UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A comparison of retained biomass anaerobic digester designs Stephenson, Robert John 1987

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1987_A7 S73.pdf [ 7.27MB ]
Metadata
JSON: 831-1.0097233.json
JSON-LD: 831-1.0097233-ld.json
RDF/XML (Pretty): 831-1.0097233-rdf.xml
RDF/JSON: 831-1.0097233-rdf.json
Turtle: 831-1.0097233-turtle.txt
N-Triples: 831-1.0097233-rdf-ntriples.txt
Original Record: 831-1.0097233-source.json
Full Text
831-1.0097233-fulltext.txt
Citation
831-1.0097233.ris

Full Text

A COMPARISON OF RETAINED BIOMASS ANAEROBIC DIGESTER DESIGNS by ROBERT JOHN STEPHENSON A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Bio-Resource Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1987 ® Robert John Stephenson, 1987 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the 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 of Bio-Resource Engineering The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 D a t e : A p r i l 1 9 8 7 ABSTRACT The principles behind anaerobic digestion are fairly well understood, but the limits of application of each digester design axe not known. Because there are significant differences in the properties of the many wastewaters requiring treatment optimal anaerobic digester performance requires the matching of feed characteristics to a digester design and mode of operation. No consensus has yet emerged on digester design, operating conditions or feed/digester match-ups. In this study, three bench scale retained biomass anaerobic digester designs were examined for their response to a sequence of varied hydraulic retention times (HRTs) and influent wastewater concentrations. The digester designs studied were the upfiow anaerobic filter, the upfiow anaerobic expanded bed and the upfiow anaerobic sludge bed. The wastewater was screened and diluted dairy cow manure obtained from the UBC dairy barn. The parameters monitored included the total and soluble chemical oxygen demand (TCOD and SCOD), volatile and suspended solids (VS and SS), total volatile fatty acids (VFAs), total Kjeldahl and ammonia nitrogen (TKN and NH3-N), pH, biogas production, and the methane (CH4) and carbon dioxide (C02) content of the biogas. Wastewater treatment efficiencies, measured in terms of TCOD, SCOD, VS, and TVFA removals, and methane productivity and methane yield for each of the digester designs were examined for the range of the operating conditions. The anaerobic filter digester effected a mean TCOD removal efficiency of 47% ± 14% at a mean 4.0 day HRT, 51% ± 9% at a mean 2.3 day HRT and 35% ± 11% at a mean 1.3 day HRT. The expanded bed digester effected a mean TCOD removal of 45% ± 15% at a mean 4.3 day HRT, 38% ± 12% at a mean 2.5 day HRT and 28% ± 9% ay a mean 1.3 day HRT. The sludge bed digester effected a mean TCOD removal of 53% ± 9% at a mean 3.8 day HRT, 45% ± 12% at a mean 2.2 day HRT and 32% ± 10% at a mean 1.2 day HRT. For all three digesters, the difference in the treatment efficiency over the range of HRTs tested, from 5 to 1.25 days was not in proportion to the change in HRT. ii Methane productivity, measured against either the removal or addition of substrate in terms of TCOD, SCOD, VS and TVFA, demonstrated considerable variability. Methane production increased with both substrate addition and substrate removal. Methane yield increased with increasing HRT. The sludge bed digester generally exhibited the greatest but most variable methane yields. It produced 0.095 L CH4/g VS added at a mean 3.8 day HRT and 0.037 L CH4/g VS added at a mean 1.2 day HRT. The anaerobic filter delivered the greatest methane yield at the intermediate HRT, 0.044 L CH4/g VS added at a mean 2.3 day HRT. The expanded bed demonstrated low methane yields over the range of feed strengths and HRTs tested. Biogas