Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The design and testing of a short HRT attached growth pre-fermenter system: SCVFA production by two media.. Dumitrescu, Alex 1998

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

Item Metadata

Download

Media
ubc_1998-0223.pdf [ 10.99MB ]
Metadata
JSON: 1.0099238.json
JSON-LD: 1.0099238+ld.json
RDF/XML (Pretty): 1.0099238.xml
RDF/JSON: 1.0099238+rdf.json
Turtle: 1.0099238+rdf-turtle.txt
N-Triples: 1.0099238+rdf-ntriples.txt
Original Record: 1.0099238 +original-record.json
Full Text
1.0099238.txt
Citation
1.0099238.ris

Full Text

THE DESIGN AND TESTING OF A SHORT HRT ATTACHED GROWTH PRE-FERMENTER SYSTEM: SCVFA PRODUCTION BY TWO MEDIA TYPES AND TWO FEED SOURCES by ALEX DUMITRESCU B.Sc, UBC, 1990 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Bio-Resource Engineering) We accept this thesis as conforming, to the required standard THE IJNIVI/RSITY OF BRITISH COLUMBIA April, 1998 © Alex Dumitrescu, 1998 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. Department The University of British Columbia Vancouver, Canada Date f\p*-\\ WF& DE-6 (2/88) ABSTRACT A pilot-scale, short Hydraulic Retention Time (HRT), fixed film fermenter system was constructed to test for Short Chain Volatile Fatty Acids (SCVFA) production performance at various retention times. The performance of two media types, Ringlace® and Kaldnes media, was monitored at intervals of 20min, 40min and 60min provided by three 20min hydraulic retention time (HRT) reactors in each media series. The UBC waste water pilot plant was used as the waste source for all experiments. Two fermentation runs were completed; the first using primary effluent as feed and the second using screened raw wastewater. The sampling and processing scheme addressed temporal variations in the following parameters: SCVFA concentration, total organic carbon (TOC), total chemical oxygen demand (tCOD), soluble chemical oxygen demand (sCOD) and suspended solids analysis. Turbidity analysis was attempted but was not continued for the duration of the research. Metabolic fingerprinting, using the BIOLOG plate assay, provided a snapshot of the biofilm's grouped microbial metabolic activities present at the time of sampling. SCVFA production in the BF-system was near zero in the reactors containing no media during both the primary effluent and raw wastewater runs. Significant levels of SCVFA's were produced by ii short HRT attached growth fermenters using both primary effluent and raw wastewater as feed. Both sets of media reactors tested produced more SCVFA's than the control reactors. TOC and sCOD analyses, which measure dissolved organics in the waste stream, indicated a net solubilization of filterable solids in both media series fed raw wastewater. Kaldnes media was deemed not suitable as a short HRT pre-fermentation media in the current BF-system configuration, due to an inability to handle the solids loading of the waste stream. Suspended solids analysis supported the choice of appropriate media for future research by demonstrating the tendency of Kaldnes media to become clogged with solids. Ringlace® media has proven itself as a suitable short HRT pre-fermentation media by resisting plugging under the solids loading conditions tested. Results of the two fermentation runs indicate that, a 60min HRT fermenter system fed raw domestic wastewater, is capable of producing significantly more SCVFA's than a system fed primary effluent, (11.2mg/l and 5.6mg/l, respectively). iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES vLIST OF FIGURES vii LIST OF ABBREVIATIONS viiACKNOWLEDGEMENT ix CHAPTER 1: INTRODUCTION 1 1.1 Fermentation1.2 Biological Nutrient Removal 3 1.3 The Bio-Fermentation System 4 1.4 Objectives 5 CHAPTER 2: LITERATURE SEARCH 7 2.1 Biological Nutrient Removal2.2 Pre-Fermentation: Role in Nutrient Removal 12 2.3 Microbial Biofilms 14 2.4 Attached Growth Media 6 2.5 BIOLOG Assay 2CHAPTER 3: MATERIALS and METHODS 24 3.1 Bio-Fermentation System Components3.1.1 Primary Clarifier 27 3.1.2 Storage Bucket3.1.3 Pumps and Plumbing 8 3.1.4 Reactors 9 3.1.5 Pre-Filter Bucket 33 3.1.6 BF-System Flow Profile 4 3.1.6.1 Run#l 6 3.1.6.2 Run #23.2 Media 33.2.1 Kaldnes Media 37 3.2.1.1 Kaldnes Surface Area and Media Loading 39 3.2.2 Ringlace® Media 41 3.2.2.1 Ringlace® Surface Area and Media Loading 43 3.3 Analytical Procedures 7 3.3.1 Sampling3.3.1.1 Preliminary experiments (June-July, 1996) 49 3.3.1.2 Run #1 (August-November, 1996) 4iv 3.3.1.3 Run #2 (December, 1996-February, 1997) 50 3.3.2 Analysis 51 3.3.2.1 Chemical Oxidation Demand 53.3.2.2 Suspended Solids 53 3.3.2.3 Total Organic Carbon 4 3.3.2.4 Turbidity3.3.2.5 Short Chain Volatile Fatty Acids 55 3.3.3 Experiments 59 3.3.3.1 Preliminary Experiments 53.3.3.2 Main Experiments3.3.3.2.1 Run#l 60 3.3.3.2.2 Run #23.3.4 Statistical Analysis 61 CHAPTER 4: RESULTS AND DISCUSSION 63 4.1 Chemical Oxygen Demand4.2 Suspended Solids 66 4.3 Short Chain Volatile Fatty Acids 69 4.4 Total Organic Carbon 77 4.5 BIOLOG Assay 82 4.6 Media Comparison 5 4.7 SummaryCHAPTER 5: CONCLUSIONS and RECOMMENDATIONS 90 5.1 Conclusions 95.2 Recommendations 1 REFERENCES 3 LIST OF APPENDICES 98 APPENDIX A: RESULTS 9 APPENDIX B: STATISTICAL ANALYSIS 112 APPENDIX C: PHOTOS 119 v LIST OF TABLES Table 2.1: Examples of Attached Growth Media 17 Table 3.1: BF-System Reactor Working Volumes 32 Table 3.2: COD Protocol 5Table 3.3: GN Microplate 8 Table 4.1: Total sCOD Production (minus bucket) 64 Table 4.2: Total tCOD Production (minus bucket) 5 Table 4.3: Run #1 SCVFA Production (minus control) 74 Table 4.4: Run #2 SCVFA Production (minus control) 5 Table 4.5: Averaged TOC Production, Runs # 1 and #2 81 Table 4.6: Averaged Biolog Plate Data 83 Table 4.7: Test Media Surface Area Loading 5 vi LIST OF FIGURES Figure 1.1: Bacterial Fermentation Pathways 2 Figure 2.1: Three Basic Types of BNR Systems 11 Figure 3.1: BF-System Initial Configuration 25 Figure 3.2: BF-System Configuration, Experimental Runs 26 Figure 3.3: BF-System Reactor Components 30 Figure 3.4: Ringlace® Cage Detail 2 Figure 3.5: Run #2 Pre-filter Bucket 33 Figure 3.6: BF-System Flow Definitions, Run #1 35 Figure 3.7: BF-System Flow Definitions, Run #2Figure 3.8: Kaldnes Media 38 Figure 3.9: Ringlace® Media 42 Figure 4.1: VSS Analysis of the 60min Reactors 67 Figure 4.2: Averaged Total Solids Analysis 8 Figure 4.3: Total SCVFA Production per Reactor (Run #1) 70 Figure 4.4: Total SCVFA Production per Reactor (Run #2) 2 Figure 4.5: Average TOC Production, Run #1 78 Figure 4.6: Average TOC Production, Run #2 80 vii LIST OF ABBREVIATIONS BF-system Biological Fermentation System BNR Biological Nutrient Removal BOD : Biochemical Oxygen Demand Cl,2,3, Control reactors: 1,2,3, COD Chemical Oxygen Demand EPS Extracellular Polymeric Substances GN Gram Negative HRT : Hydraulic Retention Time Kl,2,3, Kaldnes reactors: 1,2,3, N2 : Nitrogen gas ORP Oxidation Reduction Potential PHB poly-P-hydroxybuterate PP Polypropylene PVC polyvinyl chloride Rl,2,3, : Ringlace® reactors: 1,2,3, sCOD Soluble Chemical Oxygen Demand SCVFA Short Chain Volatile Fatty Acid SRT Solids Retention Time SS Suspended Solids tCOD Total Chemical Oxygen Demand TOC : Total Organic Carbon TSS Total Suspended Solids VFA Volatile Fatty Acid VSS Volatile Suspended Solids viii ACKNOWLEDGEMENT I wish to express my sincere appreciation to the following people for their support and assistance throughout the entire study. Dr. Donald S. Mavinic, Professor of the Civil Engineering Department, for his patience and guidance throughout the painful process of writing about a year of blood, sweat, and tears. There was a time I would have given up had it not been for his encouragement. Dr. K. Victor Lo, Professor of the Bio-Resource Engineering Department, for finding me a second project to keep my dream of a masters degree alive. Paula Parkinson, Susan Harper and Jufang Zhou, staff of the Environmental Engineering Laboratory, for teaching me how not to blow up the lab, and being kind when I nearly do. Frederic A. Koch, research associate of the Civil Engineering Department, for his guidance in and around every topic to do with being a graduate student, his willingness to teach me as well as push me, but most of all for his friendship. Guy Kirsch, machinist and technician of the Civil Engineering Machine Shop. Guy built anything and fixed everything I asked and helped me feel at home in the shop. Dwayne Doucette, Jason White and (Dr.) Hong Zhao, fellow students in the Civil Engineering Department, for the assistance in the plant and in the library, but mostly, the comradeship I needed when the going was tough. Special Thanks to my wife, Sheena Dumitrescu, for the countless hours of emotional support and the patience necessary to be married to a graduate student. Funding for this research was provided by: The Science Council of British Columbia Reid-Crowther Limited. The UBC Environmental Engineering Department NSERC grant of Dr. D S Mavinic I thank them for giving me the means with which to accomplish the research. ix CHAPTER 1: INTRODUCTION 1.1 Fermentation Anaerobic fermentation is the, microbially mediated, breakdown of carbohydrate and protein compounds, without the presence of oxygen as a terminal electron acceptor. The end products of complete fermentation are methane and carbon dioxide. Typical biochemical pathways, for the fermentation of waste water, are presented in Figure 1.1 (Barnard, 1992). The anaerobic degradation of complex organics follows a multi-step mechanism of both parallel and series reactions (Shin and Song, 1995). The degradation pathway can be divided into four distinct steps: Hydrolysis, Acidogenesis, Acetogenesis, and Methanogenesis. The first three steps are grouped as Acidification reactions, whereas the last is named Methanation. Hydrolysis consists of the conversion of complex organics (proteins, carbohydrates, and lipids) to monomelic and oligomeric compounds (amino acids, sugars, and fatty acids). Acidogenesis consists of the degradation of the products of hydrolysis, yielding the following: volatile fatty acids, lactate and alcohols. The products of Acidogenesis are further metabolized in acetogenic reactions to yield acetate, carbon dioxide, and hydrogen. Methanogenesis reactions convert the acetate and hydrogen, produced in the previous step, to methane and carbon dioxide. 1 Figure 1.1: Bacterial Fermentation Pathways Propionic Acid Complex Waste 15% Acetic Acid Other Compounds Adapted from Barnard, 1994. Pre-fermentation systems are designed to augment the acidification steps, and hence, produce volatile fatty acids. Retention times are kept short in pre-fermenters, to prevent methanation (). Methanation reactions would consume the VFA's produced in the pre-fermenter, negating the benefits of VFA addition (see section 1.2). Most current fermentation literature is focused on the methanation component of fermentation. Literature on the optimization of VFA production, outside 2 the BNR field, involves the optimization of the acidogenic stage of a two stage anaerobic sludge digester (Barnard and Rabinowitz, 1997). 1.2 Biological Nutrient Removal Conventional biological treatment plants treat domestic wastewater by utilizing the metabolic capabilities of complex bacterial populations to oxidize organic components to carbon dioxide, water and biomass, which is then settled out of the treated waste water. Suspended solids are either physically removed in a primary clarifier or metabolized biologically. Dissolved organics are principally metabolized biologically. In addition to the damaging effects of suspended solids and oxygen demand of untreated domestic wastewater, it has been shown that the presence of nitrogen and phosphorus contribute to eutrophication of receiving waters (Argaman, 1991). Nitrogen can be removed from the system via biochemically processing the nitrogen containing compounds in a two step process, nitrification-denitrification. The end product, N2 gas, evolves from the system and is therefore removed (Barnard, 1974). However, the nitrogen is not destroyed, it is simply converted to its gas form, which is environmentally non-damaging. Phosphorus, unlike nitrogen, cannot be converted to gas form and must be recovered from the waste stream as an inorganic precipitate or as phosphate bound within bacterial cells (Trevino et al., 1992; Lie et al., 1997). Biological Nutrient Removal (BNR) was developed during the mid-1970's as it became clear that the nutrients, nitrogen and phosphorus, were still present in conventionally treated effluent. This effluent was contributing to the eutrophication of receiving waters by fertilizing algal growth blooms. Early BNR systems consisted of a modified activated sludge process, with the addition of 3 anoxic and anaerobic zones (Barnard, 1974). This type of system , called the "Bardenpho" process, utilized the ability of certain bacteria to take up excessive amounts of phosphate aerobically and store it. The phosphate-rich bacteria could then be removed from the system with the 2° clarifier sludge. Modern BNR plants are designed to treat wastewater by removing these nutrients from the waste stream as well as controlling traditional components such as suspended solids and Biochemical Oxygen Demand (BOD). The first BNR plant in western Canada was built in Kelowna, British Columbia, in 1982 (Barnard et al.,1995). Two more BNR plants were built in British Columbia, Westbank and Penticton (Dawson et al., 1995). Several existing plants in western Canada have been retrofitted to BNR systems including the following. Calgary and Edmonton Alberta, Salmon Arm British Columbia (Wilson and Keller, 1995). The industrial sponsor for this research, Reid Crowther, Ltd., is currently involved with several BNR projects, world-wide, with an emphasis on improving the process efficiency and lowering the treatment costs. In addition, Reid Crowther Ltd. has been involved in various laboratory, pilot-scale and full-scale BNR research projects, in conjunction with the University of British Columbia, resulting in the exchange and implementation of knowledge between the university, engineering consultants, and treatment plant operators. 1.3 The Bio-Fermentation System A pilot-scale, short Hydraulic Retention Time (HRT), fixed film fermenter system was constructed, at the University of British Columbia pilot plant site, to test for Short Chain Volatile Fatty Acid (SCVFA) production performance at various retention times, as a measure of acidogenic potential 4 and feasibility as a retrofit to existing BNR plants. The Bio-Fermentation (BF) system consisted of nine PVC reactors grouped as three series of three reactors. Two series represent test media fermenters and the third was a negative control with no media. The fermenter system tested two different media concurrently, Ringlace® and Kaldnes media, to determine which might be better suited for short HRT partial fermentation of primary effluent and raw wastewater. The media were judged mainly by their ability to produce SCVFA's, within a set HRT of the BF-system. Three reactors in series allow sampling at time intervals of 20min and 40min, as well as the 60min system HRT. The UBC waste water pilot plant was used as the waste source for all experiments. The system was designed as "up-flow" but was inverted to yield a "down-flow" system after preliminary testing. Two fermentation runs were completed; the first using primary effluent as feed and the second using screened raw wastewater. 1.4 Objectives This research focused on three parameters in the BF system: fixed film media type, waste stream makeup, and HRT. Each variable was analyzed in terms of its effect on SCVFA production and system sustainability. After the initial startup period necessary to develop biofilm populations, and troubleshoot the operational protocol for the BF-system, two experimental runs were performed. The first assessed media performance, fermenting primary effluent, while the second assessed media performance, fermenting raw wastewater. Both runs employed the same HRT, media configurations and sampling design. 5 The following outlines the specific objectives of the two BF system sampling runs: 1. Compare the performance and suitability of two fixed film media types for the fermentation of primary domestic effluent to yield a significant increase in SCVFA levels. 2. Compare the performance and suitability of two fixed film media types for the fermentation of raw domestic wastewater to yield a significant increase in SCVFA levels. 3. Determine the feasibility of this reactor system as a producer of significant SCVFA's within the HRT constraints set up by Reid Crowther Ltd., the co-sponsor of this research project. 3. Determine the SCVFA production possible utilizing a short HRT prefermenter. 4. Describe the accumulated biofilm in each experimental condition metabolically. 6 CHAPTER 2: LITERATURE SEARCH 2.1 Biological Nutrient Removal The focus of domestic wastewater treatment has shifted from treating individually targeted components of sewage to the understanding of nutrient interactions as they relate to overall waste treatability (Stensel et al., 1992). The goal of the BNR system is to remove nitrogen and phosphorus from the waste stream as well as BOD content (Stensel and Barnard, 1992). Biological Nutrient Removal systems (BNR) consist of a modified activated sludge process, with the addition of anoxic and anaerobic zones (Barnard, 1974). The basic principles of nitrogen and phosphorus removal, presented below, were adapted from a BNR design manual written by Randall, Barnard, and Stensel (1992). Nitrogen, as ammonia-Nitrogen, is removed from the waste stream by a series of nitrification and denitrification reactions. Nitrification reactions involve the oxidation of ammonia-Nitrogen, in the presence of oxygen, producing nitrate and nitrite (equations 2.1 to 2.3). Nitrosomas 2NH4+ + 302 2N02~ + 4H+ + 2H20 Equation 2.1 Nitrobacter 2N02" + 02 2N03" Equation 2.2 Total reaction NH4+ + 202 -• N03" + 2H+ + H20 Equation 2.3 7 Nitrification reactions occur in the aerobic zone of BNR systems. The reactions are mediated by two types of bacteria, Nitrosomas m&Nitrobacter, which produce nitrite and nitrate, respectively. It has been shown that, the overall reaction (equation 2.3), is rate limited by the nitrite producing step. Denitrification reactions involve the reduction of nitrate, in the absence of free oxygen, to nitrogen gas. These reactions require an electron donor, wastewater BOD, to provide energy for the reduction reactions. Various heterotrophic bacteria are responsible for denitrification reactions, including bacteria which can utilize oxygen as well as nitrate as the final electron acceptor. Denitrification reactions occur, predominantly, in the anoxic zone of BNR systems. However, recent studies have shown that, under low oxygen conditions, denitrification can also occur simultaneously with nitrification in the aerobic zone (Koch, 1996). Phosphorus must be manually removed from the waste stream, as particulate matter, because it does not exist as a gas, as in the previous case of nitrogen (Argaman, 1991). In chemical removal systems, phosphorus is removed as a mineral precipitate. In BNR systems, the particulate matter is bacterial cells. BNR systems facilitate the removal of waste stream phosphorus through alternating anaerobic and aerobic conditions (Argaman, 1991). In the aerobic zone, bacteria capable of storing phosphate, use up some of the energy generated from metabolizing, both extracellular and stored carbon sources, storing phosphate as polyphosphate granules (Jardin and Pdpel, 1997). Under anaerobic conditions, these organisms extract energy from the release of stored phosphate, which gives them a biological advantage over "non-polyphosphate storing" bacteria. This energy is used for cell maintenance and for the uptake and storage of carbon sources, specifically acetates which are assorted as poly-p-hydroxybuterate (PHB) (Stensel and Barnard, 1992). Upon re-exposure to 8 aerobic conditions, "polyphosphate containing" bacteria metabolize the stored carbon source, and once again, take up excess phosphate. The stored phosphorus is removed from the waste stream via the wasting of sludge, collected from the aerobic zone, containing bacteria which have an excess accumulation of phosphate. The ratio of phosphorus removal to sludge removal has been calculated: lg P to 3g SS sludge (Jardin and Popel, 1997). The form of carbon, specifically Short Chain Volatile Fatty Acids, present in the influent sewage has been shown to increase nitrogen and phosphorus treatability levels in existing BNR systems (Lie et al., 1997; Munch and Koch, 1997; Oldham and Abraham, 1994; Rabinowitz and Barnard, 1997; Raper et al., 1997). Converting a larger fraction of the waste's carbon component to VFA's internally could increase this effect, and hence overall nutrient removal. The HRT of the anaerobic zone, of a BNR system, can be drastically reduced because of the VFA supplementation of sidestream primary sludge fermenters (Barnard et al.,1995). Facilities employing primary sludge fermenters have shown that if their fermenters are taken out of service, they can no longer meet their phosphorus discharge limits (Barnard and Rabinowitz, 1997). There is some discrepancy in the reported relationship between VFA content and phosphate removal. Reported values for the amount of VFA's necessary to remove one milligram of soluble phosphate range from 5mg (Barnard, 1998) to 20mg (Lie et al., 1997). However, it has been shown that, excessively high concentrations of acetic acid are inhibitory to phosphorus removal (Randall and Chapin, 1997). Lab-scale BNR experiments indicated that, acetic acid concentrations of up to 400mg/L were beneficial to the process. Above 400mg/L however, the performance of both phosphorus and COD removal dropped. 