coefficients  between groups are due to  sampling error, or may be attributed to differences between groups. The program uses the F-test to check the hypothesis that each regression coefficient is identical among all groups. Depending on the result of the F-test, the program will do one of two things: 1) If the hypothesis is rejected, that is, the regression coefficients  are different,  SLTEST will: a) find for each pair of equations the first regression coefficient in  RESULTS AND DISCUSSION / 80 which the above equations differ, or b) find, for each pair of equations, all the regression coefficients in which the above equations differ. 2) If the hypothesis is not rejected, SLTEST will test the hypothesis that a common equation can be used for all samples; that is, SLTEST will test if the intercepts with the y-axis are equal.  SLTEST was subsequently used to test for similarities in the regression coefficients and y-intercepts of the regression equations representing the overall non-filterable residue removal curves.  As part of the experimental design, each waste group, identified by a weight class and feed pellet size, was replicated. The replications of the settling trials were to be used as a check for uniformity and reproducibility of the Type II settling curves. However, due to the highly variable hydraulic conditions within the tank, the amount of accumulated solids present after 24 hours was never constant The initial solids concentration in each settling, trial was therefore different.  The degree of Type II or flocculant particle settling in a system, has been shown to be a function of the suspended solids concentration, the particle size distribution, the velocity gradient within the system, and the distance the particles are to settle. Changing the initial solids concentration, while maintaining the latter three factors constant, should therefore affect the level of Type II settling present and subsequently the rate of solids removal. We may therefore expect some differences in the rates of solids removal between replications with different initial solids concentrations.  RESULTS AND DISCUSSION / 81 Therefore, before comparisons could be made between waste groups it was first necessary to determine if variations in the initial non-filterable residue concentrations were significant The first hypothesis to be tested therefore was: 1) a single regression equation can be used to represent the replicated settling curves in each group.  If  the hypothesis is accepted, then the replicate curves in each group can be  represented by a single equation. The variation observed in the solids removal curves within the group are therefore not significant Analysis of the groups is simplified through the use of the single regression equation. However, if the hypothesis is rejected, indicating that the variability observed in the inital non-filterable residue concentrations is significant, then the analysis of the settling characteristics must also include the effects of concentration. Further analysis between groups would have to take initial solids concentration into consideration.  The linear regression equations summarized in Appendix I were subsequently compared using. SLTEST at confidence levels of 0.05 and 0.01 (95% and. 99%),, to determine, if the regression equations within each group could be pooled and represented by a single, common regression equation.  Results of SLTEST at both confidence levels indicated that a common slope of 0.011 was present in all groups. However, common regression equations for each group were not possible due to statistically significant variations in the y-intercepts (0.0116 0.0369).  A rejection of the hypothesis on this basis would appear to indicate that within the  RESULTS AND DISCUSSION / 82 range of initial concentrations observed in the settling trials (179 - 1945 mg/l), varying the concentration did statistically affect the settling rate of the waste solids. This is not surprising given the influence of concentration on the degree of flocculant settling present in a suspension. At higher concentrations the opportunity for particle contact is greater. Large, high velocity particulates can collide and coalesce with smaller less settlable  fractions,  thereby  facilitating  their  removal.  