composition averaged 62.1% methane and 17.1% carbon dioxide for the anaerobic filter, 43.6% methane and 5.3% carbon dioxide for the expanded bed. and 61.1% methane and 18.9% carbon dioxide for the sludge bed. iii TABLE OF CONTENTS iv TABLE OF CONTENTS ABSTRACT i i TABLE OF CONTENTS i v L I S T OF FIGURES v i i i L I S T OF TABLES x i ACKNOWLEDGEMENT x i i I . A COMPARISON OF RETAINED BIOMASS ANAEROBIC DIGESTER DESIGNS 1 I I . INTRODUCTION 2 I I I . OBJECTIVES 6 IV. LITERATURE REVIEW 8 A. BIOCHEMICAL BASICS 9 B. DIGESTER PERFORMANCE 12 C. INDICATORS OF DIGESTER PERFORMANCE 14 1. VOLATILE SOLIDS 14 2. TOTAL COD 15 3. SOLUBLE COD 15 4. TOTAL VOLATILE FATTY ACIDS 15 V. DESIGNING AN ANAEROBIC DIGESTER 17 A. FORM OF BACTERIAL RETENTION 23 B. FEED INLET SYSTEM 26 C. FEEDING FREQUENCY 27 D. ASPECT RATIO 29 E. MIXING 32 V I . ANAEROBIC DIGESTER DESIGNS 35 A. UPFLOW ANAEROBIC FILTER 36 v B. UPFLOW ANAEROBIC EXPANDED BED 42 C. UPFLOW ANAEROBIC SLUDGE BED 49 V I I . CONCLUSIONS FROM THE LITERATURE 56 V I I I . MATERIALS AND METHOD 58 A. EXPERIMENTAL SET-UP AND OPERATION 59 B. ASPECT RATIO 59 C. SUPPORT SURFACES 60 1. EXPANDED BED 60 2. ANAEROBIC FILTER 61 3. SLUDGE BED 61 D. METHOD OF FEEDING 61 E. MIXING AND HEATING 64 F. FEEDSTOCK 65 G. SAMPLING PROCEDURE AND ANALYSIS 67 IX. INTERPRETATION OF RESULTS 69 A. PRECISION OF THE ANALYTICAL TESTS 70 B. HYDRAULIC RETENTION TIMES OF THE'EXPERIMENTAL PERIOD 72 C. PROBLEMS WITH THE EXPERIMENTAL APPARATUS 74 1. FEED DEGRADATION AND SETTLING 74 2. DIGESTER PLUGGING PROBLEMS 77 D. FEED CHARACTERISTICS 78 1. TCOD AND SCOD 81 2. TOTAL, VOLATILE AND SUSPENDED SOLIDS 83 3. TOTAL VOLATILE FATTY ACIDS 83 4. TOTAL KJELDAHL AND AMMONIA NITROGEN 86 5. pH. 90 v i 6. SUSPENDED SOLIDS 90 E. SUBSTRATE REMOVAL 92 1. TCOD REMOVAL 94 2. SCOD REMOVAL '99 3. TVFA REMOVAL 103 4. VS REMOVAL 107 F. BIOGAS PRODUCTION 111 G. BIOGAS COMPOSITION 114 H. METHANE PRODUCTION 116 I . SUMMARY 130 X. CONCLUSIONS FROM THE EXPERIMENTS 134 X I . RECOMMENDATIONS 137 REFERENCES 141 v i i LIST OF FIGURES Figure Page 1 Simplified Process of Anaerobic Digestion 11 2 Schematic of the Anaerobic Filter Process 38 3 Schematic of the Anaerobic Expanded Bed Digester 44 4 Schematic of the Anaerobic Sludge Bed Digester 52 5 HRTs of the Experimental Period 73 6 Relationship Between the Substrate Indicators of the Feed 80 7 Total and Soluble COD of the Feed 82 8 Total, Volatile and Suspended Solids of the Feed 84 9 Volatile Fatty Acids of the Feed 85 10 TKN and NH 3-N of the Feed 87 11 TKN of the Feed and Digester Effluents 88 12 NH 3-N of the Feed and Digester Effluents 89 13 pH of the Feed and Digester Effluents 91 14 Suspended Solids of the Feed and Digester Effluents 93 15 TCOD Removal vs TCOD Addition — Anaerobic Filter 95 16 TCOD Removal vs TCOD Addition — Expanded Bed 96 17 TCOD Removal vs TCOD Addition — Sludge Bed 97 v i i i 18 Effect of HRT on TCOD Removal Efficiency 98 19 SCOD Removal vs SCOD Addition ~ Anaerobic Filter 100 20 SCOD Removal vs SCOD Addition — Expanded Bed 101 21 SCOD Removal vs SCOD Addition — Sludge Bed 102 22 TVFA Removal vs TVFA Addition — Anaerobic Filter 104 23 TVFA Removal vs TVFA Addition ~ Expanded Bed 105 24 TVFA Removal vs TVFA Addition — Sludge Bed 106 25 VS Removal vs VS Addition — Anaerobic Filter 108 26 VS Removal vs VS Addition ~ Expanded Bed 109 27 VS Removal vs VS Addition ~ Sludge Bed 110 28 . Biogas Production of the Digesters 113 29 Biogas Composition of the Digesters 115 30 CH« Production vs TCOD Addition — Anaerobic Filter 117 31 C H 4 Production vs TCOD Removal — Anaerobic Filter 118 32 CH4 Production vs TCOD Addition — Expanded Bed 119 33 CH« Production vs TCOD Removal — Expanded Bed 120 34 C H 4 Production vs TCOD Addition — Sludge Bed 121 35 C H 4 Production vs TCOD Removal « Sludge Bed 122 ix 36 CH 4 Production vs VS Addition — 37 CH4 