9 Three basic types of BNR systems are shown in Figure 2.1: Three-Stage Bardenpho, Five-Stage Bardenpho, and University of Capetown (UCT). The three-stage Bardenpho system employs anaerobic, anoxic, and aerobic reactors in series. A portion of the aerobic activated sludge is recycled to the anoxic zone to enhance denitrification. Secondary clarifier underflow is recycled to the anaerobic zone to stimulate biological phosphorus removal. A three-stage Bardenpho system first went into service in 1974 (Barnard, 1974). The five-stage Bardenpho modification to the previous system was designed to increase both nitrogen and phosphorus removal (Argaman, 1991). The five-stage modification first went into practice in 1978 (Randall, 1992). Two extra reactors were added, anoxic and aerobic, to further enhance denitrification and ensure aerobic conditions in the secondary clarifier, respectively. Anaerobic conditions in the secondary clarifier must be avoided to prevent the bacterial release of phosphate. The five-stage design is predominantly used where the nitrogen removal regulations are stringent (Barnard and Rabinowitz, 1997). The UCT system is identical to the three-stage Bardenpho, with the exception to the recycle configuration (Argaman, 1991). Activated sludge is recycled to the anoxic zone and then from the anoxic zone to the anaerobic zone. This configuration ensures that, no nitrates will enter the anaerobic zone (Barnard and Rabinowitz, 1997) and hence, no interference with SCVFA storage (Argaman, 1991). The UCT process first went into practice in 1984. 10 Figure 2.1: Three Basic Types of BNR Systems Three-Stage Bardenpho Process + 1 • Five-Stage Bardenpho Process University of Capetown (UCT) Process • • n Anaerobic I I Anoxic Q Aerobic (Adapted from Barnard and Rabinowitz, 1997) 11 2.2 Pre-Fermentation: Role in Nutrient Removal It has been shown that, wastewater must be at least partially fermented, yielding volatile fatty acids, before it can fuel enhanced phosphorus removal (Randall, 1994). This "pre-fermentation" can occur in three ways: • Fermentation can occur in sewer pipes leading to the waste facility. • The anaerobic zone of a BNR plant can provide the fermentation. • A pre-fermenter can partially ferment the waste stream, either raw wastewater or settled solids, and feed VFA's directly to the anaerobic zone of the BNR process. A prefermenter ensures a plentiful supply of VFA's to the anaerobic reactor, independent of fluctuations in raw sewage flow (Randall, 1994). High VFA's in the anaerobic zone can improve the settleability of secondary sludge, by selecting for polyphosphate organisms in the BNR system. The settleability is increased because, polyphosphate bacteria form better floes than non-polyphosphate bacteria, and the increased phosphorus content, in the floes, increases the overall sludge density (Randall, 1994). Even under conditions where the sewage is partially fermented in the sewer system, the addition of a prefermenter, will produce higher and more consistent phosphorus removal rates (Barnard, 1994). The addition of a side-stream prefermenter, provides VFA's to the BNR system independent of the main system flow, and the current "health" of the activated sludge organisms (Barnard, Stevens, and White, 1995). A side-stream fermenter can supply a concentrated amount of VFA's anywhere in the BNR process, allowing more optimization options for nitrogen and phosphorus removal. The results of a BNR facility, Beckton Sewage Treatment Works, (in London, England) demonstrated that the average effluent phosphorus without prefermentation was 4mg/L. With the addition of a side-stream prefermenter, the effluent 12 phosphorus dropped to l-2mg/L (Rabinowitz and Williams, 1995). The size of the anaerobic reactor can be reduced as much as 95% because of the VFA contribution of a prefermenter (Randall, 1994). The combination of, a smaller anaerobic zone and prefermenter, has been shown to prevent the secondary release of phosphorus in BNR systems (Barnard, 1994). Secondary phosphorus release is defined as, the anaerobic release of phosphorus, without the associated uptake of nutrients. Without the accumulation of PHB in the anaerobic zone, there will be no enhanced phosphorus uptake in the aerobic zone, due to lack of bacterial energy stores. There are four types of sludge fermenters currently employed in BNR systems: activated primary tank, complete mix fermenter, single stage fermenter/thickener, and 2-stage complete mix fermenter/thickener. Descriptions of each type, listed below, were adapted from a fermenter system summary by Barnard and Rabinowitz (1997). An activated primary tank fermenter utilizes the sludge blanket, formed in the primary clarifier, as a source of fermentable solids for the BNR system. The primary clarifier underflow is recycled to the clarifier inlet, either directly or via a mixed holding tank. A fraction of the recycled sludge is wasted to the sludge-handling system to maintain a workable sludge blanket depth on the floor of the clarifier. VFA's, produced by the sludge blanket, are carried into the BNR process with the primary effluent. An advantage of using this type of fermenter is that, no additional construction is needed, since the system employs an existing clarifier as a reactor (Barnard, 1994) A complete mix fermenter is essentially a modified version of this system. The main operational differences between the two systems involve the location of sludge fermentation. In the later system, the 13 fermentation occurs in an enlarged mixed holding tank, whereas the primary clarifier sludge blanket is minimized. The single and 2-stage fermenter systems separate the primary effluent flow from the VP A- enriched fermenter effluent. The single stage system utilizes an enlarged gravity thickener, which can operate at a flexible HRT and SRT because it is not directly dependant on the flow of raw sewage. One advantage of this type of system over the previous examples is that, WA-rich thickener effluent is added directly to the BNR anaerobic zone, protecting the VFA's from metabolic activity in the clarifier. Thickener underflow is drawn off to the sludge handling system. The 2-stage system adds a complete mix fermenter in series between the primary clarifier and the gravity thickener. This design allows recycle from the thickener back to the mixed fermenter, allowing longer SRT. 2.3 Microbial Biofilms Microbial biofilms can be defined as mixed or pure bacterial populations growing on an inert surface. It has been shown that almost any "wetted" surface is inhabited by a population of bacteria (Gjaltema and Greibe, 1995). The bacterial populations live in a layer of Extracellular Polymeric Substances (EPS) on the surface of the inert substratum. The EPS retains the bacteria and diffused nutrients providing anchorage and protection. It has been documented that a killed or inactivated biofilm surface is quickly re-colonized because of the EPS layer left behind from the previous population (Flemming and Melo, 1995). A developed biofilm layer protects its bacterial inhabitants from suspended disinfectants or poisons, which makes an established biofilm very difficult to kill (LeChevallier et al, 1988). The transport of any chemical from the bulk liquid to the biofilm layer 14 is highly rate lunited. The result is that the required contact time for a suspended disinfectant to kill the resident population is much longer for a biofilm population than a suspended population. The detailed structure of biofilms can only be theorized due to the lack of effective techniques for the in situ observation of biofilm development (Lazarova and Manem, 1995). Light microscopy has the advantage of simplicity and the ability to view the biofilm directly, without chemical or physical pre-treatment. Sophisticated monitoring techniques are not practical for use within current reactor designs (Gjaltema and Greibe, 1995). Confocal laser and electron microscopy have been used to study biofilm formation in micro-scale experiments (Wolfaardt et al., 1994). However, the biofilm analysis of larger systems occurs off-line, involving the removal of the biofilm from the substratum. The removed biofilm can then be characterized by physical or chemical parameters, such as: mass, thickness, TOC, COD, protein content, or VSS (Lazarova and Manem, 1995;Meraz et al., 1995). In practice, biofilm sampling is difficult because access to the interior of bioreactors is highly limited, especially in the case of anaerobic fermenters (Gjaltema and Greibe, 1995). The accepted theory of biofilm structure involves a series of layers, each with unique bacterial and biochemical characteristics. Recent research has identified the development of a biofilm layer as the growth of microbial clusters which mushroom up from the substratum surface (Bishop and Rittmann, 1995). The result is a non-uniform, mountainous terrain, with channels between developing colonies of bacteria. Biofilm heterogeneity can be described in terms of the following: diversity within populations, metabolic reactions, biofilm chemistry, variation in the amount of EPS and variation in the physical 15 structure (Bishop and Rittmann, 1995). Modeling biofilm development and performance is difficult and highly dependant on the specific reactor conditions. Bacterial growth dynamics in the biofilm layer have been studied to determine standardized growth models for biofilm systems (Cheng et al., 1992; Meraz et al, 1995). Models relating product inhibition, metabolic efficiency and nutrient flux into the biofilm, have been theorized (Bhaskar and Bhamidimarri, 1991; San et al., 1993). However, they are not ready to serve as guidelines for bioreactor design. 2.4 Attached Growth Media In biological sewage treatment systems, biofilm reactors have higher volumetric productivity over suspended growth systems (Bhaskar and Bhamidimarri, 1991). In order to achieve this high treatment efficiency, it is important that a high concentration of microorganisms is retained in the process. An optimal way to obtain this is to use a suspended carrier, or attached growth, in the process to increase the Solids Retention Time (SRT) relative to system Hydraulic Retention Time (HRT) (Lessel, 1991). An effective attached growth media must offer a high specific surface area, a good biofilm holding capacity and avoid clogging due to increasing biomass. Once the media clogs, the available surface area is greatly reduced, thus lowering the efficiency of nutrient transfer from the waste water to the biofilm. Abrasion is used, in fluidized media systems, to control and limit biofilm development (Wilderer et al., 1995). The following desired properties of suspended carriers have been outlined by Tsubone (et al., 1992). • The surface of the carriers must be rough, to ease biofilm colonization. • The carriers must be easily fluidized. The fluidization should not require more aeration or 16 mixing than is necessary to sustain the biological reaction in the reactor. • The carrier must be easy to separate from the effluent. The size of the carrier must be such that it is easy to screen out of the reactor outflow. • The functional lifespan of the media must be long. The media should not be susceptible to mechanical and biological breakdown In previous research, with both aerobic nitrifying reactors and anaerobic fermenters, a variety of media types have been employed (Table 2.1). The trend in media design is away from rigid media structures, due to plugging tendencies, towards fluidized carriers or other flexible media (Demerci eta!., 1993; Lessel, 1991; Tsubone et al., 1992). Table 2.1: Examples of Attached Growth Media Researcher(s) Media Type(s) Media Description Application Cheng et al., 1992 ceramic particles 0.5mm diameter, 1.7g/cm3 Anaerobic fermenter Demirci et al., 1993 pea gravel 3M-macrolite spheres PP-composite chips aluminum oxide PP+oat hulls, 2-3x3mm Anaerobic fermentei Hem, 1991 KMT media Kaldnes media pre-cursor Nitrifying reactor Lessel, 1991 Bionet Linpor Ringlace rigid polyethylene mesh floating foam cubes flexible PVC rope Nitrifying reactor Meraz et al., 1995 PE pellets 0.28-0.56mm diameter, 0.27g/cm3 Anaerobic fermentei Ney et al., 1990 porous glass rings 25x26mm, 1.7g/cm3, 55%porosit Anaerobic fermentei Tsubone et al., 1992 porous PP carrier 3x3mm, lg/cm3 Air-fluidized-bed PE - polyethylene; PP - polypropylene 17 The research of each paper referenced in Table 2.1 is summarized in the following paragraphs. Cheng (et al., 1992) performed methanation research with a pilot-scale Upflow Anaerobic Sludge Blanket (UASB) reactor, fed food processing wastewater. Ceramic carriers were fluidized, in a 21m tall tower, by a laminar flow of 25-37m/hr. System HRT was decreased from 24hr at startup to 1.5hr after methane production was established. Analysis of the carriers after one month revealed negligible bacterial growth. The biofilm development was limited to cracks in the media surface. The smooth surface, combined with high liquid velocity, was inhibiting biofilm colonization. Further analysis after two months revealed complete bacterial coverage of the carrier surface. Bacterial testing of the biofilm indicated that the biofilm consisted primarily of filamentous organisms. Cheng hypothesized that rods and cocci were absent because they were easily dislodged by the wastewater flow. Rods and cocci were present in suspension, supporting his theory. Demirci (et al., 1993) tested three types of suspended carrier in a lab-scale anaerobic fermenter, fed glucose as a carbon source. The three media tested were: 3M-macrolite spheres, pea gravel, and composite polypropylene (PP) chips. The results indicated that only the composite PP chips produced a complete biofilm growth layer. The pea gravel and macrolite spheres produced only a partial bacterial covering, consisting predominantly of filamentous organisms. Further biofilm formation testing with a variety of composite materials demonstrated that, oat hulls and soy flour, combined with PP, produce the best carrier particles. Demirci and Pometto (1995) later hypothesized that, the organic component of the composite carriers increases the affinity for bacterial adhesion, by possibly providing a carbon and/or nitrogen source. 18 Hem (1991) performed biofilm research on fluidized attached growth nitrifying reactors, lab-scale then pilot-scale. He chose a, newly developed, cylindrical carrier called KMT as the experimental media. This media was a pre-cursor of the Kaldnes media tested in this research. The original KMT media carrier differs from the current Kaldnes design by its lack of fins along the carrier's outer circumference. Hem found that, even after the biofilm was established, there was little growth on the outside of the carrier. Biofilm on this surface is unprotected and can be easily dislodged by either physical contact with other carriers or internal reactor structures. Lessel (1991) tested the performance of three media types in a full-scale nitrifying reactor. The media tested were the following. • Bionet, a rigid bio-filter material fixed in cages. • Linpor, porous cube-shaped suspended carrier media. • Ring-lace, flexible rope media attached to rigid cages. The same material is currently known as Ringlace® media. The experimental results indicated that both Bionet and Linpor media types were inefficient, as nitrifying media, due plugging and general solids buildup. Lessel recommended that very high shear was required to prevent clogging of the Bionet media, hence extra cost associated with aeration or mixing. Linpor media, once loaded with biofilm, tended to sink to the bottom of the reactor. Lessel found that, up to 50% of the Linpor media in the reactor had settled to the bottom. Ring-lace media maintained its biofilm layer, without clogging. Lessel concluded that, the swaying movement of the media in its cage, prevented excess accumulation of solids. In addition, Lessel demonstrated that, although the accumulation of biomass and solids by Ring-lace media was much lower than the solid 19 structure media and individual carrier media, the nitrification performance of all three was comparable. Meraz (et al., 1995) assessed the immobilization dynamics of a downflow fluidized bed anaerobic reactor. The media, consisting of ground polyethylene pellets (0.28-0.56mm), was fluidized with a laminar flow of 9m/h. The reactor design was similar to that of the BF-system reactors filled with Kaldnes media. Two operational differences between the two designs were that, the reactor used by Meraz, was equipped with an external trap device to filter out media that had sunk to the bottom and a higher laminar flow (approximately 3m/h for the BF-system). Biofilm performance was measured as ability to degrade acetate. The results of batch testing at mesophilic (35 °C) and thermophilic (55 °C) temperatures indicated the following. • Mesophilic biofilm activity increased for the first 70 days of the run. The amount of immobilized biofilm plateaued after 60 days. • Thermophilic biofilm activity was very low throughout the experiment (120 days). The immobilized biofilm was also poorly developed. Meraz hypothesized that, the poor thermophilic performance may have been due to the source of bacterial inoculum. The inoculum was mesophilic sludge, acclimated to acetate degradation. The thermophilic inoculum was incubated at 55 °C for two years prior to use. He noted that, during thermophilic incubation, the sludge lost its granular structure which may have hindered its ability to form a biofilm on the polyethylene pellets. Ney (et al., 1990) performed anaerobic fermentation research on a lab-scale (20L), packed bed, 20 attached growth fermenter. The media chosen consisted of porous glass rings (25mm by 26mm), with a wall thickness of 4mm and pore size range of 60-300/^m. HRT was adjusted from four days, at startup, to 12.5h after two weeks of operation. Methane production had reached a maximum after two weeks. The results indicated the following. • 82% of the total biomass in the reactor was immobilized on the glass rings. • The amount of attached biofilm was evenly divided between the outer surface of the rings and the pores inside the structure. • After eight weeks, the system clogged with biomass. The design of these reactors was quite similar to the BF-system reactors filled with Kaldnes media, which functioned more like a packed bed than a fluidized system, due to low laminar flow. Tsubone (et al., 1992) assessed attached growth media performance in an Air-Fluidized-Bed-Bioreactor (AFBBR). Resin-based media, which entraps bacteria, was produced by combining concentrated activated sludge with a photopolymerizable resin. The carriers produced were approximately 3mm in diameter. Results comparing the performance of entrapped media with non-entrapped media indicated that, there is no advantage to entrapping biomass inside the media. Tsubone hypothesized that, most of the biologically-mediated reactions occur in the developed biofilm, not the entrapped organisms. Tsubone developed list of criteria for a suitable suspended carrier (see page 16). A cylindrically shaped, expanded polypropylene carrier (CEPPC) was developed which matched the listed criteria. The media, resembled cylindrical "cheerios" in appearance (3mm by 3mm cylinder with a 1.5mm inner diameter). This media was shown to be very effective at treating high BOD wastes in an AFBBR reactor. 