MODELLING CARBON OXIDATION IN PULP MILL ACTIVATED SLUDGE SYSTEMS DETERMINING MODEL PARAMETERS by DEBORAH JANE STANYER B.Sc, The University of Victoria, 1987 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Civil Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1997 © Deborah J. Stanyer, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department DE-6 (2/88) ABSTRACT To predict the behavior of municipal wastewater treatment facilities, the Activated Sludge Model No. 1 is often used. This model has also served as the basis for the development of many other models to determine the behavior of activated sludge systems treating municipal wastewater, and has also been extended to the development of a model for petrochemical activated sludge treatment systems. Application of the model requires the determination of kinetic parameters and wastewater composition using continuous culture laboratory scale activated sludge reactors. These systems are laborious to operate and are equipment intensive. Activated sludge systems are a commonly used method of secondary treatment of pulp mill effluents in British Columbia. The application of the Activated Sludge Model No. 1 has not been extended previously to predict the performance of carbon oxidation in activated sludge systems treating pulp mill effluent. The objective of this study was to establish a database of experimental information on the characteristics of activated sludge systems treating bleached Kraft pulp mill wastewater using simple on-site batch test methods. Batch test methods involving respirometry and chemical oxygen demand measurements were successful in generating a data set of wastewater and biomass characteristics comparable to those found in the literature. ii Development of a dynamic mechanistic model based on the data generated this study will provide a framework for studying the behavior of pulp mill activated sludge systems and provide a basis for planning further experimentation. iii TABLE OF CONTENTS ABSTRACT i" TABLE OF CONTENTS iv GLOSSARY OF TERMS AND SYMBOLS vii LIST OF TABLES x LIST OF FIGURES xi ACKNOWLEDGMENTS xii 1. INTRODUCTION 1 2. THE PULPING PROCESS 5 2.1 Pulp Mill Effluent Composition 8 2.1.1 Wood Composition 8 2.1.2 Bleached Kraft Mill Effluent Composition 10 2.1.3 Mechanical Pulping Effluent Composition 12 3. THE ACTIVATED SLUDGE TREATMENT PROCESS 14 3.1 Howe Sound Pulp and Paper's Effluent Treatment System 18 3.2 Western Pulp Limited's Effluent Treatment System 21 4. ACTIVATED SLUDGE MODELLING 24 4.1 History 24 4.2 IAWPRC Model Number 1 29 4.2.1 Basic Concepts of the IAWPRC Model No. 1 30 4.2.1.1 Heterotrophic Death 30 4.2.1.2 Heterotrophic Growth 31 4.2.1.3 Influent Wastewater Fractionation 33 4.2.2 Model Design 37 5. PROCEDURES FOR THE ESTIMATION OF MODEL PARAMETERS 40 5.1 IAWPRC Methods 40 5.1.1 Heterotrophic Yield 40 5.1.2 Decay Coefficient 42 5.1.3 Readily Biodegradable Substrate 43 5.1.4 Inert Soluble Organic Matter 44 5.1.5 Inert Particulate Organic Matter 45 5.1.6 Slowly Biodegradable Substrate 45 iv 5.1.7 Maximum Specific Growth Rate and Half Saturation Coefficient ....46 5.1.8 Maximum Specific Hydrolysis Rate and Half-Saturation Coefficient for Hydrolysis of Slowly Biodegradable Organic Matter 47 5.2 Alternate Methods 47 5.2.1 Heterotrophic Growth Yield 48 5.2.2 Decay Coefficient For Heterotrophic Microorganisms 49 5.2.3 Maximum Specific Growth Rate And Half-Velocity Coefficient 49 5.2.4 Wastewater COD Fractions 53 5.2.4.1 Readily Biodegradable COD 54 5.2.4.2 Soluble Inert COD 59 5.2.4.3 Particulate Inert Substrate 60 5.2.4.4 Slowly Biodegradable Substrate 61 5.2.5 Conclusion 62 6. EXPERIMENTAL PROGRAM 63 6.1 Experimental Design 63 6.2 Analytical Methods 65 6.2.1 Chemical Oxygen Demand 65 6.2.2 Solids 65 6.2.3 Oxygen Uptake Rate 65 6.3 Experimental Procedures 67 6.3.1 Growth Yield 67 6.3.2 Decay Rate 68 6.3.3 Maximum Specific Growth Rate 69 6.3.4 Readily Biodegradable COD 70 6.3.4.1 Physical - Chemical Method 70 6.3.4.2 Aerobic Batch Method 71 6.3.5 Soluble Inert COD 71 6.3.6 Particulate Inert COD 71 6.3.7 Slowly Biodegradable COD 72 7. RESULTS AND DISCUSSION 73 7.1 Growth Yield Experiments 73 7.2 Decay Rate Experiments 78 7.3 Maximum Specific Growth Rate Experiments 82 7.4 Wastewater COD Fractions 90 7.4.1 Readily Biodegradable COD 90 7.4.2 Soluble Inert COD 95 7.4.3 Particulate Inert COD 98 7.4.4 Slowly Biodegradable COD 100 8. CONCLUSIONS AND RECOMMENDATIONS 103 9. REFERENCES 107 v Appendix 1. DETERMINATION OF TRUE HETEROTROPHIC GROWTH YIELD 114 Appendix 2. DETERMINATION OF DECAY COEFFICIENT FOR HETEROTROPHIC BIOMASS 119 Appendix 3. DETERMINATION OF THE MAXIMUM SPECIFIC GROWTH RATE 121 Appendix 4. DETERMINATION OF READILY BIODEGRADABLE COD 124 Appendix 5. DETERMINATION OF PARTICULATE INERT COD 132 Appendix 6. DETERMINATION OF INFLUENT COD COMPOSITION AT WESTERN PULP PARTNERSHIP LTD 136 Appendix 7. DETERMINATION OF INFLUENT COD COMPOSITION AT HOWE SOUND PULP AND PAPER LTD 137 vi GLOSSARY OF TERMS AND SYMBOLS / p - fraction of inert particulate formed during heterotrophic decay (mg COD/mg cell COD) \i - specific growth rate of heterotrophic microorganisms (d'1) 0 - temperature coefficient i i m a x - maximum specific growth rate for heterotrophic microorganisms (d"1) ATP - adenosine triphosphate b' H - traditional decay coefficient for heterotrophic microorganisms (d"1) b H - model decay coefficient for heterotrophic microorganisms (d-1) BKME - bleached Kraft mill effluent BOD - biochemical oxygen demand (mg/L) COD - chemical oxygen demand (mg/L) CODso, - influent total truly soluble COD (mg COD/L) DTPA - diethylene-triamine-penta-acetic acid F/M - food to microorganism ratio (mg/mg) HSPP - Howe Sound Pulp and Paper Limited IAWPRC - International Association on Water Pollution Research and Control K A - substrate transfer rate (mg COD/mg VSS«d) k h - maximum specific hydrolysis rate (d'1) vii K m p - maximum specific growth rate constant for the utilization of stored COD (mg COD/mg VSS»d) K s half saturation coefficient (mg/L) K s p - half saturation coefficient for the utilization of stored COD (mg COD/mg VSS) K x - half saturation coefficient for hydrolysis of slowly biodegradable organic matter (mg/L) MCRT - mean cell retention time (d) MLVSS - mixed liquor volatile suspended solids Mw - molecular weight NCASI - National Council for Air and Stream Improvement NZFP - New Zealand Forest Products OUR - oxygen utilization rate PAPRO - Pulp and Paper Research Organization of New Zealand q - specific substrate utilization rate (d"1) RAS - return activated sludge RFA - resin and fatty acids Sj - soluble inert COD in the wastewater (mg COD/L) SRT - solids retention time (d) S s - readily biodegradable substrate (mg COD/L) TMP - thermomechanical pulp Vm i - volume of mixed liquor (L) VSS - volatile suspended solids Vww - volume of wastewater (L) viii Western Pulp Partnership Limited heterotrophic biomass (mg COD/L) particulate inert COD in the wastewater (mg COD/L) slowly biodegradable COD in the wastewater (mg COD/L) growth yield for heterotrophic microorganisms (mg cell COD/mg COD) observed growth yield (mg cell COD/mg COD) ix LIST OF TABLES Table 1. Process Kinetics And Stoichiometry For Heterotrophic Bacterial Growth In An Aerobic Environment (Henze et al., 1987) 38 Table 2. Parameters For The Model And The Order In Which They Should Be Determined 41 Table 3. Methods For Estimating Organic Fractions In Wastewater 54 Table 4. Growth Yield Values Obtained From Literature 77 Table 5. Variation In The Slope Of The OUR Curve With Changes In The Ratio Of Substrate To Biomass Used 86 Table 6. Comparison Of Readily Biodegradable COD Results Using The Ekama etal. (1986) And Mamais etal. (1993) Methods 95 Table 7. Summary Of Results Of Model Parameter Determination 105 Table A 1. Experimental Growth Yield Results 116 Table A 2. Experimental Traditional Decay Coefficient Results 120 Table A 3. Experimental Maximum Specific Growth Rate Results 123 Table A 4. Experimental Aerobic Batch Test Results 128 Table A 5. Experimental Results For Influent Wastewater Fractionation At Western Pulp Partnership Ltd 136 Table A 6. Experimental Results For Influent Wastewater Fractionation At Howe Sound Pulp And Paper Ltd 137 x LIST OF FIGURES Figure 1. The Chemical Components Of Wood (Smook, 1989) 9 Figure 2. Effluent Treatment Flow Schematic For Howe Sound Pulp And Paper Limited (Strang, 1992) 19 Figure 3. Effluent Treatment Schematic For Western Pulp Partnership Limited (Taylor, 1996) 22 Figure 4. Influent COD Fractions 34 Figure 5. Schematic Diagram Of The Respirometer Used In The Experimental Studies 66 Figure 6. Growth Curve For Western Pulp Sample Collected November 14, 1995 74 Figure 7. Death - Regeneration Cycle Under Aerobic Conditions (Lishman And Murphy, 1994) 79 Figure 8. Decreasing OUR In Batch Aerobic Digestion Test With Western Pulp Return Activated Sludge Collected November 28, 1995 80 Figure 9. Batch Test To Estimate u,max For The Western Pulp Sample Collected October 17, 1995 83 Figure 10. Logarithmic Form Of The Relative OUR For The Western Pulp Sample Collected October 17, 1995 83 Figure 11. Batch Test To Estimate The Maximum Specific Growth Rate For Howe Sound Sample Collected February 23^ 1996 85 Figure 12. Logarithmic Form Of The Relative Oxygen Uptake Rate For Howe Sound Sample Collected February 23, 1996 85 xi ACKNOWLEDGMENTS I would like to thank Professor E. R. Hall for his guidance and support throughout this research project. I would also like to thank the laboratory staff of the Civil Engineering department for their assistance and use of laboratory equipment and Scott Jackson for his computer and instrumentation assistance. At the Pulp and Paper Center my thanks go to Rita Penco for her assistance with the literature review and to Peter Taylor for constructing the reactors used in the study. My thanks also to Jeanne Taylor and the staff of the technical department at Western Pulp Partnership Limited for their assistance during the project. They were always helpful and the use of their mill laboratory was very much appreciated. At the Howe Sound Pulp and Paper mill I would to thank Al Strang for use of the UNOX laboratory, space to set up my experimental reactors and the endless supply of COD chemicals. I would like to thank Siew Sim for sharing her knowledge and data regarding the operation of the UNOX system, Dave Shepard and Julie Bidulka for their chemical expertise, Brad Palm for his knowledge of the bleaching process and Lome Piercey for his knowledge of the TMP process. My thanks also to Sonja Fallis for her invaluable word processing tips. My biggest thanks go to my daughter Abby and husband Bill who gave me the motivation to complete this project through all the challenges we met with during its completion. To one and all my appreciation and thanks. xii 1. INTRODUCTION The activated sludge process has been used to treat domestic and industrial wastewater since its conception in England in 1914 (Mines and Sherrard, 1989). Over time, the operation of activated sludge plants has evolved to deal with increasingly complex and variable wastes. In the beginning, plant engineers lacked an understanding of the biochemical reactions involved and developed solutions by trial and error for their particular situations. From this, several original concepts and process modifications emerged and gained acceptance in the practice of sewage treatment. The operational difficulties encountered in the operation of treatment plants led to a need for mathematical models incorporating the fundamental microbial mechanisms into a rational engineering description of the process. A significant evolution in modelling practice was experienced beginning in 1962 with the single-component model developed by McKinney (Orhon and Artan, 1994). In 1983, the International Association on Water Pollution Research and Control (IAWPRC) task group reviewed existing models and developed one simple mechanistic model incorporating carbon oxidation, nitrification and denitrification. The result was the multi-component IAWPRC Activated Sludge Model No. 1 (Henze etal., 1987). This model has been in use extensively and successfully in the design, control and optimization of municipal activated sludge l systems. Recently, the more complex Activated Sludge Model No. 2 has been published (Gujerefa/., 1995). In order to use the IAWPRC models to predict system behavior, a number of parameters which specify the characteristics of the influent, must be determined. Additionally, various stoichiometric and kinetic parameters that characterize the behavior of the microorganisms in the activated sludge system must be specified. Between 1990 and 1994 the Pulp and Paper Research Organization of New Zealand (PAPRO) conducted research to develop a computer model to simulate the performance of aerated lagoons treating Kraft mill effluent. This was pioneering work as the computer models to date had focused on municipal wastewater and not pulp and paper mill effluent. The research was conducted at the New Zealand Forest Products (NZFP) Kinleith mill. The mill produces approximately 450,000 air dried tonnes per annum of mainly softwood pulp, of which approximately 160,000 tonnes are bleached. The treatment system consists of two parallel systems, one treating chlorination stage bleaching effluents and general mill effluents, the other treating alkali extraction bleach plant effluents and foul condensates. Previous attempts to model the chlorination and general effluent lagoon system using the National Council for Air and Stream Improvement (NCASI) Aerated Stabilization Model (NCASI, 2 1985) were not successful, so it was decided to model the lagoon using the IAWPRC Activated Sludge Model No. 1. PAPRO conducted a chemical and biological characterization of the effluent entering the aerated lagoon (Slade et al., 1991). Parameter estimates were made using biomass withdrawn from laboratory scale activated sludge reactors based on the methodology recommended for the IAWPRC Activated Sludge Model No. 1. Although the methodology used to determine these parameters was developed on municipal wastewaters, it was found to be applicable to Kraft mill effluents. The results obtained were generally compared to those of municipal wastewaters. Efforts were made to evaluate the IAWPRC Model No. 1 using these determined parameters. The results were inconsistent in the prediction of measured substrate removal and solids concentration. These findings were thought to be the result of either inadequate parameter determination or failure of the model to include a parameter or process that is vital to the adequate modelling of the biological treatment of bleached Kraft mill wastewater (Dare, 1995). The modelling efforts of PAPRO have not been replicated. As most of the existing kinetic information is limited to domestic sewage, application of the IAWPRC model to industrial wastewater treatment still requires substantial additional information. A continuous experimental setup of laboratory scale activated sludge reactors is required to generate the model parameters as described by the IAWPRC Model No. 1. However, since steady state must be achieved, a long measuring period is required. These systems have been criticized for their cost and difficulty of operation (Wentzel ef al., 1995). The effort of many researchers has been directed toward the determination of the input parameters using simple on-site batch tests. Batch measurements are considered valuable by many researchers for the determination of wastewater and sludge characteristics (Henze, 1992; Kappeler and Gujer, 1992; Spanjers and Keesman, 1994; Spanjers and Vanrolleghem, 1995; Templeton and Grady, 1988). In general, there is excellent agreement between the batch test results and full-scale performance (Poduska, 1979). Based on the Activated Sludge Model No. 1, the objective of this research was to use simple on-site batch tests to establish a database of influent wastewater characteristics and biological constants of heterotrophic organisms which could provide the basis for the development of a dynamic mechanistic mathematical model for the prediction of carbon oxidation in an activated sludge system treating pulp mill wastewater. 4 2. THE PULPING PROCESS Many pulping processes are used to produce different pulp and paper products. Both Western Pulp Partnership Limited (WP) and Howe Sound Pulp and Paper (HSPP) use the Kraft process. This process is the most common pulping technology because of its efficient recovery of pulping chemicals and its ability to generate energy from the combustion of solubilized lignin while producing a high strength pulp. Both mills produce bleached softwood Kraft market pulp. Howe Sound produces 1000 tonnes per day and Western Pulp produces 715 tonnes per day. In the Kraft process, wood chips are cooked in digesters at an elevated temperature and pressure in an aqueous solution of sodium hydroxide and sodium sulfide (white liquor). Cooking dissolves the lignin which binds the cellulose fibres together. Western Pulp operates five batch digesters and one continuous digester. Howe Sound operates a single continuous digester. In a continuous digester an uninterrupted flow of wood chips and cooking liquor enters from the top and pulp is withdrawn continuously from the bottom into a blow tank. In a batch digester, chips and liquor are added to the top of the digester. The digester is capped and the chips are cooked under pressure and high temperature. Once cooked, the blow valve at the bottom of the vessel is opened and the contents are discharged into a blow tank at atmospheric pressure. From the blow tank, the cooked pulp proceeds to washers which 5 recover the spent liquor (black liquor) and through screens which remove uncooked fibre bundles and debris from the pulp. At Howe Sound, further delignification is accomplished with oxygen. In this process, the pulp is mixed with a caustic solution and oxygen under high pressure and temperature. Magnesium sulphate is added to protect the fibers. The pulp is then passed through two washing stages and transferred to the bleach plant. Pulp bleaching imparts whiteness or brightness to the pulp, in addition to yielding certain desirable physical and chemical properties. The bleach plant consists of a series of different stages, each one being separated by a rotary vacuum filter or washer. At Howe Sound the bleaching sequence is DCE 0 D E D and Western Pulp uses a CDE 0 D E D sequence. The first chlorination stage causes the degradation of lignin using chlorine (C) and/or chlorine dioxide (D). This initial step involves an electrophilic attack on sites of high electron densities followed by nucleophilic reactions. The chlorine acts in its molecular form both as an oxidizing agent and as a substituting agent. Delignification is achieved by oxidative depolymerization, in conjunction with chlorination and hydroxylation, resulting in the fragmentation of lignin. Lignin is degraded to alkali-soluble compounds which are then removed by alkaline extraction (Fahmy, 1992). The extraction stage, E 0 , uses oxygen to further delignify the stock and dissolve and 6 extract the degradation products. The bleaching process is typically finished by one or more oxidative stages with intermediate alkali extraction. In the D E stage, delignification with chlorine dioxide, extraction of degradation products and neutralization to a target pH occurs. The D stage represents the final chlorine dioxide bleaching step. Howe Sound also operates a thermomechanical pulp (TMP) mill the product of which is used, in conjunction with a portion of the Kraft pulp production, to produce high grade newsprint. The defibrilization process produces a high strength pulp with yields ranging from 93 - 96%. In the TMP process, wood chips are steamed under pressure and sent through six conical disk refiners and one rejects refiner operated under pressure that reduce wood chips to pulp. Latency, the tendency for the fibres to curl, is removed in the latency chest under elevated temperature and agitation. The pulp is then screened, cleaned and thickened. Diethylene-triamine-penta-acetic acid (DTPA) is added to the pulp before the bleaching stage. DTPA is a synthetic chelating agent added to bind metal ions which would otherwise catalyze the decomposition of hydrogen peroxide. Bleaching is carried out using hydrogen peroxide to produce fully bleached pulp or sodium hydrosulphite to produce semi-bleached pulp. Other chemicals added in the process are 7 sodium silicate to stabilize the hydrogen peroxide and sulphurous acid to achieve pH control. 2.1 Pulp Mill Effluent Composition The composition of wood, species used, and the pulping and bleaching processes employed, influence the composition of the effluents from pulp and paper mills. In addition, factors such as extent of wood seasoning, water usage per tonne of pulp/paper produced, mill operating conditions and in-plant control measures can produce wide variations in effluent composition even for mills of the same category. 