Production vs VS Removal — 38 CH* Production vs VS Addition — 39 CH* Production vs VS Removal — 40 CH 4 Production vs VS Addition — 41 CH* Production vs VS Removal — Anaerobic FilteT 124 Anaerobic Filter 125 Expanded Bed ° 126 Expanded Bed 127 Sludge Bed 128 Sludge Bed 129 x L I S T O F T A B L E S Table Page 1 Cycle Modes of the Experiments 62 2 Precision of the Analytical Measurements 71 3 Changes in the Influent Arising From Pumping 76 4 Summary of Feed Characteristics 79 5 Summary of Substrate Removal Efficiency 112 6 Summary of Methane Productivity 131 xi I. ACKNOWT FTxTFMFNT A number of individuals greatly assisted me with this work. I gratefully acknowledge their generous donations of both time and talent I'm grateful towards the Department of Bio-Resource Engineering for offering me an open learning opportunity and towards the students of the Department for sharing their knowledge and their friendship. I appreciate the considerable laboratory support and tolerance of James Cheng, Ken Lee, Adeline Chen and Dr. Ping Liao. I am thankful towards Dr. Victor Lo, Dr. Ross Bulley and Dr. Richard Branion, my thesis committee members, for their efforts, support and advice. This work was financially supported by the Natural Sciences and Engineering Research Council of Canada and by the B.C. Science Council GREAT Awards with Mr. R.G. McDonald of Envirocon Ltd. acting as the corporate sponsor. For their interest in my work I am appreciative. Most of all, I'm grateful for the caring support offered by Sandra, without whom this wouldn't have been possible. x i i A COMPARISON OF RETAINED BIOMASS ANAEROBIC DIGESTER DESIGNS 1 INTRODUCTION II. INTRODUCTION Anaerobic digestion is one of the main routes for the decomposition of organic matter in nature. It is an intricate and essential process where complex organic materials are degraded to gaseous products. This fermentation occurs in rivers, swamps, fresh water lakes, marine environments, and in the rumen and in the gastrointestinal tract of humans and many other animals. Anaerobic treatment can also take place in the bottom of open lagoons, or in closed systems such as a septic tank or an anaerobic digester. By controlling the conditions under which anaerobic digestion operates, the process is accelerated, thus becoming a useful technology for wastewater treatment Previously threatened fossil fuel shortages, present on-going investigations of cost trimming measures, an increased dependence on water re-use, and legislation protecting the environment have promoted a growing interest in anaerobic processes and provided a springboard for intensive research and development Whereas in the past, wastes had no value and disposal costs were not significant, now, wastes are not only useful residues for energy conversion but they also present increasingly costly disposal problems. There is emerging a need for greater efficiency and economy in the treatment of industrial, agricultural and municipal waste streams. Making biological processes instrumental in the abatement of water pollution has been activated by industrial growth and agricultural intensification and the concomitant increase of their waste streams. These wasterwaters are rich in readily degradable compounds. If discharged into fresh water bodies, the overall effect is to remove dissolved oxygen by biochemical reactions, and if discharged into municipal sewerage system, existing wastewater treatment plant facilities may be overloaded. In industrialized countries, the incentive for developing anaerobic digesters is primarily in the effective stabilization of concentrated organic wastes, with the methane byproduct reducing net costs considerably (Jech and Brautigam, 1983). In the developing countries, the emphasis lies more in the supplemental energy supply derived from unwanted organic