21 2.5 BIOLOG Assay The Biolog system (Biolog, Inc., Hayward, California) is a bacterial identification method which identifies pure isolates based on their sole source carbon utilization pattern. The microplates each contain 96-wells; 95 wells containing a single carbon sources and one well with no carbon source. Carbon source utilization (oxidation) is measured colorimetrically by the reduction of Tetrazolium violet dye from its colorless form to violet. The intensity of this color formation reflects the level of oxidation and can be measured either by manual color comparison of the test wells with the control or by using a MicroStation automated spectrophotometer unit marketed by Biolog Inc. GN-type (Gram Negative) plates were chosen for the assay. The color patterns and resulting computer database based identifications (Microlog computer database) produced from BF-system samples cannot be interpreted as true identifications because the samples were not purified cultures. The Biolog system is designed for identification of pure cultures and hence the database reflects the biochemical diversity of individual organisms only. The assay will however, yield a map of biochemical activity for the mixed population. This map or fingerprint represents the prevalent biochemistry of the reactor populations as a group. The Biolog system has been previously used for the purpose of defining the bacterial activity profile of soil populations and how this profile varies with habitat (Garland and Mills, 1991). The pattern of positive and negative reactions has been used to indicate the level of biodiversity in soil populations in terms of functional diversity and temporal development (Zak et. al.,1994). The extent and time required for dye formation has been monitored in several publications with dye intensity counts taken at various intervals for 20 hours (Haak et al., 1995) to 72 hours (Zak et. 22 al.,1994). Functional diversity in this experiment is defined as the ability, as a group, to utilize a large number of substrates. Since the starting materials and populations were the same for all systems, differences in biodiversity indicate the effects of conditional differences within each media system. Once a pattern typical for a productive short HRT acidogenic fermenter is determined, it is hoped that Biolog testing of individual systems can be used to assess the overall health of fermenter's biofilm. In the BF-system the temporal pattern of positive reactions of each system can be used for future comparison with other measurable biochemical parameters, such as SCVFA production, to assist in the optimization of the fermenter system design. It has been shown that the detection of functional bacterial groups in fermenter systems is more valuable than the detailed description of species present (deBeer and Muyzer, 1995). 23 CHAPTER 3: MATERIALS and METHODS 3.1 Bio-Fermentation System Components A bio-fermentation (BF) reactor system was developed to allow intensive monitoring of fixed film fermenters fed parallel waste streams. Because the goal was to assess SCVFA production within an HRT of one hour, triplicate reactors were connected in series for each experimental condition and operated as a single plug flow system with a total HRT of one hour. The flow through each reactor was adjusted to yield an HRT of twenty minutes. Figure 3.1 represents the original design implemented and Figure 3.2 represents the configuration employed for the two experimental runs. The system operated at ambient temperature for both runs. No attempt was made to control the temperature profile because both the two "media-containing series" and the control series were subjected to the same temperature variations. UBC Pilot plant temperature data indicates that the average seasonal temperature ranges during the two experimental runs are, 17-24 °C and 10-16 *C (Doucette, 1998). Each component of the system is discussed individually and the total flow profile calculations are also included. 24 Figure 3.1: BF-System Initial Configuration TO DRAIN X a) Primary Clarifier 1A2 B2 XC2 B v Q Reactor Feed Pumps XXX XXX XXX PLAN VIEW A,B,ci6 7A §i X til ELEVATION VIEW ft W¥ Controllers A,B,C-10 A,B,C-11 A,B,C-6 A.B.C-7 A,B,C-3 13 [9 XA 'B 'C "5 TO DRAIN PI 3Xt{ To Sludge jThickener Primary Clarifier DRAIN t——1><3 —^> DRAIN b) A B P5 (jsample) Pre-filter Bucket Overflow Filter Influent -> Sludge Pump Reactor Feed Pumps-;ure 3.2: BF-System Configuration, Experimental Runs Outlet Gas Release Ports HI ELEVATION VIEW Run#l Controllers Pre-filter Bucket Overflow Filter Sample Ports a) TO DRAIN Influent —> -Influent Pump- s,udge Pump Outlet Y Gas Release Ports JL JL JL ELEVATION VIEW /--Sample Ports b) Run #2 TO DRAIN =CXt Influent Influent Pump DRAIN Reactor Feed Pumps Controllers A B Pre-filter Bucket Screen Bag Holding Bucket 'Overflow _ DRAIN B) Reactor Feed Pumps 3.1.1 Primary Clarifier A cylindrical clarifier, operated with a functional volume of 600L, was plumbed upstream of the process for run #1 to provide primary effluent as feed. The operational HRT range of the clarifier was approximately 85-100min. The HRT cannot be calculated more accurately due to unaccountable flows within the system (due to a worn sludge pump and leaks between the outlet weir and main tank). The solids retention time (SRT) was minimized to reduce insitu fermentation by pumping out bottom solids frequently, resulting in a negligible sludge blanket. The effluent level within the clarifier outlet chamber was raised with a fifteen centimeter length of/2" pipe to just below the weir level, to reduce aeration of the primary effluent. Prior to the pipe addition, the effluent dropped fifteen centimeters over the weir as a thin film. 3.1.2 Storage Bucket Wastewater during both runs was gravity fed to a 20 L storage bucket fitted with an overflow, rinsing drain and the BF system inlet for the duration of both runs. All samples representing system influent were taken from this bucket once it was deemed that samples taken directly from the reactor inlet pipe were not representative. During run #1, the BF system inlet was fitted with a cylindrical metal mesh filter (3mm pore size) to prevent entry of larger fragments from entering the system. This mesh was removed prior to the second run due to severe mesh plugging when employed to screen raw wastewater. An inner mesh with a larger pore size (5mm) was retained for the raw sewage run. A mixer was added to the bucket to reduce solids settling along the bottom. The bucket had an HRT of about 3.5min, assuming the designed overflow rate was maintained during a 24hr cycle. An overflow rate of 10% of the designed influent flow rate was maintained to reduce scum 27 layer formation at the bucket surface. A low fluid level sensor was added for run two, which turned off the BF-system pumps if the influent level in the bucket was less than one third. 3.1.3 Pumps and Plumbing Three identical screw pumps (Moyno 500) were employed to provide continuous pressurized flow, one for each of the three reactor lines of the BF-system. Pump speed was controlled by three controller units (SP500) allowing individual adjustment. One pump pressurized each reactor series, allowing fine tuning of flows, as well as the ability to maintain different specific HRT settings on parallel reactor series within a single run in the future. Each pump was adjusted to deliver a flow of 1.4 L/min through each reactor line. A rigid Vi PVC head works plumbed from a single storage bucket outlet split the flow three ways and into each pump. Downstream of the pumps, each reactor line had a separate and, for the most part, parallel feed line constructed of Vz" PVC and Vz" flexible PVC hose. Ball valves and unions were installed between each reactor to allow a variety of sampling and maintenance options as well as the removal of an individual reactor without disassembling the rest of the system. An outlet manifold was mounted at the end of each series, fifteen centimeters above the outlet level prior to dropping the effluent to a communal 2" PVC drain. The raised outlet assured that the system would flow under positive pressure and the reactors would remain full. Gas buildup was periodically released through a valve at the top of each reactor. BF system effluent was wasted directly to drain for both runs. Designed flow settings were calculated for a reactor volume of 28L, to account for interior plumbing and minor construction differences between the reactors. Actual flows were measured and optimized manually at the outlet of each series, to assure similar flows 28 for each series. Two other pumps (Moyno 33101) were used to operate the BF-system; one to pump system feed and a second to pump out primary sludge from the clarifier. System feed flowed continuously at 6-7L/min into the primary clarifier during run one and 5-6L/min into the filter screen bucket during run two, depending on the sewage levels in the pilot plant storage tanks. The sludge pump was operated intermittently to maintain a minimal sludge blanket in the clarifier; 1.5mm on, 2.5min off. 3.1.4 Reactors The reactor system consisted of nine reactors connected in three series of three reactors and designed to be fed equal portions of feed water. The three series (K, R and C) were similar with respect to hardware and designed flow. Each series contained a different attached growth media: series K contained Kaldnes media; series R contained Ringlace® media; and series C contained no media. The two media types are characterized later in this chapter (section 3.2). Series C was the control or baseline for the system, reflecting chemical changes to the influent due to insitu fermentation by the naturally occurring suspended and fixed biomass. This parallel reactor design allowed for a true control reactor component, since each system within a run was fed the same waste stream to compensate for periodic influent make up fluctuations. 29 Figure 3.3: BF-System Reactor Components A - Kauldness media B - Ringlace® media C - intake jet D - Diffuser plate E - outlet / sample port F - vent / sample port G - inlet port H - drain Each reactor had a total volume of 30L and was divided into three compartments each separated by a diffuser plate: inlet, media, and outlet chambers (Figure 3.3). Mixing chambers were added to the reactor design to separate the plug flow active reactor zone from the turbulent entry zone of each reactor. Prior to the BF system assembly, dye tests were performed with a larger scale clear PVC reactor to determine the mixing and flow characteristics of the mixing chambers. The results were used to determine design specifications of the mixing chamber, diffuser plate and nozzle size of the inlet ports. Each chamber was defined as the area between the end surface of the reactor and a diffuser plate with 1/4" holes drilled at Vi centers. The holes were chamfered to 3/8" on both sides of the plate to provide better flow. The outlet chamber and inlet mixing chamber were identical, 30 except for a 3/8" diameter jet adapter, which was added to the inlet pipe of the inlet mixing chamber. The dye tests had indicated that a jet smaller than V2" was necessary to stimulate turbulence within the mixing chamber. The media zone was defined as the large central area between the diffuser plates and was designed as plug flow. The two media types resided in this zone either in bags (Kaldnes) or on PVC racks (Ringlace®). Kaldness media was encased in 1/4" mesh sacks, 20-21" in length, during the primary effluent run. However, for the raw wastewater run, it was loaded without sacks in an attempt to reduce plugging within the bags. Cylindrical PVC cages were constructed to provide a removable support structure for the Ringlace®. The units consisted of two 1" long cylindrical sections mounted on four 10" by Vi" hollow spacers by 11" long steel rods (Figure 3.4). By sizing the cages to yield multiple units per reactor, destructive biofilm sampling was possible without disturbing the bottom % of each reactor. In addition to sampling considerations, the small unit size allowed the option of running the system with different amounts of media per unit Of reactor volume. Although the nine reactors were designed to be identical, there were some structural and plumbing differences which have resulted in non-identical volumes (Table 3.1). For all HRT calculations, the volumes of the upper mixing chamber, lower mixing chamber and media zone are defined as IL, IL and 28L respectively. 31 Figure 3.4: Ringlace® Cage Detail Loaded Cage Detail (side view) 29cm omit-Cage Structure Loaded Cage Table 3.1: BF-System Reactor Working Volumes Media Kaldnes Ringlace Control Reactor #1 (L) 21.3. 26.2 28.4 #2 22.0 26.3 27.5 #3 22.5 26.1 28.5 Average (Mean) 21.9 26.2 28.1 Media Volume 6.2 1.9 na Media / reactor 28% 7% na 32 3.1.5 Pre-Filter Bucket A second reservoir (Figure 3.5) was placed upstream of the original filter bucket to house a screen bag (4mm pore size) to catch the larger solids particles when the system feed was switched to raw wastewater for run two. This reservoir was designed to run almost empty to reduce the retention time within it, and hence, reduce insitu fermentation. The disposable screen bags were changed periodically when they filled up with solids and impeded flow. Within one day, the 4mm bags were shown to be inadequate and were replaced with larger pore size bags,7mm. Figure 3.5: Run #2 Pre-filter Bucket 33 3.1.6 BF-System Flow Profile Each experimental run utilized a different pool of equipment. As a result, the flow profile of each run is unique and is discussed separately. For a schematic view of the flow profile for runs one and two, see Figures 3.6 and 3.7, respectively. Definitions of each flow component are listed below. XL raw sewage into the system X4ABC BF-system influent flow X2 clarifier sludge waste flow to the thickener X5 bucket overflow X3 primary effluent flow Xg pre-filter bucket overflow 34 Figure 3.6: BF-System Flow Definitions, Run #1 Figure 3.7: BF-System Flow Definitions, Run #2 35 3.1.6.1 Run#l The design flow for each reactor series was 1.5L/min (X4ABC)which translates to a minimum of 4.5L/min entering the influent pump head works. After trial and error testing, it was determined that the clarifier sludge waste flow (X2) had to be maintained at 1.2L/min at a 15% duty cycle. In order to regain a full bucket each duty cycle , the minimum raw sewage flow (XJ was calculated to be 6.5L/min. Periodic overflow (X5) on each duty cycle served the purpose of scum removal from the top of the bucket reservoir. Flows were checked daily to account for flow loses caused by hardware inefficiencies and waste stream supply fluctuations. 3.1.6.2 Run #2 The design flows for run two were similar to run one so few changes were made to the system flows. Without the sludge pump running, X] only needed to exceed the X4ABC flow requirements to ensure some scum removal with the bucket overflow (X5). The raw sewage flow was operated between 5-6L/min for the duration of the run. 3.2 Media Biomass in a fixed film system is attached to the media surface, and not floating freely as in activated sludge systems. As a result, the fixed biomass has a greater tendency to remain in the system and reduce the effluent solids level. Thus the total amount of sludge retained by the system during normal operations is significantly increased (Lessel, 1991). Fixed media vary with respect to design but share the following goal: to provide a high surface area per unit volume attachment surface for bacterial growth. The large amount of surface area provided by the media allows for 36 development of a population of biomass with more "active surface" and hence greater metabolic capability than conventional floe solids systems. This enables the organic waste material to be broken down and digested much more quickly than a typical activated sludge process. Decreased retention time translates to smaller system size and therefore, significant cost savings. Both media types tested are being used aerobically in full-scale waste treatment systems (Sen et al, 1993; Hem, 1991). 3.2.1 Kaldnes Media Kaldnes media was developed as a neutrally buoyant attached growth media with the goal of increased nitrification efficiency in aerobic reactors (Hem, 1991), similar to that of Ringlace®. The Kaldnes media for this research was supplied by Anglian Water (Cambridge, England). The patented Kaldnes Process is based on the biofilm which grows on the surfaces of free-floating plastic carriers made of polyethylene (PURAC Engineering Inc, 1998). The carriers, as shown in Figure 3.8, consist of plastic cylinders, 7mm by 9mm, with internal cross-braces and ridges around their circumference. The ridges were designed to protect surface biofilm from abrasion in a fluidized reactor (Hem, 1991). 37 Figure 3.8: Kaldnes Media a) W= 9mm, L= 7mm b) Media in CSTR (PURAC Engineering Inc.) Kaldnes carriers have a density, without biofilm, of just under lg/cm3,which allows them to be easily suspended by a reactor's aerators, in an aerobic process, or by mechanical mixers in non-aerated systems. Research with a polypropylene carrier similar to Kaldnes stated that, because of the protected nature of the interior of the carrier, a biofilm layer was established in three days (Tsubone et el., 1992). The exterior surfaces required twelve days for biofilm establishment. The manufacturer (Kaldnes Miljoteknologi AS, Norway) recommends a maximum media loading of 67% of the reactor volume. This gives an effective media surface area loading of approximately. 