The  rate  of  removal of  non-filterable residue would subsequently be higher in the beginning, declining over time  as  the  concentration  is  reduced. A high  correlation between  the  initial  concentration and the level of the y-intercept might therefore be expected. However, when the y-intercepts and the initial concentrations ate compared within waste groups and between groups (Figure 5.27), no correlation is evident (r =0.077). Results of the 2  regression indicate approximately 8% of the variability in the y-intercepts is explained by the concentration effect  The low correlation observed between the level of the y-intercept and the initial concentration may be an indication that; 1) the range, of. concentrations observed in the settling trials may be below the level at which Type II settling occurs, or 2) if Type II settling is occurring at higher concentrations, it's contribution to the overall rate of removal is insignificant. In either case it must be assumed that the variation observed in the initial concentrations are not responsible for the variations in the y-intercepts.  In comparing the replications within each waste group, it has been assumed that, given the feed pellet size and fish size were kept constant, wastes in each replication should also be relatively uniform. However, it is impossible to control the amount of fragmentation  of  faeces  and  unconsumed  feed  pellets  resulting  from  repeated  RESULTS AND DISCUSSION / 83  FIGURE 5.27  Variability of the y-intercept of the linearized overall non-filterable residue removal curves as a function of the initial non-filterable residue concentration  oo 0.5-. ui 1  .•> or r> o  _J  o 3  0.4H  UJ  or Ul  ZD  a to  0.3H  CG  < or ui  =1 0.2  OO Q  o  y  o.H o  or ui 0.0  500  1000  1500  2000  2500  INITIAL N O N - F I L T E R A B L E RESIDUE REMOVAL CONCENTRATION [mg/i  RESULTS A N D resuspension might  and massication  therefore  y-intercepts.  be  by fish. Some  expected.  This  could  variation account  DISCUSSION / 84  in the range of particle for variations  observed  a lower  common  proportion  of the same  size  particles. However,  slope within the settling trials indicates  pattern of the waste solids. Given as  discrete  in the  Wastes with a higher proportion of larger, faster settling particulates would  show a high level of removal within the early stages of settling, compared with  sizes  entities, the proportional  to a waste  the presence  of a  there is a similarity in the settling  the assumption that the waste solids are behaving distribution  of particle  sizes  and their respective  settling velocities should therefore be similar.  One  method  of determining  equations of the individual  the validity of this assumption particle settling velocity  is to test the regression  distribution  curves for a common  regression equation.  The  individual  particle settling velocity  using the relationship SLTEST  distribution  curves were  subsequently  linearized  in equations 5.1 - 5.2.3 and compared using SLTEST. Results of  at a confidence level  of 0.01 indicated  that the pooled  regressions can be  represented by a single, common equation (Equation 5.3).  Y - 92.59X/(2.01 + X)  Where  The  (5.3)  X  =  settfing velocity of particles (cm/sec)  Y  =  fraction of particles with settling velocities less than X  proportional  distribution  of particle sizes  within  waste groups and between  waste  RESULTS AND DISCUSSION / 85 groups must therefore be considered identical.  