2.1.1 Wood Composition Wood consists of three main components: (i) carbohydrates, (ii) lignin, and (iii) extractives. At least 60% of the wood components are high-polymeric compounds: cellulose, hemicellulose and lignin. The remaining 5-15% of the components are extractives including resins, terpenes, fats, inorganic compounds and proteins (Fahmy, 1992). Figure 1 outlines the chemical composition of wood. Cellulose is a component of the cell wall and makes up about 45% of the total dry wood weight. Cellulose is a polysaccharide made up of 600-1500 glucose units. Hemicellulose is another component of the cell wall. Hemicellulose is a 8 polymer of five different sugars: glucose, mannose, galactose, xylose and arabinose (Smook, 1992). 21% Hardwoods 25% Softwoods LIGNIN CARBOHYDRATES 35% Hardwoods 25% Softwoods Glucose Terpenes Resin acids (softwoods) Phenols Unsaponifiables Glucose Mannose Galactose Xylose Arabinose Figure 1. The Chemical Components Of Wood (Smook, 1989) Lignin is the second most abundant molecule (Fengel and Grosser, 1975). Lignin binds plant fibres together and provides structural support. Lignin is a complex aromatic polymer built up from three different monomers. The proportions of the monomers varies with wood species. In lignin, chromophoric systems are present: quinones, quinone methides and biphenyls. These components impart the brown color to pulp (Fahmy, 1992). 9 Extractives are wood constituents which can be extracted by neutral solvents. The length of time that wood is seasoned prior to pulping, the form in which it is stored, i.e., logs or chips, and the wood species are important factors affecting the extractives content and composition of wood. Some of the major constituents of softwood extractives are fatty acids, fatty acid esters, higher molecular weight phenolics, resin acids, juvabione and compounds belonging to the terpenoid class. The terpenoids are found only in softwood species and are constituents of turpentine, an important by-product of the Kraft process recovered from digester relief gases. Juvabione and its derivatives are major extractives of fir wood species. Juvabione compounds inhibit the maturation of the larval stages of certain insects and protect the tree from infestation. The fatty acids in wood extractives are saturated and unsaturated carboxylic acids usually ranging from 12 to 24 carbons (Leach et al., 1976). 2.1.2 Bleached Kraft Mill Effluent Composition In the production of bleached Kraft pulp, lignin and extractives are removed from the wood in order to uncover and purify the cellulose fibers. In the chemical pulping process, 90 - 95% of the lignin in the wood is removed and is subjected to a high degree of degradation and transformation. Deploymerization occurs and results in the solubilization of the degraded lignin products in the spent liquors from the chlorination and extraction stages of the bleaching sequence. Chlorine reacts with the solublilzed lignin producing chlorinated organic material 10 in the effluent. Chlorinated constituents originating in bleach plant caustic extraction effluents are more resistant to degradation than are naturally occurring wood extractives (Leach et al., 1977). Modified cooking (extended delignification) and oxygen delignification may decrease the amount of residual lignin entering the bleach plant by as much as 50% (Dillner et al., 1990). More than 300 chlorinated compounds have been identified in spent liquors from the bleaching of chemical pulp (Springer, 1993). These compounds can be classified by molecular weight. Degraded residual lignin products comprise a large portion of the high molecular weight (Mw > 1000) chlorinated organic material in bleached Kraft mill effluent (BKME) (Springer, 1993). The high Mw portion of the effluent carries most of the chemical oxygen demand (COD), adsorbable organic halogen (AOX) and color and is presumably biologically inactive and non-toxic because of its molecular size (Dahlman and Morck, 1993). The fraction of organic material in BKME between 500-1000 Mw is also mainly composed of oxidized degradation products from the residual lignin, and may be structurally similar to the material in the high molecular weight fraction (Dahlman and Morck, 1993). i i The low molecular weight compounds found in BKME are comprised of residues of the extractives carried over to the bleaching plant with the pulp. The nature of the pulping processes employed at a particular mill has an effect on the levels of fatty acids in the effluent. In Kraft cooking, the hot alkaline conditions saponify (hydrolyze) the fatty acids (Jones, 1975). The common fatty acids are oleic, linolenic, palmitoleic and linoleic, which are straight chain carboxylic acids. Free resin acids are dissolved as sodium salts in the cooking liquor (Edde, 1984). Some of the common resin acids found in pulp mill effluents are pimaric, isopimaric, palustric, dehydroabietic, neoabietic, sandaracopimaric and abietic acids. These acids all have a ring structure. Resin and fatty acids (RFA) are generally considered to be the principal toxicants in Kraft pulping wastewater (Springer, 1993). The principal heavy metal ions of concern in BKME may include aluminum, chromium, copper, nickel, titanium, iron, mercury and zinc. These metals may originate from chemicals and additives used in the pulping process or from equipment corrosion (Springer, 1993). 2.1.3 Mechanical Pulping Effluent Composition The wood furnish, age of wood, pulp processing conditions and bleaching chemicals used affect mechanical pulping effluent composition. TMP effluent is generally very high in biochemical oxygen demand (BOD), COD and suspended solids concentrations due to extensive water recycling (Lo et al., 1994a). 12 Elevated levels of wood extractives such as the resin acids: abietic, dehydroabietic, isopimaric and palustric are predominant in mechanical pulping effluent (Leach et al., 1976). These resins along with sandaracopimaric, neoabietic and pimaric acid, account for 60 - 90% of the overall toxicity of mechanical pulping effluents. Minor toxic factors include alcohols related to the above acids, in particular pimarol and isopimarol, and an insect regulator, juvabione and some of its derivatives (Leach and Thakore, 1976). Other potentially toxic compounds include diterpenes, aldehydes, unsaturated fatty acids and (epi)manool. The pH during pulping can be an important factor controlling the release of extractive compounds to the mechanical pulping wastewaters (Walden and Leach, 1975). Hydrogen peroxide and DTPA from bleaching, and metals from the chemical additives used to impart desired qualities to the paper product are also present in the effluent. Alum is added to the papermaking system to reduce pitch problems. Excess alum in the effluent effectively binds phosphorus and reduces the amount available for cell synthesis in the activated sludge reactor. 13 3. THE ACTIVATED SLUDGE TREATMENT PROCESS The activated sludge process relies on the removal of pollutants from wastewater through a series of biochemical reactions. The pollutants that are removed by these reactions constitute the biodegradable substrate components of the wastewater. A substrate may be metabolized by either of two different major processes: it may be channelled into a sequence of degradation reactions converting it first into a number of intermediate products and then to stable end products with energy generation. These reactions are called catabolism or dissimilation. The substrate may otherwise be part of the anabolic or assimilative reactions where major biochemical components of cells are manufactured from the intermediate products by means of a variety of biosynthetic pathways. The biochemical reactions in these processes occur because the biodegradable wastewater components act as substrates to form enzyme-substrate complexes with specific enzymes. A complex is generated by the attachment of substrate on the active site on the surface of the enzyme molecule. When the biochemical reaction is completed, the active site is liberated and returns to its cyclic function to receive another substrate molecule. An enzyme becomes saturated when all its active sites are full at high substrate concentrations, yielding maximum reaction rates. Sometimes, enzymatic functions may be retarded or even 14 stopped by inhibitors which may either compete for, or simply block, the active sites (Alberts et al., 1983). A number of pollutants present in wastewaters, such as heavy metals, non biodegradable organics, etc., may exert significant inhibitory action upon biological treatment processes (Ganczarczyk, 1983). Activated sludge systems are traditionally designed for the removal of carbonaceous organic compounds in wastewaters. Heterotrophic microorganisms require organic compounds for their carbon source. Energy is obtained by the heterotrophs through enzyme-mediated reactions in the glucose fermentation pathway and the citric acid cycle (Mines and Sherrard, 1989). Almost complete removal may be attained in these systems because a fraction of the organic carbon is incorporated into new cells through biosynthesis, while the rest is oxidized to CO2 by means of catabolic reactions. Nitrogen- and phosphorus-containing compounds are also removed as part of the synthesized cellular material but at lower rates, dictated by the chemical composition of microorganisms. Process modifications may facilitate the oxidation of nitrogenous compounds to nitrate or, in the absence of oxygen, the conversion of nitrates into nitrogen gas resulting in complete nitrogen removal. Different process modifications may also provide enhanced biological phosphorus removal. 15 Several process modifications have been developed to permit existing plants to treat large flows and loads while maintaining high effluent quality. These modifications are essentially hydraulic in nature and include complete mix, plug flow, contact stabilization, step feed, extended aeration, oxidation ditch and high rate oxygen activated sludge. The results of intensive development by the Union Carbide company have resulted in an oxygen activated sludge system known as UNOX. The UNOX system is particularly applicable where the available space for a treatment facility is limited and where a relatively strong wastewater is being treated (Ganczarczyk, 1983). For these reasons, the UNOX system has been installed at many British Columbia coastal pulp mills. The UNOX system uses a covered and staged oxygenation basin for contact of oxygen and mixed liquor. The particular configuration of tankage selected for a given system such as the size and number of compartments per aeration tank, and the number of parallel aeration tanks, is determined by several parameters including wastewater characteristics, plant size, land availability and treatment requirements. A prime advantage of the UNOX aeration tank design is its ability to transfer oxygen rapidly across the gas-liquid interface at low turbulence and to meet high oxygen uptake demands per unit volume of mixed liquor (Huang and Mandt, 1978). 16 High purity oxygen (90-100 per cent by volume) enters the first stage of the system and flows concurrently with the wastewater being treated. The pressure in the oxygenation tank is maintained between 5 and 10 cm of water column to prevent backmixing from stage to stage (Huang and Mandt, 1978). Mass transfer and mixing within each stage are accomplished either with surface aerators or with submerged sparging turbine aerators. Typical UNOX treatment technology calls for dissolved oxygen levels in the aeration tanks of 1.5-4 mg/l, mixed liquor suspended solids levels of 2500-6500 mg/l and hydraulic retention times of about 4 hours (Metcalf and Eddy, 1991). The comparative values for air aerated systems (conventional activated sludge systems) are dissolved oxygen levels of 1-2 mg/l, mixed liquor suspended solids levels of 1500-2500 mg/l and a retention time of about 6 hours (Ganczarczyk, 1983). Effluent mixed liquor from the aeration tank is separated by gravity clarifiers and the concentrated biological solids are then recycled back to the first stage in the aeration tank to maintain a concentrated active population of microorganisms. Excess biological solids are removed from the system by wasting from either the clarifier or the aeration tank. 17 3.1 Howe Sound Pulp and Paper's Effluent Treatment System There are four main process sewer lines in the mill complex: the general sewer, the acid sewer, the alkaline sewer and the clean water sewer. The general, acid and alkaline sewers collect all of the contaminated effluent generated by the pulp mill processes. The clean sewer collects uncontaminated cooling water which does not require treatment and which can be discharged directly into the ocean. The effluent treatment system consists of a pumping station, primary clarifier, nutrient feed system, spill pond, oxygen activated sludge reactor (UNOX), two secondary clarifiers, a sludge return system and a sludge de-watering system. A schematic of the effluent treatment system is given in Figure 2. The effluent spill pond protects the secondary effluent treatment system from upsets which may affect its efficiency. The spill pond has a capacity of 33,000 cubic meters. Effluent can be diverted to the spill pond from the primary clarifier inlet. Once in the spill pond, effluent is returned to the treatment system on a controlled basis either to the primary clarifier or to the secondary treatment system. The primary clarifier has a diameter of 56.4 meters and a side wall depth of 5.5 meters. Effluent collected in the overflow trough flows by gravity to the rest of the treatment system. 18 61 3 c 3 IS) c c. o I CO o c 3 Q. *D C "5" a a J (D r-(D a To The bleach plant acid and caustic sewers are mixed in the bleach effluent mix tank which is then mixed with the clarified effluent from the primary clarifier. The combined effluent is cooled in a cooling tower and pumped to the UNOX reactor. The pH of the cooled effluent is measured and automatically neutralized by adding sulphuric acid to the primary clarifier overflow or 50% caustic soda to the bleach effluent mix tank. The temperature of the effluent from the cooling tower is 32°C. Ammonium polyphosphate is added to the combined effluent sewer at a point between the effluent cooling system and the UNOX reactor. The UNOX oxygen activated sludge bio-reactor is comprised of two trains with four compartments per train. The system has a hydraulic retention time of approximately 6.5 hours. The first two compartments of each reactor have been made into one double-sized chamber by the removal of the first stage baffles. Pure oxygen is automatically metered to the reactor to maintain a constant static pressure in the first chambers. Oxygen is obtained from the on-site cryogenic plant. The biologically treated effluent from the UNOX reactor discharges into a splitter box which divides and directs the effluent into two parallel 67 meter diameter clarifiers which separate the biological solids from the treated effluent. The 20 clarified effluent flows by gravity to a foam tower and through an outfall and diffuser to the ocean. About 85% of the biological solids removed by the clarifiers is recycled back to the UNOX reactor. The waste solids from the primary clarifier and the solids from the secondary clarifiers are combined in the sludge blend tank and pumped to the sludge press. In the sludge press, the combined sludge is dewatered in two parallel lines each consisting of a rotary screen thickener followed by a sludge dewatering screw press. Effluent from the thickeners and screw presses is returned to the general sewer pump station for transfer to the treatment system. 3.2 Western Pulp Limited's Effluent Treatment System The main components of the secondary treatment facility are a 30 meter diameter primary clarifier, lift station, mix tank, a 32,500 cubic meter three train UNOX activated sludge reactor, two 55 meter diameter secondary clarifiers and a sludge handling system consisting of two rotary screen thickeners and screw presses. A schematic of the treatment system is given in Figure 3. The mill effluent flows into three major streams: fibre bearing, acid and caustic. The fiber bearing streams originate from the digester and bleach plant area and are pumped to the primary clarifier. The clarified effluent is then mixed with the remaining streams. Prior to secondary treatment the effluent is cooled and 21 II neutralized. The evaporator combined condensate, bleach plant acid, caustic extraction filtrate and the digester foul condensate are diverted through four plate heat exchangers. Additional direct cold water injection is available on a controller. Effluent pH is controlled by adding lime kiln scrubber slurry mixture to the acid effluent before the mix tank. The pH at the mix tank can also be adjusted with 93% sulphuric acid or 50% caustic. Anhydrous ammonia and phosphoric acid are added as nutrient sources for biomass growth. The UNOX is operated as a two train reactor providing 7 hours hydraulic retention time; the third train is left empty. In the event of a major spill the mill effluent could flow to the third train and then to the mix tank for treatment in the other UNOX trains. The clarified effluent from the secondary clarifiers flows through a foam trap and then to a deep water mulitport diffuser in the ocean. 23 4. ACTIVATED SLUDGE MODELLING 4.1 History Activated sludge was recognized as a feasible wastewater treatment method shortly after the laboratory experiments of Arden and Lockett in 1914. In their original work, it was recognized that physical, chemical and biological mechanisms were responsible in varying degrees for the purification of wastewater, although no attempts were made to identify their existence or relative importance (Mines and Sherrard, 1989). In the years that followed, explanations of the mechanisms of organic matter removal by activated sludge were based on incomplete experimental evidence, leading to considerable controversy. Many theories were proposed, though few gained support. One hypothesis, the coagulation theory, advocated the scrubbing action of the suspended solids in sewage. The adsorption theory advocated adsorption of colloidally dispersed matter in sewage. The colloid theory suggested that negatively and positively charged sewage colloids mutually coagulated. The biozeolite theory described a series of base exchange reactions with organic compounds and the activated sludge. In 1923 a biological mechanism was proposed as an alternate theory by Buswell and Long (as cited by Orhon and Artan, 1994). They observed filamentous and unicellular bacteria, as well as various protozoa and metazoa, and suggested that the purification was accomplished by synthesis of the organic matter from sewage into the living 24 material of the floe. The biological theory gained wide acceptance and the adsorption theory prevailed as the accepted mechanism to describe the initial step in organic removal, however, a lack of agreement on the relative importance of each process in the overall removal mechanism remained (Orhon and Artan, 1994). Models were first developed to mathematically express the different hypotheses proposed to explain the removal of organic matter by activated sludge. The most commonly recognized model for microbial growth is the one proposed by Monod in 1949, as cited by Tebbutt and Aziz (1980), from pure culture studies which implied a hyperbolic relationship between growth and substrate utilization. In the 1950's, researchers began to investigate the composition of sewage. Helmers et al. as cited by Orhon and Artan (1994), reported in 1951 that the rate of activated sludge growth for three industrial wastes was observed to be proportional to the BOD reduction. Using BOD as a measure, Garrett and Sawyer, as cited by Tebbutt and Aziz (1980), in 1952, determined that the kinetics of soluble substrate removal in a complex mixture by a mixed culture could be represented by the same type of expression proposed by Monod for pure cultures. Substrate removal was described as a two phase discontinuous function. 25 Also in 1952, Hoover and Porges , as cited by Goodman and Englande (1974), developed an empirical formula, C5H7N02,for the composition of activated sludge microorganisms. This formula and the metabolic balances presented by these researchers were important contributions to the understanding of the activated sludge process and subsequent modelling efforts (Goodman and Englande, 1974). In 1954, Eckenfelder and O'Conner, as cited by Goodman and Englande (1974), proposed a mathematical model for activated sludge wastewater treatment. This model was subsequently modified and expanded in publications by Eckenfelder during the period of 1960-1971. These publications established a nomenclature and mathematical approach to activated sludge process design that is still widely used today (Goodman and Englande, 1974; Tebbutt and Aziz, 1980). In 1962, McKinney, as cited by Orhon and Artan (1994), advanced another mathematical model for completely mixed activated sludge treatment systems. McKinney developed process kinetics for a completely mixed reactor using a rate expression which is first order with respect to substrate and zero order with respect to biomass. Between 1968 and 1970, further publications by McKinney followed, establishing a second nomenclature and approach to activated sludge process design that is also widely used today (Orhon and Artan, 1994). 