350 m2/m3 reactor. Kaldness media was encased in 1/4" mesh sacks to ease insertion into the reactors and reduce the chance of plugging at the diffuser plate, by wedging up against the chamfered holes. Bags were used during the primary effluent run only. The Kaldnes media was loaded without sacks during the raw wastewater run in an attempt to reduce plugging within the bags. The loaded sacks were 20-21" in length, roughly % the height of the media zone. Although this ratio was suggested by the manufacturer for fluidized systems, it was chosen as the ratio for our plug-flow system. The media 38 and bags had a displacement volume of approximately 4.4 L 3.2.1.1 Kaldnes Surface Area and Media Loading All surfaces on the Kaldnes media, including edges, were included in the surface area calculations performed. The surface area calculated is an indicator of the theoretical maximum area possible. Once a biofilm is formed, the areas between ridges and the interior of the cylinder, fill up with growth. Therefore, a better approximation of the "working" surface area is between the above calculation and that of a solid cylinder. Equation 3.1 calculates the surface area of the media's cylindrical walls. Cylinder + outside + Inside Kaldnes Edge Cylinder Cylinder Cylinder (Equation 3.1) Surface area (pi*4.52)-(pi*42) + (2*pi*4.5*7) + (2*pi*3.75*7) = 376mm2 The cross-braces inside the Kaldnes cylinder and the external ridges are accounted for in Equation 3.2. Cross + Cross + Ridges = Kaldnes sides ends Ridges (Equation 3.2) Surface area 8*(7*3.75) + 2[(0.5*8)+(0.5*7.5)] + 18(2*0.5*7) = 352mm2 39 The two solutions above are combined, in Equation 3.3, to yield the maximum surface area possible per Kaldnes particle. Cylinder + Ridges = Kaldnes Surface area Surface area Media total (Eq uation 3.3) Surface area 376mm2 + 352mm2 = 728mm2 7.3cm2 The calculations for the minimum working surface area for the Kaldnes media, representing clogged media, is shown in Equation 3.4. Cylinder wall + Top and Bottom = Kaldnes Full Cylinder (Equation 3.4) Surface area (2 x pi x 4.5x7) + (2 x pi x 4.5x4.5) = 32.1 mm2 The surface area loading per reactor was calculated by, counting the number of Kaldnes units in a IL container, and then extrapolating to determine the total number per reactor. The counting results, conducted by a fellow researcher, indicated that approximately 925 Kaldnes carriers occupy the volume of 1L. By considering the maximum individual surface area and bulk media volume, the relative surface area loading per volume of media was calculated, 676m2 media/m3 media. The Kaldnes media bags occupied approximately 63% of the reactor volume, resulting in a reactor loading of 430m2/m3. If the clogged media model is considered, the calculated minimum reactor 40 loading would be reduced to 19m2/m3. The previous calculations refer to media loading as a function of biofilm surface area. However, since one of the goals of attached growth systems is to increase the SRT, this researcher also calculated the volume loading of the media. Media volume loading was defined as the volume of solids retained by the media per unit volume of the reactor. This value reflected a maximum value, and assumed that the media carriers are completely full of solids. It was assumed that the volume of the media structure is negligible in its impact on the maximum calculated volume of solids retained by each carrier, 445mm3. The Kaldnes media volume loading, in the BF-system, was calculated to be 0.4m3/m3. The above calculations do not include the solids trapped between adjacent carriers, only the solids attached to a free carrier such as in a fluidized system. 3.2.2 Ringlace® Media Ringlace® biomedia was developed in Japan in the 1970's and became available in the United States in 1990. Ringlace® Products, Inc., the supplier of the media for this project, has been the exclusive distributor for North America since 1992. Ringlace® rope media was first used in sewage reactors by Lessel (1991) in an attempt to upgrade the nitrification performance of an aeration tank. He found that the use of such an attached growth media effectively increases the sludge age relative to a suspended growth system with the same HRT. 41 Figure 3.9: Ringlace® Media. a) Ringlace Strands, b) Full-scale Ringlace Cages (Ringlace® Products Ltd., Website) Ringlace® media consists of polyvinyl chloride plastic fibers knitted into continuous strands containing many loops arranged in rows along a 6mm wide ribbon. The ribbon is wound in a tight spiral with the loops on the outside, resulting in the appearance of a rope (Figure 3.9a). The media must be mounted to a fixed lattice, such as the full-scale cages pictured in Figure 3.9b, as Ringlace® media is not self supporting. Once mounted, it is stretched to 110-115% of its relaxed length to allow some strand movement but to reduce the chance of entanglement between parallel sections as suggested by the manufacturer (Ringlace® Products Ltd., Oregon). The established ratio of media to reactor volume for Ringlace® is 18 - 2 8% to maintain homogeneity (Lessel, 1991). However, this range is designed for aerobic conditions and should not apply to this research. Full-scale Ringlace® installations have been in operation world-wide since 1978, with over 300 Ringlace® installations in Japan (www. Ringlace®. com). All of these systems employ Ringlace® in an aerobic environment, this research however, focuses on Ringlace® performance in an anaerobic environment. 42 Ringlace® media requires a rigid mounting surface to maintain the strand spacing recommended by its manufacturer. Custom fabricated cylindrical cages were constructed from PVC and stainless steel which fit, length-wise, three deep into each of the three Ringlace® media reactors (Figure 3.3). Each cage housed thirty-eight strands mounted from six PVC rungs on each end plate for a total length of eleven meters of Ringlace®. Strands were aligned parallel to the plug flow and affixed at one inch (2.5cm) centers with PVC zap straps. Previous research used 4.5cm spacing between strands of Ringlace®, mounted inside nitrification/denitrification reactors (Lessel, 1991). Total length per reactor was therefore 33m and total length per 60min FfRT of 99m. The media loading for the Ringlace® was 3500 m/m3 This is much higher than the recommended media loading range, which is 250-500m Ringlace®/ m3 cage volume (Lessel, 1991). The recommended loading values refer to an aerobic system, which requires aeration between strands. One of the goals of this research was to maximize the biofilm contact, so a tighter packing distance than the recommended value was chosen. Pictures of the cage structure loaded with media are displayed in Appendix C. 3.2.2.1 Ringlace® Surface Area and Media Loading The Ringlace® media strand was unrolled to yield a flat 6mm wide ribbon, woven from both loop material and other support fibers. Four bundles of loop strands are woven in across the width of the ribbon (Figure 3.10). The loop bundles each consist of ten similar length loop strands and protrude from only one surface of the ribbon. When the ribbon is allowed to form a tight helix, the loop-side faces outward. The Ringlace® rope was stretched to approximately 110% of its static length and several five inch sections were marked with a felt pen. Each section was unrolled and re-measured 43 to determine the "ribbon to rope" length ratio. The ratio was found to be approximately 1.1:1, ribbon to rope, respectively. Several loop bundles were cut off, as close to the ribbon surface as possible, to determine the average length. Figure 3.10: Ringlace® Media, Surface Detail Inner Side Outer Side Stretched Unrolled Unrolled Side view Flat view Loop Detail 44 For the purpose of surface area determination, the following assumptions were made: • The active biofilm matrix forms on the loops only, therefore only the loop area was counted. • Loops contribute to the surface area equally and are of identical functional length. • Area was calculated based on the media being stretched to 110% of its natural length. • 1 cm of stretched Ringlace® unrolls to 1.1 cm of ribbon. • The strand thickness is 0.1mm. The following equations were used to calculate the surface area per centimeter of Ringlace® stretched on the cage lattice. Equation 3.5 relates the total individual strand length of each fiber in 1cm of stretched Ringlace® media. loop number of number of Stretch Strand length x loops per x bundles x correction = Length (Equation 3.5) (cm) bundle per cm per cm 2.6cm x 10 x 20 x 1.1 572cm 5720mm 45 Equation 3.6 relates the total strand surface area in 1cm of stretched Ringlace® media. Strand Strand Strand 2 x pi x Radius x Length = Surface area (Equation 3.6) per cm per cm 2 x 3.1416 x 0.05mm x 5720mm = 1797mm2 18cm2 Since each reactor contains 33m of Ringlace®, the surface area per reactor can also be calculated. The maximum Ringlace® surface area per reactor was calculated to be approximately 5.9 m2. The loading per cubic meter of reactor was 208 m2/m3. At the cessation of the runs, it was noted that the Ringlace® media resembled a solid rope surface, although not smooth. Therefore, the functional surface area and media loading, of this media fall somewhere between the above calculation and the calculated minimum surface area of the media, shaped like a cylinder, with a 9mm radius. The calculated minimum surface area for Ringlace® media in the current BF-system configuration is 67 m2/m3 (1.9 m2/reactor). The media volume loading of Ringlace® media was calculated, in a similar manner as the minimum surface area calculation above, and assumed that the media is shaped like a rope with a 9mm diameter when full of solids. The calculated maximum volume of solids retained by Ringlace® media was 2.5xl0"4m3/m. The maximum media loading for Ringlace®, in the current BF-system configuration, was 0.3m3/m3. 46 3.3 Analytical Procedures The sampling and processing scheme was designed to address temporal variations in the following parameters: VFA concentration, total organic carbon(TOC), total chemical oxygen demand(tCOD), soluble COD components and turbidity. 3.3.1 Sampling Initially, all sampling was done to coincide with the HRT of the system, in order to better represent real changes to the primary effluent. Samples were taken "forward" starting from the system influent and ending with the effluent of the third reactor in each series. This sampling scheme was designed to overcome temporal variations in sewage stream makeup by sampling from the "same" slug of material by timing the sampling to coincide with the HRT of each reactor. However, this sampling scheme was abandoned due to the disruptive effects of sampling on system flow continuity. Each reactor sampling port was allowed to run for 30 to 60 seconds prior to sample collection which channeled the system flow away from the downstream reactors. The sampling scheme was reversed for the experimental runs. Samples were collected from the outlets of the third reactors and last from the reservoir bucket. This method resulted in no downstream disruptions occurring prior to sampling. Unless otherwise noted, the pre-filter bucket and each reactor outlet flows were sampled for each parameter. The crude sample consisted of 40-50ml aliquots of liquid. Samples were extracted from each sample port, following flushing, to ensure that the samples represented the makeup of the active process stream not the idle liquid present in the sampling pipe. Each sample was processed and preserved 47 at the pilot plant within twenty minutes of sampling. Sub-samples requiring unfiltered aliquots were dispensed directly from the crude sample bottles and preserved, with the exception of suspended solids samples (which require no preservative). Approximately 10-20ml of each crude sample was vacuum filtered through 1.5 urn glass fibre filters (Whatman 934-AH,). The resulting filtrate was dispensed in multiple l-3ml aliquots for laboratory analysis. All samples were chemically preserved and refrigerated to prevent degradation before analysis. Since the 24 hour sampling experiment of the pilot plant influent revealed regular temporal changes, sampling was done once during each sampling day during a specified time period when the influent make-up was constant. Samples were taken three times per week during both runs, except where operational problems prevented regular sampling. The following section lists and discusses relevant sampling and maintenance changes made during the BF-system runs. Reactor flushing protocol was constantly adjusted to try to optimize solids removal while reducing media exposure to oxygen. The importance of a strict regimen of flushing became evident during run #2 because of the high solids loading of a BF-system using raw wastewater as feed. Details of this progression and an outline of the best protocol are included in the following section. 48 3.3.1.1 Preliminary experiments (June-July, 1996) • BF-system was operated as an up-flow system. • Initial sampling window was set to 8-9:30am due to the results of the 24h sampling. • Samples were taken from all sampling points at once and processed at the UBC Pilot Plant. Sampling order was as follows: K3, R3, C3, K2, R2, C2, KI, Rl, CI, storage bucket and primary effluent. • A four inch impeller type mixer was added to the storage bucket after it was noted that a sludge blanket (2-5cm) formed on a daily basis. • The up-flow design was shown to result in the accumulation of fine solids within the reactors. Within eight weeks of operation, several of the BF reactors were completely filled with fine solids. 3.3.1.2 Run#l (August-November, 1996) • BF-system re-configured as a down-flow system. • Due to results of the particle analysis experiment, the Whatman no. 4 filter was dropped in favor of Whatman glass fiber filters for all analyses. • Sampling order was adjusted to reduce the time interval between sampling and processing each sample. Samples were sampled and processed in four groups as follows: 60min samples, 40min samples, 20min samples and finally the bucket and primary effluent samples. • Sampling window was moved to 3-5pm (September 12,1996) due to personal inefficiencies with early morning sampling. 49 3.3.1.3 Run #2 (December, 1996-February, 1997) • Began flowing raw wastewater December 10,1996 • Switched filter bags from 4mm to 7mm pore to alleviate plugging • Reactor Al plugged (December 17), cleaned out Kaldness media by hand. • Noted that the raw influent contained few solids for approximately one week due a pilot plant reservoir mixer breakdown (December 26 - January 4). • As of the January 14 sampling, a "bucket" sample is taken by scooping material directly from the storage bucket. • Reactor KI plugs often and must be force flushed with tap water forwards and backwards under high pressure to release built up solids (January 23 and 30). 3.3.1.4 BF-System Reactor Flushing Protocol • Keep pumps on running during flushing. • Open drain into a bucket, with valve to top and next reactor closed • Momentarily open the top valve to introduce SOME air into top (also speeds flow out and helps dislodge solids.) • Periodically close the drain valve for 10-20s to build some pressure and help dislodge solids, only after some air has entered the mixing chamber at the top of the reactor. • After repeating above until the drain effluent is no longer opaque, open valve to next reactor and close the drain. • Bleed out top air by opening the top valve while closing the valve from the downstream reactor • Repeat the above series for each downstream reactor. 50 3.3.2 Analysis Details of each analytical technique utilized during this research is discussed in this section individually. Preserved BF-system samples were refrigerated for up to three weeks and processed in batches at the Civil Engineering laboratory, either by this researcher or the resident technicians, with the exception of TOC; this assay was performed at the Bio-Resource Engineering laboratory. In addition to regular liquid sampling, samples of each media were removed at the end of each run to assess biofilm activity by determining its crude microbial fingerprint. Metabolic fingerprinting utilized the BIOLOG plate assay. 3.3.2.1 Chemical Oxidation Demand Chemical Oxidation Demand (COD) analysis of the BF-system attempted to describe variations in components of the system feed as it is processed by the reactors. Total COD (tCOD) is defined as the colorimetrically determined COD in an aliquot of the whole, unfiltered, BF-system sample. Soluble COD (sCOD) is defined as the COD which passes through a 1.5um glass fiber filter. Particulate COD is defined as the difference between total and soluble COD measurements for the purposes of these experiments. Both tCOD and sCOD were measured to determine which feed components are being metabolized to yield VFAs. Two techniques were employed for the processing of COD samples (Table 3.2). Both methods were chosen from the established COD protocol of the Environmental Engineering Laboratory at UBC (Hach method, digestion at 150 °C for 2h). Total COD samples were processed by the "high range" method and soluble COD by the "low range". 51 Preliminary samples were processed with both methods to determine the working range of each sample and select the appropriate method. COD sampling was done concurrently with VFA sampling during startup and experimental runs. Table 3.2: COD Protocol Hi Range (20-900) Low Range (0-200) Wavelength used 600 nm 420 nm Standards (mg COD/L) 0, 50, 100, 200, 400, 800 0, 50, 100, 200 Zero" 0 200 COD reagent: (10.217g K2Cr207) + (167ml H2S04) + (833ml H20) (2.27g K2Cr207) + (167ml H2S04) + (833ml H20) - volume dispensed 1.2 ml 1.2 ml Acid reagent: (5.5g Ag2S04 / 1kg H2S04) (S.Sg Ag2S04 / 1kg H2S04) - volume dispensed 2.8 ml 2.8 ml Sample volume 2 ml 2 ml 52 3.3.2.2 Suspended Solids Solids analysis of the BF-system was limited to suspended solids, specifically components filterable by glass fiber filter (Whatman 934-AH, 1.5um). Results are defined by the following parameters. • TSS was defined as the fraction of total sample solids filterable by glass fiber filter which does not evaporate after heating to 105°C. • DSS was defined as the fraction of total sample solids which are not filterable by glass fiber filter and pass through with the filtrate. DSS was not measured during the BF-system experiments. • FSS was defined as the fraction of filterable suspended solids which remains after heating to 550°C. • VSS was defined as the fraction of the TSS which oxidizes and is driven off in gaseous form by heating to 550°C. This value is obtained by the subtraction of FSS from TSS for each sample. A separate 250ml bottle of BF-system sample was taken at each sampling point and refrigerated at the pilot plant, prior to processing for suspended solids in the Environmental Engineering laboratory. A volume of each sample, between 10ml and 100ml, was vacuum filtered using pre-dried and weighed glass fiber filters. All filtrate was discarded and the filters were dried for 12-24 hours at 105^ to determine TSS, then flamed in a muffle furnace for one hour at 550°C to determine FSS. SS analysis was performed during part of run #1 only. 53 3.3.2.3 Total Organic Carbon Total Organic Carbon (TOC) analysis was added near the completion of run one and continued throughout run two. The chosen TOC sampling and processing protocol resulted from verbal consultation with Environmental Engineering laboratory staff. Disposable 4ml Pipettes were modified to serve as both the transfer devices and storage chambers for the 3ml filtered TOC samples. Approximately 2cm of the tapered open end of each pipette was removed to enlarge the aperture sufficiently to allow easier insertion of the preservative, 0.1ml of 0. IN sulfuric acid using a pipettman. The sample aliquot is drawn up by suction prior to addition of preservative. Each filled plastic pipette was sealed by melting approximately 1cm of the open end and pinching it off with pliers. The preserved samples were refrigerated until being processed in two or three week batches. 