Analysis of the linearized overall non-filterable  residue removal curves and the  subsequent determination of the source of the variations in the y-intercepts, has been based on the premise that the difference in the y-intercepts is significant. It was therefore  concluded that the  significantly  different  settling  characteristics of the  wastes must also be  However, the submission of a common regression equation,  characterizing the pooled linearized individual particle settling velocity distribution curves, indicates that the proportional distribution of particle sizes within the wastes is identical. It should therefore  be assumed that if all extraneous  factors such as  turbulence , short circuiting and upwelling within the column are kept constant, wastes from different size fish or different size feed pellets should behave in a similar manner. The variability observed within the y-intercepts must therefore be considered the result of factors outside of those controlled in the settling trials.  Evaluation of the results of the SLTEST has also been based on the premise- that given no common regression equation is possible due to statistically significant variations in the y-intercepts, that these variations should be considered important when designing a sedimentation basin. The variations observed in the settling behavior of the waste solids might be considered significant, assuming a clarifier could be ideally designed with no turbulence or short circuiting. However, given the nature of the design process in which the settling curves are to be used, and the general inclusion of a large safety factor (Warrer-Hansen, 1982), any differences  observed between the settling  curves could be ignored for design purposes. It may therefore be assumed that within the retention time (24 hours), fish sizes (llg-311g), feed type (OMP), and feed pellet  RESULTS AND DISCUSSION / 86 sizes (3/32-3/16) used, there is no significant difference in the settling behavior of the waste solids produced.  A single curve can therefore be used to characterize the settling behavior of the waste solids produced in the study. The individual non-filterable residue removal percentiles and individual particle settling velocity distribution fractions were subsequently pooled in order to produce the single overall non-filterable residue removal curve (Figure 5.28) and the individual particle settling velocity distribution curve (Figure 5.29).  Analysis of Figure 5.28 shows that a residence time of 10-15 minutes is sufficient to reduce the non-filterable residue by 82%. However, the curve also shows that holding the effluent longer than 15 minutes, does not provide any appreciable increase in solids removal. Increasing the residence time to 30 minutes, effectively doubling the retention time, only increased the level of solids removal an additional 4 %. While holding the effluent for 60 minutes produced a total increase of only 7%. Therefore, selecting, a. retention, time greater than 15. minutes does not effectively increase the level of solids removal. Selecting a retention time greater than 15 minutes also increases, unnecessarily the size of the system, and consequently, the cost  ln a summary of capital, operating and maintenance costs as a function of surface area (Figure 5.30), Underwood McLellan (1979) show that for small sedimentation basins (<1000 m ), these three costs increase linearly with increasing surface area. 2  However, as the surface area of the sedimentation basin increases (>1000 m ), capital, 2  operating and maintenance costs begin to increase at a faster rate. It is therefore desirable to select a residence time which minimizes the size of the treatment system,  RESULTS AND DISCUSSION / 87  RESULTS AND DISCUSSION / 88  FIGURE 5.29  Individual particle settling velocity distribution curve for Rainbow trout fed an Oregon Moist Pellet diet  RESULTS AND DISCUSSION / 89  FIGURE 5.30  Summary  of  sedimentation  capital,  operating  and  basin as a function  maintenance  of surface  costs  for a  area (Underwood  McLellan, 1979)  1,000,000  in or <  O  100,000,  in O o  10,000.  1,000  Jt  .3  .4 .5 .6 .7 J J I  SURFACE  2  3  4  AREA  3 6 78  9|0  ( m x 2  2  100)  3  4  3 6 7*  »100  RESULTS AND DISCUSSION / 90 while still achieving the desired level of solids removal.  