26 McCabe, in 1963, as cited by Orhon and Artan (1994), suggested that the biological activities in the activated sludge process could be divided into two phases: (i) a log growth where the growth rate is only limited by the minimum generation time of the organism and (ii) a declining growth phase in which the BOD concentration starts to limit the growth rate. He pointed out that the transition from one phase to the other was continuous, implying the validity of a hyperbolic function in the overall description of the process (Orhon and Artan, 1994). The transformation of the numerous observations into process parameters that could be interpreted both from an operational and microbiological standpoint was made by Pearson in 1966, as cited by Goodman and Englande (1974). The incorporation of the long-recognized concepts into process material balances enabled him to quantify two basic parameters: the specific substrate removal rate and the net growth rate. These expressions allowed a graphical evaluation of the growth and decay coefficients for data produced under different steady-state conditions. This led to the adoption of sludge age as one of the key modelling and design parameters for the activated sludge process. Attempts have been made to establish a functional relationship between the overall specific growth rate of a culture and the substrate concentration. One of 27 the first groups to recognize the importance of metabolic products in activated sludge effluents was Chudoba et al. through their research efforts between 1967 and 1969 (Chudoba, 1985a). Their experiments revealed metabolic product production in response to 17 specified compounds (Chudoba, 1985b). Metabolic intermediates were also observed for continuous flow systems at steady state degrading multi-component substrates (Grady and Williams, 1975). It was believed that a portion of the assimilated organic matter removed from the waste stream during initial contact was stored by the microorganisms. The soluble COD in the effluent appeared not to be a fraction of the influent substrate which is depleted, but is presumably organic matter released through microbial activities (Rittman etai, 1987). Marais and Ekama (1976) stated that at steady state, the rate of substrate storage is balanced by the rate of utilization of stored substrate. They deduced that substrate removal must therefore rely on the mechanism of microbial growth. A new concept of substrate composition was proposed: readily biodegradable substrate, composed of soluble compounds which can be readily taken up by the cells for metabolic activity, and slowly biodegradable substrate consisting of soluble and colloidal compounds with large and complex structures requiring enzymatic hydrolysis prior to adsorption (Dold etai, 1980). A three-step mechanism involving hydrolysis, adsorption and synthesis was proposed for the utilization of slowly biodegradable substrate. The rate-limiting step was identified as hydrolysis. 28 The contributions of Dold, Marais and Ekama for aerobic systems consolidated much of the previous research. The results of their research formed the basis for the IAWPRC Model No. 1 which is outlined in the following section. 4.2 IAWPRC Model Number 1 The IAWPRC formed a task group in 1983 to promote the development of, and facilitate the application of practical models for the design and operation of biological wastewater treatment systems. The goal was to review existing models and to develop a simple model having the capability to predict the performance of single sludge systems carrying out carbon oxidation, nitrification, and denitrification. The group chose to cite only those papers that consolidated the work of previous investigators or those individual contributions which represented important new concepts. There has been a long transition period between the promotion of the activated sludge method of waste treatment and the establishment of a theoretical framework that both quantitatively describes the process, and provides a rational basis for its design. The development of the IAWPRC model dates back 12 years. From its start, wastewater from municipal sources, settled and unsettled, served as the substrate source. The conflicting nature of the many hypotheses for the mechanistic explanation of the process, the difficulty of expressing them in discrete mathematical models, and the contrived nature of the systems on 29 which the models were developed were the main reasons for this slow transition. The following is a review of the IAWPRC Model No. 1. 4.2.1 Basic Concepts of the IAWPRC Model No. 1 4.2.1.1 Heterotrophic Death Biomass is lost by decay, which incorporates a large number of mechanisms including endogenous metabolism, death, predation and lysis. These factors are generally considered together as a single parameter, the decay coefficient (bH). This parameter is important for the prediction of sludge production and oxygen requirements within a system. The original concept of death followed the endogenous respiration approach of Pirt (1965) and Herbert (1958), as cited by Lishman and Murphy (1994), that organisms feed on other microorganisms in order to maintain their basic energy requirements when there is no other biodegradable substrate available. This process generates an non-biodegradable portion which accumulates in the reactor as inert endogenous residue. Typically about 20% of the biomass formed is considered to contribute to the inert residue (Henze et al., 1987). The IAWPRC model has adopted the death regeneration concept of Dold et al. (1980). Decay is assumed to result in the conversion biomass to a combination of inert particulate products, / p , and slowly biodegradable substrate which re-enters the cycle of hydrolysis, growth, etc. The loss of biomass by decay is 30 assumed to occur at a rate which is independent of the nature or concentration of the electron acceptor present, but hydrolysis of the resultant slowly biodegradable substrate to readily biodegradable substrate occurs only under aerobic conditions. In order for the fraction of inert products formed per net unit mass of mixed liquor volatile suspended solids (MLVSS) to equal 0.20, the fraction formed during each passage around the cycle must be less than 0.20. This follows from the fact that the observed fraction equals: 0.20 -1-YH(1-/ ') Equation 4-1 where Y H = the heterotrophic growth yield for heterotrophic microorganisms (mg cell COD/mg COD) (Henze et al., 1987). If the observed fraction is 0.20, then the value of / p for the model must be 0.08 mg COD/mg cell COD (Dold and Marais, 1986). This fraction does not vary between wastewaters as it is a characteristic of the biomass (Henze et al., 1987). 4.2.1.2 Heterotrophic Growth Microbial growth occurs by metabolism of available substrate. A portion of this substrate is converted to cell components, whereas the remainder is oxidized to provide energy for synthesis and cell maintenance. The growth yield defines the proportion of available substrate that is converted to cell mass. The maximum 31 rate of heterotrophic growth is defined by u,max. This parameter also determines the maximum oxygen requirements of a biological system. The half saturation coefficient (Ks) is the concentration of substrate at which the growth rate is half the maximum growth rate. It indicates the readiness with which microorganisms can assimilate their substrate. A low value of Ks indicates that a particular substrate is readily biodegradable. Biomass growth is assumed to follow Monod kinetics by the following equation: Ss JU - //max* Ks + Ss Equation 4-2 where \i = specific growth rate of heterotrophic microorganisms (d'1); S s = readily biodegradable substrate concentration (mg/L). The specific growth rate is related to the specific substrate utilization rate by the following relationship: Equation 4-3 where q = specific substrate utilization rate (d'1). 32 The IAWPRC activated sludge model assumes that the growth rate is a function of only the readily biodegradable substrate. It is therefore important to fractionate the influent substrate into degradable and non-degradable components. 4.2.1.3 Influent Wastewater Fractionation The carbonaceous strength of the influent is defined in terms of its electron donor capacity, in practical terms expressed in equivalent form by the COD (Dold and Marais, 1986). Influent COD can be divided into two main fractions, based on biodegradability as outlined in Figure 4. The non-biodegradable matter is biologically inert and exits the activated sludge system unchanged. The physical state of the non-biodegradable material defines the next sub-division. Inert soluble organic matter S i ( leaves the system at the same concentration it enters. Inert particulate matter, Xi, becomes enmeshed in the activated sludge and is removed from the system through wastage (Dold and Marais, 1986). The biodegradable COD also has two sub-fractions: slowly biodegradable and readily biodegradable. This division is made on the basis of an observed biological response, not on physical separation (Ekama etal., 1986). All of the influent particulate material, biodegradable and non-biodegradable, is assumed to be enmeshed immediately in the sludge matrix and is completely removed from the effluent (Dold et al., 1980). The two biodegradable COD 33 fractions are utilized for synthesis of new cell mass. The rates of utilization of the two, however, differ greatly. INFLUENT COD BIODEGRADABLE COD 1 NON-BIODEGRADABLE COD READILY BIODEGRADBLE COD 1 SLOWLY BIODEGRADABLE COD SOLUBLE INERT COD PARTICULATE INERT COD Figure 4. Influent COD Fractions Readily biodegradable matter consists of simple molecules that may be taken up directly by the heterotrophic bacteria. Readily biodegradable material is considered to be the only substrate used for growth of new biomass and removal of readily biodegradable substrate is considered to be proportional to growth. A portion of the energy associated with the substrate molecules is incorporated into new biomass, and the balance is expended to provide the energy needed for the synthesis. The electrons associated with that portion are transferred to the exogenous electron acceptors (oxygen or nitrate). Using the absorption-synthesis concept, growth rate in terms of readily biodegradable substrate utilization is as expressed as: fJSs Umax Ss = * A H dt Ks + Ss Equation 4-4 34 where X H = heterotrophic biomass (mg COD). This reaction is very rapid (Dold and Marais, 1986). Slowly biodegradable matter, X s , is composed of a mixture of solid, colloidal and soluble organic particulate of complex organic structure. This material cannot pass through the cell wall and is enmeshed immediately in the sludge matrix. The enmeshed particulate biodegradable material is adsorbed and stored on the active heterotrophic organism mass at a finite rate to yield stored substrate, Xs (Dold and Marais, 1986). The rate of particulate biodegradable adsorption is expressed as a mass transfer process: = K A X E S X H (1-XS/XH) dt Equation 4-5 where KA = substrate transfer rate (mg COD/mg VSS«d); X E s = enmeshed particulate biodegradable material (Dold and Marais, 1986). The stored material is hydrolyzed by the extracellular enzymes fixed to the surface of the organisms. The resulting hydrolyzed material is absorbed directly for synthesis. The conversion (hydrolysis) involves no energy utilization so there is no utilization of electron acceptor associated with it. 35 Although three processes are involved: hydrolysis, absorption and synthesis; the three are treated as one, in the same fashion as absorption and synthesis are treated as one in the Monod growth rate expression. The governing rate is that of hydrolysis; based on Levenspiel's active site surface reaction kinetic theory (Dold and Marais, 1986). The rate of stored particulate substrate utilization is expressed as: dXs dt Equation 4-6 where K m p = maximum specific growth rate constant for the utilization of stored COD (mg COD/mg VSS«d); K s P = half saturation coefficient for the utilization of stored COD (mg COD/mg VSS). The specific rate of hydrolysis of slowly biodegradable substrate is usually much slower than the specific rate of utilization of readily biodegradable substrate so that it becomes the rate-limiting factor in the growth of biomass when Xs alone is present as substrate. The division of substrate into two forms provides a built-in lag in uptake of electron acceptor which allows space-time dependent variations in oxygen and nitrate utilization to be modelled. Kmp(Xs/XH) K S P + ( X S / X H ) J * X H 36 4.2.2 Model Design To be mathematically traceable while providing realistic predictions, reactions must be representative of the most important fundamental processes occurring within a system. The term process is used to mean a distinct event acting upon one or more system components. The model is based on material balance equations. These equations quantify both the kinetics (rate-concentration dependence) and the stoichiometry (relationship that one component has to another in a reaction) of each process. The fundamental processes considered in the model for heterotrophic biomass are growth of biomass, decay of biomass and hydrolysis of slowly biodegradable organics. To describe the model a simple matrix is presented from the IAWPRC report (Henze et al., 1987) in Table 1. The matrix describes the simplest heterotrophic activities that are modelled. The fundamental process of aerobic growth according to Monod kinetics and decay according to first order kinetics are described. These have the index, / The components heterotrophic biomass, dissolved organic substrate and dissolved oxygen are included. These have the index, /'. and are introduced in the form of an array of concentrations in the first row of the matrix combined with their measurement in the last row. The elements within the matrix comprise the stoichiometric coefficients, vy, which set out the mass relationships between the components in the individual processes. From the matrix presentation of these components and processes, kinetic 37 expressions can be developed and are listed in the right hand side of the matrix with the index, p;. Table 1. Process Kinetics And Stoichiometry For Heterotrophic Bacterial Growth In An Aerobic Environment (Henze et al., 1987) Component 1 2 3 j Process I X H S s So Process Rate, pj \ [ML:3I!I 1 Growth 1 -1/YH -1 - Y H Y H M-may Ss * X H Ks + Ss ! 2 Decay -1 -1 bH X H I Observed | Conversion I Rates ML ' 3-J-1 n = j j Biomass [M(COD)/L-3] Substrate [M(COD)/L-3] Oxygen [M(-COD)/L-3] Stoichiometric Parameters: true growth yield : Y H Kinetic Parameters: maximum specific growth rate : p:max half-velocity coefficient : Ks specific decay coefficient : bH The IAWPRC Model No. 1 describes the processes of carbon oxidation, nitrification and denitrification. To this end, the model matrix is more complex than the simple matrix described here. A total of seven dissolved components, six particulate components and nine processes are included to characterize the 38 wastewater and activated sludge. To apply the model for use in the design and operation of wastewater treatment systems, parameter values which are wastewater specific and concentrations of important components in the influent must be evaluated. 39 5. PROCEDURES FOR THE ESTIMATION OF MODEL PARAMETERS 5.1 IAWPRC Methods The following section describes the procedures recommended by the IAWPRC task group (Henze et al., 1987) to determine the parameters required to describe carbon oxidation using the Model No. 1. The parameters to be determined as inputs to the Model No. 1 to describe carbon oxidation are listed in Table 2. The evaluation should proceed in the order given as some of the parameter values are required before others can be evaluated. Henze et al. (1987) advocated the use of laboratory scale completely mixed continuous flow activated sludge reactors to generate the model parameters. When operated in an aerobic mode at a number of different solids retention times (SRTs), the effect of the net specific growth rate on the electron acceptor requirement and the sludge production rate can be determined. The data obtained can be used with other tests to characterize the wastewater and evaluate the stoichiometric coefficients. 5.1.1 Heterotrophic Yield The heterotrophic growth yield is estimated by observing the mass of cell material formed during removal of soluble substrate in a batch reactor. An 40 Table 2. Parameters For The Model And The Order In Which They Should Be Determined Symbol Name Prior Information Needed Y H Yield for heterotrophic biomass bH Decay coefficient for heterotrophic biomass Y H , fp S s Concentration of readily biodegradable COD in wastewater Y H Si Concentration of soluble inert COD in wastewater Xi Concentration of inert suspended organic matter in wastewater fp, bn, Ss, Si X s Concentration of slowly biodegradable organic matter in wastewater Ss, Si, Xi Umax Maximum specific growth rate for heterotrophic biomass Y H , fp, X s , Si, Xj Ks Half-saturation coefficient for heterotrophic biomass Y H , fp, X s , Si, Xj kh Maximum specific hydrolysis rate Kx Half-saturation coefficient for hydrolysis of slowly biodegradable substrate aliquot of filtered wastewater is seeded lightly with acclimated biomass from a completely mixed continuous flow reactor. The growth yield is estimated from the analysis of periodic aliquots of both total and soluble COD: Cell COD = Total COD - Soluble COD Equation 5-1 41 A cell COD Y H = A soluble COD Equation 5-2 If this is done several times, an approximate Y H value may be determined. 5.1.2 Decay Coefficient The decay coefficient is very important for predictions of sludge production and oxygen requirements, so it must be determined for the sludge in use. Sludge is removed from a continuous flow reactor and put into a batch reactor where the oxygen utilization rate (OUR) can be measured many times over a period of several days (Ekama etal., 1986). Nitrification is inhibited during the test by the addition of 20 mg/L of allylthiourea and the pH is maintained at a constant value near neutrality. The batch aerobic digestion method provides no external substrate. Under conditions where there is no accumulation of intermediates, the rate of decline in measured OUR is equivalent to the net endogenous respiration rate, or traditional decay coefficient, b'H (Dold etal., 1980). The following expression may be written for the OUR, if X H is described in terms of the volatile suspended solids concentration (VSS): O U R = 1.42(1-/ P)b'HXH Equation 5-3 The process rate expression for endogenous decay can be used to define X H in terms of its initial concentration, XHo, at the beginning of the experiment: 42 X H - X H o e b H ' Equation 5-4 where t = time (d). Substituting the value of X H above, the equation is rearranged as: In OUR = ln[l .42(1 - /„)b' H X H o ] - b ' H t Equation 5-5 (Lishman and Murphy, 1992). Equation 5-5 shows that for batch digestion of biosolids a plot of the natural logarithm of the oxygen uptake rate versus time yields a straight line with slope b'H and intercept [(1-fp )b'H XHo ]• The model decay coefficient, bH, which incorporates the death-regeneration concept, can be derived from the traditional decay coefficient using the following equation: b „ = b ' „ / l - Y „ ( l - / , ) Equation 5-6 once the growth yield has been determined. 5.1.3 Readily Biodegradable Substrate The concentration of readily biodegradable substrate in the influent can be estimated using the continuous flow activated sludge process method of Ekama et al. (1986). The method involves measuring the change in OUR in a single completely mixed reactor operated at an SRT of 2 days under a daily cyclic square wave feeding pattern (12h with feed; 12h without feed). The procedure is based on the hypothesis that the specific substrate utilization rate associated with heterotrophic biomass is high enough to deplete all of the readily biodegradable substrate supplied to the system. The OUR level monitored 43 during feeding is controlled by the readily biodegradable substrate entering the reactor through the influent stream plus that released by the hydrolysis of slowly biodegradable substrate. After feed termination there is a rapid drop in the oxygen uptake rate. This is because any accumulated readily biodegradable substrate is rapidly used. The OUR will not drop to zero, however, because the accumulated slowly biodegradable substrate will continue to be used at the same rate for a period of time. Thus, the immediate drop in OUR is associated only with the readily biodegradable material and can be used to find its concentration: _ AOUR*V as = 7 r-Q(1-YH) Equation 5-7 where AOUR = change in OUR following feed termination; V = reactor volume; Q = feed flow rate prior to termination. The above expression reflects the basic principle that 1 g of biodegradable COD is consumed at the expense of (1 - Y H ) g of dissolved oxygen. The value of Y H must be estimated before the calculation of S s . 