3.3.2.4 Turbidity Turbidity measurements were taken to assess temporal variations in suspended solids and color at each stage of the system. The goal was to determine the correlation between turbidity and SS using preliminary data and to replace SS analysis for the duration of the experiments in favor of the quicker of the two analyses, turbidity. However, early in run one, turbidity measurements were dropped from the protocol due to inconsistency of results. 54 3.3.2.5 Short Chain Volatile Fatty Acids For the purposes or the BF-system experiments, only short chain volatile fatty acid (SCVFA) concentration was measured. SCVFA results are expressed as total acetic acid equivalents and were calculated from the measured concentrations of propionic acid (CH3COOH) and acetic acid (CH3CH2COOH) by the following equation: acetic s propionic x acetic acid + I [ 0.81 x acid J _ acid Equation 3.7 concentration ^ concentration equivalent Aliquots for VFA analysis were taken from the vacuum filtrate of all BF-system samples. Each 1ml sample was preserved with 0.1ml of 2-3% Phosphoric acid, sealed in glass ampules and refrigerated until being processed in weekly batches by Environmental Engineering laboratory personnel. Samples during preliminary testing were taken from the pre-filter bucket and the sampling ports of the third reactor in each series only. During the actual runs, SCVFA samples were taken from all reactor sampling ports, as well as the pre-filter bucket and raw influent. Sampling was done once per week during the start up period of run one and three times per week during both runs. 3.3.2.6 BIOLOG Assay Samples from the third reactor in each of the tested media series were removed at the cessation of both the first and second runs. These samples, consisting of biofilm and media, were processed to release a portion of the affixed population into suspension. Samples of media were shaken to release biofilm components into suspension. In an attempt to separate the microbial biomass from nutrients 55 in suspension, the samples were centrifuged, washed and re-suspended in sterile distilled water, prior to inoculation of the BIOLOG plates. The procedure resulted in significant solids present in the final dilutions so the protocol was modified for the second run. The media was removed prior to centrifugation step in run two to reduce solids transfer to the dilution series. Some effort was made to preserve an anaerobic environment for the samples but the bulk of the analyses were performed aerobically, according to the recommended modified Biolog analysis for anaerobic organisms in the published user manual (Marello and Bochner, 1989). The modified protocol requires inoculation of a higher concentration of organisms than with aerobic samples to compensate for lack of growth by anaerobic cultures under aerobic conditions. One difference between the recommended protocol and the protocol of this researcher was the reliance on turbidity standards and strict controls on inoculate concentration. In an attempt to reduce exposure to oxygen, this author chose to use several dilutions for each sample, rather than adjust the turbidity to the recommended range. The second series of Biolog analysis employed the use of an anaerobic chamber for incubation and part of the sample processing to reduce the samples' exposure to oxygen. It was hoped that these modifications would increase the biochemical response of anaerobic populations. In addition to manual notation of well pigmentation, the photography of incubated BIOLOG plates was added to the protocol for run two. Two light sources were used; a fluorescent light box was used to back light the plates for all time points and natural sunlight was used for the last time point only. The first analytical run consisted of six gram negative (GN) type plates (Table: 3.3). Three plates 56 were used for each media type. Ringlace® and Kaldness media. Three different concentrations of each sample were inoculated and processed for 24hrs. All of the plates were observed at 4, 8, and 24hrs; distinguishable well patterns were noted. The second analytical run consisted of fourteen GN type plates were inoculated with three concentrations of sample suspensions and processed for 36hrs; seven plates were used for each media type. All of the plates were observed at 17, 24, and 36hrs; distinguishable well patterns were noted. Pigment intensity present in each of the 95 test wells on the Biolog plate relative to the control well intensity was recorded by manual observation and assigned a number between "0-4" representing the range from control well intensity to strongly pigmented wells. Photographs were also taken of all plates throughout run two. 57 Table 3 .3: GN Microplate 1 2 3 4 5 6 7 8 9 10 11 12 A water a-cyclodextrin dextrin glycogen tween 40 tween 80 N-acetyl-D-N-acetyl-D-adontitol L-arabinose D-arabitol cellobiose A galactosamine glucosamine B l-erythritol D-fructose L-fructose D-galactose gentiobiose a-D-glucose m-inositol a-lactose a-D-lactose lactulose maltose D-mannitol D-mannose C D-melibiose B-methyl D-glucoside psicose D-raffinose L-rhamnose D-sorbitol sucrose D-trehalose turanose xylitol methyl pyruvate mono-ethyl succinate D acetic acid cis-aconitic citric acid formic acid D-galactonic D-D-gluconic D-giucosaminic D-glucuronic acid a-hydroxy B-hydroxy G-hydroxy acid acid lactone galacturonic acid acid acid butyric acid butyric acid butyric acid E p-hydroxy phenylacetic acid itaconic acid a-keto butyric acid a-keto glutaric acid a-keto valeric acid D,L-lactic acid malonic acid propionic acid quinic acid D-saccharic acid sebacic acid succinic acid F bromo succinamic glucurona alaninamide D-alanine L-alanine L-alanyl-L-asparginine L-aspartic acid L-glutamic glycyl-L-aspartic glycyi-L-succinic acid acid mide glycine acid acid glutamic acid G L-histidine hydroxy L-proline L-leucine L-omithine L-phenyl alanine L-proline L-pyro glutamic acid D-serine L-serine L-threoline D,L-carnitine G-amino butyric acid H urocanic acid inosine uridine thymidine phenyl ethylamine putrescine 2-amino ethanol 2,3-butanediol glycerol DL-a-glycerolphos phate glucose-1-phosphate glucose-6-phosphate 1 2 3 4 5 6 7 8 9 10 11 12 3.3.3 Experiments 3.3.3.1 Preliminary Experiments Prior to establishing a sampling schedule, a 24 hour sample of raw influent and primary effluent was characterized to determine the validity of grab samples or necessity of sampling within a daily time window for consistent results. The sampling was performed by Frederic Koch, Research Associate, Jason White, summer student and this author on July 4,1997 from 7am until 6am July 4. Raw wastewater and primary effluent were sampled hourly for a 24hr period and processed to reveal daily fluctuations in the following: suspended solids, nitrates, phosphates, COD and VFA's (data presented in Appendix A). Sampling windows were established at times when the makeup of the influent was most consistent: 8-10am and 3-5pm. 3.3.3.2 Main Experiments The first run compared the performance of Ringlace® and Kaldness as fixed film media for the acidogenic fermentation of primary domestic effluent in terms of VFA production and resistance to plugging. The HRT was set to 20min per reactor, yielding a one hour HRT for each reactor series. The first two series were Kaldness and Ringlace® containing reactors respectively and the third series was the control; this series contains no media and reflected changes to the influent resulting from time and naturally occurring biological activity present on the plumbing surfaces and in the bulk liquid. The second run focused on VFA production performance using screened raw wastewater as feed and utilized the same media as the previous run. The BF-system headworks had to be re-worked to 59 accept the raw wastewater. Other than feed related hardware changes, the second run was designed to mimic the conditions in the later half of run one: down-flow design and afternoon sampling. 3.3.3.2.1 Run#l The first run was initiated on June 6,1997 and operated for several weeks as an up-flow system, without sampling to establish the biofilm and standardize operating protocol. After a false start on June 30, the first sample was processed on July 4. All subsequent samples were obtained between 0800hr and 0945hr until September 12. In the hope of processing fresher sewage, the sampling time interval was moved to late afternoon, 1600hr to 1830hr, for the rest of the run. Because the principal filling time for the raw waste water reservoirs was lOOOhr, it was decided that an afternoon sampling would yield results of fermenting fresher sewage than the original morning samples. Run one was halted on July 28, when severe solids buildup was noted within the reactors. The BF system design was changed from up-flow to down-flow, to circumvent the solids build up. The re configured down-flow system was re-started on July 31 with cleaned media. Since all media and reactors were washed during the stoppage, no BF-system samples were taken until Aug 16. Samples were taken approximately bi-weekly until October 15. 3.3.3.2.2 Run #2 The renovated BF-system was started December 13 (a Friday) but was halted within 24hr due to pre-filter plugging. The filtering effect of packed Kaldnes media resulted in a dense mass of solids accumulating in the KI reactor within 24 hours of the start-up of run two. Modifications to the 60 protocol and system head works, mentioned in the sampling section (3.3.1.3), only delayed the inevitable plugging, resulting from running raw wastewater through a packed Kaldnes media bed. The Kaldnes series was operated with a positive pressure of 20-30 psi for most of the second run. The system was operated on and off as modifications were being made during the run itself, until December 30, 1996. During this period plugging was rampant and resulted in two system modifications: usage of a larger pore pre-filter bag and removal of the Kaldnes media bags from within the reactors. Samples for the second run were taken tri-weekly from January 5 to February 17 with the exception of the first week. Due to the distinct possibility of complete plugging, five samples were taken during the first week. 3.3.4 Statistical Analysis All statistical calculations were performed using Quattro Pro spreadsheet software (version 6.0). Statistical analysis of this research addressed the following aspects of the experimental results. Firstly, the sampling and processing error for the SCVFA samples was considered. Samples taken January 5th (1997), were collected and processed in duplicate to test variability in the sampling collection procedure, as well as the variability in sample processing. Secondly, Student t-tests were conducted to compare the two test media SCVFA production trends, against the control series production (0.95%CL). In addition, Student t-tests were utilized as the vehicle for the comparison of SCVFA production differences between the media types. Sequential 61 t-tests were conducted with increasing "hypothesized mean difference" values, until the obtained "t" value was negative, to determine the maximum, statistically valid, difference in SCVFA production. Table 4, in Appendix B, provides an example of the testing method, beginning with a hypothesized difference of OmgHAc/L and ending with a negative "t" value at 18mgHAc/L. The t-tests were conducted on the data sets from both experimental runs. Finally, linear regression analyses were performed to compare the SCVFA production trends, with the trends in TOC, sCOD and tCOD production. Correlation between the data sets was expressed numerically, in terms of the calculated Pearson Coefficient, and graphically by considering the calculated least squares line. The Pearson Coefficient is a dimentionless indicator of overall correlation between two data sets, with a value range of zero to one. The correlation between two data sets in the least squares method is represented by the slope of the calculated least squares line: a positive slope implies positive correlation, a negative slope implies an inverse correlation, and a slope of zero implies no correlation. The regression analyses were performed on data from both media types, but only from the results of experimental run #2. 62 CHAPTER 4: RESULTS AND DISCUSSION Each experimental parameter assayed during the research is discussed individually in the following sections. Trends affecting more than one parameter are expressed with regard to the main goal of the research, SCVFA analysis of the BF system. Tabulated data is included with each section, where appropriate. Data for all experiments is located in Appendix A. 4.1 Chemical Oxygen Demand Variability in COD analysis results of both experimental runs was large. Tables 4.1 and 4.2 summarize the averaged total production values of sCOD and tCOD for both experimental runs. The disparity between the mean and median values of each tCOD data set illustrate the skewed nature of the sample sets. Much of this variability can be attributed to the varying solids load to each reactor, due to system plugging and general solids building up through both runs. Due to the inconsistent results, tCOD analysis was not chosen as an indicator of either the BF system or individual media performance. The high variance number reported for all three reactor systems, in both runs, demonstrates the inconsistency of the tCOD data. Soluble COD data, while still somewhat erratic, provided useful information of BF system performance. Run one sCOD total production indicated a net loss of the measurable sCOD component from the waste stream. This may be due to adsorption and metabolic degradation by the insitu biofilm flora. The results from the control fermenters indicate that, when fed primary wastewater, the sCOD degradation pattern was independent of media within the reactors; hence, the 63 increased working biofilm surface area, provided by the growth media made no detectable sCOD profile changes to the waste stream. Table 4.1: Total sCOD Production (minus bucket) Media type Kaldnes Ringlace Control Run#l Mean (mg/L) -10.3 -8.8 -9.4 Standard Error 4.9 5.0 4.1 Median -6.0 -2.5 -6.5 Standard Deviation 21.8 22.5 18.5 Variance 475.6 504.6 341.2 Range 96.0 87.0 75.0 Minimum -83.0 -68.0 -52.0 Maximum 13.0 19.0 23.0 Count 20 20 20 Confidence Level(0.95) 9.6 9.8 8.1 Run #2 Mean (mg/L) 21.7 10.2 -1.6 Standard Error 2.2 2.9 2.6 Median 21.0 11.0 -1.0 Standard Deviation 8.9 12.2 10.7 Variance 80.1 147.8 114.0 Range 32.0 55.0 50.0 Minimum 5.0 -29.0 -33.0 Maximum 37.0 26.0 17.0 Count 17 17 17 Confidence Level(0.95) 4.3 5.8 5.1 64 Table 4.2: Total tCOD Production (minus bucket) Media type Kaldnes Ringlace Control Run#l Mean (mg/L) 98.3 -22.8 -0.1 Standard Error 52.1 30.4 32.8 Median 28.0 -1.0 -5.0 Standard Deviation 226.9 132.4 143.1 Variance 51487.3 17531.3 20477.8 Range 1020.0 594.0 736.0 Minimum -122.0 -315.0 -261.0 Maximum 898.0 279.0 475.0 Count 18 18 18 Confidence Level(0.95) 104.8 61.2 66.1 Run #2 Mean (mg/L) 127.6 -54.5 149.3 Standard Error 53.2 24.3 53.7 Median 106.0 -53.0 42.0 Standard Deviation 219.3 100.3 221.6 Variance 48094.8 10068.8 49091.1 Range 779.0 396.0 689.0 Minimum -133.0 -252.0 -43.0 Maximum 646.0 144.0 646.0 Count 17 17 17 Confidence Level(0.95) 104.2 47.7 105.3 65 Run two sCOD total production results indicated increased levels in both media reactor streams. Increased sCOD levels indicate a net increase of dissolved organics in the waste stream at each stage of the fermentation. The control reactor series demonstrated similar trends to the control series fed primary effluent in run one, ie, a net loss of dissolved organics. The comparison of control series performance during the two runs indicates that, increasing the feed strength and solids content alone, does not increase the rate of particulate organics breakdown in a short HRT fermenter. Results from the "media-containing series" indicate that, approximately lOmg/1 to 20mg/l of sCOD can be produced from the fermentation of raw wastewater for sixty minutes in a BF system, depending on the attached growth media, Ringlace® or Kaldnes media, respectively. 4.2 Suspended Solids Suspended solids analysis, consisting of TSS and VSS measurements, was performed during part of run one, from August 9 to September 9. Results were found to be highly sensitive to differences in BF-system maintenance procedures, specifically, flushing routines. The solids results could be adjusted up or down, respectively, by increasing or decreasing the volume of wastewater flushed each sampling day. The assay results, while varied, do illustrate the behavior of attached media fermenters relative to excessive solids retention. Figure 4.1 expresses the VSS results for the third reactor of all series. VSS samples from mid August demonstrate the effects of poor system maintenance on solids content in the reactors. Sustainable system maintenance protocols were still being determined during this time interval and the large VSS peaks in all reactor series prove that the SRT of reactors was not being tightly controlled. 66 Figure 4.1: VSS Analysis of the 60min Reactors BF-System VSS Analysis 60min reactor -1000 -i Aug 9 Aug 16 Aug 19 Aug 21 Aug 23 Aug 26 Aug 30 Sep 3 Sep 6 Sep 9 Date 67 Figure 4.2: Averaged Total Solids Analysis 300 250 200 BF-System Solids Analysis Averaged values, Run #1 VSS FSS E '150 o tn 100 50 + + + Inf Buck K1 K2 K3 R1 R2 R3 C1 C2 C3 Reactors 68 The second half of the SS samples processed demonstrate more stability in the solids profile of the system and a much lower SRT. The two media series responded differently to solids throughput. The Ringlace® series SS profile matched that of the control reactors. The VSS graph reporting averaged VSS for all reactors (Figure 4.2), shows the similarity between Ringlace® reactors and their corresponding control counterpart. The third Kaldnes reactor yielded twice the SS content of the other two series. This result is consistent with the finding that, by the cessation of run one, the K3 reactor was almost completely full of solids. 