Holding the effluent in contact with the settled solids for long periods is also undesirable as decomposition and leaching of the organic solids reduces the level of dissolved oxygen (0 ) 2  and increases the level of dissolved nutrients in the clarified  effluent  Figure 5.28, illustrating the distribution of individual particle settling velocities within the waste solids, shows the solids particulates settled at significantly higher rates than reported by Walden & Birkbeck (1977) or Stechey (1986). Walden & Birkbeck (1977) report that the waste solids exhibited two main settling fractions. Over 50% of the solids settled faster than 0.2 cm/sec, while 40% had settling rates faster than 0.51 cm/sec. Using these velocities for comparison, the waste solids generated from feeding the Oregon Moist Pellet ® feed, showed over 80% of the solids had settling rates faster than 0.2 cm/sec, while 72% had settling rates faster than 0.51 cm/sec.  The purpose of this study was; 1) to characterize the settling behavior of waste solids from a trout culture unit as a function of fish size and feed pellet size, and 2) relate the results to current gravity solids separation system design proceedures.  Using the overall non-filterable residue removal curves, and individual particle settling velocity distribution curves, it was possible to show that solid wastes retained for 24 hours exhibited uniform settling behavior. Differences in settling rates of fresh excreta, reported as a function of fish size (Wheaton, 1977; Warrer-Hansen, 1979) or pellet size (Walden & Birkbeck, 1974) were negated by physical and chemical processes  RESULTS AND DISCUSSION / 91 occurring in the holding tank.  The curves produced in this study provide the basis for the selection of a suitable retention time and depth for a gravity solids separation system to treat the cleaning flows from a fish farm or hatchery employing a 24 hour cleaning cycle.  Appendix II summarizes the use of the overall non-filterable residue removal curve and individual particle settling velocity distribution curve for sizing a gravity solids separation system for a fish farm producing 50-75 tonnes of fish a year, utilizing approximately 500 liters of water per second.  6. CONCLUSION The  findings of this study can be summarized as follows:  II settling  curves were  individual  particle  used  settling  to derive the velocity  overall non-filterable  distribution  curves  gairdneri) culture cleaning wastes. Analysis of the Test  (S:SLTEST),  showed  that  there  Flocculant particle or Type  was  no  for  residue removal and  Rainbow  curves using the  significant  difference  Trout {Salmo Equality of Slope in  the  settling  behavior of the particulate solids when wastes from different size fish or different size feed pellets were compared. A single curve was therefore used to characterize each of the  non-filterable  residue removal curves (Figure 5.28)  and individual  particle settling  velocity distribution curves (figure 5.29) derived in the study.  The  overall non-filterable  solids  removal  residue removal curve showed  (>80%) occurs  waste longer than 15 minutes  within  the  first  15  that the  minutes  did not contribute significantly  of  bulk of the waste  settling.  Holding  the  to the overall level  of  solids removal.  The  individual particle settling velocity distribution curve showed the solids produced in  the study using the Oregon Moist Pellet ® feed settled at a faster rate than reported in the literature for non-fresh excreta, but slower than for fresh excreta.  Differences due to fish size and feed pellet size in the settling behavior of the waste solids,  reported  in  the  literature,  were  not  retained within the culture tank for 24 hours.  92  apparent  when  the  waste  solids  were  7. RECOMMENDATIONS FOR FUTURE WORK  7.1. RESIDENCE TIME Muir (1978) and Warrer-Hansen (1979;1982) have shown that the length of time the waste solids are retained within the culture unit can significantly reduce the settleability of those solids. In our study, we selected a retention time of 24 hours as being fairly representative of the length of time waste solids would be retained in a commercial fish farm or government hatchery employing a daily cleaning schedule of the ponds or raceways. The overall solids removal and individual particle settling velocity distributions were  subsequently  higher  than  generally  reported  in  the  literature  where  longer  retention times are involved. One area for further work would therefore be to quantify the effects of retention time on the settleability  of waste solids. The results to be  used as a management tool in scheduling tank cleaning intervals for optimum solids removal efficiency within the gravity solids separation system.  