5.1.4 Inert Soluble Organic Matter To estimate the concentration of inert soluble organic matter an aliquot of the reactor contents from a completely mixed continuous flow reactor treating the 44 wastewater at an SRT in excess of 10 days is aerated in a batch reactor. Samples are removed periodically and analyzed for soluble C O D until the concentration remains constant. The final soluble C O D is the inert material which is assumed to be equal to the concentration in the feed. 5.1.5 Inert Particulate Organic Matter The concentration of C O D contributed by inert suspended organic material is evaluated as a parameter for fitting the model to data showing the effect of S R T on the sludge production. This is done by using a one dimensional search routine which chooses Xj to minimize the error sum of squares when predicted sludge production rates are compared to measured rates as a function of SRT. This fitting acts to tune the model to the particular wastewater under study and compensates for any errors made in the estimation of the growth yield or decay coefficients. 5.1.6 Slowly Biodegradable Substrate The concentration of slowly biodegradable substrate in the influent, Xs, can be calculated by difference once the other fractions have been determined using the following formula: Total Influent COD = S s + X s + X s + S, Equation 5-8 45 5.1.7 Maximum Specific Growth Rate and Half Saturation Coefficient The maximum specific biomass growth rate is difficult to evaluate accurately, but its value is not critical because the model is not very sensitive to this parameter; i.e. an error factor of 2 or 3 will have little impact on model predictions (Henze et al., 1987). The main function of u m a x is to allow the maximum OUR to be predicted. The procedure of Chudoba etai. (1985) is suggested to evaluate the maximum specific growth rate for heterotrophs. Acclimated biomass from a laboratory activated sludge reactor is placed in a closed respirometer. The vessel is oxygenated until a high dissolved oxygen concentration is obtained. The oxygen supply is turned off and the endogenous respiration rate is measured. After establishment of the endogenous rate, appropriate dilutions of the wastewater are contacted with the biomass. In response to the substrate, the biomass increases its respiration rate temporarily until the substrate is exhausted. By proper choice of the biomass concentration, the change in cell mass may be negligible so that the oxygen consumption rate returns to the original endogenous rate upon substrate exhaustion. The difference between the endogenous rate and the oxygen consumption rate immediately after introduction of the substrate is the net endogenous respiration rate associated with a particular substrate concentration. It can be seen that when a single substrate is added the OUR is constant until the substrate is almost exhausted. Furthermore, the net rate is a measure of the rate of substrate consumption 46 associated with the concentration added, and thus is proportional to the specific substrate removal rate associated with that concentration. Injections are repeated using different substrate concentrations until the entire curve of respiration rate versus substrate concentration has been defined. The data can then be analyzed to obtain the maximum respiration rate and the half-saturation coefficient. 5.1.8 Maximum Specific Hydrolysis Rate and Half-Saturation Coefficient for Hydrolysis of Slowly Biodegradable Organic Matter Once all the above parameters are determined, the maximum specific hydrolysis rate, kh, and half-saturation coefficient for hydrolysis of slowly biodegradable organic matter, K x, are estimated by curve-fitting techniques. These match the response of the model to the oxygen uptake pattern produced under the square-wave feeding pattern used for the determination of the concentration of readily biodegradable substrate in the influent. 5.2 Alternate Methods The authors of the Model No. 1 report (Henze et al., 1987) note that some of the techniques recommended for parameter determination are provisional and, that as more experience is gained in the use of the model, modifications to these techniques will develop. Numerous investigators have proposed alternate methods to those in the Model No. 1, both prior to and since its publication. Of interest to this research project were the procedures presented using batch tests 47 on full scale or pilot plant activated sludge systems. The following section reviews some of the alternate methods proposed for model parameter determination. 5.2.1 Heterotrophic Growth Yield To determine the true growth yield, Y H , Slade etai. (1991) measured a series of observed growth yields, Y o b s , from which the true growth yield was calculated. A sample of filtered wastewater was inoculated with a small amount of acclimated biomass from a laboratory activated sludge reactor. The batch reactor was stirred continuously and wall growth was regularly resuspended. The total COD and soluble COD were measured immediately and then at regular intervals. By 150 hours, both COD concentrations had reached a plateau. The observed growth yield was calculated by the following equation: final cell COD - initial cell COD o b s ~ initial soluble COD - final soluble COD Equation 5-9 The initial cell COD is the difference between the total and soluble COD at the beginning of the test. As the sample was filtered before seeding, the particulate material is solely due to the added biomass. The final cell COD is the difference between the total and soluble COD at an appropriate time during the test. The difference in the soluble COD is simply the difference between the initial soluble COD and the soluble COD at that time. 48 Following the linearisation method of Fieschko and Humphrey (1984), the reciprocal of the observed growth yield values was plotted against time and the true growth yield, Y H , was calculated as the reciprocal of the y-intercept. 5.2.2 Decay Coefficient For Heterotrophic Microorganisms Slade etal. (1991) followed the method recommended by Henze etal. (1987) to examine the effect of solids retention time on bH. Three reactors were run simultaneously at three different solids retention times: 8, 12 and 15 days. Mixed liquor wasted from each reactor was transferred to a respirometer and oxygen uptake rates were measured twice daily for 11 days. The natural logarithms of the oxygen uptake rates were plotted against time and the slope of the line taken as b'H l the traditional decay coefficient. Their results showed no effect of SRT on decay rate. 5.2.3 Maximum Specific Growth Rate And Half-Velocity Coefficient The kinetic parameters for maximum substrate removal and the half-velocity coefficient are typically evaluated by measuring the active biomass and the rate limiting substrate concentration for continuous flow steady state reactors operated at various solids retention times. Because a considerable lag time is required to reach steady state, periods of up to several months are required to gather sufficient data for accurate estimates of the kinetic coefficients. (Knudson etal., 1982). 49 The use of the nonspecific parameters VSS and COD has been the limitation of the conventional approaches for the estimation of kinetic constants. The bulk parameter COD does not permit the determination of growth limiting substrate. VSS cannot be used to measure accurately the active biomass in an activated sludge process as factors such as raw waste characteristics, wastewater temperature and SRT can cause the ratio of active to inactive solids to change appreciably from plant to plant, or within a given plant, as changes occur in the waste or treatment environment (Huang et al., 1985). A number of other parameters have been proposed as viability measurements. For example, plate counts (Walker and Davies, 1977), adenosine triphosphate (ATP), and dehydrogenase enzyme activity (Weddle and Jenkins, 1971) have been proposed. However, the conversion of these viability and activity parameters into viable biomass concentrations in the terms used in modelling, incorporated significant uncertainties. These difficulties in the interpretation of COD and VSS measurements encourage the development of experimental procedures using respirometric measurements for the assessment of kinetic constants. Slade et al. (1991) used a modified form of the procedure of Chudoba et al. (1985) described in the IAWPRC report (Henze et al., 1987). A continuous flow activated sludge reactor was operated on a "square wave" pattern consisting of 12 hours with feed, 12 hours without feed (Henze et al., 1987). Mixed liquor from the reactor was transferred to a respirometer and the background oxygen 50 uptake rate recorded. The required volume of substrate was injected and the change in the dissolved oxygen recorded for approximately two minutes. To prevent a change in the concentration of the biomass in the respirometer, the substrate was freeze concentrated prior to injection. The Monod equation was linearized to produce a direct relationship between the inverse of the growth rate and the inverse of the substrate concentration: H S s Equation 5-10 The specific change in OUR, AOUR s, was measured for each substrate concentration and converted to a growth rate by the following relationship: // = - ^ - * A O U R , 1 - Y Equation 5-11 The value of u, for each S s value was represented by a line on a graph passing through - S s on the Ks axis and p. on the p m a x axis. The values of u.max and Ks were determined from the corresponding values on axes at the intersection of the lines. The p.max values determined using this method compared well with those generated by inserting the values of u.max and Ks into the Monod equation, however, poor reproducibility in the estimated p m a x values was noted. Kappeler and Gujer (1992) developed another batch test method based on respirometry involving centrifuged wastewater and a very small amount of 51 biomass obtained directly from a pilot plant. A COD:VSS ratio of 4 was used. In this method it is assumed that unlimited heterotrophic growth can be sustained in a batch reactor for a period over which the readily biodegradable substrate concentration remains sufficiently high. During the first period of the batch test, oxygen respiration increases due to the unlimited heterotrophic growth. The rate then decreases to a lower level because of limiting concentrations of readily biodegradable substrate. At this level oxygen respiration is dominated by growth on substrate which is released by hydrolysis. Oxygen respiration in a batch-test with neither substrate nor oxygen limitation is: In this experiment the oxygen uptake rate depends only on heterotrophic biomass. The mass balance for the heterotrophic biomass in a batch-test with neither substrate nor oxygen limitation is as follows: ro2(t) = - [ ( l - Y H ) / Y H ] > * / " m » * X H ( t ) - ( l - / p ) * b H * X H ( t ) Equation 5-12 dX, H dt Equation 5-13 Integration of Equation 5-13 with XH(t0) = XHo leads to: X„(t) = X H o * e l >n»-bH)*0 Equation 5-14 52 The above equation can be substituted into equation 5-14 to calculate the oxygen respiration at any time without limitation: r o 2 M = "[((I- Y „ ) / Y H ) * ^ m a x - ( l - / p ) * b H ] * X H / e « ^ - b » ^ Equation 5-15 The respiration of heterotrophic biomass at any time without limitation can be compared with the initial respiration: ro 2(t)/r O j(t 0) = e ^ - b » H Equation 5-16 The logarithmic configuration of Equation 5-16 is: Mro2(0/ro2(O] = ( / W - b H ) * t Equation 5-17 This equation represents a straight line with (u.max - bH) as slope. The maximum specific growth rate of heterotrophic organisms depends on reactor configuration, temperature and SRT (Kappeler and Gujer, 1992). Results generated for domestic wastewater using this method compared well to the values of u.max reported by Henze et al. (1987) in the Model No. 1. 5.2.4 Wastewater COD Fractions There are many methods available in the literature to fractionate the COD in wastewater. Table 3 provides a partial listing of the various methods in the literature. 53 Table 3. Methods For Estimating Organic Fractions In Wastewater Organic Test \ Reference Soluble Inert Batch Batch/Continuous Batch Continuous Batch Henze etal., 1987 Germirli etal., 1991 Chudoba, 1985a Ekama etal., 1986 Lesouef et al., 1992 Readily Biodegradable Dynamic Continuous OUR Continuous OUR Batch OUR Specific Compounds & Mw Batch OUR Batch Chemical - Physical Ekama and Marais, 1977 Sollfrankand Gujer, 1991 Kristensen etal., 1992 Henze etal., 1992 Kappeler and Gujer, 1992 Mamais etal., 1993 Slowly Biodegradable Continuous OUR Soluble-Inert Soluble Dynamic Continuous OUR [ Batch OUR | Calculation from Mass Balance Sollfrankand Gujer, 1991 Lesouef et al., 1992 ! Ekama etal., 1986 I Kappeler and Gujer, 1992 | Henze etal., 1987 Particulate Inert | Calibrated with Sludge Production \ Batch \ Continuous & Model Calibration 1 Batch | Henze et al., 1987 | Kappeler and Gujer, 1992 ! Ekama etal., 1986 j Lesouef et al., 1992 5.2.4.1 Readily Biodegradable COD The fraction of readily biodegradable COD defines the mass of substrate that is immediately available to the microorganisms for growth. The measurement of readily biodegradable substrate can be made in two ways: indirectly, by respiration measurements as originally done by Ekama and Marais (1977) and further developed by Ekama et al. (1986) and Sollfrank and Gujer (1991) or directly, through filtration methods (Dold etal., 1980; Mamais etal., 1993). 54 The continuous flow activated sludge process method of Ekama et al. (1986) was outlined in the IAWPRC Model No. 1 (Henze etal., 1987) for determination of the influent readily biodegradable COD concentration. The procedure involves observing the OUR profile in a completely mixed activated sludge system with sludge recycle, operated at a sludge age of around two days under daily cyclic square-wave feeding conditions (i.e., 12 hours feed; 12 hours without feed). It is based on the hypothesis that the specific substrate utilization rate associated with heterotrophic biomass is high enough to deplete all the readily biodegradable substrate supplied to the system. The aerobic batch test method (Ekama et al., 1986) consists of running an aerated batch reactor with a mixture of preselected volumes of wastewater and mixed liquor, and monitoring the OUR profile approximately every 10 minutes for about four to five hours. During the first phase of the test the OUR level is high as the microorganisms consume the readily biodegradable substrate initially present in the wastewater. The OUR remains approximately constant for a period after which the OUR drops to a second plateau which correlates to the consumption of readily biodegradable COD generated by hydrolysis. The readily biodegradable concentration may be calculated from the equivalent oxygen consumption. This concentration is equal to the area between the OUR curve and a horizontal line projected to the vertical axis at the level of the second OUR plateau: 55 (1 -Y H ) V „ Equation 5-18 where A 0 2 = mass of oxygen consumed by S s (area under OUR curve); Vww = volume of wastewater in the mixture; Vmi = volume of mixed liquor in the mixture. The horizontal line implies that the oxygen consumption attributed to hydrolyzed substrate and endogenous respiration stays constant during the test. This is not an entirely correct assumption. Both the rate of hydrolysis and endogenous respiration are likely to drop as a result of the decrease in the amount of slowly biodegradable COD and biomass in the course of the experiment (Kappeler and Gujer, 1992). The interpretation of the OUR curve is very dependent on the accuracy of determining the point where Ss is depleted. The accurate measurement of S s also depends upon the selection of an appropriate value for the initial food to microorganism (F/M) ratio. If the F/M is too high, it may not be possible to accurately depict the initial high OUR level, whereas if it is too low, the difference between the high and the low OUR levels may not be accurately quantified (Ekama etai., 1986). 56 Slade et al. (1991) examined both the flow through and the aerobic batch method but reported only the results of the latter in their paper. The authors state that the aerobic batch test was found to produce consistent results. An F/M ratio of 0.1 was used in their experiments with pulp mill effluent. Readily biodegradable COD consists of relatively small molecules, whereas slowly biodegradable COD comprises larger and more complex molecules. Though the separation of these two fractions in the IAWPRC model is based on a biological response and not on a physical separation (Ekama etai., 1986), the use of filtration methods has been proposed by several researchers as an approximation of the biokinetic division. Dold etai. (1980) evaluated the use of 0.45 u,m filters on domestic wastewater and noted that a fraction of the slowly biodegradable COD passed through the filter causing overestimation of the readily biodegradable COD concentration. The use of ultramembranes or gelfiltration to separate COD into a molecular weight fraction < 10,000 appears to produce good estimates of the readily biodegradable substrate concentration in domestic waste streams (Dold et al., 1986). Facilities, for this type of ultrafiltration, however, are not widely available. To overcome these problems, Mamais etai. (1993) investigated flocculation of the slowly biodegradable COD before filtration through 0.45 urn filters. The batch method described by Mamais et al. (1993) is based on the rationale that 57 membrane filtration of a sample that has been flocculated will produce a filtrate containing only truly soluble organic matter. The method is based on the following assumptions: 1) influent total truly soluble COD consists of a readily biodegradable fraction and a non-biodegradable fraction (Henze etai., 1987) 2) the non-biodegradable fraction is equal to the truly soluble effluent COD from an activated sludge treatment plant treating the influent at a mean cell retention time (MCRT) > 3 days (Ekama et al., 1986) The procedure involves removal by flocculation and precipitation of colloidal matter that normally passes through a 0.45um membrane filter. Readily biodegradable COD is related to the truly soluble influent COD by the equation: S s = CODSoi - Si where S s = influent readily biodegradable soluble COD; CODsoi = influent total truly soluble COD; Sj = influent non-biodegradable soluble COD. The flocculation method determines COD S Oi in the equation. Results from four domestic wastewaters yielded the same results using both the physical-chemical method (Mamais etai., 1993) and the flow-through activated sludge method (Ekama etai., 1986). 58 5.2.4.2 Soluble Inert COD Ekama et al. (1986) stipulated that the influent inert soluble COD should be equivalent to the COD of the filtered effluent from a laboratory completely mixed reactor system operated at sludge ages between 10 and 20 days. Lesouef et al. (1992) proposed that samples be withdrawn from the influent stream to an activated sludge system and aerated in batch reactors for several days until a final soluble COD level is achieved. For municipal wastewater an aeration time of 10 days is suggested. These types of methods have the major drawback of not differentiating between soluble inerts in the effluent and soluble residual fraction of microbial products. They are measured together. This simplified approach, however, may be accepted for domestic wastewaters provided that the existence of residual products is accounted for in the determination of other organic fractions (Germirli etal., 1991). The soluble inert component may be negligible in some industrial wastewaters and in the effluent from the chemical treatment step of multistage treatment schemes commonly used for complex strong wastes. For such wastes, the procedure proposed by Germirli et al. (1991) may be followed. It consists of running parallel aerated batch reactors. One containing filtered wastewater and glucose seeded with a very small amount of biomass previously acclimated to the glucose-wastewater mixture, and one containing only wastewater with a very 59 small amount of acclimated biomass. Glucose is the central compound in the metabolic pathways of the biodegradation of all organic matter has been used in a number of previous experiments to prove the existence and significance of the generation of microbial products (Chudoba, 1985; Daigger and Grady, 1977; Grady and Williams, 1975). At the end of the experiment, the soluble COD concentration in the wastewater-glucose reactor decreases to a concentration which is equal to the sum of the initial soluble inert material present in the wastewater and the soluble inert microbial products produced. The soluble COD concentration in the glucose reactor decreases to a concentration equal only to the produced soluble inert microbial products since it contains no initial soluble inert material. The initial soluble inert fraction of the wastewater is calculated as the final soluble COD concentration of the wastewater-glucose reactor less the final concentration of the glucose reactor. 