4.3 Short Chain Volatile Fatty Acids The BF system attached growth reactors produced significant SCVFA's with both primary effluent and raw wastewater as feed. The control series in both run one and two produced less than 1.5mg/L and 2.5mg/L SCVFA as acetic acid, respectively. This indicates that the increased bacterial concentration present in attached growth systems are beneficial for the production of SCVFA's under short HRT fermentation conditions. The primary effluent run produced more SCVFA's, after 60 minutes of fermentation, in both "media-containing" series than in the control. However, the production profiles differed for each media (Figure 4.3). The Ringlace® media series produced a total of 5.6mg/L of SCVFA's, but almost 90% of the production occurred in the first 40 minutes. This demonstrates that a reactor designed with Ringlace® media as the attached growth media and a 40 minute HRT is capable of producing 5mg/L of SCVFA's, using primary domestic effluent as feed. The Kaldnes media series produced 69 Figure 4.3: Total SCVFA Production per Reactor (Run #1) Total SCVFA Production Median values, per reactor, Run #1 2.5 - -2 - -E <1.5 u. > o CO 1 --0.5 --Kaldnes Ringlace antral 2 reactor 70 3.7mg/L after 60 minutes, but less than 50% was produced in the first 40 minutes. Solids analysis performed during the first third of run one indicated that the organic solids content in reactor K3 was the highest among the reactors. At the cessation of solids sampling, it appeared that the VSS content in the K3 reactor was double the amount in any of the other media reactors. The effects of twice as much VSS in this reactor may have been inflated the SCVFA production numbers. The correlation between increased reactor VSS and increased VFA production has been documented by Anderson et al. (1994). VFA production in their acidogenic pre-fermenter rose 400% following a 30% increase in system VSS. The two highest production values for reactor K3,11.4mg/L and 7.8mg/L (Table 4.3), coincide with high solids contents at up to five times the VSS levels noted later in the run. Similar peaks occurred in the other reactor series, although only the "media-containing series" displayed noticeable increases in SCVFA production. Reactor R3 displayed both a solids peak from August 16th to 19th, at up to two times the normal solids content, and SCVFA production up to ten times higher than its calculated median value for the remainder of the run. Statistical testing of the median production values for both media, in comparison with the control, revealed significant differences of 3mgHAc/L and 4mgHAc/L for Kaldnes and Ringlace® medias, respectively (Tables 3 & 5, Appendix B). The raw sewage feed run, run two, produced substantially more SCVFA's in all reactor series than in run one (Figure 4.4). The Ringlace® and control series produced approximately double the SCVFA's produced in run one after 60 minutes, 11.2mg/L and 2.5mg/L respectively. 71 Figure 4.4: Total SCVFA Production per Reactor (Run #2) 72 However, the Kaldnes series produced over six times the SCVFA's produced in run one, yielding 20mg/L after 60 minutes. Statistical testing of the median production values for both media, in comparison with the control series, revealed significant differences of 17mgHAc/L and 8mgHAc/L for Kaldnes and Ringlace® medias, respectively (Tables 4 & 6, Appendix B). Of the SCVFA production in the Kaldnes reactors, almost 70% of the acid production occurred in reactor KI (Table 4.4). Similar to the K3 reactor in run one, solids buildup in the KI reactor contributed to the high 20 minute acid production in the second run. The production rate was lower in each subsequent reactor, 5.4mg/L and 1.4mg/L, for K2 and K3, respectively. The K2 acid production level is significant for a 20 minute HRT fermentation, even though the waste stream was essentially filtered by the KI reactor. With this reactor, some solids buildup was present, but certainly not the sole source of the acids production. Reactor K2 did not require the high influent pressure necessary for KI, indicating much less clogging. Therefore, the production in reactor K2 represented attached biofilm activity, not the result of suspended sludge floe fermentation. 73 Table 4.3: Run #1 SCVFA Production (minus control) HAc mg/L Kaldnes Ringlace Date KI K2 K3 Rl R2 R3 July 11 2.9 0.1 2.5 1.4 -5.4 7.8 July 17 -0.1 3.7 -0.3 -1.3 0.7 -2.4 Aug 16 3.5 -5.3 15.7 7.0 -5.2 8.4 Aug 19 -5.0 5.9 6.7 -1.1 0.1 5.1 Aug 21 1.2 1.0 4.2 -0.0 3.4 -0.1 Aug 23 -7.7 9.3 1.6 -7.1 11.1 -0.3 Aug 26 -0.2 -2.2 3.6 1.5 -0.2 2.1 Aug 30 4.8 -0.3 -3.0 0.7 3.0 -0.7 Sep 3 1.6 1.1 8.3 4.1 0.1 5.9 Sep 6 -0.2 1.1 -0.0 1.2 1.8 -0.0 Sep 9 -1.9 2.2 1.1 0.6 1.4 -0.7 Sep 12a -0.4 -0.3 4.4 2.1 0.5 0.3 Sep 12b 2.0 -0.2 3.2 3.3 1.1 0.7 Sep 20 -0.4 0.1 4.8 1.4 1.4 2.6 Sep 23 0.9 -0.3 1.4 2.2 0.9 0.5 Sep 25 -0.8 0.8 0.4 0.9 2.7 -0.3 Sep 27 0.3 -1.3 1.6 2.3 4.7 -1.8 Oct 6 0.3 0.0 -1.2 3.8 -2.1 1.1 Oct 7 -0.8 -1.2 0.7 -1.3 3.0 1.3 Oct 8 0.2 -0.9 -0.5 3.2 1.0 0.3 Oct 15 -0.2 -0.6 0.1 3.5 -0.2 1.2 Mean -0.2 0.5 2.8 1.5 1.5 1.4 Standard Error 0.6 0.7 1.0 0.7 0.7 0.6 Median -0.2 -0.2 1.6 1.5 1.1 0.5 Standard Deviation 2.7 3.0 4.2 2.9 3.2 2.6 Variance 7.4 9.1 17.5 8.1 10.0 6.6 Range 12.5 14.6 18.7 14.1 16.2 10.2 Minimum -7.7 -5.3 -3.0 -7.1 -5.2 -1.8 74 Table 4.4: Run #2 SCVFA Production (minus control) HAc mg/L Kaldnes Ringlace Date KI K2 K3 Rl R2 R3 Jan6 8.2 6.1 1.1 2.8 0.5 3.3 Jan7 9.5 1.0 6.8 2.1 -1.1 3.8 Jan9 9.4 3.3 3.6 4.7 1.7 3.7 JanlO 8.3 1.7 4.5 5.1 0.1 4.1 Janl4 6.4 4.3 1.2 2.8 0.1 -1.2 Jan 15 12.2 3.5 -3.3 3.8 5.1 -4.5 Janl7 15.9 6.1 -1.8 7.5 4.2 1.1 Jan20 12.8 10.7 -1.5 2.9 6.9 3.5 Jan22 12.1 8.0 0.4 8.3 0.8 3.8 Jan24 10.3 6.8 3.3 0.4 6.2 3.6 Jan27 10.9 2.0 3.5 2.5 -3.3 4.3 Jan29 9.0 15.4 -4.7 4.2 6.2 0.3 Jan31 19.9 1.0 -0.4 13.8 -5.4 -1.1 Feb3 17.0 2.2 -1.1 3.5 0.8 1.0 Feb5 14.5 3.5 8.0 3.8 1.2 2.2 Feb7 12.1 7.0 -1.2 6.2 0.6 3.4 Feb 10 14.5 1.1 0.8 3.0 7.2 -3.7 Febl2 14.8 2.8 -1.3 1.6 5.7 2.0 Feb 14 12.3 6.8 -1.7 3.2 4.5 -2.1 Febl7 14.3 4.2 -1.3 3.5 2.4 7.1 Feb 19 15.3 2.1 0.1 3.8 2.7 1.9 Mean 12.4 4.7 0.7 4.3 2.2 1.7 Standard Error 0.7 0.8 0.7 0.6 0.7 0.6 Median 12.2 3.5 0.1 3.5 1.7 2.2 Standard Deviation 3.3 3.6 3.2 2.9 3.4 2.9 Variance 11.2 12.8 10.4 8.1 11.4 8.4 Range 13.5 14.4 12.7 13.4 12.6 11.5 Minimum 6.4 1.0 -4.7 0.4 -5.4 -4.5 75 The Ringlace® series produced 2 - 4.5mg/L of SCVFA's in all reactors. Approximately 50% of the total acids was produced in the Rl reactor, the first 20 minutes of fermentation. Subsequent reactors demonstrated production rate drops of 36% and 31%, respectively, for R2 and R3, yielding 2.9mg/L and 2mg/L production. Other published research, utilizing lab-scale, acetogenic stirred batch-fermenters, indicated that a VFA production peak occurs after one to four hours of fermentation, dependant on the process temperature (Kozuchowska and Evison, 1995). Kozuchowska and Evison identified the most efficient temperature for the fermentation of synthetic coffee waste as 451, with an HRT of one to three hours, producing a maximum of 358mg VFA/1. Because the BF-system operated at ambient temperature, direct comparison of SCVFA production is impossible. The above results support the theory that short HRT acidogenic fermentation efficiency relies on the presence of readily degradable components in the waste stream to fuel large SCVFA production. As these feed components are metabolized, and hence removed from the waste stream, the SCVFA production rates drop. The results of run #2 indicate that a BF-system, consisting of Ringlace® media reactors fueled by raw wastewater, will lose half its SCVFA production efficiency if operated at an HRT of over 20 minutes. Error analysis was done to determine variability in SCVFA sample collection and processing. The results indicated that the error associated with collecting duplicate samples was ±3% and the error associated with processing duplicate samples was ±7%. 76 4.4 Total Organic Carbon The results of Total Organic Carbon (TOC) analysis of the BF-system samples can best be interpreted by assessing either production or consumption of measurable TOC, at each stage of the process. Since SCVFAs are also measured as TOC, one would have expected to see a net production of TOC at each stage of the process and to link this production to that of SCVFAs. TOC samples represent dissolved and fine colloidal organic carbon in the system because they were filtered prior to processing. The net production of TOC, in the BF-system, represents the fermentation of suspended organics, adsorbed into the biofilm, and processed by the established flora to yield dissolved organic products. Run one TOC levels remained approximately the same in the Ringlace® and control reactor series (Figure 4.5). However, the Kaldnes reactor series demonstrated high variability and a net loss of approximately 2mg/l in the K2 reactor. This indicates a net loss of free dissolved organics, probably by bacterial metabolism. Reactor K3 results indicate the opposite metabolic conditions, ie, a net production of TOC, due to the metabolic breakdown of suspended organic solids directly extracted from the bulk liquid or stored organics adsorbed to the attached biofilm. Solids buildup in the K3 reactor, noticed near the end of run one, may explain the increase in TOC production. By effectively increasing the solids loading to that reactor, more stored nutrients were provided for the biofilm to break down. TOC sampling early in the operation of run one could have answered this question by indicating K3 production levels prior to solids buildup. 77 Figure 4.5: Average TOC Production, Run #1 Average TOC Production Run #1 20 40 60 time (min) Kaldnes Ringlace Control Total 78 Run two "averaged" TOC production levels were much higher than the production in run one, due to the higher nutrient loading of a raw wastewater feed with respect to primary effluent as feed (Figure 4.6). The control reactor series produced the least increase in measurable TOC of the three reactor series, 2mg/l at sixty minutes (Table 4.5). The Kaldnes reactor series produced the greatest increase in TOC concentration at approximately six times the control increase. The Ringlace® reactor series produced approximately two times the control increase. Production rates were highest in the first reactor of each media series. The third reactor in each series yielded a net loss of TOC in all but the control reactor. This indicates that, after approximately forty minutes of fermentation, the biofilm populations prefer to metabolize the dissolved organics in the waste stream than the remaining particulate organics. Both media systems consumed approximately lmg/1 of measurable organic carbon between forty to sixty minutes of fermentation. Conversely, the control series results indicated no significant change in the TOC of reactor C3 relative to reactor C2, probably due to the lower nutrient consumption in the previous reactors, relative to the "media-containing" series. The ideal fermentation HRT indicated by the results of the BF-system, relative to TOC production, is less than forty minutes for a system fed raw wastewater. No ideal fermentation HRT for a system fed primary effluent can be derived from the first run, due to the lack of visible trends in the data. 79 Figure 4.6: Average TOC Production, Run #2 12 10 Average TOC Production Run #2 E o 2 4 20 40 60 time (min) Kaldnes Ringlace Control Total 80 Table 4.5: Averaged TOC Production, Runs #1 and #2 mg/L Kaldnes Ringlace Control Run#l 20min 0.15 0.34 0.29 40 -1.97 0.65 -0.31 60 0.87 -0.07 0.35 Total -0.86 0.91 0.33 Run #2 20mm 7.41 3.40 0.98 40 4.72 1.90 1.12 60 -0.93 -0.68 -0.10 Total 11.24 4.71 2.00 81 4.5 BIOLOG Assay The experiments successfully produced carbon source utilization patterns on Biolog GN plates. The patterns produced from the first experimental analysis using samples from reactors K3 and R3 at the end of run one showed little correlation between media types and within dilutions of the same sample. These results reflect sample handling and general protocol faults. A significant amount of biofilm solids were present throughout the plating dilution series; this is believed to have affected both, the ability to monitor pigment formation in the undiluted plates, and the intensity of pigment generated. The latter effect resulted from the utilization of biofilm solids as a carbon source in the plate wells, in addition to or instead of the packaged specific carbon source. Adjustments to the protocol for the second experiment, using samples from reactors K3 and R3 at the end of run two, produced reproducible patterns, both within each plate dilution series and between the two media types. The averaged well intensities for all plates processed during both experiments is presented in Table 4.6. The results of the first experiment were not considered representative of the BF-system biofilm, due to the high variance in pigment intensity between the two media and within each dilution series of a single media. The patterns produced in the second experiment show strong similarities within each dilution series, as well as between the patterns produced by each media type. A correlation analysis was performed with the cumulative plate data from the two media types tested. The resulting correlation coefficient, 0.914, shows the strong correlation between the two sets of patterns across the 95 test wells. 82 Table 4.6: Averaged Biolog Plate Data a) Kaldnes, Run #1 (24hr) 1 2 3 4 5 6 7 8 9 10 11 12 A 1 3 3 A B / / 1 1 / B C 1 1 1 C D 1 1 2 2 D E / / 1 1 1 1 E F 1 / / 1 1 2 2 2 F G / / / 1 1 2 2 G H 1 2 1 1 1 1 1 2 2 1 2 H 1 2 3 4 5 6 7 8 9 10 11 12 c) Ringlace®, Run #1 (24hr) 1 2 3 4 5 6 7 8 9 10 11 12 A 1 1 1 2 2 1 1 1 1 1 A B 1 / / / / 1 1 / / 1 B C 1 1 / 1 1 1 / C D 1 / 1 1 1 D E / 1 / / / 1 / E F 1 / 1 1 1 / / F G 1 / 1 1 G H / 1 1 / 1 / / / H 1 2 3 4 5 6 7 8 9 10 11 12 b) Kaldnes, Run #2 (36hr) 1 2 3 4 5 6 7 8 9 10 11 12 A 2 3 / / 2 4 1 2 A B 1 2 1 1 2 1 1 1 B C 2 3 / 1 / 2 3 3 1 C D 1 2 1 3 3 1 D E 1 / 2 / / E F / 1 2 1 / / F G / 2 2 1 G H / 1 / 2 1 2 3 H 1 2 3 4 5 6 7 8 9 10 11 12 d)Ringlace®, Run #2 (36hr) 1 2 3 4 5 6 7 8 9 10 11 12 A / 3 4 / / 1 4 1 / 3 A B 2 1 / 1 1 2 2 1 B C 2 4 / 2 / 3 4 4 2 C D 2 2 / 2 4 4 1 D E / / 3 / 1 2 E F 1 / / 2 1 2 / F G / / 3 3 / G H / 2 / / 1 / 1 2 H 1 2 3 4 5 6 7 8 9 10 11 12 Metabolic fingerprinting of the mixed population from the BF-system, using the Biolog identification system, can yield consistent patterns within a sampling run. However, repetition of the second experiment is necessary before any conclusions can be reached about the validity of the obtained metabolic fingerprint. Future microbiological research is necessary to isolate, purify and re-test the sample organisms, to determine if the metabolic variation is consistent within a given fully functional system. In addition, testing must be done with both aerobic and anaerobic components of an existing sewage system to compare the fermenter biochemical fingerprints with fingerprints produced by other sewage treatment environments. Long term studies are necessary to study the variations in the nutrient fingerprint produced under both ideal performance conditions and conditions which result in low overall performance. Temporal microbial analysis of the prevalent organisms, thriving in the studied fermenter system under varying conditions, would link the BIOLOG results to actual population trends within the reactor and parallel temporal changes in other measurable biochemical components such as volatile fatty acids. The development of a standardized quantitative technique for assessing the pigment producing trends of mixed bacterial populations would allow the sharing of information between facilities that are geographically isolated. If a technique can be developed using scanned images of incubated BIOLOG plates processed by readily available imaging software, the pool of information would be larger than techniques involving costly colorimetric equipment. A technique such as this would also allow the operators of full-scale sewage fermenters to assess the ongoing biochemical health of their systems by comparing "nutrient fingerprints". 84 4.6 Media Comparison The two media tested, Kaldnes and Ringlace®, were compared physically and in terms of SCVFA production. The results of the surface area analysis revealed that, the physical differences between the two media types, translate to differences in usable surface area. The calculations revealed that the maximum and minimum surface area loading possible, in the current BF-system configuration, were significantly different between the two media (Table 4.7). Table 4.7: Test Media Surface Area Loading. Media Type Kaldnes Ringlace Maximum Surface Area Loading 430 m2/m3 208 m2/m3 Minimum Surface Area Loading 19 m2/m3 67 m2/m3 The media volume loading calculations indicated that, in the current configuration, the Kaldnes reactors had the capability to retain, or entrap, 25% more solids than the Ringlace® media reactors. During run #2, this figure was certainly much higher for reactor KI, due to solids entrapment 85 between Kaldnes carriers. Suspended solids analysis, performed during run #1, revealed that the Ringlace® series SS profile matched the control reactor series profile. Therefore, Ringlace® does not contribute to excessive solids deposition in BF-reactors. Conversely, the third reactor in the Kaldnes series SS content was twice that of the corresponding control (and Ringlace®) reactor at the cessation of SS sampling. This researcher believes that the solids buildup in reactor K3 originated from "sluffed" off biofilm from previous reactors, not solids in the primary effluent. Primary effluent solids would have been entrapped in the KI reactor first; this pattern was evident during run #2 where raw waste solids clogged the KI reactor. The results indicate that Kaldnes media is not practical as an attached growth media in the BF-system, due to clogging tendencies. When the BF-system was fed primary effluent, neither media type had any effect on the sCOD concentration of the waste stream. Conversely, when fed raw wastewater, the Kaldnes and Ringlace® reactor series produced 10mg/l and 20mg/1 of sCOD, respectively. The control series in the two runs demonstrated similar trends, indicating that the sCOD production was a direct product of the attached biofilm of both media types. The two media differed in production of SCVFA's, not only in the amount produced, but also in the timing of production within the 60min HRT. The Ringlace® series, fed primary wastewater, produced 5mg/l of SCVFA's in the first 40min. By comparison, the Kaldnes series produced less than 2mg/l of SCVFA's in 40min. Statistical testing revealed that, the difference in median production values between Kaldnes and Ringlace®, was not statistically significant (Table 1, Appendix B). Run #2 SCVFA data for Kaldnes data, specifically the KI reactor, was not considered in the media performance comparison because the excessive solids buildup contributed much of the 86 reactor fermentation. Statistical comparison of the median production values for run #2, revealed a significant difference of 9mgHAc/L between the two media. Correlation analyses were performed between the SCVFA production of the control series, and the two media series, for both experimental runs. The Pearson Coefficients, calculated in Appendix B, were significantly higher for Ringlace® media (0.825 and 0.638), than the values calculated for Kaldnes media (0.450 and 0.499). These calculations indicate that, variability in the Ringlace® series production values followed the variability trends in the control series, more closely than did the Kaldnes series. Ringlace® has proven itself to be the better media for the BF-system in the current plug-flow configuration, especially when the system was fed raw wastewater. The Ringlace® series displayed double the SCVFA production in run #2 compared with run #1, producing almost 5mgHAc/L in 20min. 4.7 Summary SCVFA production in the BF-system was