A. further, area of research, associated with residence time, is the amount of physical shearing, hydration and resuspension of the waste solids within the culture units. These parameters all act to reduce the particle size and density of the waste solids. However, the  magnitude  of these processes will  be  a function of the flow rate, hydraulic  characteristics of the tank and the stocking density, all of which are site specific.  The  waste solids  collected  in our study were subjected  to physical agitation and  resuspension as a result of the high stocking densities used in the small 200 litre tanks. However, it is difficult to judge if these conditions would be comparable to those found in the larger production units. In some raceways and circular ponds, dead  93  RECOMMENDATIONS FOR FUTURE WORK / 94 zones provide a collection point for waste solids, which may then undergo little physical shearing. While some solids may be subjected to extreme physical agitation as the effluent cascades from one raceway to another, or is circulated by a mechanical aeration unit Given that these are all site specific factors, one recommendation is to compile a series of settling curves for a variety of farms and hatcheries, to determine if the waste solids can be characterized by the holding system used (eg. raceway, Burrows pond, circular pond etc.).  7.2.  FEED TYPE  This research has shown that the waste solids produced from Rainbow trout fed the Oregon Moist Pellet ® diet, settle at a higher rate than reported in the literature for dry pellet feeds (Walden & Birkbeck, 1974; Stechey, 1986).  An attempt was made to compare the OMP feed with a dry pellet feed (Purina Trout Chow ®). However, the hydraulic conditions within the holding tank units used, were not suitable and. very little waste solids were retained. It is therefore recommended that a different holding tank be used which will retain the solid wastes, or use cleaning wastes from farms or hatcheries using the dry feed, in the settling trials.  Similiar settling trials could also be carried out on other pellet feeds such as EWOS ®, Rangen ®, and BIO-Dyne ®. Other feed types which might be investigated are trash fish, and the wet feeds which are produced on site at the farms, now being used extensively in the salmon farming industry.  REFERENCES Bergheim, A. & A.R.,Selmer-01sen 1978. "River Pollution from a Large Trout Farm in Norway". Aquaculture (14):267-270 Bergheim, A. & A.,Silversten 1981. "Oxygen Consuming Properties of Effluents from Fish Farms". Aquaculture (22): 185-187 Bergheim, A.,H.,Hustveit, A.,Kittelsen, & A.R.,Selmer-01sen  1984. "Estimated Pollution  Loadings from Norwegian Fish Farms II: Investigations 1980-81". Aquaculture (36): 157-168 Bhattacharyya, G.K., & R.A., Johnson 1977. "Statistical Concepts and Methods". Wiley Series in Probability and Mathematical Statistics., John Wiley & Sons, New York. Bodien, D.G. 1970. "An Evaluation of Salmonid Hatchery Wastes". U.S. Dept Inter. Fed. Water Qual. Adm. Portland, Oregon PB 215496, 53pp. Brisbin, K.J. 1970. "Report on the Pollutional Aspects of Trout Hatcheries in British Columbia". Prepared for Fish and Wildlife Branch, Dept. Recreation Cons., Underwood McLellan & Associates. Brockway,  D.R. 1950.  "Metabolic  Products  and Their Effects".  Prog. Fish-Cult  12(3): 127-129. Brown, CM. & C.E.,Nash 1979. "Aquaculture as a Method for Meeting Hatchery Discharge  Standards".  