5.2.4.3 Particulate Inert Substrate Significant research effort has been devoted to the assessment of the particulate inert organic substrate concentration. The measurement of this fraction can be done through a batch test (Kappeler and Gujer, 1992), by calibration of the model with respect to the observed sludge production (Henze et al., 1987), or by the tedious continuous technique of Ekama et al. (1986). Ekama et al. (1986) proposed calculating particulate inert COD by comparing the measured MLVSS concentration to the calculated value on the basis of process 60 kinetics. A similar approach based upon the comparison of observed and calculated sludge production is outlined in the Model No. 1 report (Henze et al., 1987). These procedures require that three kinetic constants, Y H , bH and / p be correctly determined by independent experimental methods. An estimation based on standard laboratory analyses of these fractions is preferable. A simplified method to determine the refractory influent compartments is the measurements of COD in the batch test procedure of Lesouef et al. (1992). Two aerated batch reactors, one started up with unfiltered wastewater and the other with filtered wastewater are required. In each reactor, total and soluble COD are monitored for a period long enough to ensure the depletion of all biodegradable substrate. The inert particulate matter is the difference between the final total COD of the non-filtered sample, from which is deducted the soluble inert fraction and the produced biomass. This simple test yields values of non-biodegradable soluble and particulate matter comparable to those proposed in Model No. 1 (Lesouef et al., 1992). 5.2.4.4 Slowly Biodegradable Substrate The measurement of slowly biodegradable substrate can be accomplished through iterative curve fitting to match the response of the OUR curve produced using the flow through activated sludge method of Ekama et al. (1986) or the batch method of Kappeler and Gujer (1992) described for the determination of 61 readily biodegradable substrate. However, since influent wastewater is made up of readily biodegradable substrate, slowly biodegradable substrate, inert suspended substrate and inert soluble organic matter, the simplest method to calculate the slowly biodegradable content is by a COD mass balance, assuming the other organic concentrations have been estimated as recommended by Henze et al. (1987). 5.2.5 Conclusion Due to the increased use of models for activated sludge systems since the inception of the Model No. 1, several methods for characterizing the input parameters have been developed. The use of simple on-site batch tests was of interest to this research. Based on the review of the methods presented in this and the previous section, several batch test methods described were chosen to determine the model input parameters to describe carbon oxidation as outlined in the Activated Sludge Model No. 1 (Henze etal., 1987). 62 6. EXPERIMENTAL PROGRAM 6.1 Experimental Design The research was set up in two phases. In the first phase, the experiments were conducted in the laboratory at the Pulp and Paper Research Centre on the University of British Columbia campus. During this phase, the reactor designs were developed and the batch test methodologies were refined using wastewater collected from the Western Pulp Partnership Limited operation in Squamish, British Columbia. In the second phase, the experiments were conducted at the Howe Sound Pulp and Paper mill site in Port Mellon, British Columbia. During this phase an extensive database of wastewater and biomass characteristics was developed. Western Pulp wastewater was collected every two weeks between July and November 1995. Samples for filtered, unfiltered, and readily biodegradable COD were prepared at the mill site and preserved in acid for transportation to the University for digestion and analysis. Primary effluent, mixed liquor and return activated sludge (RAS) samples were collected in Nalgene sample bottles and transported in a cooler during the three hour car trip back to the University. The tops of the bottles containing the return activated sludge samples were fitted with an aeration port which was utilized in conjunction with a portable air pump connected to the vehicle's battery. 63 Upon arrival at the University, the decay, growth yield and particulate inert COD experiments were set up. The biomass for the maximum specific growth rate determination was aerated and maintained at room temperature overnight. The wastewater was also maintained at room temperature overnight. The following day, the wastewater was reheated to the original temperature before the experiments were conducted. On occasion, the mill shipped samples to the University laboratory. These samples were not temperature-controlled or aerated during shipment. Upon receipt, the samples were maintained in the same manner as the above samples for testing the following day. In the second phase, experiments were run at the Howe Sound Pulp and Paper mill site from February to August 1996. The samples were collected from the process and immediately placed into the batch reactors for experimentation. Primary influent and foul condensate are introduced separately to the head of the UNOX reactor at Howe Sound. Based on flow, the foul condensate makeup is approximate 10% of the total flow to the reactor. This mixture was replicated for the experiments conducted at the mill. The experimental methods used were the same as those developed using the Western Pulp effluent. The addition of a respirometric method for the determination of readily biodegradable COD was made feasible by the elimination of sample transportation time at the Howe Sound mill. 64 6.2 Analytical Methods Model parameters were determined using the methodologies described by the IAWPRC Model No. 1 procedures (Henze etal., 1987) and by alternate batch test methods proposed by a variety of researchers for use with the Model No. 1. 6.2.1 Chemical Oxygen Demand All samples were analyzed for COD according to Standard Methods (APHA, 1985). The COD analysis followed the chemistry procedure outlined in method 5220D with the exception that the digested samples were measured at 620 nm wavelength on a HACH DR/2000 spectrophotometer. 6.2.2 Solids The total and volatile suspended solids were measured by following procedures 2540D and 2540E respectively in Standard Methods (APHA, 1985). 6.2.3 Oxygen Uptake Rate The measuring system shown in Figure 5 was used to carry out the batch tests for the determination of the decay rate, maximum specific growth rate and readily biodegradable COD concentration by respirometery. The oxygen utilization rate of a sample was measured in a two litre water-jacketed respirometer. The respirometer was constructed from acrylic tubing sealed to an acrylic base. The respirometer was topped with a conical lid fitted for insertion of a dissolved oxygen probe and connection of an air supply. A laboratory stir plate and magnetic stir bar were used to maintain the respirometer contents in 65 suspension. The water jacket was connected to a water bath (VWR Scientific Model 1130A) to maintain the contents at a constant temperature. DISSOLVED OXYGEN AIR SUPPLY METER I— VENT DATALOGGER COMPUTER z MAGNETIC STIR PLATE Figure 5. Schematic Diagram Of The Respirometer Used In The Experimental Studies To determine the OUR of a sample, the air supply was shut off and the linear decrease in dissolved oxygen was recorded. The concentration of dissolved oxygen was monitored on a dissolved oxygen meter (YSI Model 54). The signals from the dissolved oxygen meter were registered on a data logger and downloaded directly to a laptop computer every 3-5 seconds depending on the experiment conducted. Because the measuring chamber was airtight, it was assumed that the actual respiration rate of the tested biomass at any time during the batch test did not depend on oxygen input. The gradient of the dissolved oxygen concentration therefore represented the actual oxygen uptake rate of the biomass. 66 6.3 Experimental Procedures The following is a brief description of the experimental procedures used to determine the model parameters. Detailed methodology, worked examples of the calculations used and tables of results for each parameter are listed in Appendices 1-6. 6.3.1 Growth Yield The method used to determine the heterotrophic growth yield was that outlined by Slade et al. (1991) with the exception that the influent wastewater and biomass samples were obtained from a full scale activated sludge treatment system. The heterotrophic growth yield, Y H , was estimated by observing the mass of cell material formed during removal of soluble substrate. The process was tracked by measuring the changes in COD over a series of days and determining the observed growth yields (Y0bS) from which the true growth yield was calculated. A simple water-jacketed batch reactor was constructed from acrylic tubing sealed to an acrylic base. The volume of the reactor was approximately 2 litres. Compressed air was bubbled through the reactor. The reactor was stirred continuously using a laboratory stir plate and a magnetic stir bar. Wall growth was minimized by scraping the sides of the reactor and aeration tubing with a bottle brush regularly to resuspend the cells. The water jacket was connected in 67 series to a water bath to maintain the contents at the temperature of mill's activated sludge reactor. Influent wastewater was vacuum filtered through a Whatman GF/C glass fibre filter mat to remove the particulate material. The filtered wastewater was placed in the reactor. Based on BOD values provided by the mills, sodium phosphate monobasic and urea were added as nutrients to provide a BOD:nitrogen: phosphorus ratio of 100:5:1 for microbial growth. The initial soluble COD of the reactor contents was determined on triplicate samples. The reactor was seeded with 2 ml of return activated sludge and the initial total COD of the reactor contents was determined. Duplicate samples were removed from the reactor twice daily for approximately 150 hours and analyzed for soluble and total COD. The linearisation method described by Fieschko and Humphrey (1984) was used to calculate Y H from the observed yield data. The reciprocal of the observed growth yield measurements was plotted against time. The true growth yield was calculated as the reciprocal of the y-intercept. 6.3.2 Decay Rate The decay coefficient was determined by the batch respirometric technique outlined in Model No. 1 (Henze etal., 1987). Instead of mixed liquor from a bench scale reactor, return activated sludge from the secondary clarifiers at the mills was used. 68 The sludge was aerated, stirred and kept at a constant temperature in the respirometer. Nitrification was inhibited by the addition of 20 mg of allylthiourea. The pH was monitored periodically to ensure that it was constant and remained near neutrality. The OUR was measured over several days. Evaporative losses were made up daily with distilled water. The natural logarithm of the OUR was plotted against time to determine the traditional decay coefficient, b'H. The model decay coefficient, bH, was calculated from the traditional decay coefficient using Equation 5-6: b H= b' H /1 - Y h * ( 1 - / P ) where fp was assumed equal to 0.08 mg/mg cell COD (Dold and Marais, 1986). 6.3.3 Maximum Specific Growth Rate The maximum specific growth rate was determined in the respirometer using the batch method described by Kappeler and Gujer (1992). Influent wastewater, biomass and nutrients were added to the respirometer. The mixture was kept at a constant temperature and stirred. The OUR was measured immediately. The dissolved oxygen probe was then removed and the respirometer contents were reaerated to maintain a dissolved oxygen concentration above 4 mg/L. The oxygen uptake rate was measured periodically until a discernible decrease in rate was observed. The results were transformed and the slope of the relative 69 OUR versus time line was related to the maximum specific growth rate according to the equations outlined by Kappeler and Gujer (1992). Several different CODA/SS ratios were investigated. 6.3.4 Readily Biodegradable COD Readily biodegradable COD was assessed using the physical-chemical batch method described by Mamais et al. (1993) at the Western Pulp mill and by the physical-chemical method and the aerobic batch method of Ekama ef al. (1986) at the Howe Sound mill. 6.3.4.1 Physical - Chemical Method Samples of influent wastewater to and effluent from the activated sludge system were flocculated by the addition of a zinc sulfate solution and mixed vigorously for approximately one minute. The pH of each sample was adjusted to 10.5 with sodium hydroxide. The influent sample was allowed to settle for a few minutes and the effluent sample was spun in a centrifuge for 5 minutes to speed the separation of the supernatant and colloidal material. The supernatants from both samples were filtered through 0.45 urn membrane filters and the CODs of triplicate samples of the filtered supernatants were determined. The readily biodegradable COD concentration was calculated from the difference between the COD of the influent and effluent samples. 70 6.3.4.2 Aerobic Batch Method A volume of wastewater of known COD strength was mixed with a volume of biomass of known VSS concentration to a predetermined F/M ratio in the respirometer. The mixture was kept at a constant temperature and stirred. The OUR was measured immediately. The dissolved oxygen probe was then removed and the respirometer contents were reaerated to maintain a dissolved oxygen concentration above 4 mg/L. The oxygen uptake rate was measured periodically until a discernible decrease in rate was observed. The area under the OUR curve corresponding to the readily biodegradable COD concentration of the influent wastewater was then calculated. Several F/M ratios were investigated. 6.3.5 Soluble Inert COD Following the method of Lesouef etai. (1992), the soluble inert COD was calculated as the final soluble COD concentration at the end of the growth yield test. At the end of the growth yield test, the soluble COD was constant and it was assumed that the all of the soluble biodegradable COD had been utilized. 6.3.6 Particulate Inert COD The batch method of Lesouef et al. (1992) was used to determine influent particulate COD. Water-jacketed batch reactors similar to the growth yield reactors were used. Two samples of influent wastewater were collected. One sample was filtered using a Whatman GF/C filter. The filtered and non-filtered 71 samples were placed in reactors and aerated continuously for several days. Samples were collected periodically for COD determination until a final soluble COD level was attained. The inert particulate matter was determined by the difference between the final total COD of the non-filtered sample less the biomass growth, minus the soluble inert COD. 6.3.7 Slowly Biodegradable COD The slowly biodegradable COD, Xs, was determined by difference once all of the other parameters were determined. 72 7. RESULTS AND DISCUSSION This chapter presents and discusses the results of the experimental program designed to collect the database for the development of a pulp mill activated sludge model. 7.1 Growth Yield Experiments The most important stoichiometric parameter in activated sludge modelling is the growth yield coefficient, Y H . Generally, yield is a coefficient which reflects a quantitative relationship between selected parameters of a biochemical process. The heterotrophic growth yield defines the amount of biomass formed per unit amount of organic matter removed for heterotrophic growth and is important for the accurate evaluation of sludge production and oxygen utilization in biological wastewater treatment systems. Substrate removal occurs as a result of two simultaneous reaction pathways. In the first pathway, substrate is utilized both as the energy and carbon source of biosynthesis. In the second, it is expended to derive energy for cell maintenance (Jones, 1973). Experimentally, it is not possible to separately identify these two reactions and only the net result of the overall reaction may be observed. 73 A typical growth yield curve from the experiments conducted is given in Figure 6. The observed growth yield decreases (1/ Y o b s increasing) with time. The y-intercept is calculated as the reciprocal of the true growth yield, 1/YH . This method of analysis follows the recommendation of Fieschko and Humphrey (1984) from their statistical analysis of the determination of true growth yield coefficients. Figure 6. Growth Curve For Western Pulp Sample Collected November 14, 1995 As the cell population ages, more microorganisms die, releasing further substrate for metabolism (endogenous respiration). Although further growth continues, the yield decreases as the proportion of substrate used for synthesis decreases, and that for maintenance increases. 74 Stoichiometry sets a material balance between all reactants and products of a mechanism. The stoichiometric expression of a mechanism is independent of the reaction rate and is valid for every reaction as it proceeds. The true yield may change for various carbon and energy sources because rates of energy generation and utilization are quite substrate-specific, resulting in different stoichiometric expressions for each substrate (Orhon and Artan, 1994). The growth yield of heterotrophic microorganisms in this study was found to be (at the 95% level of confidence): WP 0.71 (± 0.17) mg cell COD / mg COD @ 39°C n=4 HSPP 0.53 (± 0.09) mg cell COD / mg COD @ 30°C n=14 The + values represent the 95% confidence limits for the Y H at each mill. The effect of temperature on the rate of a biochemical reaction is generally expressed by the Arrhenius equation: rt = r20e(t-20) This expression can be empirically adopted to describe the effect of temperature on substrate utilization and microbial growth. 0 is a constant, known as the temperature coefficient and is a function of several factors including substrate concentration, F/M and type of substrate (Medronho and Russo, 1983). At a 75 initial COD concentration of 2.2 kg/m3, a good correlation for Y with temperature was obtained with an linear regression correlation coefficient of 0.995 (Medronho and Russo, 1983). A typical 0 value of 1.04 is proposed for the activated sludge process (Metcalf and Eddy, 1991). To simulate the temperatures at which the activated sludge systems were operated, the growth yield experiments were run at 39°C for Western Pulp effluent and 30°C for Howe Sound effluent. Correcting the measured values to 20°C, the growth yield becomes 0.34 mg cell COD/mg COD for the heterotrophic microorganisms in the Western Pulp activated sludge system and 0.36 mg cell COD/mg COD for the heterotrophic microorganisms in the Howe Sound system. The temperature-corrected growth yield values compare well with the literature values for pulp mill effluents reported in Table 4. The absolute values of kinetic coefficients depend to some extent on how the associated treatment system is operated. This in turn affects population dynamics of bacteria in the process (Springer, 1993). For activated sludge systems, the yield coefficients do not reflect a global, fixed stoichiometry because they involve the combined effect of growth and endogenous mechanisms which have variable relative impacts for different operating conditions. For each set of operating parameters there will be a different overall stoichiometry defining an observed yield for the actual 76 operation. The variable nature of the observed yield and the overall stoichiometry have significant effects on sludge production rates and nutritional requirements of activated sludge (Orhon and Artan, 1994). Table 4. Growth Yield Values Obtained From Literature Wastewater Source Heterotrophic Growth Yield (mg cell COD / mg COD) @20°C Researcher BKME 0.42 Slade etai., 1991 TMP 0.561 Liver et al., 1993 BKME 0.40 Springer, 1993 CTMP 0.542 Uu etai., 1993 Bisulfite Pulping 0.46-0.482 Lo and Valade, 1980 TMP 0.28-0.352 Lo etai., 1994b Municipal 0.53-0.66 Metcalf & Eddy, 1979 Municipal 0.46-0.69 Henze etai., 1987 Urban 0.67 Ekama etai., 1986 Urban 0.64 Sollfrank & Gujer, 1991 Urban 0.61 Torrijos etai., 1994 1 g MLVSS/g BOD5 2gMLSS/gBOD 5 77 7.2 Decay Rate Experiments The decay coefficient is an important kinetic coefficient used to quantify sludge production and oxygen consumption in the activated sludge process. The decay coefficient is a characteristic of the microorganism population (Henze et al., 1987). The traditional approach to microorganism decay follows the concepts of Pirt and Herbert (Lishman and Murphy, 1994). They observed that, in the absence of an external substrate, a decrease in microbial mass occurs. They concluded that a fraction of the viable organism mass was used directly as an energy source by the remaining living organisms for cell maintenance. The specific rate of change of viable mass due to endogenous respiration was expressed as the net endogenous respiration rate, b'H . The concept of decay adopted in the IAWPRC model is different from the traditional approach because of the death-regeneration concept of Dold et al. (1980) illustrated in Figure 7. Decaying biomass is assumed to be converted to both particulate products and slowly biodegradable substrate resulting in substrate recycling. The conversion results in no loss of COD or use of electron acceptor. Decay continues at the same rate whether aerobic, anoxic or anaerobic conditions prevail. The slowly biodegradable substrate is hydrolyzed to form readily biodegradable substrate which can be used to form more cells under aerobic or anoxic conditions. The other portion of the decaying organisms 78 is assumed to be non-biodegradable and adds to the residue of endogenous mass in the system. The overall rate of the death-regeneration cycle is therefore dependent on the relative rate of death (lysis of cell membranes), hydrolysis Figure 7. Death - Regeneration Cycle Under Aerobic Conditions (Lishman and Murphy, 1994) (enzymatic degradation of degradable cell protoplasm) and growth (oxidation of the decay products and the synthesis of new organism) (Lishman and Murphy, 1994). The net effect of this cycle in a batch test is a decline in the active 79 organism mass which can be measured by the decline in the oxygen utilization rate. 4 5 6 7 Time (days) 10 11 Figure 8. Decreasing OUR in Batch Aerobic Digestion Test With Western Pulp Return Activated Sludge Collected November 28, 1995 Figure 8 shows a typical result from the batch decay tests in this study. The decay coefficient, b'H, for heterotrophic microorganisms was found to be (at the 95% level of confidence): WP 0.20 ± 0.03 d'1 @ 39°C n=4 HSPP 0.19 ± 0.05 d"1 @ 30°C n=13 The decay coefficient can be adjusted for temperature by the Arrhenius equation where 9 = 1.029 (Lishman and Murphy, 1994). At 20°C, the decay coefficient for Howe Sound becomes 0.14 d'1, and for Western Pulp becomes 0.12 d' 1. 