near zero in the reactors containing no media during both the primary effluent and raw wastewater runs. Significant levels of SCVFA's were produced by short HRT attached growth fermenters using both primary effluent and raw wastewater as feed. Both sets of media reactors tested produced more SCVFA's than the control reactors. TOC and sCOD analyses, which measure dissolved organics in the waste stream, indicated a net solubilization of filterable solids in both media series fed raw wastewater. The correlation analysis of the SCVFA production, of both media types, with the other measured 87 parameters, yielded both positive and negative results. Least squares lines calculated for the correlation of "soluble" parameters, sCOD and TOC, with SCVFA production revealed positive correlations (Figures 1-4, Appendix B). Conversely, the results of the tCOD correlation analyses revealed a slightly negative correlation (Figures 5 & 6, Appendix B). However, since the Pearson Coefficients generated for the tCOD correlations were low in both runs, one cannot justify the negative correlation. Therefore, the tCOD data was concluded to have no correlation with the SCVFA production data. Kaldnes media was not suitable as a short HRT pre-fermentation media in the current BF-system configuration, due to inability to handle the solids loading of the waste stream. Suspended solids analysis supported this researcher's choice of appropriate media for future research by demonstrating the tendency of Kaldnes media to become clogged with solids. Ringlace® media has proven itself as a sustainable short HRT pre-fermentation media by resisting plugging under all solids loading conditions tested. Results of the two fermentation runs indicate that, a short HRT fermenter system fed raw domestic wastewater, is capable of producing significantly more SCVFA's than a system fed primary effluent. SCVFA production in the BF-system, by Ringlace® media, would have a significant impact on the phosphorus removal efficiency of a BNR facility. The BF-system was designed as an inline prefermenter, which results in little dilution of the SCVFA's produced, relative to the production of a side-stream prefermenter system. A typical side-stream fermenter is designed to provide 5-10% of the flow of raw sewage influent. For example, a production of lOOmg SCVFA/L in a side-stream fermenter, operating at 5% of the system flow, is equivalent to the production an inline prefermenter, such as the BF-system, producing 5mg 88 SCVFA/L. The two BF-system fermentation runs produced 6mg SCVFA/L*h and 9mg SCVFA/L*h, fermenting primary effluent and raw wastewater, respectively. If it is assumed that, 5mg SCVFA/L are needed to remove lmg/L of phosphorus (Barnard, 1998), the effect of an inline BF-system on phosphorus removal, in downstream BNR facility, may be estimated at an extra l-2mg/l, fermenting primary effluent and raw wastewater, respectively. However, it is important to consider that, the retrofit of a BF-system to an existing BNR facility, may result in the removal in the "last" milligram of phosphorus in the waste stream. As phosphorus effluent regulations tighten, the removal of the "last" milligram of phosphorus, will be of paramount importance. The design of a short HRT prefermenter system can significantly reduce the costs associated with reactor construction and operation, relative to a longer HRT facility (Demirci and Pometto, 1995). The BF-system design combines the space savings, associated with a short HRT system, with the low operating costs of a reactor with no moving parts. Considering the additional savings associated with reducing the size of the anaerobic zone, permitted by the VFA contribution of a prefermenter system (Randall, 1994), the retrofit of a BF-system to an existing BNR facility is a cost effective alternative to sidestream sludge fermentation. 89 CHAPTER 5: CONCLUSIONS and RECOMMENDATIONS 5.1 Conclusions The main objective for this thesis was to assess the SCVFA production performance of two fixed film media types in the BF-system, fed primary effluent and raw wastewater in separate fermentation runs. After reviewing the results of the two experimental runs, the following conclusions were made. BIOLOG analysis was found to produce a reproducible metabolic fingerprint of the prefermenter biofilm. It is therefore possible to qualify the metabolic activity of the fermentation organisms in the BF-system. However, no determination of the bacterial species, active within the BF-system, can be performed using this technique because all samples represent mixed cultures. The BF-system, attached growth fermenters series, produced more SCVFA's than the suspended growth control series. This indicates that a short HRT prefermenter design benefits from the addition of suspended growth media, to entrap solids and stimulate biofilm formation. The two media tested, Kaldnes and Ringlace®, produced 3.7mg SCVFA/L*h and 5.6mg SCVFA/L*h, respectively, during the primary effluent fermentation run. However, the results were shown to be statistically similar. Soluble COD and TOC analyses demonstrated a net increase in dissolved organics over a 60min HRT, during the raw wastewater run. This indicates a net conversion of particulate matter, in the waste stream, to dissolved matter. 90 Kaldnes media demonstrated plugging problems at the cessation of the primary effluent run and early during the raw sewage run. Because of the solids loading problem, Kaldnes was deemed unsuitable for use in a plug-flow, BF-system. A Kaldnes media system must therefore, be fluidized to avoid clogging, either due to entrapped solids or excessive biofilm growth. Ringlace® media produced 5mg HAc/1 after 20min and 11.2mg HAc/1 after 60min when the system was fed raw effluent. The SCVFA production rates of the BF-system, configured as an inline prefermenter, would have a significant impact on the phosphorus removal performance of a BNR system. The BF-system, with an HRT of 60min, could theoretically provide enough SCVFA's to treat l-2mg/L of phosphate, dependant on the type of feed. The production of significant SCVFA's with an HRT of 60min or less, translates to a large savings in construction costs associated with prefermenter design. The shorter HRT design requires a smaller reactor and hence, a smaller footprint at a treatment facility. 5.2 Recommendations Future pre-fermentation research should focus on the following areas: Future research on Kaldnes media as a suitable carrier should be performed in a mixed, or otherwise fluidized, system to prevent the plugging problems encountered in this research. A reactor system for short HRT pre-fermentation research should be designed to accommodate the solids loading of the waste-stream, without the need for "maintenance flushing". 91 In order to answer the question of biofilm behavior, particle adsorption vs particle entrapment, biofilm analyses of both media types should be performed. This would pertain to biofilm thickness, biofilm formation rate, and particle analysis. Analysis of "truly" soluble COD by the FLOC-SOL method should be performed to better understand the flow of nutrients, from suspended to dissolved organics, in a prefermenter system. Long term microbiological studies are necessary to study the BIOLOG generated nutrient fingerprint produced under both ideal fermenter performance conditions and conditions which result in low overall performance. This would link the BIOLOG results to actual population trends within the reactor and ideally parallel temporal changes in other measurable biochemical components, such as volatile fatty acids. 92 REFERENCES Anderson G K, Kasapgil B and Ince O (1994). "Microbiological study of two-stage anaerobic digestion during start-up". Wat Res, vol 28(11) 2382-2392. Argaman Y (1991). "Biological Nutrient Removal". In Biological Degradation of Wastes. Ed. A M Marul, M.U.N., Newfoundland, 85-101. Barnard J L (1974). " Cut P and N without Chemicals." Water and Wastes Engineering, vol 11(7) 33-36. Barnard J L (1992). "Design of Prefermentation Processes." In Design and Retrofit of Wastewater treatment Plants for Biological Nutrient Removal. Eds C W Randall, J L Barnard, and H D Stensel, Water Quality Management Library-Volume 5, 1-23. Barnard J L (1994). "Alternative Prefermentation Systems". In Use of Fermentation to Enhance Biological Nutrient Removal. Water Environmental Federation, Va, USA. Barnard J L (1998). Verbal communication at quarterly research group meeting, at the Environmental Engineering Department, University of British Columbia. Barnard J L and Rabinowitz B (1997). Course notes for CIVIL 566, Civil Engineering Department, University of British Columbia. Barnard J L, Stevens G M, and White J (1995). "Kelowna Demonstrates the Art of Optimization". WatQuallnt, vol.1, 17-19. De Beer D and Muyzer G (1995). "Multispecies biofilms: report from the discussion session". Wat Sci & Tech, vol. 32, 269-270. Bhaskar G V and Bhamidimarri S M R (1991). "Approximate analytical solutions for a biofilm reactor model with monod kinetics and product inhibition." Can J Chem Eng, vol. 69(Apr) 544-47. Bishop P L and Rittmann B E (1996). "Modeling heterogeneity in biofilms: report of the discussion session". Wat Sci & Tech, vol 32(8) 263-265. Cheng S S, Huang S Y, Lay J J, Tsai P S and Cho L T (1992). "Population dynamics of attached 93 biofilm in anaerobic fliudized bed pilot plant." Wat Sci Tech, vol. 26(3-4), 503-10. LeChevallier M W, Cawthorn C D and Lee R G (1988). "Inactivation of biofilm bacteria". Appl & Envir Micro, vol 54(10) 2492-2499. Dawson R N, Abraham K N, Oldham W, and Udala B (1995). "Performance Improves at Penticton." Wat Qual Int, vol. 1, 20-22. Demirci A, Pometto III A L and Johnson K E (1993). "Evaluation of biofilm reactor solid support for mixed-culture lactic acid production." Appl Envir Biotech, vol. 38, 728-33. Demerci A, Pometto III (1995). "Repeated-batch fermentation in biofilm reactors with plastic-composite supports for lactic acid production." Appl Micro Biotech, vol. 43, 585-9. Doucette D (1998). Personal communication regarding UBC Pilot Plant temperature data, Environmental Engineering Pilot Plant, University of British Columbia. Flemming H C and Melo L (1995). "Unwanted biofilms: report from the discussion sessions". Wat Sci & Tech, vol. 32, 267-268. Garland J L and Mills A L (1991). "Classification and Characterization of Heterotrophic Microbial Communities on the Basis of Patterns of Community-Level Sole-Carbon-Source Utilization." Applied and Environmental Microbiology, vol. 57(8), p2351-2359. Gjaltema A and Griebe T (1995). "Laboratory biofilm reactors and on-line monitoring: report of the discussion session." Wat Sci Tech, vol. 32(8), 257-261. Haack K S, Garchow H, Klug M J and Forney L J (1995). "Analysis of Factors Affecting the Accuracy, Reproducibility and Interpretation of Microbial Carbon Source Utilization Patterns." Applied and Environmental Microbiology, vol. 61(4), pl458-1468. Hem L J (1991). Nitrification in a Moving bed Biofilm Reactor. PhD. Thesis, Division of Hydraulic & Sanitary Engineering, The University of Trondheim, Norway. Jardin N and Popel (1997). "Waste Activated Sludge Production of the Enhanced Biological Phosphorus Removal Process." Wat Envir Res, vol.69(3), 375-381. 94 Knight B P, McGrath S P and Chaudri A M (1997). "Biomass Carbon Measurements and Substrate Utilization Patterns of Microbial Populations from Soils Amended with Cadmium, Copper, or Zinc." Applied and Environmental Microbiology, vol. 63(1), p39-43. Koch F A (1996). Verbal communication during research group quarterly meeting, Environmental Engineering Pilot Plant, University of British Columbia. Kozuchowska J and Evison L M (1995). "VFA production in pre-acidification systems without pH control." Envir Tech, vol 16, 667-675. Lazarova V andManem J (1995). "Biofilm characterization and activity analysis in water and wastewater treatment." Wat Res, vol 29(10), 2227-2245. Lessel T H (1991). "First Practical Experiences with Submerged Rope-Type Bio-Film Reactors for Upgrading and Nitrification." Wat Sci Tech, vol. 23(Kyoto), 825-834. Lie E, Christensson M, Jonsson K, 0stgaard K, Johansson P, and Welander T (1997). "Carbon and Phosphorus Transformations in a Full-Scale Enhanced Biological Phosphorus Removal Process." Wat Res, vol. 31(11), 2693-2698. Marello T A and Bochner B R (eds) (1989). Biolog Reference Manual: Metabolic Reactions of Gram-Negative Species. Biolog Inc., Hayward, California. Science Tech Publishers, Madison, Wisconsin. Mauchline W S and Keevil C W (1991). "Development of the BIOLOG Substrate Utilization System for Identification of Legionella spp." Applied and Environmental Microbiology, vol. 57(11), p3345-3349. Mendenhall W and Sincich T (1988). Statistics for the Engineering and Computer Sciences. 2nd ed. Dellen Publishing Co., San Fransisco, California. Meraz M, Monroy O, Noyola A and Ilangovan K (1995). "Studies on the dynamics of immobilization of anaerobic bacteria on a plastic support." Wat Sci Tech, vol. 32(8), 243-250. Ney U, Macario A J L, Conway De Macario E, Aivaidis A, Schoberth S M and Sahm H (1990). "Quantitative microbial analysis of bacterial community shifts in a high-rate anaerobic 95 bioreactor treating sulfite evaporator condensate." Appl Envir Micro, vol. 56(8), 2389-98. Oldham W K and Abraham K (1994). "Overview of Full-Scale Fermenter Performance". In Use of Fermentation to Enhance Biological Nutrient Removal. Water Environmental Federation, Va, USA. Rabinowitz B and Barnard J L (1995). "Biological Nutrient Removal in Western Canada." Proceedings of the First Conference on Biological Nutrient Removal, 22-28. Rabinowitz B and Barnard J L (1997). "The Use of Primary Sludge Fermentation in Biological Nutrient Removal Processes." Proceedings of the Third Conference on Biological Nutrient Removal, Brisbane, Australia, 478-485. Randall C W (1992). "Introduction." In Design and Retrofit of Wastewater treatment Plants for Biological Nutrient Removal. Eds C W Randall, J L Barnard, and H D Stensel, Water Quality Management Library-Volume 5,1-23. Randall C W (1994). "Why Use Fermentation?". In Use of Fermentation to Enhance Biological Nutrient Removal. Water Environmental Federation, Va, USA. Randall C W and Chapin R W (1997). "Acetic Acid Inhibition of Biological Phosphorus Removal." Wat Envir Res, vol 69(5), 955-960. Raper W G C, Bridger J S, Crockett J A, Burton K G, and Lockwood G A (1997). "High Performance Biological Nutrient Removal via Prefermentation." Proceedings of the Third Conference on Biological Nutrient Removal, Brisbane, Australia, 369-376. Ruiz-Trevino F A, Gonzales-Martinez S, Doria-Serrano C and Hernandes-Esparza M (1992). "Phosphorus release kinetics in biofilm reactors." Wat Sci Tech, vol. 26(3-4), 567-76. San H A, Tanik A and Orhon D (1993). "Micro-scale modelling of substrate removal kinetics in multicomponent fixed-film systems." JChem Tech Biotech, vol. 58, 39-48. Shin H-S and Song Y-C (1995). "A model for evaluation of anaerobic degradation characteristics of organic waste: focusing on kinetics, rate-limiting step." Envir Tech, vol. 16,775-784. Stensel H D (1992). "Principles of Biological Nutrient Removal." In Design and Retrofit of Wastewater treatment Plants for Biological Nutrient Removal. Eds C W Randall, J L 96 Barnard, and H D Stensel, Water Quality Management Library-Volume 5, 1-23. Tsubone T, Ogaki Y, Yoshiy Y and Takahashi M (1992). "Effects of biomass entrapment and carrier properties on the performance of an air-fluidized-bed biofilm reactor." Wat Envir Res, vol. 64(7), 884-889. Wilderer P A, Cunningham A and Schindler U (1995). "Hydrodynamics and Shear Stress: Report from the Discussion Session." Wat Sci Tech, vol. 32(8), 271-72. Wilson W and Keller W (1995). "Fermenting Ideas Aid Nutrient Removal at Bonnybrook." Wat Quallnt, vol. 1,24-26. Zak J C, Willig M R, Moorhead D L and Wildman H G (1994). "Functional Diversity of Microbial Communities: A Quantitative Approach." Soil Biology and Biochemistry, vol. 26(9), pi 101-1108. Zehnder A J B (1988). Biology of anaerobic microorganisms. Chapters 9.2-9.3. A Wiley-Interscience Publication, John Wiley & Sons, Toronto. Miscellaneous References: Anglian Water, Cambridge, England., supplier of Kaldnes media. Ringlace Products, Inc., Portland, Oregon. PURAC Engineering, Inc., 4550 New Linden Road, Suite 500, Wilmington, DE 19808. Tel: 302-996-0545; Fax: 302-0996-0544 www.Ringlace.com , Ringlace® products website. 97 LIST OF APPENDICES APPENDIX A: RESULTS 99 APPENDIX B: STATISTICAL ANALYSIS 112 APPENDIX C: PHOTOS 119 98 APPENDIX A: RESULTS Table 1: sCOD, Raw Values 100 Table 2: sCOD, Corrected for Bucket Level 101 Table 3: tCOD, Raw Values 102 Table 4: tCOD, Corrected for Bucket Values 104 Table 5: TOC, Raw Values 106 Table 6: TOC, Corrected for Bucket Values 107 Table 7: SCVFA, Preliminary and Run #1 8 Table 8: SCVFA, Run #2 109 Figure 1: 24hr Sample, COD Analysis 110 Figure 2: 24hr Sample, Ortho-P AnalysisFigure 3: 24hr Sample, Ammonia Analysis Ill Figure 4: 24hr Sample, Ammonia Analysis99 1.1 Analytical Data Table 1: sCOD, Raw Values Run#l Influent Buck K R C mg/1 1 2 3 1 2 3 1 2 3 Aug 9 233 150 165 201 Aug 30 142 162 147 143 137 Sep 3 188 205 187 157 153 Sep 6 136 139 142 141 152 126 141 137 140 142 143 Sep 9 138 146 151 148 149 Sep 12a 149 147 133 143 138 Sep 12b 148 150 144 148 142 Sep 20 148 152 150 163 152 Sep 23 167 159 156 156 144 Sep 25 147 146 155 158 159 Sep 27 150 147 142 154 157 158 152 163 157 151 153 Oct 6 178 153 146 148 147 Oct 7 145 143 137 162 145 Oct 8 Oct 15 148 147 139 157 141 Average 155 163 155 156 152 167 158 154 162 164 153 RUN #2 Influent Bucket K R C mg/1 1 2 3 1 2 3 1 2 3 Jan5 165 168 173 173 160 Jan6 143 147 158 168 158 Jan7 143 144 155 149 163 142 148 160 156 163 151 Jan9 142 154 176 165 150 JanlO 167 164 179 183 190 170 182 185 168 173 165 Jan 14 167 174 188 171 167 JanlS 150 156 168 158 169 157 166 162 155 149 155 Janl7 152 155 182 171 155 Jan20 157 163 190 179 163 Jan22 148 159 167 165 180 162 164 170 152 153 154 Jan24 157 153 190 179 155 Jan27 185 173 206 187 190 Jan29 168 174 190 183 173 Jan31 169 167 187 217 202 189 178 177 174 178 173 Feb3 175 174 202 187 163 Feb5 179 189 203 160 156 Feb7 nd 190 211 200 189 Average 151 165 171 174 187 164 168 175 161 163 163 100 Table 2: sCOD, Corrected for Bucket Level RUN#1 Influent K R C mg/1 1 2 3 1 2 3 1 2 3 Aug 9 -233 -83 -68 -32 Aug 30 -20 -15 -19 -25 Sep 3 -17 -18 -48 -52 Sep 6 -3 3 2 13 -13 2 -2 1 3 4 Sep 9 -8 5 2 3 Sep 12a 2 -14 -4 -9 Sep 12b -2 -6 -2 -8 Sep 20 -4 -2 11 0 Sep 23 8 -3 -3 -15 Sep 25 1 9 12 13 Sep 27 3 -5 7 10 11 5 16 10 4 6 Oct 6 25 -7 -5 -6 Oct 7 2 -6 19 2 Oct 8 Oct 15 1 -8 10 -6 Average -18 -12 -11 -11 -0 -9 -9 -5 -4 -10 RUN #2 Influent K R C mg/1 1 2 3 1 2 3 1 2 3 Jan 5 -3 5 5 -8 Jan6 -4 11 21 11 Jan7 -1 19 16 7 Jan9 -12 22 11 -4 JanlO 3 15 19 26 6 18 21 4 9 1 Janl4 -7 14 -3 -7 Janl5 -6 12 2 13 1 10 6 -1 -7 -1 Janl7 -3 27 16 0 Jan20 -6 27 16 0 Jan22 -11 8 6 21 3 5 11 -7 -6 -5 Jan24 4 37 26 2 Jan27 12 33 14 17 Jan29 -6 16 9 -1 Jan31 2 20 50 35 22 11 10 7 11 6 Feb3 1 28 13 -11 Feb5 -10 14 -29 -33 Feb7 -190 21 10 -1 Average -14 14 19 22 8 11 10 1 2 -2 101 Table 3: tCOD, Raw Values RUN#1 Influent Bucket K R C 1996 mg/1 1 2 3 1 2 3 1 2 3 Aug 16 202 360 420 1100 220 279 481 258 294 330 Aug 19 400 308 335 468 650 410 415 323 332 412 262 Aug 21 625 318 315 620 315 293 310 1300 1400 1100 Aug 23 410 291 262 310 298 580 316 300 635 313 295 Aug 26 340 315 355 301 300 Aug 30 312 298 405 279 270 Sep 3 382 305 500 340 300 Sep 6 403 