Presented  at  Bio-Engineering  Symposium,  Traverse City, Michigan. October, 1979. Butz, I. & B.,Vens-Cappell 1982. "Organic Load from the Metabolic Products of Rainbow Trout Fed with Dry Food". In: Report of the EIFAC Workshop on Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Chesness, J.L., L.J.,Fussel & T.K.,Hill  1973. "Mechanical Efficiency  Aerator". Transactions of the ASAE 16(1):67-8,71.  95  of a Nozzel  / 96  Chesness, J.L, W.H.,Poole & T.K.,Hill 1975. "Settling Basin Design for Raceway Fish Production Systems". Transactions of the ASAE:Paper No. 74-5005. pp326-331. Clark, J.W., V.,Warren & M.J.,Hammer 1977. "Water Supply and Pollution Control". Third Edition. Harper & Row, New York, pp 351-359. DFO, unpublished. Department of Fisheries and Oceans internal survey of Federal hatcheries.  Environmental assessment of  down  stream  effects of  hatchery effluents (1978-1981). EIFAC (European Inland Fisheries Advisory Council), 1982.  "Report of the EIFAC  Workshop on Fish-Farm Effluents". Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Faure,  A.  1977.  "Mise au point  sur la  pollution  engendreE  les  pisciculture".  Piscic.Fr.(50):33-35.In: Report of the EIFAC Workshop on Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Gunther, D.C.,D.E.,Brune & G.A.E.,Gall 1981. "Ammonia Production and Removal in a Trout Rearing Facility". Transactions of the ASAE. ppl376-1380,1385. Huber, R.T. & J.J.,Valentine 1978. "Analysis and treatment of Fish Hatchery Effluents". Report from Lamar National Fish Hatchery Development Center, Lamar Pennsylvania. Hydroscience Inc.1977. In: Underwood McLellan, 1979. "Salmon Rearing Facilities,Waste Water Study, Treatment Technology".  Underwood McLellan (1977)  Ltd., Consulting Engineers and Planners, Vancouver, B.C. Klontz, G.W. & J.G.,King 1975. "Aquaculture in Idaho and Nation Wide". Idaho Dept. of Water Resources Research Technical Completion report. IDWR Project 45-080, 86pp. Klontz, G.W.,I.R.,Brock & J.A.,MacNair 1978. "Aquaculture Techniques Water Use and Discharge Quality". Moscow, Idaho. University of Idaho, Office of  / 97 Water Research and Technology. Project A.054-1DA, 88pp. Knbsche, R. & E.Tscheu 1974. "Some Initial Aquaculture Experiments in Recirculating Water Systems". Aquaculture (ll):323-328. Liao, P.B. 1970.  "Pollution Potential of Salmonid Fish Hatcheries". Water Sewage Works 117(8):291-7.  Liao, P.B. 1970a. "Salmonid Hatchery Wastewater Treatment". Water Sewage Works 117(2):434-43. Liao, P.B. 1971. "Water Requirements of Salmonids". Prog.Fish-Cult 33:210-15. Liao, P.B. & R.,Mayo 1972.  "Salmonid Hatchery Water Reuse systems". Aquaculture  (l):317-335. Liao, P.B. & R.,Mayo 1974. "Intensified Fish Culture Combining Water Reconditioning with Pollution Abatement". Aquaculture (3):61—85. Liao, P.B., R.D.,Mayo & W.,Williams 1972.  "A Study for Development  of Fish  Hatchery Water Treatment System". Report Prepared for the Walla Walla  District  Corps,  of  Engineers.  Kramer,  Chin  & Mayo,  Consulting Engineers; Seattle, Washington. Mantle, G.J. 1982. "Biological and Chemical Changes Associated With the Discharge of Fish Farm Effluents".  In: Report of the  EIFAC Workshop on  Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Mayo, R.D. & P.B.,Liao  1969. "A  Study  of  the  Pollutional  Effects  of Salmonid  Hatcheries". A Report Prepared for the States of Wisconsin, New York and Michigan., Kramer, Chin & Mayo, Consulting Engineers, Seattle, Washington. Muir, J.F. 1978. "Aspects of Water Treatment and Reuse in Intensified Fish Culture". Ph.D. Thesis. Glasgow, Scotland. University of Strathclyde, 451pp.  / 98 Muir, J.F. 1982. "Economic Aspects of Waste Treatment in Fish Culture". In: Report of the EIFAC Workshop on Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Muir, J.F. & R.I.,Lipper 1970. From: Muir, J.F. 1982. "Economic Aspects of Waste Treatment in Fish Culture". Parker, N.C. & M.C.,Broussard 1977. "A Selected Bibliography of Water Reuse Systems for Aquaculture". Texas Agric. Exp. StaL 33pp. Petit, J. 1978. "Treatment Experiments of the Water Discharged by a Fish Farm". Technical Report of the Mission in UK. Jony en Josas, INR A, 30pp. Querellou, J.,A.,Faure & C.,Faure 1982. "Pollution Loads from Rainbow Trout Farms in  Brittany, France". In: Report of the EIFAC  Workshop on  Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Samis, S. 1983 Personal Communication. Scherb,  K.  &  F.,Braun  1971.  "Erfahrungen  mit  der  Regenbogenforellen  bei biologischer Reinigung  belebten  Schlamm  Wasser-Abwasser-Forsch".(4):118-24  Intensivhaltung  des Abwassers mit  im In:  von  Kreislaufsystem, Report  of  the  EIFAC  Workshop on Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Solbe, J.F. de L.G. 1982. "Fish Farm Effluents: A United Kingdom Survey". In: Report of the EIFAC Workshop on Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Solberg,  S.  &  F.,Bregnballe  1977.  "Dambrug  som  forurenigskilde".  ForsogsdambrugeL (58) In: Report of the EIFAC  Medd.  Workshop on  Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Solberg, SO. & F.,Bregnballe 1982. "Pollution from Farmed Trout Fed with Minced  / 99 Trash Fish". In: Report of the EIFAC Workshop on Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Sparrow, R.A.H. 1981. '"Hatchery Effluent Water Quality in B.C.". In:Bio-Engineering Symposium for fish culture. FCS Publ. 1:162-166. Speece, R.C. 1973. "Trout Metabolism Characteristics and the Rational Design of Nitrification Facilities for Water Reuse in Hatcheries". Trans. Am. Fish. Soc. 102:323-334. Stechey, D. 1986. Personal Communication Sumari, O. 1982. "A Report on Fish-Farm Effluents in Finland". In: Report of the EIFAC Workshop on Fish-Farm Effluents. Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Tchobanoglous,  G. 1979. "Wastewater Engineering: Treatment,Disposal,Reuse".  Second  Edition. Metcalf & Eddy, Inc. McGraw Hill Book Company. Underwood McLellan & Associates Ltd 1970. "Report on Pollutional Aspects of Trout Hatcheries in B.C.". Prepared for: Department of Recreation and Conservation. Fish and Wildlife Branch Underwood McLellan & Associates 1977. ImUnderwood McLellan & Associates 1979. "Salmon Rearing Facilities,  Water Study,  Treatment Technology".  March 1979. Underwood McLellan (1977) Ltd., Consulting Engineers and Planners, Vancouver, B.C. Underwood McLellan & Associates 1979. "Salmon Rearing Facilities, Waste Water Study, Treatment Technology". March, 1979. Underwood McLellan (1977)Ltd., Consulting Engineers and Planners, Vancouver,B.C. Walden,  CC. & A.E.,Birkbeck  1974. "Feasibility  and Design  Studies for Water  Treatment at a Trout Hatchery". Prepared for British Columbia Department of Public Works by B.C. Research, Vancouver, 234pp.  / 100 Warrer-Hansen, I. 1979.  "Fish Farming and Pollution Problems-Methods to Reduce  Discharged Matter". Paper Presented at the Fish Farming Meeting, Oban, 22-23 February, 1979. Warrer-Hansen, I. 1979a. "Operation Clean-Up". Fish Farming Int 6(2):32-34. Warrer-Hansen, I. 1982. "Methods of Treatment of Waste Water from Trout Fanning". In:  Report  of  the  EIFAC  Workshop  on  Fish-Farm  Effluents.  Silkeborg, Denmark 26-28 May, 1981. EIFAC Technical Paper (41). Wheaton, F. 1977. "Aquaculture Engineering". John Wiley & Sons. Inc. Willoughby, H., H.N.,Larsen & J.T. Bowen 1972.  "The Pollutional  Effects of Fish  Hatcheries". American Fishses & U.S. Trout News. 17(3): pp6-7,20-1.  APPENDIX I  TABLE V  SLTEST  generated  regression  equations  for  overall  non-filterable residue removal curves (Figures 5.5 - 5.15)  Group 9 1  2  3  4  5  6  Fish Size  Feed P e l l e t N o n - F i l c e r a b l e S i z e ( i n c h ) Residue (cone)  11.0  17.0  21.0  21.0  41.0  54.0  * Fish size  3/32  3/32  3/32  1/8  1/8  1/8  Linear Trans f o r m a t i o n  Hyperbolic EquatiL on  65 I  Y-0 .036 + 0.011X  Y-89,3X/ [3. 19+X)  542  Y-0.026 + 0 .01IX  Y-89 .6X/( 2.35+X)  738  Y-0 .037 + 0.01 IX  Y-92 .7X/( 3.41+X)  584  Y-0.026 + 0.011X  Y-90.8X/<[2.33+X)  402  Y-0.022 + 0 .011X  Y-88.8X/< l.95+X)  1007  Y-0.020 + 0.011X  Y-90.7X/ [1.78+X)  676  Y-0.032 + 0.011X  Y-93.6X/ [2.9 8+X)  179  Y-0.023 + 0 .011X  Y-88.6X/ [2.03+X)  903  Y-0.022 + 0.011X  Y-88.3X/ [1.9 3+X)  794  Y-0.027 + 0 .011X  Y-93.5X/ [2.49+X)  584  Y-0.026 + 0 .011X  Y-94.6X/ [2.45+X)  1755  Y-0.021 + 0 .011X  Y-95.IX/ [2.03+X)  1148  Y-0.021 + 0.011X  Y-93.7X/ [1.9 4+X)  422  Y-0.020 + 0.011X  Y-94.3X/ [1.91+X)  329  Y-0.012 + 0.011X  Y-90.9X/ (1.06+X)  k  1945  Y-0.031  0.011X  Y-92.9X/ (2.92+X)  1308  Y-0.025 + 0.011X  Y-95 .OX/ (2.4 +X)  616  Y-0.017 + 0.011X  Y-96.2X/ (1.67+X)  i n grams  ** I n i t i a l n o n - f i l t e r a b l e  X = Residence  time  Y = Non-fi1terable  residue concentration  i n mg/litre  (minutes)  residue  removed a t r e s i d e n c e  101  time  X (%)  I  TABLE V  (cont.)  