80 Slade et al. (1991) found the average decay coefficient for BKME at three different SRT values was 0.1 d' 1 at 25°C (0.09 d"1 at 20°C) and demonstrated that SRT had no effect on the decay coefficient. Dold et al. (1980) also demonstrated this independence with domestic sewage between the sludge ages of 2.5 and 20 days. Springer (1993) reported a decay coefficient value of 0.12 d"1 for BKME. For a bisulfite pulping effluent, the decay coefficient was found to be 0.12 d"1 (Lo and Valade, 1980). Liver et al. (1993) measured the endogenous decay coefficient of a TMP effluent as 0.11 d'1. For municipal wastewaters, Marais and Ekama (1976) reported the endogenous decay rate was 0.24 d"1. Henze et al. (1987) reported a range of decay coefficients from 0.05 d"1 for domestic sewage to 1.6 d"1 for food processing wastes. In general, the traditional decay coefficients determined at Howe Sound and Western Pulp mills agree well with the reported values in the literature. The IAWPRC model decay coefficient, bH, can be derived from the rate of OUR decline, b'H l using Equation 5-6: bH = b j 1 -yH (1-/P) where fp = 0.08 mg/mg (Henze et al., 1987). The model decay coefficient at 20°C for Howe Sound is 0.21 d"1 and 0.17 d"1 for Western Pulp. The model decay coefficient determined by Slade et al. (1991) for BKME was 0.19 d"1, a 81 value comparable to those found in this study. Henze et al. (1987) give a typical model decay coefficient value for municipal wastewater of 0.62 d"1. 7.3 Maximum Specific Growth Rate Experiments The parameters that describe biomass growth are the maximum specific growth rate , p.max ,and the half-velocity constant, Ks. The main function of |xm a x in the IAWPRC Model No. 1 is to allow the maximum OUR to be predicted, however, the model is not very sensitive to these values (Henze et al., 1987). The procedure used in this study was the respiratory technique of Kappeler and Gujer (1992). The procedure allows the evaluation of the maximum specific growth rate but not the half-velocity constant. A typical result using the Western Pulp effluent and biomass is given in Figures 9 and 10. A very high OUR followed by a sharp decrease in OUR within the first 60-90 minutes was typical of the Western Pulp samples. When cells are transferred to a batch reactor where all restrictions on growth are removed, the growth rate is limited initially. The cells cannot instantaneously shift their growth rate from \x to u.max; rather they require time for the transition to occur (Templeton and Grady, 1988). The biomass in this case was taken from the mill's activated sludge system operated at 39°C and stored at room 82 Figure 9. Batch Test to Estimate for the Western Pulp Sample Collected October 17,1995 Time (m in.) Figure 10. Logarithmic Form of the Relative OUR for the Western Pulp Sample Collected October 17,1995 temperature with aeration and no substrate source for approximately 16 hours before introduction into the batch reactor filled with influent at 39°C. The cells would probably be in an endogenous phase when introduced into the reactor. The high initial OUR observed may have been the result of the ability of some of 83 the cells in the biomass sample to thrive at the lower temperature, substrate limited conditions. When introduced to the batch reactor, however, the selective pressure of the new conditions, i.e. higher temperature, unlimited substrate, may not have favored these cells and resulted in a decrease in the observed OUR of the reactor. The delay in the second increase in respiration rate, which corresponded to the expected response of the biomass in this experimental procedure, may therefore have been the result of the shift in the growth rate of the remaining biomass from u, to umax as described by Templeton and Grady (1988). To deal with this problem in the measurement of the maximum specific growth rate, the slope was calculated from the time period when there appeared to be a steady increase in OUR to a maximum followed by a sharp decrease. In this example (Figure 10), the slope was calculated from the values recorded between 98 and 281 minutes. This lag was not apparent in the Howe Sound experiments, perhaps because the influent wastewater and biomass were taken from the process and directly entered into the batch reactors to test. A typical result is shown in Figures 11 and 12. 84 450 -i 0 50 100 150 200 Time (min.) Figure 11. Batch Test to Estimate the Maximum Specific Growth Rate for Howe Sound Sample Collected February 23, 1996 3 i 0 4 1 1 1 1 0 50 100 150 200 T ime (min.) Figure 12. Logarithmic Form of the Relative Oxygen Uptake Rate for Howe Sound Sample Collected February 23,1996 85 Kappeler and Gujer (1992) suggested that a COD:VSS ratio of 4:1 be used to estimate ixmax for domestic sewage. Table 5 outlines the COD:VSS ratios used in this study and the corresponding results. Table 5. Variation In The Slope Of The OUR Curve With Changes In The Ratio Of Substrate To Biomass Used Date Slope @ 20°C COD:VSS ratio Western Pulp Sept-20-95 4.3 4:1 Oct-17-95 1.8 4:1 Nov-14-95 1.8 13:1 Nov-28-95 0.9 23:1 Howe Sound Feb-22-96 6.6 11:1 Feb-23-96 14.1 5:1 Feb-27-96 4.0 3:1 Feb-28-96 2.5 13:1 Mar-01-96 17.6 11:1 86 From Table 5 it appears that the effect of varying the F/M ratio is unclear and that batch tests using the same ratio often produced substantially different results. There were also significant differences in the slopes measured at the two mills. Slade etai. (1991) also reported poor repeatability between batches using their experimental procedure. These discrepancies were thought to reflect changes in the wastewater composition due to mill process changes and spills. Reviewing the Western Pulp mill operations from the sampling days showed that only during the November 14, 1995 sampling period was there an evaporator leak resulting in black liquor entering the activated sludge system. From the results however, there was no apparent effect on u.max. No apparent operational difficulties resulting in a change in effluent quality occurred during the period of sampling at Howe Sound. The maximum specific growth rate was calculated using the relation: slope = / / m a x - b H Equation 7-1 (Kappeler and Gujer, 1992) and a decay coefficient of 0.2 day"1. The values of u-max were found to be (at the 95% confidence level): 87 WP 4.8 (± 3.0) d"1 @ 39°C n=4 HSPP 10.5 (± 5.8) d"1 @30°C n=5 The effect of temperature on i w is given by the Arrhenius equation in which 0 = 1.054 (Medronho and Russo, 1983). Adjusting to 20°C, the value of u.max becomes 1.8 d"1 for Western Pulp and 6.2 d'1 for Howe Sound. The maximum specific growth rate describes the growth of biomass on readily degradable substrate. It is very dependent on the nature of the wastewater being treated, temperature, reactor configuration and SRT, resulting in large ranges of values reported in the literature (Kappeler and Gujer, 1992). For municipal wastewater, values of fjm a x are reported from 3.0 to 13.2 d"1, with an average of 6 d' 1 (Henze etal., 1987). Springer (1993) gives a value of 2.2 d"1 for the maximum specific growth rate of heterotrophs in an integrated bleached Kraft mill effluent. Slade etal. (1991) found the maximum specific growth rate on BKME to be 0.52 d' 1 @ 25°C (0.43 d"1 @ 20°C). Liver et al. (1993) found i w to be 13 d"1 at 20°C for a TMP effluent. The larger ixmax value reported for TMP effluent may explain the larger values determined for Howe Sound versus Western Pulp as the Howe Sound mill treats TMP effluent in addition to the BKME in their activated sludge system. 88 In general, the results of the experiments fall within the wide range of values reported in the literature, though the large variation in results may not prove useful for the purposes of determining a representative value for modelling. Several authors allude to the difficulties in determining and interpreting maximum specific growth rate values. Henze etai. (1987) describe these values as hard to evaluate. Wentzel et al. (1995) state that the calculated value of p m a x from a batch test is unlikely to be of value in the modelling of activated sludge systems because the organism population that develops in the activated sludge system differs from that which develops in the batch reactor. Consequently the two populations probably have different kinetic constants. Additionally, biomass growth in different reactor configurations exhibits different values of |xmax even though the reactors are operated at the same solids retention time (Cech et al., 1984; Dold and Marais, 1986). The value of u.max is also generally lower for biomass grown in a completely mixed reactor with constant feed input than for biomass grown in a reactor which incorporates changes in substrate concentration. This is probably due to differences within the biomass brought on by different selective pressures in the two types of reactors. This suggests that care must be used in the collection and interpretation of kinetic data (Henze et al., 1987). 89 7.4 Wastewater COD Fractions The identification of influent characteristics with regard to the organic content is quite useful from the standpoint of process kinetics and operating strategy for biological wastewater treatment. One of the most widely used lumped parameters for wastewater characterization is COD. This parameter is often preferred to others because it permits an electron and energy balance to be drawn between organic substrate, biomass and oxygen utilized (Gaudy and Gaudy, 1971). Tables A5 and A6 in the Appendix list the detailed results of the wastewater COD fractionation experiments for both mills. 7.4.1 Readily Biodegradable COD The subdivision of the degradable COD in the Model No. 1 originally relates to the bisubstrate model of Dold et al. (1980) which identifies two major fractions: readily biodegradable COD and slowly biodegradable COD. The differentiation is based upon experimental observations showing a significant difference of approximately an order of magnitude between the rate of biodegradation of the two fractions. Within each fraction there are a number of compounds associated with a range of biodegradation rates. Readily biodegradable organic matter consists of simple organic molecules such as volatile fatty acids and low molecular weight carbohydrates, alcohols, amino 90 acids etc., that can be metabolized directly (Henze et al., 1992). The volatile fatty acids, especially acetic acid, are responsible for the major part of this fraction in domestic wastewaters (Henze, 1992). The readily biodegradable fraction usually comprises 20% of the total COD of municipal wastewater (Henze et al., 1992) and 10 - 35% of that in settled sewage (Orhon and Artan, 1994). The readily biodegradable portion of pulping wastewaters consists mainly of wood sugars which are cellulose and hemicellulose degradation products, oxidized fragments originating from the residual lignin in the pulp and residues of extractives that are carried over to the bleaching plant with the pulp (Dahlman and Morck, 1993). TMP effluent contains elevated levels of wood extractives (Lo et al., 1994a). No chlorinated organic material is formed in the TMP process as hydrogen peroxide is used as the bleaching agent. Constituents derived from wood extractives are more readily biodegradable than chlorinated compounds (Leach et al., 1977). Biodegradation of pimaric acid and pimarol proceeds to completion within 3 days (Leach etai., 1977). In laboratory fermentation studies, palustric, dehydroabietic and abietic acid salts were more readily attacked by microorganisms than were isopimaric or pimaric acid salts (Leach et al., 1977). Well operated treatment 91 systems should be capable of eliminating normal levels of the naturally occurring wood extractives: dehydroabietic acid, pimaric acid and pimarol (Leach et al., 1977). Microbiological degradation of chlorinated organics is often slow and this material may survive aerobic treatment. Leach et al. (1977) found that dichlorodehyrdoabietic acid and trichloroguaiacol were the most resistant to biodegradation of the compounds that they tested. Toxic levels of dichlorodehydroabietic acid may arise if resin acids are not efficiently washed from the brownstock prior to bleaching. Trichloroguaiacol is probably derived from the lignin that remains in the brownstock entering the bleach plant. Tetrachloroguaiacol was not readily degraded under laboratory conditions until the microorganisms had adapted. This could mean that a mill changing from the manufacture of unbleached pulp to bleached pulp might therefore discharge toxic levels of tetrachloroguaiacol during the lag phase needed for adaptation of the microbial population in the treatment system. The method used to determine the readily biodegradable COD of the total influent COD was the rapid physical-chemical method of Mamais et al. (1993). The fraction of readily biodegradable COD in the total influent COD was found to be (at the 95% level of confidence): 92 WP 0.24 (± 0.09) mg/mg Range: 0.07 - 0.49 n=8 HSPP 0.44 (± 0.04) mg/mg Range: 0.23 - 0.60 n=24 The sample containing only 7% readily biodegradable COD from the Western Pulp mill corresponded to the sample collected when the black liquor leak from the evaporators was occurring. Shock loading due to black liquor spills have been put forward as an explanation for detoxification failures (Leach et al., 1977), though no toxicity failure occurred at Western Pulp on this day. Modified Kraft cooking (extended delignification) and oxygen delignification, when used in combination, can reduce the amount of residual lignin entering the bleach plant by as much as 50% (Dillner et al., 1990). The use of chlorine dioxide as the only delignification reagent in the first bleaching stage results in a significant reduction in the amount of organically bound chlorine measured as AOX in the bleach plant effluent (Dillner et al., 1990). This implies that the formation of highly chlorinated low molecular weight compounds such as tri- and tetrachlorinated phenolic compounds and polychlorinated dibenzo-p-dioxins and dibenzofurans is more or less eliminated (Dahlman and Morck, 1993). The Howe Sound mill uses oxygen delignification in combination with chlorine dioxide and this may explain the high readily biodegradable COD values in the influent versus the Western Pulp samples. 93 The respirometric aerobic batch method of Ekama ef al. (1986) was also applied to the Howe Sound effluent. The results yielded an average value of 0.10 (± 0.11) mg/mg. The method was time consuming and produced highly variable results. Slade ef al. (1991) found their influent to contain 15% readily biodegradable COD as determined by the respirometric method and stated that their results showed remarkable consistency. Henze et al. (1987) found 24 - 32% by the same method in municipal wastewater. Ekama etal. (1986) reported a range of 13 - 27% for readily biodegradable COD in municipal wastewaters. Kappeler and Gujer (1992) measured a value of 7% for readily biodegradable substrate in primary domestic effluent and 11 % in a combined domestic and industrial wastewater. The results produced using the respirometric method in this study were comparable to those found in the literature, but not to the results generated by the Mamais et al. (1993) method using the same influent samples. The larger fraction of readily biodegradable COD determined for the Howe Sound influent using the Mamais ef al. (1993) method agrees, however, with the findings of Franta ef al. (1994) i.e., the majority of constituents of papermill wastewater are readily biodegradable. The results produced using the physical-chemical methods were also much more consistent than those produced using the 94 respirometric method. Table 6 outlines a comparison of the results generated by the two methods. Table 6. Comparison Of Readily Biodegradable COD Results Using The Ekama et al. (1986) And Mamais et al. (1993) Methods Date F/M Ekama Method Mamais Method % % 27- Feb-96 0.45 11 40 28- Feb-96 1.05 7 48 Mar-01-96 1.04 13 37 Average 10 42 7.4.2 Soluble Inert COD Soluble inert substrate passes through a biological treatment system in an unchanged form. There is evidence in the literature to show that the majority of the soluble organics in the effluent from biological treatment is not original substrate, but products generated through microbial metabolism (Grady and Williams, 1975; Orhon era/., 1989; Germirli etai., 1991; Sollfrank et al., 1992). A significant portion of this soluble organic matter is refractory, or at least very slowly degradable for the normal hydraulic and sludge retention time ranges 95 used in the activated sludge process (Chudoba, 1985a). The compounds produced can be classified into three categories: (i) compounds excreted by microorganisms in response to interactions with their environment; (ii) compounds produced as a result of substrate metabolism and bacterial growth; (iii) compounds released during lysis and degradation of microorganisms (Chudoba, 1985a). Rittman ef al. (1987) found that soluble microbial products are composed of a wide range of compound types and molecular weights. Microbial products are formed in proportion to the substrate utilization rate and in proportion to the concentration of accumulated biomass. To ignore the production of microbial products is to neglect a major contributor of organic matter to the effluent from microbial reactors (Daigger and Grady, 1977). The activated sludge process is therefore both a consumer and producer of COD and the concentration of inert soluble COD in the effluent may be higher than in the influent (Henze, 1992). The IAWPRC Task Group (Henze ef al., 1987) provided an acceptable basis to develop the concept of soluble residual products generation. In the model, the active biomass was postulated to be partially converted into inert particulate matter and slowly degradable matter, the latter yielding soluble microbial products through a hydrolysis step (Germirli ef al., 1991). 96 A large portion of the soluble inert organic material found in influent pulping wastewater has a high molecular weight (Mw < 1000). It is composed of highly oxidized degradation material originating from the residual lignin in pulp and is rich in olefin structures and carboxylic acid groups. The material is mainly aliphatic but contains minor amounts of phenolic structure. This fraction contains coloured constituents (Franta et al., 1994). This portion of the influent also contains oligo- and/or polysaccharides mainly originating from hemicelluloses (Dahlman and Morck, 1993). Approximately 10% of the effluent COD concentration is of microbial origin (Franta etai., 1994). Following the method of Lesouef et al. (1992), the soluble inert COD content of the total influent COD was found to be (at the 95% level of confidence): WP 0.36 ( + 0.07) mg/mg Range: 0.19 - 0.53 n=9 HSPP 0.32 (± 0.02) mg/mg Range: 0.23 - 0.41 n=16 Literature values for the soluble inert COD content in primary domestic wastewater range from 3 -10% (Sollfrank and Gujer, 1991; Henze, 1992; Lesouef et al., 1992). In the Model No. 1 report (Henze era/., 1987), the typical content of soluble inert COD in settled domestic wastewater is reported as 8-11 % of the total COD. Kappeler and Gujer (1992) give a value 20% for a combined domestic and industrial wastewater. It appears that the results 97 generated using the Howe Sound and Western Pulp mill influents are comparable to each other, close to the combined wastewater value of Lesouef ef al. (1992), but differ greatly from the literature values reported for municipal wastewater. This difference may be due to the unreactive, high molecular weight, lignin-derived portion of pulping influent that would not necessarily be present in domestic wastewater. 