281 292 290 361 703 319 280 292 299 292 Sep 9 352 542 606 282 281 Sep 12a 386 281 nd 318 281 Sep 12b 422 422 300 288 265 Sep 20 369 301 503 350 272 Sep 23 375 315 270 250 282 Sep 25 400 339 297 235 264 Sep 27 522 320 270 337 286 335 317 263 295 284 302 Oct 6 420 342 370 353 388 Oct 7 420 280 282 422 286 Oct 8 Oct 15 372 280 270 290 332 290 231 261 297 297 275 Average 393 336 301 347 419 408 310 313 487 471 336 102 RUN #2 Influent Bucket K R C mg/1 1 2 3 1 2 3 1 2 3 Jan5 380 525 460 520 482 Jan6 390 363 317 325 525 Jan7 326 362 468 241 402 Jan9 640 500 440 248 658 JanlO 462 460 585 500 442 Janl4 430 442 555 295 443 Janl5 425 400 690 1000 900 372 850 Janl7 450 385 560 340 612 Jan20 375 390 352 302 515 Jan22 340 498 365 315 540 Jan24 340 318 400 265 301 Jan27 398 380 312 308 482 Jan29 378 357 735 355 355 Jan31 420 400 645 278 408 Feb3 398 360 612 483 1000 Feb5 393 354 1000 498 1000 Feb7 391 408 365 331 425 Average 408 406 534 352 555 Note: shaded values were originally shown as ">900 or >1000" etc. 103 Table 4: tCOD, Corrected for Bucket Values RUN#1 Influent K R C mg/1 1 2 3 1 2 3 1 2 3 Aug 16 158 218 898 18 77 279 56 92 128 Aug 19 92 27 160 342 102 107 15 24 104 -46 Aug 21 -307 -310 -5 -310 -332 -315 675 775 475 Aug 23 119 -29 19 7 289 25 9 344 22 4 Aug 26 25 40 -14 -15 Aug 30 14 107 -19 -28 Sep 3 77 195 35 -5 Sep 6 122 11 9 80 422 38 -1 11 18 11 Sep 9 -190 64 -260 -261 Sep 12a 105 nd 37 0 Sep 12b 0 -122 -134 -157 Sep 20 68 202 49 -29 Sep 23 60 -45 -65 -33 Sep 25 61 -42 -104 -75 Sep 27 202 -50 17 -34 15 -3 -57 -25 -36 -18 Oct 6 78 28 11 46 Oct 7 140 2 142 6 Oct 8 Oct 15 92 -10 10 52 10 -49 -19 17 17 -5 Average 67 -29 18 98 78 -20 -23 157 142 -0 104 RUN #2 Influent K R C mg/1 1 2 3 1 2 3 1 2 3 Jan5 -145 -65 -5 -43 Jan6 27 -46 -38 162 Jan7 -36 106 -121 40 Jan9 140 -60 -252 158 JanlO 2 125 40 -18 Janl4 -12 113 -147 1 Janl5 25 290 600 500 -28 450 Janl7 65 175 -45 227 Jan20 -15 -38 -88 125 Jan22 -158 -133 -183 42 Jan24 22 82 -53 -17 Jan27 18 -68 -72 102 Jan29 21 378 -2 -2 Jan31 20 245 -122 8 Feb3 38 252 123 640 Feb5 39 646 144 646 Feb7 -17 -43 -77 17 Average 2 128 -54 149 105 Table 5: TOC, Raw Values RUN#1 Influent Bucket K R C mg/1 1 2 3 1 2 3 I 2 3 Sep 20 29.6 31.2 31.7 25.3 28.4 28.4 33 27.6 26.7 33.7 28.1 Sep 23 29.3 30.9 30 30 29.3 30.8 32.5 31.7 30.6 30.1 29.7 Sep 25 28.5 28.7 29.3 26.6 29.1 28.8 30.7 29.7 29.3 28.3 31 Sep 27 36.6 37.1 36.4 40 40.9 39.3 41 41.9 43.1 41.3 39.8 Oct 6 51 46.5 45 41.4 40.8 46.3 44.7 43.8 47.5 42.4 41.6 Oct 7 36.6 40.2 41.7 38 40.9 41.3 40.9 43.7 40.4 39.4 43.5 Oct 8 40.4 37.4 38.7 36.7 nd 39.6 36.8 40.5 36.8 36.5 39.7 Oct 15 ** 26.5 25.2 25.6 24.6 23.5 25.4 25.5 25.6 25.1 25.3 26.4 Average 34.81 34.65 34.8 32.83 29.11 34.99 35.64 35.56 34.94 34.63 34.98 Sample processed in January RUN #2 Influent Bucket K R C mg/1 1 2 3 1 2 3 1 2 3 Jan5 43.55 40.5 nd nd 41.35 nd nd 42.2 nd nd 36.45 Jan6 32.9 33.9 43.6 47.1 45.9 39.9 39.8 33.7 36.3 41.2 39.8 Jan7 26.6 26.3 29.8 33.1 35.7 27.1 30.2 29.5 28.5 37 28.9 Jan9 26.7 29.5 33.1 35.7 36.4 30 32.7 30.7 28.4 28.5 22.6 JanlO 27.9 29.6 34.9 37.9 41.1 33.1 35.5 38.4 30.5 32.6 40.6 Janl4 31.1 35.9 43.4 43.6 43.1 32.1 34.6 35.5 32 34 40.6 Janl5 nd 41.6 46.4 44.9 49.1 43.8 43.8 45.2 41.4 38.4 45.7 Janl7 40 40 50.6 50.8 51.9 36.5 45.3 46.3 41.6 41 29.2 Jan20 27.1 39.5 49 52.6 51.9 44 46.1 47 32.3 46.1 41.1 Jan22 35.1 37.5 37.9 42.2 46.5 36.3 39.5 43.5 26.5 35.6 37.1 Jan24 28.5 28.4 35.4 46.5 42.3 40.9 41.3 37.5 39.9 36.5 38.6 Jan27 45.5 46.9 55.8 58.5 59.1 50 47.5 48.1 46.8 50 51.4 Jan29 42.2 45.7 53 68.1 54.9 45.1 49.4 53 47.5 45.7 50.4 Jan31 39.6 34.3 55.5 65.7 53.9 56 50 48.6 49.9 33.4 40.1 Feb3 35.2 43.8 51.3 62.5 58.7 41.1 50.8 51.1 47.7 41.3 45.7 Feb5 41.6 45.3 53.5 56.1 53.5 45 49.2 31.2 33.1 42 33.4 Feb7 47.2 38.5 46.9 58.4 60.5 54.5 52 53.8 51.2 49.8 49.7 FeblO 28.3 29.8 39.7 41 43.5 32.3 34.2 36.5 30.6 31.2 31.6 Febl2 31.6 28.3 43 41.4 40.5 31.9 34.2 37.6 31 30.1 29.7 Febl4 29 42 50.1 40.1 51.1 29.6 33.5 35.3 40.6 30.1 30.2 Feb 17 40.4 41.7 52 56.8 50.2 45 45.6 49.6 42.9 42.1 43.1 Feb 19 41.8 44.7 50.7 54.9 54.4 45.1 46.7 48.1 42.8 40.2 35.9 Average 33.72 37.44 43.44 47.18 48.43 38.15 40.09 41.93 36.43 36.67 38.27 106 Table 6 TOC, Corrected for Bucket Values RUN#1 Influent K R C mg/1 1 2 3 1 2 3 1 2 3 Sep 20 -1.6 0.5 -5.9 -2.8 -2.8 1.8 -3.6 -4.5 2.5 -3.1 Sep 23 -1.6 -0.9 -0.9 -1.6 -0.1 1.6 0.8 -0.3 -0.8 -1.2 Sep 25 -0.2 0.6 -2.1 0.4 0.1 2 1 0.6 -0.4 2.3 Sep 27 -0.5 -0.7 2.9 3.8 2.2 3.9 4.8 6 4.2 2.7 Oct 6 4.5 -1.5 -5.1 -5.7 -0.2 -1.8 -2.7 1 -4.1 -4.9 Oct 7 -3.6 1.5 -2.2 0.7 1.1 0.7 3.5 0.2 -0.8 3.3 Oct 8 3 1.3 -0.7 nd 2.2 -0.6 3.1 -0.6 -0.9 2.3 Oct 15 ** 1.3 0.4 -0.6 -1.7 0.2 0.3 0.4 -0.1 0.1 1.2 Average 0.1625 0.15 -1.825 -0.863 0.3375 0.9875 0.9125 0.2875 -0.025 0.325 RUN #2 Influent K R C mg/1 1 2 3 1 2 3 1 2 3 JAN5,97 3.05 0.85 1.7 -4.05 JAN6,97 -1 9.7 13.2 12 6 5.9 -0.2 2.4 7.3 5.9 JAN7,97 0.3 3.5 6.8 9.4 0.8 3.9 3.2 2.2 10.7 2.6 Jan9 -2.8 3.6 6.2 6.9 0.5 3.2 1.2 -1.1 -1 -6.9 JanlO -1.7 5.3 8.3 11.5 3.5 5.9 8.8 0.9 3 11 Janl4 -4.8 7.5 7.7 7.2 -3.8 -1.3 -0.4 -3.9 -1.9 4.7 Janl5 -41.6 4.8 3.3 7.5 2.2 2.2 3.6 -0.2 -3.2 4.1 Janl7 0 10.6 10.8 11.9 -3.5 5.3 6.3 1.6 1 -10.8 Jan20 -12.4 9.5 13.1 12.4 4.5 6.6 7.5 -7.2 6.6 1.6 Jan22 -2.4 0.4 4.7 9 -1.2 2 6 -11 -1.9 -0.4 Jan24 0.1 7 18.1 13.9 12.5 12.9 9.1 11.5 8.1 10.2 Jan27 -1.4 8.9 11.6 12.2 3.1 0.6 1.2 -0.1 3.1 4.5 Jan29 -3.5 7.3 22.4 9.2 -0.6 3.7 7.3 1.8 0 4.7 Jan31 5.3 21.2 31.4 19.6 21.7 15.7 14.3 15.6 -0.9 5.8 Feb3 -8.6 7.5 18.7 14.9 -2.7 7 7.3 3.9 -2.5 1.9 Feb5 -3.7 8.2 10.8 8.2 -0.3 3.9 -14.1 -12.2 -3.3 -11.9 Feb7 8.7 8.4 19.9 22 16 13.5 15.3 12.7 11.3 11.2 FeblO -1.5 9.9 11.2 13.7 2.5 4.4 6.7 0.8 1.4 1.8 Febl2 3.3 14.7 13.1 12.2 3.6 5.9 9.3 2.7 1.8 1.4 Febl4 -13 8.1 -1.9 9.1 -12.4 -8.5 -6.7 -1.4 -11.9 -11.8 Febl7 -1.3 10.3 15.1 8.5 3.3 3.9 7.9 1.2 0.4 1.4 Feb 19 -2.9 6 10.2 9.7 0.4 2 3.4 -1.9 -4.5 -8.8 Average -3.72 8.2095 12.129 10.993 2.6714 4.7 4.4864 0.8714 1.1238 0.825 107 Table 7: SCVFA, Preliminary and Run #1 1°E fflue nt B ucke t KI K2 K3 Rl R2 R3 CI C2 C3 Prelim HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot July 4th nd 13 2 15 19 4 22 19 3 22 24 4 27 21 4 24 21 4 25 16 3 18 July 11 nd 23 4 26 29 5 33 29 4 32 31 4 35 28 4 31 25 3 28 29 5 33 27 4 30 29 4 32 26 4 29 July 17 nd 20 3 23 25 3 28 32 4 36 30 4 33 24 3 26 29 4 32 24 3 26 25 4 28 29 4 32 27 4 29 RUN#1 Aug 16 nd 27 3 29 27 nn 27 26 2 27 34 6 38 28 3 30 28 3 31 32 4 35 23 nn 23 27 2 29 25 nn 25 Aug 19 28 3 30 27 3 29 29 3 31 32 5 36 38 6 44 31 5 35 30 5 34 36 5 40 33 4 36 32 4 35 32 5 36 Aug 21 nd 29 4 32 28 4 31 28 4 31 31 6 36 27 4 30 29 5 32 29 5 32 27 4 30 26 4 29 26 4 29 Aug 23 25 3 28 28 4 31 28 5 32 29 5 33 29 6 34 28 6 33 31 6 36 30 6 34 36 5 40 28 4 32 27 5 31 Aug 26 28 3 31 28 4 32 29 4 32 29 5 33 31 6 35 30 5 34 32 6 37 33 6 37 29 4 32 32 5 36 31 4 34 Aug 30 22 4 26 26 2 28 27 3 30 30 5 34 2 24 3 26 30 4 34 27 4 30 23 2 25 28 3 30 25 2 27 Sep 3 29 6 33 35 6 40 32 6 37 37 7 42 39 8 46 34 6 39 38 7 44 39 7 45 31 5 35 35 5 40 33 3 35 Sep 6 27 2 28 29 3 32 29 4 32 31 4 34 33 5 37 31 4 34 33 4 36 36 4 39 30 4 32 30 4 33 33 4 36 Sep 9 21 3 23 31 3 34 32 3 35 35 4 38 40 5 44 34 4 37 36 5 40 40 5 44 34 3 37 35 4 38 40 4 43 Sep 12a 15 4 18 25 4 28 23 4 26 24 4 27 27 5 31 25 4 29 27 5 31 27 5 31 24 4 27 25 3 28 25 3 28 Sep 12b 19 3 21 22 3 25 21 3 24 23 3 25 24 4 27 22 4 25 24 4 28 24 4 28 20 3 22 21 3 23 20 3 23 Sep 20 17 2 19 17 2 18 18 2 20 18 2 20 22 2 24 19 3 22 21 3 23 23 3 25 18 2 20 18 2 20 18 2 20 Sep 23 15 1 16 16 2 17 16 2 17 16 2 17 18 1 19 17 2 18 18 2 20 18 2 20 15 2 16 15 2 17 15 2 17 Sep 25 15 2 17 17 2 18 16 2 18 17 2 18 18 2 20 17 2 19 20 3 22 20 3 22 17 2 18 16 2 18 17 2 19 Sep 27 11 2 12 11 2 13 14 2 15 14 nn 14 16 nn 16 15 2 17 20 3 22 18 3 21 13 2 15 14 2 15 14 2 16 Oct 6 7 nn 7 10 2 12 12 3 14 14 3 16 15 2 16 15 3 18 16 3 18 17 3 20 12 3 14 14 3 16 15 3 17 Oct 7 5 nn 5 10 2 12 13 2 15 15 2 17 18 2 20 13 2 15 18 3 20 21 4 24 14 3 16 16 3 19 18 4 21 Oct 8 13 nn 13 12 nn 12 13 nn 13 13 nn 13 14 2 16 15 2 17 16 2 18 13 nn 13 13 nn 13 14 nn 14 Oct 15 18 0 18 19 2 20 19 2 20 19 2 20 20 2 21 20 2 22 20 2 22 22 2 24 19 2 20 19 2 21 20 2 22 Average 17 2 18 22 3 24 22 3 25 24 3 26 25 4 30 23 3 26 26 4 29 27 4 30 23 3 25 23 3 26 23 3 26 o 00 Table 8: SCVFA, Run #2 R aw WW B uck et KI K2 K3 Rl R2 R3 CI C2 C3 RUN #2 HA Pr Tot HA Pr rot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot HA Pr Tot Jan5 S 2 10 11 3 13 nd nd nd nd 25 8 31 nd nd nd nd 19 6 23 nd nd nd nd 15 5 19 Jan6 14 2 16 14 3 16 21 7 26 27 9 35 29 10 37 17 4 21 19 5 24 22 7 28 15 4 18 16 5 20 17 6 21 Jan7 11 2 12 13 2 14 19 6 24 21 8 28 25 9 32 15 2 17 15 4 18 17 3 19 13 2 15 15 3 17 12 3 14 Jan9 10 3 12 13 4 16 18 7 24 22 8 28 24 8 30 16 4 19 18 5 22 19 5 24 12 3 15 12 4 15 11 3 13 JanlO 17 3 19 17 4 20 23 7 29 27 9 35 30 11 39 21 6 26 24 7 30 26 10 33 18 4 21 21 5 25 20 6 24 Janl4 18 2 19 18 2 20 22 5 26 26 7 31 27 8 34 21 3 23 21 3 24 21 3 24 18 2 20 19 2 21 20 3 22 Janl5 11 1 12 13 2 14 20 7 26 23 8 29 24 9 31 15 3 18 19 4 23 20 5 23 12 2 14 12 2 14 16 3 19 Janl7 20 0 20 20 20 28 9 36 32 12 42 32 12 42 23 5 27 25 8 31 28 9 35 20 20 20 20 19 3 22 Jan20 13 0 13 15 15 24 9 31 31 12 41 31 11 40 18 4 21 22 6 27 25 8 32 18 18 17 17 16 3 18 Jan22 14 0 14 14 14 21 6 26 26 10 34 27 10 35 18 5 22 19 5 23 21 7 27 14 14 14 14 14 14 Jan24 9 0 9 12 12 19 6 24 24 9 30 26 10 33 14 14 17 4 20 20 5 23 14 14 13 13 13 13 Jan27 17 2 18 12 2 13 23 9 30 27 12 36 30 13 40 19 3 21 20 3 23 24 5 28 17 2 19 20 4 23 20 5 24 Jan29 14 2 16 11 2 13 19 9 27 28 12 38 25 12 35 18 4 22 19 6 24 20 8 26 16 3 18 11 3 14 13 3 15 Jan31 10 2 12 15 3 17 24 11 33 24 14 36 25 14 36 19 10 27 18 6 23 17 7 23 11 3 13 12 4 15 12 4 16 Feb5 23 3 25 25 4 28 30 10 38 33 12 43 36 13 47 23 5 27 25 6 29 24 6 28 20 3 23 21 4 24 18 4 21 Feb7 19 2 21 17 2 18 28 8 34 34 11 42 32 11 41 25 4 28 26 5 30 29 6 33 20 2 22 21 2 23 21 2 23 Feb 10 16 0 16 19 2 20 27 9 34 28 9 35 33 10 41 19 4 22 26 5 30 27 6 31 17 2 19 18 2 20 23 3 25 Feb 17 17 2 19 18 2 20 28 9 35 32 11 40 31 11 39 21 4 24 24 5 28 29 8 35 19 3 21 20 3 22 20 3 22 Feb 19 17 2 18 16 2 18 26 9 33 29 11 38 30 11 38 18 4 21 23 5 27 24 7 29 16 2 18 18 3 21 18 3 21 Average 14 2 16 15 2 17 22 8 30 26 10 36 28 10 37 18 4 22 20 5 25 22 6 27 15 2 18 16 3 18 16 4 19 o Figure 1: 24br Sample, COD Analysis 24hr Sample COD Analysis 3 50 -.1_..r-..1f- p.....,..™- i .T~ Figure 2: 24hr Sample, Ortho-P Analysis 24hr Sample Ortho-P Analysis time 110 Figure 3: 24hr Sample, Ammonia Analysis Figure 4: 24hr Sample, Ammonia Analysis 24hr Sample NOx Analysis time 111 APPENDIX B: STATISTICAL ANALYSIS Table 1: Media SCVFA Production Comparison, Run #1 113 Table 2: Media SCVFA Production Comparison, Run #2 11Table 3: Kaldnes Media SCVFA Production Relative to the Control, Run #1 114 Table 4: Kaldnes Media SCVFA Production Relative to the Control, Run #2 114 Table 5: Ringlace Media SCVFA Production Relative to the Control, Run #1 115 Table 6: Ringlace Media SCVFA Production Relative to the Control, Run #2 115 Figure 1: Kaldnes Media Correlation Analysis, SCVFA Production vs TOC Production .... 116 Figure 2: Ringlace Media Correlation Analysis, SCVFA Production vs TOC Production .... 116 Figure 3: Kaldnes Media Correlation Analysis, SCVFA Production vs sCOD Production ... 117 Figure 4: Ringlace Media Correlation Analysis, SCVFA Production vs sCOD Production ... 117 Figure 5: Kaldnes Media Correlation Analysis, SCVFA Production vs tCOD Production .... 118 Figure 6: Ringlace Media Correlation Analysis, SCVFA Production vs tCOD Production ... 118 112 Table 1: Media SCVFA Production Comparison, Run #1 Run#I Ringlace Kaldnes t-Test: Paired Two-Sample for Means Kaldnes Ringlace Date total total Mean 5.90392 5.09931 Aug 16 6.1848 9.8324 Variance 11.10831 14.02844 Aug 19 11.1837 14.707 Observations 17 17 Aug 21 0.3256 3.4013 Pearson Correlation 0.73602 Aug 23 3.2158 2.726 Pooled Variance 12.56838 Aug 26 5.7412 3.4788 df 16 Sep 3 4.7458 5.6109 Hypothesized Mean Difference 0 Sep 6 7.3143 5.1754 t 1.27587 Sep 9 10.1985 10.3962 P(T<=t) two-tail 0.22021 Sep 12a 2.9214 3.7017 t Critical two-taii 2.11991 Sep 12b 2.9612 2.8073 Sep 20 6.7968 5.8902 Sep 23 2.7302 1.2022 Sep 25 4.1613 1.2813 Sep 27 8.1754 3.6011 Oct 6 7.8913 4.216 Oct 7 12.4383 8.1034 Oct 15 3.381 0.557 Table 2: Media SCVFA Production Comparison, Run #2 Run #2 Kaldnes Ringlace t-Test: Paired Two-Sample for Means Kaldnes Ringlace Data total total Mean 19.97114 10.48941 Jan5 18.169 10.73 Variance 10.03812 17.92647 Jan6 20.551 11.697 Observations 22 22 Jan7 17.489 4.934 Pearson Correlation 0.64839 Jan9 14.34 8.058 Pooled Variance 13.98229 Jan 10 18.408 13.141 df 21.00000 Jan 14 14.646 4.515 Hypothesized Mean Difference 0 Jan 15 16.946 9.006 t 13.67993 Jan 17 22.644 15.333 P(T<=t) two-tail 0.00000 Jan20 25.834 17.137 t Critical two-tail 2.07961 Jan22 20.657 13.089 Hypothesized Mean Difference 4 Jan24 21.895 11.745 t 7.90886 Jan27 27.129 14.249 P(T<=t) two-tail 9.89103E-08 Jan29 21.8 12.879 t Critical two-tail 2.07961 Jan31 19.234 6.145 Hypothesized Mean Difference 8 Feb3 21.733 8.911 t 2.13779 Feb5 18.476 -0.442 P(T<=t) two-tail 0.04445 Feb7 22.785 15.078 t Critical two-tail 2.07961 Feb 10 21.123 11.197 Hypothesized Mean Difference 9 Feb 12 16.308 9.292 t 0.69502 Feb 14 19.333 7.492 P(T<=t) two-tail 0.49466 Febl7 19.642 15.398 t Critical two-tail 2.07961 Febl9 20.223 11.183 113 Table 3: Kaldnes Media SCVFA Production Relative to the Control, Run #1 Run#l Kaldnes Control t-Test: Paired Two-Sample for Means Kaldnes Control Data total total Mean 5.09931 1.63208 Aug 16 9.8 -4.065 Variance 14.02844 18.48901 Aug 19 14.7 7.0766 Observations 17 17 Aug 21 3.4 -3.0105 Pearson Correlation 0.44972 Aug23 2.7 -0.4671 Pooled Variance 16.25873 Aug 26 3.5 2.3196 df 16 Sep 3 5.6 -5.3592 Hypothesized Mean Difference 0 Sep 6 5.2 4.3331 t 3.36655 Sep 9 10.4 8.8946 P(T<=t) two-tail 0.00393 Sep 12a 3.7 -0.0035 t Critical two-tail 2.11991 Sep 12b 2.8 -2.1214 Hypothesized Mean Difference 2 Sep 20 5.9 1.432 t 1.42463 Sep 23 1.2 -0.8361 P(T<=t) two-tail 0.17348 Sep 25 1.3 0.8434 t Critical two-tail 2.11991 Sep 27 3.6 2.9888 Hypothesized Mean Difference 3 Oct6 4.2 5.0826 t 0.45366 Oct7 8.1 9.4374 P(T<=t) two-tail 0.65616 Oct 15 0.6 1.2 t Critical two-tail 2.11991 Table 4: Kaldnes Media SCVFA Production Relative to the Control, Run #2 Run #2 Kaldnes Control t-Test: Paired Two-Sample for Means Kaldnes Control Data total total Mean 19.97114 2.40882 Jan5 18.2 6.22 Variance 10.03812 12.51800 Jan6 20.6 5.087 Observations 22 22 Jan7 17.5 0.172 Pearson Correlation 0.49944 Jan9 14.3 -2.062 Pooled Variance 11.27806 JanlO 18.4 3.901 df 21.00000 Janl4 14.6 2.815 Hypothesized Mean Difference o Jan 15 16.9 4.572 t 24.44127 Janl7 22.6 2.454 P(T<=t) two-tail 0.00000 Jan20 25.8 3.911 t Critical two-tail 2.07961 Jan22 20.7 0.2 Hypothesized Mean Difference 16 Jan24 21.9 1.5 t 2.17426 Jan27 27.1 10.706 P(T<=t) two-tail 0.04126 Jan29 21.8 2.172 t Critical two-tail 2.07961 Jan31 19.2 -1.185 Hypothesized Mean Difference 17 Feb3 21.7 3.615 t 0.78257 Feb5 18.5 -7.619 P(T<=t) two-tail 0.44261 Feb7 22.8 4.862 t Critical two-tail 2.07961 FeblO 21.1 4.648 Hypothesized Mean Difference 18 Feb 12 16.3 -0.038 t -0.60912 Febl4 19.3 1.929 P(T<=t) two-tail 0.54898 Febl7 19.6 2.448 t Critical two-tail 2.07961 Feb 19 20.2 2.686 114 Table 5: Ringlace Media SCVFA Production Relative to the Control, Run #1 Run#l Ringlace Control t-Test: Paired Two-Sample for Means Ringlace Control Data total total Mean 5.66473 1.44765 Aug 16 6.2 -4.065 Variance 10.80676 16.92190 Aug 19 11.2 7.0766 Observations 19 19 Aug 21 0.3 -3.0105 Pearson Correlation 0.82474 Aug 23 3.2 -0.4671 Pooled Variance 13.86433 Aug 26 5.7 2.3196 df 18 Aug 30 1.7 -1.4 Hypothesized Mean Difference 0 Sep 3 4.7 -5.3592 t 7.89365 Sep 6 7.3 4.3331 P(T<=t) two-tail 0.00000 Sep 9 10.2 8.8946 t Critical two-tail 2.10092 Sep 12a 2.9 -0.0035 Hypothesized Mean Difference 3 Sep 12b 3.0 -2.1214 t 2.27817 Sep 20 6.8 1.432 P(T<=t) two-tail 0.03514 Sep 23 2.7 -0.8361 t Critical two-tail 2.10092 Sep 25 4.2 0.8434 Hypothesized Mean Difference 4 Sep 27 8.2 2.9888 t 0.40634 Oct6 7.9 5.0826 P(T<=t) two-tail 0.68928 Oct 7 12.4 9.4374 t Critical two-tail 2.10092 Oct 8 5.6 1.16 Oct 15 3.4 1.2 Table 6: Ringlace Media SCVFA Production Relative to the Control, Run #2 Run #2 Ringlace Control t-Test: Paired Two-Sample for Means Ringlace Control Data total total Mean 10.48941 2.40882 Jan5 10.7 6.22 Variance 17.92647 12.51800 Jan6 11.7 5.087 Observations 22 22 Jan7 4.9 0.172 Pearson Correlation 0.63780 Jan9 8.1 -2.062 Pooled Variance 15.22224 JanlO 13.1 3.901 df 21 Janl4 4.5 2.815 Hypothesized Mean Difference 0 Janl5 9.0 4.572 t 11.25709 Jan 17 15.3 2.454 P(T<=t) two-tail 0.00000 Jan20 17.1 3.911 t Critical two-tail 2.07961 Jan22 13.1 0.2 Hypothesized Mean Difference 6 Jan24 11.7 1.5 t 2.89848 Jan27 14.2 10.706 P(T<=t) two-tail 0.00859 Jan29 12.9 2.172 t Critical two-tail 2.07961 Jan31 6.1 -1.185 Hypothesized Mean Difference 7 Feb3 8.9 3.615 t 1.50537 Feb5 -0.4 -7.619 P(T<=t) two-tail 0.14712 Feb7 15.1 4.862 t Critical two-tail 2.07961 FeblO 11.2 4.648 Hypothesized Mean Difference 8 Febl2 9.3 -0.038 t 0.11227 Febl4 7.5 1.929 P(T<=t) two-tail 0.91167 Febl7 15.4 2.448 t Critical two-tail 2.07961 Febl9 11.2 2.686 115 Figure 1: Kaldnes Media Correlation Analysis, SCVFA Production vs TOC Production VFA Production vs TOC Production Kaldnes Media, Run #2 10 15 20 VFA (mg/L) Slope: 0.627, Int.: -1.53, Pearson Coef.: 0.4505 Figure 2: Ringlace Media Correlation Analysis, SCVFA Production vs TOC Production SCVFA Production vs TOC Production Ringlace Media, Run 82 10 VFA (mg/L) Slope: 0.889, Int.: -4.84, Pearson Coef: 0.5840 116 Figure 3: Kaldnes Media Correlation Analysis, SCVFA Production vs sCOD Production SCVFA Production vs sCOO Production Kaldnes Media, Run #2 15 VFA (mg/L) Slope: 1.257, Int.: -3.64, Pearson Coef.: 0.4902 Figure 4: Ringlace Media Correlation Analysis, SCVFA Production vs sCOD Production Q O O SCVFA Production vs sCOD Production Ringlace Media, Run 82 5 10 VFA (mg/L) Slope: 1.806, Int.: -8.54, Pearson Coef: 0.6856 117 Figure 5: Kaldnes Media Correlation Analysis, SCVFA Production vs tCOD Production -J 4O0-' O O 9 200 •--SCVFA Production vs tCOD Production Kaldnes Media, Run #2 H 1 1 1 r-10 15 20 VFA (mg/L) 25 30 Slope: -14.70, Int.: 423.9, Pearson Coef.: 0.2335 Figure 6: Ringlace Media Correlation Analysis, SCVFA Production vs tCOD Production Q o O -100 SCVFA Production vs tCOD Production Ringlace Media, Run 82 5 10 VFA (mg/L) Slope: -3.122, Int.: -22.1, Pearson Coef: 0.1436 118 APPENDIX C: PHOTOS Figure 1: BF-System Influent Processing Reactors, Runs #1 and 2 120 Figure 2: Ringlace Media Biolog GN Plate, Run #2 12Figure 3: BF-System Reactors 121 Figure 4: BF-system Influent PumpsFigure 5: Reactor Mixing Chambers 122 Figure 6: Ringlace Reactor, Outlet Mixing Chamber 12Figure 7: Ringlace cage, clean 123 Figure 8: Ringlace cage, carrying solids 12Figure 9: Kaldnes media bag 124 119 Figure 1: BF-System Influent Processing Reactors, Runs #1 and 2 Figure 2: Ringlace Media Biolog GN Plate, Run #2 1 * • • • # • UF # f # # « ii- ? •si, J& & ft * - ,«« I :j IP' 'iH W 1 f • - ft '# 120 Figure 3: BF-System Reactors Figure 5: Reactor Mixing Chambers Figure 7: Ringlace cage, clean Figure 8: Ringlace cage, carrying solids 123 Figure 9: Kaldnes media bag 124 

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
China 9 39
United States 7 0
Japan 6 0
Guatemala 1 0
City Views Downloads
Beijing 9 0
Tokyo 6 0
Ashburn 4 0
Unknown 2 2
Redmond 1 0
Sunnyvale 1 0

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

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0099238/manifest

Comment

Related Items