SLTEST  generated  regression  equations  for  102  overall  non-filterable residue removal curves (Figures 5.5 - 5.15)  Group it  Fish Size  Feed P e l l e t Size (inch)  54.0  7  5/32  73.0  8  5/32  93.0  9  10  5/32  89.0  3/16  3/16  311.0  11  * Fish ** X  Non-Filterable Residue (cone)  size  Initial  Linear Transformation  Hyperbolic Equation  792  Y = 0 .016 + 0 .011X  Y=94.2X/(1 .53+X)  744  Y=0 .014 + 0 .01IX  Y=93.6X/(1 .30+X)  595  Y=0.021 + 0.011X  Y=91.7X/(1 .96+X)  262  Y = 0 .012 + 0 .011X  Y=9 3.2X/(1 .14+X)  792  Y = 0 .017 + 0 .011X  Y=94.9X/(1 •62+X)  693  Y=0.017 + 0 .011X  Y=95 .9X/(1 .67+X)  657  Y=0.014 + 0 .011X  Y=93.0X/(1 .34+X)  543  Y=»0.021 + 0.011X  Y=91.4X/(1 .96+X)  594  Y=0 .016 + 0 .011X  Y-92 .9X/(1 .46+X)  503  Y=0.022 + 0 .011X  Y=94.3X/(2 .07+X)  397  Y=0.019 + 0 .011X  Y=94.0X/(1 .75+X)  332  Y=0 .022 + 0.011X  Y=91.2X/(2 .0 3+X)  1254  Y = 0 .020 + 0 .01IX  Y=94.1X/(1 .87+X)  889  Y=»0.02 9 + 0.011X  Y=94.9X/(2 .75+X)  286  Y-0 .015 + 0.011X  Y-93.9X/(1 .43+X)  977  Y=0.020 + 0 .011X  Y=94.8X/(1 .91+X)  i n grams  non-filterable  = Residence  time  Y = Non-fi1terable  residue  concentration  i n mg/litre  (minutes) residue  removed  at residence  time  X  (%)  APPENDIX II  Figures 5.28 (overall non-filterable residue removal curve) and 5.29 (individual particle settling  velocity  distribution curve) can be  used  to  estimate the  overflow rate  (mVmVhour) and residence time for various per cent solids removal.  In the example to be discussed, a fish farm producing 50-75 tonnes per year uses an estimated 500 1/sec or 1800 mVhour. From the observations and discussions of the overall non-filterable residue removal curve (Figure 5.28) it was concluded that holding the effluent longer than 15 minutes did not effectively contribute to the overall level of solids removal. Therefore, assuming a maximum residence time of 15 minutes will provide for 82% removal of the non-filterable solids.  An overflow rate or surface loading rate can thus be obtained by selecting a suitable settling velocity (V ) from the individual particle settling velocity distribution curve o (Figure 5.29) such that 82% of the solid particulates will be removed. From Figure 5.29 it can be. observed, that 82% of the particles have settling, velocities greater than 0.2 cm/sec (7.2 m/hour).  From the discussion of Equations 3.1 and 3.2 it was shown that the settling velocity is the equivalent of the overflow rate or surface loading rate. A settling velocity of 7.2  m/hour is  therefore  the  equivalent of  a loading rate of 7.2 mVmVhour.  Introducing a safety factor of 1.5 to reduce turbulence (Warrer-Hansen, 1982), yields a surface loading rate of 5 m /mVhour. 3  Given the water consumption of the farm is 1800 mVhour, the necessary basin area 103  / 104 will be:  Sur f a c e  a r e a of  clarifier  Flow  rate  Overflow  1800  rate  m/hour  3 2 5 m /m / h o u r  '=  360 m  Therefore, a farm utilizing 1800 mVhour, requiring 82% of non-filterable solids removal, will require a sedimentation basin with a depth of 1.25 .m and a surface area of 360 m . A width of 10-20 meters will maintain the water velocity in the J  settling zone within the 2-4 an/sec maximum velocity suggested to prevent scouring and resuspension of the already settled solids (Warrer-Hansen, 1982).  

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 5 0
Japan 4 0
China 3 1
Mexico 2 0
Russia 1 0
Vietnam 1 1
City Views Downloads
Tokyo 4 0
Ashburn 3 0
Shenzhen 2 1
Unknown 2 8
Mountain View 1 0
Cho Ha 1 1
Beijing 1 0
Saint Petersburg 1 0
Seattle 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058685/manifest

Comment

Related Items