7.4.3 Particulate Inert COD Particulate inert substrate becomes entrapped and accumulates in the activated sludge system leaving only through sludge wasting. Depending on its origin, wastewaters may contain a significant fraction of biodegradable organic compounds in particulate form. Domestic wastewater, even after primary sedimentation, typically contains more organic carbon in colloidal and suspended form than in dissolved form. In the biological treatment of sludges, organic material is present exclusively in particulate form. Activated sludge usually has the ability to efficiently flocculate all particulate matter present in the reactor. Particulate organic material in primary effluents of domestic wastewater treatment plants is rapidly adsorbed by into the activated sludge but its degradation is very slow (Lesouef ef al., 1992). The degradation rate of particulate material depends primarily on the type and on the exposed 98 surface area of the material (surface limited reactions). Because of its slow degradation rate, particulate material substantially increases the observed activated sludge yield coefficient and consequently depresses oxygen requirement (Lesouef et al., 1992). Not all particulate material is degraded at a constant rate. A full spectrum of degradation rates exist from slow to fast for large and small particles to dissolved organic compounds. It is equally possible that degradation rates depend to some extent on sludge age (Lesouef ef al., 1992). The particulate inert organic fraction exhibits a similar pattern to its soluble counterpart. Some inert material is present in raw wastewater and some inert suspended organics are produced during activated sludge metabolism (Henze, 1992). The suspended inerts produced are dealt with as a separate fraction (endogenous residue) in the Activated Sludge Model No. 1 and thus the amount of raw wastewater inert suspended material is unchanged during treatment. The inert suspended organics produced during activated sludge metabolism are modelled as a fraction of the net biomass decay. The method used to determine the influent particulate inert COD was that of Lesouef ef al. (1992). The fraction of particulate inert COD in the total influent COD was found to be (at the 95% level of confidence): 99 WP 0.07 (± 0.05) mg/mg Range: 0.03 - 0.11 n=3 HSPP 0.03 (+0.01) mg/mg Range: 0.01 - 0.09 n=10 Using this method, Lesouef et al. (1992) reported a value of 7 -10% for domestic wastewater. Literature values for the inert particulate COD content in primary domestic wastewater range from 4-11 % (Sollfrank and Gujer, 1991; Henze, 1992; Kappeler and Gujer, 1992). In the Model No. 1 report (Henze et al., 1987), the typical fraction of inert particulate COD for domestic wastewater is given as 11-20%. Kappeler and Gujer (1992) give a value of 20% for a combined domestic and industrial effluent. The values determined for the Howe Sound and Western Pulp mill influents compare well with each other and with the literature values for domestic wastewaters. Values for pulp and paper influents not were found in the literature reviewed for this paper. 7.4.4 Slowly Biodegradable COD Slowly biodegradable COD is defined as the organic fraction which needs to be hydrolyzed before utilization by the biomass. The numerous compounds in wastewater have varying hydrolysis rates. Some soluble compounds hydrolyze slowly and some suspended compounds hydrolyze rapidly. Colloidal compounds disappear from the liquid phase as soon as they are brought into contact with the activated sludge following a very rapid 100 physical pathway, due to adsorption on the floes. During the first few hours, they participate only very little in the total oxygen demand for the oxidation of the biodegradable organic matter (Torrijos, et al., 1994). In municipal wastewater, the slowly biodegradable portion is thought to be composed of solid, colloidal and soluble organic matter of complex organic structure. Slowly biodegradable COD can account for 15-25% of raw municipal wastewater (Henze, 1992). Kappeler and Gujer (1992) reported a value of 53% for a mixed domestic and industrial wastewater and 60% for primary domestic effluent. In the IAWPRC report (Henze et al., 1987) a range of 43 - 49% is given for slowly biodegradable COD in settled domestic sewage. For pulping effluent, the slowly biodegradable portion may be composed of the oligomeric fraction of the wastewater (Mw 500-1000). This fraction contains oxidized degradation products from the residual lignin and may be structurally similar to the material in the high molecular weight fraction (Dahlman and Morck, 1993). In studies of pulping effluents, Mao and Smith (1995) determined that after the concentration of the fragment with Mw < 1000 becomes limited, a portion of the microbial population is induced to excrete extra-membrane enzymes which may be able to attack larger molecules. Most probably these enzymatic processes are non-selective and the process may be kinetically 101 controlled in the overall biochemical reactions involved in the metabolism of lignins and their derivatives. The slowly biodegradable COD fraction of the influent total COD was determined by difference once the other fractions had been experimentally determined. The fraction of the total influent COD that was slowly biodegradable was found to be (at the 95% confidence level): WP 0.42 (±0.16) mg/mg Range: 0.27 - 0.55 n=3 HSPP 0.23 (+0.05) mg/mg Range: 0.10 - 0.35 n=9 The results show that BKME and domestic wastewater contain a similar fraction of slowly biodegradable material. 102 8. CONCLUSIONS AND RECOMMENDATIONS A model provides a valuable tool to investigate the performance of full scale treatment systems without operational changes. This reduces any potential negative impact that a change in an operating parameter may have on the efficiency of the system. By manipulating key parameters, changes in process discharges, and their likely impact on the treatment system and effluent quality can be investigated. Modelling therefore provides a cost effective method for evaluating treatment strategies and system configurations. The IAWPRC Model No. 1 incorporates a number of parameters that specify the characteristics of the influent wastewater and system biomass. Accurate estimates of these parameters are essential if the model is to provide reasonable estimates of effluent quality, sludge production, oxygen utilization and dynamic process behavior. Continuous culture techniques, in which lab scale activated sludge reactors are operated at a number of specific growth rates, are recommended. There are, however, numerous difficulties associated with these systems. Operation of continuous culture systems is extremely laborious and time consuming; therefore, experiments conducted with them are expensive. Additionally, an extended time period is required to achieve steady state, which may result in changes in the characteristics of the microbial community, thereby affecting the observed kinetics. These difficulties, among others, have promoted an interest in rapid techniques for determining biodegradation kinetics, such as 103 those using batch reactors. The ease with which batch reactor procedures can be performed makes them attractive. The goal of this study was to determine the major parameters required to use the IAWPRC model to describe carbon oxidation in a pulp mill activated sludge system using on-site batch test methods. Table 7 summarizes the results found at the two bleached Kraft pulp mills studied: Western Pulp Ltd. and Howe Sound Pulp and Paper. Municipal sewage values provided in the Model No. 1 report (Henze etal., 1987) and values generated by Slade etal. (1991) in their investigation on bleached Kraft effluent from the Kinleith pulp mill in New Zealand have been included for comparison. Although the methodology adopted was developed for municipal wastewaters, it was found to be applicable to pulp mill effluent. The results obtained were generally comparable to those reported for municipal wastewaters and to the data set generated by Slade et al. (1991). Inconsistency was noted between the values of the maximum specific growth rate and the readily biodegradable COD fraction of the influent wastewater determined for the two mills examined in this study. The results of the maximum specific growth rate experiments fall within the range of values reported in the literature, though the large variation in the results may render these values useless for the purposes of determining a representative value for modelling. The values for the 104 influent readily biodegradable COD concentration from the Howe Sound mill were almost twice those from the Western Pulp mill. This may be explained by the fact that the Howe Sound mill employs oxygen delignification in combination with chlorine dioxide bleaching, and operates a TMP mill which may produce a more biodegradable effluent. Table 7. Summary of Results of Model Parameter Determination Parameter Western Pulp Ltd. (39°C) Howe Sound Pulp and Paper (30°C) NZFP Kinleith (25°C) Municipal Sewage (20°C) Y H (mg cell COD/mg COD) 0.71 ±0.17 0.53 ± 0.09 0.51 ±0.09 0.46 - 0.69 Umax (d'1) 4.8 ±3.0 10.5 ±5.8 0.52 ±0.11 3-13.2 b H (d 0 ) 0.21 ±0.03 0.17 ±0.05 0.19 ±0.04 0.62 S S (mg/mg) 0.24 ± 0.09 0.44 ± 0.04 0.15 ±0.02 0.24 - 0.32 Si (mg/mg) 0.36 ± 0.07 0.32 ± 0.02 0.08-0.11 Xi (mg/mg) 0.07 ± 0.05 0.03 ±0.01 0.11 -0.20 Xs (mg/mg) 0.42 ±0.16 0.23 ± 0.05 0.43 - 0.49 Note: ± values indicate the 95% confidence intervals for the mean values reported The experimental database generated in this study provides a basis for further studies and increases the understanding of the behavior of pulp mill activated 105 sludge systems. Additional studies in the following areas would provide valuable insight in the behavior of pulp mill activated sludge systems: • impact of process upsets (spills, leaks, etc.) on the biomass kinetics and influent wastewater composition, • chemical composition of the influent COD fractions identified in this study, • analysis of influent wastewater composition and biomass kinetics for activated sludge systems treating effluents from different pulping processes, and • further investigation into batch respirometric procedures to evaluate the maximum specific growth rate of heterotrophic microorganisms. This initial experimental database provides the basic information to describe carbon oxidation in a pulpmill activated sludge system using simple batch test methods. It is anticipated that the experimental data from this study will be used to develop a model of a pulp mill activated sludge system similar to the IAWPRC model for municipal wastewater treatment. 106 9. REFERENCES Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J. D. 1983. Molecular Biology of the Cell. Garland Publishing inc., New York, NY. APHA. 1985. Standard Methods for the Examination of Water and Wastewater. 16th Ed. American Public Health Association, New York, NY. Cech, J. S., Chudoba, J. and Grau, P. 1984. Determination of Kinetic Constants of Activated Sludge Microorganisms, Wat. Sci. Tech., Vol. 17, No. 2/3, pp. 259-272. Chudoba, J. , Cech, J. S., Farkac, J. and Grau, P. 1985. Control of Activated Sludge Filamentous Bulking: Experimental Verification of a Kinetic Selection Theory. Wat. Res., Vol. 20, No. 2, pp. 191-196. Chudoba, J. 1985a. Quantitative Estimation in COD Units of Refractory Organic Compounds Produced by Activated Sludge Microorganisms. Wat. Res., Vol. 19, No. 1, pp. 37-43. Chudoba, J. 1985b. Inhibitory Effect of Refractory Organic Compounds Produced by Activated Sludge Micro-organisms on Microbial Activity and Flocculation. Wat. Res., Vol. 19, No. 2, pp. 197-200. Daigger, G. T. and Grady, C. P. L. Jr. 1977. A Model for the Bio-oxidation Process Based on Product Formation Concepts. Water Research, Vol. 11, pp. 1049-1057. Dahlman, O. and Morck, R. 1993, In Bleached Pulp Mill Effluents: Composition, fate and effects in the Baltic Sea, Ed. A. Sodergren. Swedish Environmental Protection Agency Report No. 3756. Dare, P. 1995. Personal Communication. New Zealand Forest Research Institute, Wood Technology Division, Rotorua, NZ. Dillner, B., Larsson, L.-O. and Tibbling, P. 1990. Non-chlorine bleaching of pulp produced by the modified continuous cooking process. Tappi Journal 73:167-172. Dold, P. L , Ekama, G. A. and Marais, G. v. R. 1980. A General Model for the Activated Sludge Process, Prog. Wat. Tech., Vol. 12, pp. 47-77. 107 Dold, P. L. and Marais, G. v. R. 1986. Evaluation of the General Activated Sludge Model Proposed by the IAWPRC Task Group, Wat. Sci. Tech., Vol. 18, pp. 63-69. Edde, H. 1984. Environmental Control for Pulp and Paper Mills. Pollution Technology Review No. 108. Noyes Publications, New Jersey. Ekama, G. A., Dold, P. L. and Marais, G. V. R. 1986. Procedures for Determining Influent COD Fractions and the Maximum Specific Growth Rate of Heterotrophs in Activated Sludge Systems, Wat. Sci. Tech., Vol. 18, pp. 91-114. Ekama, G. A. and Marais, G. v. R. 1977. The Activated Sludge Process. Part II. Dynamic Behavior. Water SA, Vol. 1, pp. 18-50. Fahmy, M. K. 1992. Treatment of Sulphite Pulp Bleaching Effluents in Anaerobic/Aerobic Fluidized Biofilm Reactors. Doctoral Thesis. Swiss Federal Institute of Technology, Zurich. Fengel, D. and Grosser, D. 1975. Chemical Composition of Softwoods and Hardwoods - A Review of the Literature, Holz Roh-Werkstoff 33, No. 1, pp. 32-34. Fieschko, J . and Humphrey, A. E. 1984. Statistical Analysis in the Estimation of Maintenance and True Growth Yield Coefficients, Biotech, and Bioeng., Vol. 26, pp. 394-396. Franta, J. , Helmreich, B., Pribyl, M., Adamieta, E. and Wilderer, P. A. 1994. Advanced Biological Treatment of Papermill Wastewater; Effects of Operating Conditions on COD Removal and Production of Soluble Organic Compounds in Activated Sludge Systems. Wat. Sci. Tech. Vol. 3., No. 3, pp. 199-207. Ganczarczyk, J. J . 1983. Activated Sludge Process: Theory and Practice, Marcel Dekker, Inc., New York. Gaudy, A. F. and Gaudy, E. T. 1971. Biological Concepts for Design and Operation of the Activated Sludge Process. US Environmental Protection Agency Water Pollution Research Series, Report No. 17090, FQJ, 09/71. US EPA, Washington, DC. Germirli, F., Orhon, D. and Artan, N. 1991, Assessment of the Initial Inert Soluble COD in Industrial Wastewaters. Wat. Sci. Tech., Vol. 23, No. 4-6, pp. 1077-1086. Goodman, B. L. and Englande, A. J. 1974. A Unified Model of the Activated Sludge Process. Journal WPCF, Vol. 46, No. 2, pp. 312-332. 108 Grady, C. P. L. Jr. and Williams, D. R. 1975. Effects of Influent Substrate Concentration of the Kinetics of Natural Microbial Populations in Continuous Culture, Wat. Res., Vol. 9, pp. 171-180. Gujer, W., Henze, M., Mino, T., Matsuo, T., Wentzel, M. C. and Marais G. V. R. 1995. The Activated Sludge Model No. 2: Biological Phosphorus Removal. Wat. Sci. Tech. Vol. 31, No. 2. Henze, M., Grady, C. P. Jr., Gujer, W., Marais, G. V. R. and Matsuo, T. 1987. Activated Sludge Model No. 1, IAWPRC Scientific and Technical Report No. 1. IAWPRC, London. Henze, M. 1992. Characterization of Wastewater for Modelling of Activated Sludge Process, Wat. Sci. Tech., Vol. 25, No. 6, pp. 1-15. Huang, J. Y. C. and Mandt, M. G. 1978. Pure Oxygen Activated Sludge System Transfers Oxygen Rapidly, Wat. Sew. Works, Vol. 125, No. 6, pp. 98-100, 102-103. Jones, G. L. 1973. Bacterial Growth Kinetics: Measurement and Significance in Activated Sludge Processes, Wat. Res., Vol. 7, pp. 1475-1492. Jones, D. 1975. Training Manual on the Basic Technology of the Pulp and Paper Industry and its Waste Reduction Practices. Environment Canada. EPS 6-WP74-3, Ottawa, Ontario. Kappeler, J . and Gujer, W. 1992. Estimation of Kinetic Parameters of Heterotrophic Biomass under Aerobic Conditions and Characterization of Wastewater for Activated Sludge Modelling, Wat. Sci. Tech., Vol. 25, No. 6, pp. 125-139. Kristensen, G. H., Jorgensen, P. E. and Henze, M. 1992. Characterization of Functional Groups and Substrate in Activated Sludge and Wastewater by AUR, NUR, and OUR. Wat. Sci. Tech., Vol. 25, No. 6, pp. 43-57. Knudson, M. K., Williamson, K. J . and Nelson, P. O. 1982. Influence of Dissolved Oxygen on Substrate Utilization Kinetics of Activated Sludge. Journal WPCF, Vol. 54, No. 1, pp. 52-60. Leach, J. M., Mueller, J. C. and Walden, C. C. 1976. Identification of Toxic Materials from Kraft and Groundwood Pulp Mill Effluent. Process. Biochem., Vol. 11, No. 1, pp. 7-10. 109 Leach, J. M. and Thakore, A. N. 1976. Toxic Constituents in Mechanical Pulping Effluents, Tappi, Vol. 59, No. 2, pp. 12-132. Leach, J. M., Mueller, J . C. and Walden, C. C. 1977. Biodegradability of Toxic Compounds in Pulp Mill Effluents, Preprints from the 63rd Annual Meeting of the Technical Section, CPPA, Book A, pp. 135-140. Lesouef, A., Payraudeau, M., Rogalla, F. and Kleiber, B. 1992. Optimizing Nitrogen Removal Reactor Configuration by On-Site Calibration of the IAWPRC Activated Sludge Model, Wat. Sci. Tech., Vol. 25, No. 6, pp. 105-123. Lishman, L. A. and Murphy, K. L. 1994. The Significance of Hydrolysis in Microbial Death and Decay. Wat. Res., Vol. 28, No. 11, pp. 2417-2419. Liu, H. W., Lo, S. N. and Lavallee, H. C. 1993. Study of the Performance and Kinetics of Aerobic Biological Treatment of a CTMP Effluent. Pulp and Paper Canada, Vol. 94, No. 12, pp. 172 -177. Liver, S. F., Miyamoto, H. K. and Black, S. A. 1993. Bench-Scale Testing of Activated Sludge Treatment for a TMP Wastewater, Water Poll. Res. J. Canada, Vol. 28, No. 3, pp. 571-596. Lo., S. N., Lavallee, H. C , Rowbottom, R. S., Meunier, M. A. and Zaloum, R. 1994a. Activated Sludge Treatment of TMP Effluents Part 1: Effluent Characterization and Treatability Study, Tappi Journal, Vol. 77, No. 11, pp. 167-178. Lo., S. N., Lavallee, H. C , Rowbottom, R. S., Meunier, M. A. and Zaloum, R. 1994b. Activated Sludge Treatment of TMP Effluents Part 2: Biokinetic Parameters and Effluent Treatment Strategies, Tappi Journal, Vol. 77, No. 11, pp. 179-184. Lo, S. N. and Valade, J. L. 1980. Biodegradation of High-Yield Spent Bisulfite Pulping Liquor in Two Volume-Varied Chemostats in Series. Can. Jour. Chem. Eng., Vol. 58, pp. 241-248. Mamais, D., Jenkins, D. and Pitt, P. 1993. A Rapid Physical-Chemical Method for the Determination of Readily Biodegradable Soluble COD in Municipal Wastewater, Wat. Res., Vol. 27, No. 1, pp. 195-197. Mao, H. and Smith, D. W. 1995. A Mechanistic Model for Assessing Biodegradability of Complex Wastewaters, Wat. Res., Vol. 29, No. 8, pp. 1957-1975. no Marais, G. V. R. and Ekama, G. A. 1976. The Activated Sludge Process Part 1 -Steady State Behavior, Water SA, Vol., 2, No. 4, pp. 163-200. Medronho, R. A. and Russo, C. 1983. Influence of Temperature on the Kinetic Parameters of the Activated Sludge Process, Env. Tech. Letters, Vol. 4, pp. 1-8. Metcalf and Eddy, Inc. 1991. Wastewater Engineering: Treatment, Disposal, and Reuse. 3rd Edition. Revised by George Tchobanoblous and Frank Burton. McGraw-Hill Inc., Toronto. Mines, R. O. Jr. and Sherrard, J. H.1989. In Encyclopedia of Environmental Control Technology. Vol. 3: Waste-Water Treatment Technology. Cheremisinoff, P. N., ed., Gulf Publishing Co., pp. 203-228. NCASI. 1985. User Manual for NCASI Aerated Stabilization Model (SASBV2) and Guidance for its use. National Council for Air and Stream Improvement Technical Bulletin No. 458, New York. Orhon, D. and Artan, N. 1994. Modelling of Activated Sludge Systems, Technomic Publishing Co., Lancaster, PA. Orhon, D., Artan, N. and Cimsit, Y. 1989. The Concept of Soluble Residual Product Formation in the Modelling of Activated Sludge, Wat. Sci. Tech., Vol. 21, pp. 339-350. Poduska, R. A. 1979. Operation, Control and Dynamic Modeling of the Tennessee Eastman Company Industrial Wastewater Treatment System. Proc. Ind. Waste Conf. 34th, Purdue Univ., Lafayette, pp. 167-183. Rittman, B. E., Bae, W., Namkung, E. and Lu, C. -J . 1987. A Critical Evaluation of Microbial Product Formation in Biological Processes, Wat. Si. Tech., Vol. 19. pp. 517-528. Slade, A. H., Dare, P. H. and Leonard, A. M. 1991. Modelling Aerated Lagoons Treating Bleached Kraft Mill Effluents - Determination of Model Inputs, In Proc. 1991 TAPPI Envir. Conf., San Antonio, Texas, pp. 223-233. Smook, G. A. 1989, Handbook for Pulp and Paper Technologists, Joint Textbook Committee of the Paper Industry, Montreal. Sollfrank, U. and Gujer, W. 1991. Characterization of Domestic Wastewater for Mathematical Modelling of the Activated Sludge Process, Wat. Sci. Tech., Vol. 23, No. 4-6, pp. 1057-1066. i n Sollfrank, U., Kappeler, J . and Gujer, W. 1992. Temperature Effects on Wastewater Characterization and the Release of Soluble Inert Organic Material. Wat. Sci. Tech., Vol. 25, No. 6, pp. 33-41. Spanjers, H. and Keesman, K. 1994. Identification of Wastewater Biodegradation Kinetics, Proc. 3rd IEEE Conf. Contrl. Appl., Vol. 2, pp. 1011-1016. Spanjers, H. and Vanrolleghem, P. 1995. Respirometry as a Tool for Rapid Characterization of Wastewater and Activated Sludge, Wat. Sci. Tech., Vol. 31, No. 2, pp. 105-114. Springer, A. M. 1993. Industrial Environmental Control - Pulp and Paper Industry. 2nd Edition., TAPPI Press, Atlanta, GA. Strang, A. T. 1992. Secondary Effluent Treatment with High-Rate Activated Sludge at Howe Sound Pulp and Paper Ltd. Pulp and Paper Canada. Vol. 93, No. 5, pp. 65-68. Taylor, J . 1996. Activated Sludge Treatment of Kraft Pulp Mill Effluent. Pulp and Paper Canada, Vol. 97, No. 11, pp. 38-42. Tebbutt, T. H. and Aziz, J , A. 1980. An Assessment of Simple Mathematical Models for the Activated Sludge Process. Env. Tech. Letters, Vol. 1, pp. 440-449. Templeton, L. L. and Grady, C. P. L. Jr. 1988. Effect of Culture History on the Determination of Biodegradation Kinetics by Batch and Fed-Batch Techniques. Journal WPCF, Vol. 60., No. 5, pp. 652-658. Torrijos, M., Cerro, R. -M., Capdeville, B., Zeghal, S., Payraudeau, M. and Lesouef, A. 1994. Sequencing Batch Reactor: A Tool for Wastewater Characterization for the IAWPRC Model, Wat. Sci. Tech., Vol. 29, No. 7, pp. 81-90. Walden, C.C. and Leach, J. M. 1975. Identification and Treatment of the Toxic Materials in Mechanical Pulping Effluents. CPAR Project Report No. 149-3, Environment Canada, Ottawa. Walker, I. and Davies, M. 1977. The Relationship Between Viability and Respiration Rate in the Activated Sludge Process. Wat. Res., Vol. 11, pp. 575-578. Weddle, C. L. and Jenkins, D. 1971. The viability and Activity of Activated Sludge. Wat. Res., Vol. 11, pp. 575-578. 112 Wentzel, M. C , Mbewe, A. and Ekama, G. A. 1995. A Batch Test for Measurement of Readily Biodegradable COD and Active Organism Concentrations in Municipal Waste Waters. Water SA, Vol. 21, No. 2, pp. 11 124. 113 Appendix 1. DETERMINATION OF TRUE HETEROTROPHIC GROWTH YIELD To determine the true growth yield a series of observed yields must be measured from which the true growth yield is calculated. The true growth yield is calculated as the reciprocal of the y-intercept when the reciprocal of the observed yields are plotted against time. METHOD 1. Filter the influent wastewater to the activated sludge aeration basin with vacuum through a Whatman GF/C glass fiber filter mat to remove any suspended matter. 2. Fill the reactor with 500 ml of the filtered wastewater and agitate using a stir plate and magnetic bar. Aerate using a glass frit diffuser. 3. Add urea and sodium phosphate monobasic as nutrients for microbial growth. 4. Withdraw triplicate samples for the determination of the initial soluble COD. 5. Add 2 ml of return activated sludge from the treatment system. 6. Stir the reactor contents vigorously for a few minutes to disperse the biomass and remove triplicate COD samples using a wide mouth pipette to determine the initial total COD. 7. At regular intervals remove duplicate samples for measurement of total and soluble COD. Remove approximately 6 ml of the reactor contents with a syringe and filter through a Sartorius cellulose nitrate filter unit (0.45 pm pore size). Take duplicate samples for soluble COD from the filtrate. Determine the total COD as in (6) above. Stir the reactor contents vigorously until all biomass floes are dispersed sufficiently well to allow representative sampling. 114 8. Repeat (7) at regular intervals for at least 150 hours. 9. For each sampling time calculate the observed yield (Y0bS): Y o b s = A particulate COD A soluble COD 10. Plot the reciprocal of the observed growth yield values against time. The true growth yield is calculated as the reciprocal of the y-intercept (Fieschko and Murphy, 1994). 115 Table A 1. Experimental Growth Yield Results Source Date Y H (mg cell COD / mg COD) Western Pulp 17-Oct.-95 0.67 @39°C 14-Nov-95 0.48 28-Nov-95 0.87 28-Nov-95 0.80 30°C Howe Sound 15-Apr-96 0.69 15-Apr-96 0.43 22-Apr-96 0.62 22-Apr-96 0.83 29-Apr-96 0.40 29-Apr-96 0.44 15-May-96 0.49 15-May-96 0.39 21-May-96 0.44 21-May-96 0.39 04-Jun-96 0.79 04-Jun-96 0.63 12-Jun-96 0.30 12-Jun-96 0.64 116 WORKED EXAMPLE DATE: November 14, 1995 MILL: WP Time (mg/L) 0 15.3 49.2 74.5 86.0 135.7 166.9 212.2 Total COD (mg/y 1215 911 755 698 685 716 735 695 Soluble COD [mgA) 1349 1184 1001 948 916 945 939 885 STEP 1 Calculate the changes in particulate and soluble COD, and calculate Y0bS for each time. Time A Soluble COD ( 1 ) A Particulate COD ( 2 ) Yobs T'Tobs (hrs) ...(mg/y (mg/L) 0 0 0 15.3 304 138 0.46 2.20 49.2 460 111 0.24 4.13 74.5 517 115 0.22 4.48 86.0 530 96 0.18 5.50 135.7 499 94 0.19 5.29 166.9 480 69 0.14 6.92 212.2 520 55 0.11 9.39 (1) A Soluble COD (2) A Particulate COD = Soluble COD at time 0 - Soluble COD at time t (Total - Soluble) COD at time t -(Total - Soluble) COD at time 0 117 (3) Y o b s = A Particulate COD A Soluble COD e.g. at time 74.5 hours A Soluble COD = (1215 - 685) = 530 mg/L A Particulate COD = (948 - 698) - (1349 -1215) = 2 5 0 - 1 3 4 = 116 mg/L Yobs = 116 530 = 0.22 A linear regression of 1/Y 0b S versus time gives: y-intercept: 2.09 r2: 0.92 Y H 1 intercept = 1 2.09 = 0.48 118 Appendix 2. DETERMINATION OF DECAY COEFFICIENT FOR HETEROTROPHIC BIOMASS The decay coefficient is calculated from the rate of change of the oxygen utilization rate of biomass over several days. METHOD 1. Collect 2L of return activated sludge from the treatment system and transfer to a respirometer. 2. Add 20 mg of allylthiourea to the respirometer to prevent nitrification. 3. Measure the oxygen uptake rate periodically for several days. 4. Top the respirometer contents daily with distilled water to compensate for evaporative losses. 5. Plot the natural logarithm of the oxygen uptake rate against time. The slope of the line defines the traditional decay coefficient, b'H. 6. Determine the model decay coefficient using the following equation: bH = bjH 1-Yh(1-/P) where: / p = fraction of biomass that forms particulate decay products (/p = 0.08) assumed by the IAWPRC model 119 Table A 2. Experimental Traditional Decay Coefficient Results Date b'H (day1) Western Pulp @39°C 17-Oct-95 0.23 25-Oct-95 0.16 14-Nov-95 0.20 28-Nov-95 0.20 Howe Sound @30°C 02-Apr-96 0.20 0.20 10-Apr-96 0.06 0.12 22-Apr-96 0.18 01-May-96 0.30 0.21 15-May-96 0.24 0.11 28-May-96 0.36 0.26 12-Jun-96 0.09 0.08 120 Appendix 3. DETERMINATION OF THE MAXIMUM SPECIFIC GROWTH RATE The maximum specific growth rate of heterotrophs is determined in a batch test with filtered wastewater and a very small amount of biomass. The oxygen utilization rate is recorded until a distinct drop in the rate occurs. Using the slope of the natural logarithm of the relative respiration rate versus time, the maximum specific growth rate is calculated. METHOD 1. Collect influent wastewater to the activated sludge aeration basin and filter using a Whatman GF/C filter. Add the filtered influent and a very small amount of return activated sludge from the treatment system to a respirometer to produce a mixture of activated sludge and wastewater in a predetermined ratio (approximately 1:20 for domestic wastewater). 2. Agitate the respirometer contents using a stir plate and a magnetic stir bar. 3. Add urea and sodium phosphate monobasic as nutrient for microbial growth. Add 20 mg allylthiourea to inhibit nitrification. 4. Aerate the respirometer contents using a glass frit diffuser until the wastewater dissolved oxygen concentration is approximately 6 mg/l. 5. Insert a calibrated dissolved oxygen probe linked to a data logger. Allow the probe to equilibrate for 30 seconds and then record the change in dissolved oxygen concentration for 5 minutes or until the dissolved oxygen reaches 1 mg/l. 6. Remove the probe and reinsert the aeration frit. 121 7. Measure the OUR every 5 minutes. Ensure the dissolved oxygen probe is inserted into the respirometer at least 30 seconds before the OUR is measured to allow it to come to equilibrium. 8. The oxygen respiration should increase due to unlimited heterotrophic growth until an abrupt decrease in the OUR is detected when concentration of readily biodegradable substrate becomes limiting. 9. Plot the relative oxygen uptake rate, In r(t)/r(t0), against time. The maximum specific growth rate is determined by the addition of the decay coefficient and the slope. 122 Table A 3. Experimental Maximum Specific Growth Rate Results Date Slope Western Pulp @39°C Sept-20-95 4.81 Oct-17-95 3.76 Nov-14-95 3.80 Nov-20-95 1.77 Howe Sound @30°C Feb-22-96 9.76 Feb-23-96 20.92 Feb-27-96 5.91 Feb-28-96 3.66 Mar-01-96 11.30 123 Appendix 4. DETERMINATION OF READILY BIODEGRADABLE COD Two methods were evaluated: the rapid physical-chemical method (Mamais et al., 1993) and the aerobic batch test method (Ekama et al., 1986). The physical-chemical method involves removal of the colloidal matter that normally passes through a 0.45 jum filter by flocculation and precipitation. Membrane filtration of samples prepared in this way produces a filtrate containing total influent soluble COD from which the non-readily biodegradable soluble COD is subtracted to yield the readily biodegradable COD. For the aerobic batch method, a volume of wastewater of known COD strength is mixed with a previously aerated volume of settled mixed liquor. The oxygen utilization rate is monitored until it drops to a lower plateau level. The area beneath the OUR curve is used to calculate the readily biodegradable COD. PHYSICAL - CHEMICAL METHOD 1. Collect a sample of influent wastewater to the activated sludge aeration reactor and a sample of effluent mixed liquor from the activate sludge aeration reactor. 2. Flocculate the samples by adding 1 ml of 100 g/L zinc sulfate solution to a 100 ml sample of each wastewater sample. Mix vigorously with a magnetic stirrer for approximately 1 minute. 3. Adjust the pH of the samples to 10.5 using 6M sodium hydroxide and allow the samples to settle for a few minutes. 4. Withdraw the clear supernatant from each sample with a syringe and filter through a Sartorius cellulose nitrate filter unit (0.45 um pore size). Determine the COD of each sample in duplicate. 5. The COD of the influent filtered supernatant is the influent total truly soluble COD, CODsol, and the COD of the effluent filtered supernatant is the influent non-readily biodegradable COD, Si. 124 The readily biodegradable COD is determined by the following equation S s = CODsol - Si 125 WORKED EXAMPLE DATE: MILL: COD TOTAL: February 27, 1996 H S P P 1457 mg/L S T E P 1 Determine the COD of the influent and effluent filtered supernatant samples CODsol = 1224 Si = 640 STEP 2 Determine the readily biodegradable COD concentration: S s = CODsol - Si = 584 mg/L The fraction of readily biodegradable COD in the total influent COD is: 584 =0.40 1457 126 AEROBIC BATCH TEST METHOD 1. Collect a sample of mixed liquor from the outlet of the activated sludge aeration train and aerate for 1 hour. 2. Allow the aerated mixed liquor sample to settle. 3. Collect a sample of influent wastewater and determine the COD concentration. Determine the VSS concentration of the mixed liquor sample and calculate the F/M ratio. 4. Mix a known volume of mixed liquor with a known volume of wastewater in a respirometer to obtain the F/M ratio desired. Add urea and sodium phosphate monobasic as nutrients for microbial growth. Aerate until the wastewater dissolved oxygen is approximately 6 mg/ L using an aeration frit. 5. Insert a calibrated dissolved oxygen probe linked to a datalogger into the respirometer. Allow the probe to equilibrate for at least 30 seconds and then record the change in the dissolved oxygen concentration for 5 minutes, or until the dissolved oxygen reaches 1 mg/L. 6. Remove the probe and reinsert the aeration frit. 7. Measure the OUR as often as possible (every 10-15 minutes). 8. Continue measurements until the OUR has dropped to a new plateau. Take measurements for at least 20 minutes after the OUR drops to ensure that an accurate measurement of the new plateau has been made. 9. Calculate the area under the OUR curve that is attributable to the readily biodegradable COD. 127 Table A 4. Experimental Aerobic Batch Test Results Date Y H Fraction of Readily Biodegradable COD in Total Influent COD Feb-27-96 0.53 0.11 Feb-28-96 0.53 0.07 Mar-01-96 0.53 0.13 128 WORKED EXAMPLE DATE: MILL: VOLUME WASTEWATER: VOLUME BIOMASS: Y H : COD TOTAL: February 27, 1996 HSPP 2040 ml 500 ml 0.53 1457 mg/L Time OUR Area Calculated (min.) (mg..02/L!hr) 0 39 593 15 40 638 29 51 847 46 49 514 62 16 396 82 24 504 105 20 518 136 14 216 162 3 TOTAL 2689 Step 1 Calculate the area under the OUR curve. The plateau must be determined subjectively. The plateau in this example is 3 mg 02/L*hr. The area is calculated by the following equation: Area = (OURi + OUR2)/2 * (Time2- Timei) e.g. for Time 0, OUR = 39 Time 15, OUR = 40 Area =(39 + 40) *(15-0) 2 =593 Sum the areas to give the total area under the curve. Step 2 Calculate the area due to the readily biodegradable COD. This is the total area minus the area due to background respiration as indicated by the plateau (OUR = 3 mg 02/L*hr. For this example, the background area is 3 mg 02/L*hr x 162 minutes. The area due to readily biodegradable COD is: Area =4225-3x162 = 3739 ma O? x min. Lhr Step 3 Calculate the readily biodegradable COD concentration. S s = AO? x V m , + Vww ( 1 - Y H ) V w w where A 0 2 = area under the oxygen uptake rate curve due to readily biodegradable COD 130 Vww = volume of wastewater (ml) Vmi = volume of biomass (ml) Y H = growth yield S s = 1 x 3739 mg 0? * min. x 1 hr x (2040 + 500) ml 1-0.53 Lhr 60 min. 2040 ml = 165 mg 0 2 /L Oxygen units are assumed to be equivalent to COD units. S s = 165 mg COD/L The fraction of readily biodegradable COD in the total influent COD is: 165 =0.11 1457 131 Appendix 5. DETERMINATION OF PARTICULATE INERT COD After isolating residue on a glass fiber filter, refractory portions can be measured after biodegrading all the filtrate. The inert soluble substrate can be determined after continuous aeration of a filtered sample for several days until a final COD level is achieved. The matter remaining in the filtrate contains rapidly and slowly biodegradable substrate. If a non-filtered sample is aerated continuously in parallel, the inert particulate matter can be estimated from the difference between the total COD of the aerated non-filtered sample, less the created biomass and the soluble inert COD. METHOD 1. Collect a 2L sample of influent wastewater to the activated sludge aeration reactor. 2. Filter 500 ml of the wastewater sample with vacuum through a Whatman GF/C glass fiber filter mat to remove any suspended matter. 3. Fill a water-jacketed reactor with 500 ml of the filtered wastewater and agitate using a stir plate and a magnetic stir bar. Aerate using a glass frit diffuser. 4. Add urea and sodium phosphate monobasic to the reactor as nutrients for microbial growth. 5. Determine the initial soluble COD. Remove approximately 6 ml of the reactor contents with a syringe and filter through a Sartorius cellulose nitrate filter unit (0.45 urn pore size). Remove triplicate samples of the filtrate for soluble COD determination. 6. Add 2 ml of return activated sludge from the treatment system. 132 7. Fill a second water-jacketed reactor with 500 ml of unfiltered influent wastewater and agitate using a stir plate and magnetic stir bar. Aerate using a glass frit diffuser. 8. Add 2 ml of return activated sludge from the treatment system. 9. Add urea and sodium phosphate monobasic to the reactor as nutrients for microbial growth. 10. Stir the reactor contents vigorously for a few minutes to disperse the biomass and remove triplicate COD samples using a wide mouth pipette to determine the initial total COD. 11. Connect the reactors to a water bath to maintain the contents at the activated sludge aeration reactor temperature. 12. Aerate for several days and periodically withdraw samples from both reactors for the measurement of total and soluble COD as in (1) above until a final COD level is reached. Stir the reactor contents vigorously until all biomass floes are dispersed sufficiently well to allow representative sampling. 13. The inert particulate matter is the difference between the final total COD of the non-filtered sample, minus the soluble inert COD and the produced biomass. 133 WORKED EXAMPLE DATE: June 4, 1996 MILL: HSPP Step 1 Determine the total and filtered COD of the influent wastewater sample. Total COD 1302 Filtered COD 1077 Step 2 Determine the total and filtered COD of both the filtered and unfiltered sample after a final COD level has been attained. Unfiltered Sample Total COD 603 Filtered COD 390 Filtered Sample Total COD 458 Filtered COD 329 Step 3 Determine the particulate inert COD concentration of the influent wastewater. 134 Soluble Inert COD: Si = Soluble COD of the Total sample after aeration = 390 mg/l Biodegradable Matter Remaining in the Filtrate: Initial Filtered COD of Total Sample - Si = 1077 - 390 = 687 mg/l Particulate COD formed by Degradation of Matter Remaining in the Filtrate: Total COD of Filtered Sample after Aeration - Soluble COD of Filtered Sample after Aeration = 458 - 329 = 129 mg/l Apparent Yield for Degrading the Matter Remaining in the Filtrate: Particulate COD Formed / Biodegradable Matter Remaining in the Filtrate = 129/687 = 0.188 Inert Particulate COD: Final COD = Si + * + Yield (Initial Total COD - S, - X,) = 51 mg/l 135 9£T CD f n I cn co # P- >l O O CQ O CD • 3 < CO -vl O CO CO CD O O O b k) CD -U o o p "o CO O 0) CD 00 0 O O b o b 01 O) o o o o 05 M oo oo Z Z o < CD OI K) O CO CO O CD TJ CD c o CO c , , ^. w TJ CQ CQ cDCDcbcocbcbcbcb c n c n o i o i c n c n o i c n oo > K) cn co on o i c n w ^ w o i u ^ . * . - U - ^ N i O - U C D N i - t k - J i . c o c o r o c D C o c o o o o O O CO M -fc. -»• -fck co r*o -fc»- co co ->j -si -»• -»• cn o -»• co o oo o CD CO co cn -»• cn ->J co -»• NJ o o o o o o p p NJ ^ b co ho ho C O C O - * J C D - * O . U C O c o - k c o c o - N i o c n c o c T ) ODOOCnOO-^ M-^ O p p o p p p p p p co co cn cn co co —co co l o ^ u N i c n i o c D i o r o co cn o cn o o o b b - j co co oo O) cn cn o CD CD CD O O O ro cn js» NJ cn CD == % CD CD SI ° - CD O H o a O 91 CO O o O c D 2 CD CO cn CO o CO O CO 0) CO O o" -n CO o 3 X CO X o CO o 3 H CO IT CD > cn m x TJ CD 3' CD 13 ID 73 CD CO C_ CO c CD 0) co (—»-CD CD CD CO 2-o CO 5' CO CD - i "D C_ T J "O CO rj CD CO :x > TJ T J CD Q . X CD O m —i m 73 S o z o T l z T l I— cz m z H O o o o o TJ o CO CD < m co —I m 70 z TJ CZ r~ TJ TJ > z m CO o ro CD • i 0 J i 1 c c c CD CD CD CD i CD CD ro cL c D o ro ro -»• ^ ^. j*. o ro ro ->• o CD CD N) cn - 1 o o CD cn to ro ro ro ro co co -vj co ro c 3 0) 03 CU ^ > > > > 5) "O T J TJ TJ TJ DO DO DO T> —i —i CD CD cr cr T l T l CD CD cr cr i T l CD D" CD Cb ( ' n t'n CO m ^ ^ ^ ^ ^ CD CD CD CD ^ CD CD C D O i g g ^ ^ C D C D C J l C D c D C D C D C D C D C D C D C D ^ ^ ( f l O ) 0 3 0 c n c n ^ ^ * . o ^ * ; C n w M ^ - ' ^ i o c n c n t n c n - ' N o p c o o u ^ . A c o o i r o c o c n s s r o c n t o r o c o ^ ^ cn j i j*. o cn cn o co s s oo ro ro to J i CO -vl cn CD cn CO CD CO CD CD CD CO CD CD -U o o p p p Ji. CO Ji. J i J i —»• ji. ro CD CD CD CD — C O CO GO 00 --4 C D ^ C D ^ C D C ^ I O C D C D J ^ C » - ^ O l C D C O o ^ ^ - ^ o i t o c n o o o o o s j j j O i c n a o o c D - ^ J C D J i . c n c n o o J i ^ J W C D J i C n C D p p o o p p p p p p p p p p cn J i J i cn J i cn ro j i J^ co £ £ j*. co o o o o o o - i c n w j i - i o i s co o N s c n j i J i J ^ c o j ^ c D J i C o j i J i r o r O O O C D C D C O C O O C D C O r O C D C O J i - i C D O O ^ t D O S S W O O p o o o o o p p p p p p U U C O W C O ^ U U M U C o k ) N U U O O - l ( O O U O - i ^ j i co co cn cn CD ( D O N - 1 - 1 - 1 o o p p p p CD CD CD CD CD CD co ro ro co J i CD to J i co co ro CD co J i cn CD M Ol CO CD N o o o o o co ro ro ro 00 O CO -»• -»• co =-CD fS. O CD CD S Q. DO CD O H O 2-05 O O O c CD CO CO o 3 DO co o o 3 00 X g o' X co CU X O co =r. O DO cr CD > p m X TJ CD 3' CD 5T 73 CD CO c_ TO TJ CD Q-X o m H m 73 3 O z o T | CZ m z —i o o o o o TJ o CO § o DO Z o CD CO o c Q. TJ c_ TJ DO Q. TJ 00 TJ CD o m CO O c TJ C r -TJ > TJ > TJ m 73 CO 03 > cn -2 CQ o o o CD < ro CO CD CA) o o o 4 ^ C O J v J i CD O O O b o w ro cn ro to o o o b o b >• ro co CA) CA) CA) O O O b b k) cn oo co CA) tO -»• O CO CT) c c c 0 CT) c C Q I I I 1 CO CO CO CD CD CD CD CD CD co cn cn —-4 — C O CD ->- co oo -vj j i ro ro co -» cn co co (O O N O CO CD CD 00 O O CA) cn CD co J ^ o o o j i ro -si co J i CD o o cn cn j i j i oo -»• cn cn S O) S M p o o p co co co ro cn co o co ro co J^ to o co o o o o o b b b b CO CO ro j i j i -^J CD CO CD co J i ro o o o o o co to -si cn oo o O o CD a CD a. 0) CO »-CD 0) CD O H o a CO O O O c CD CO CO 0) CO o co o 3 CO 0) CO O o ' 3 cu o o ' 3 CO cu o 3