residues. In converting a nuisance to a fuel, basic needs of sanitation and water 3 4 supply can be addressed while decreasing the reliance on centralized and capital intensive power generating facilities and on imported fossil fuels. Methane fermentation has been used for municipal and industrial sludge stabilization for nearly 100 years (Chandler et al., 1980). Yet, application of this process for the purpose of producing energy is still in its infancy. According to Hawkes and Horton (1981), one of the drawbacks to progress in anaerobic digester development is perhaps that it has been regarded as an "alternative energy" process, and therefore disregarded as a serious possibility for energy production and wastewater treatment in the developed world. Having been regarded as low technology, generally digesters have been poorly designed and constructed. Anaerobic processes have in the past not generally been considered suitable due to their low rates of organic removal, requiring long detention times for good waste stabilization, and their reported extreme sensitivity to fluctuations in process parameters (Bull et al., 1984c). The methane gas produced by the process can be used to fulfill energy needs; hence the process has been limited to the treatment of wastes with high potential for gas production (Kobayashi et al., 1983). Consequently, anaerobic treatment has been rarely considered for low strength, low temperature wastes (Genung et al., 1978). Numerous investigations are slowly remolding the prevailing opinion, however. The results of retained biomass processes and the improved knowledge of the effects of the environmental factors and of the biochemistry and microbiology of anaerobic digestion, have led researchers such as Lettinga et al. (1983b) to expect the effective treatment of an extended range of wastewaters. The scope of process applications is broadening likewise. Oleszkiewicz and Koziarski (1982b) indicated that anaerobic digestion can be applied as both a roughing pre-treatment of strong organic effluents and as a polishing treatment of dilute effluents at ambient temperatures. How anaerobic digestion works is fairly well understood in a biochemical sense, but the limits of application of each digester configuration are not known (Hall and Jovanovic, 1982 and van den Beld, 1984). The variety and number of investigations points to the 5 ultimate potential of anaerobic digestion but development efforts are fragmented; no consensus has yet emerged on digester design, operating conditions, or feed/digester matchups. Studies in anaerobic digestion have been designed to meet a variety of objectives. Consequently, there lacks similarity between the investigators' digester design and testing conditions. The absence of scaling factors with which to translate the results of individual experiments to some standard means that it is not possible to identify from these data key operating parameters useful for developing criteria for designing full scale facilities. A comparison of the processes from the literature is difficult given the range of wastewaters and conditions under which the tests have been performed. Such comparative studies have yielded widely conflicting conclusions. Jech and Brautigam (1983) reported that no clear cut choice seems evident among the various digester configurations. Because there are significant differences in the properties of many wastewaters, optimal anaerobic digester performance would require matching the feed characteristics to a digester design and mode of operation. For the potential user, an assessment of the relative performance of the available systems would be beneficial. Only a direct comparison of the performance of the digester designs under the same influent conditions and detention times might demonstrate the important advantages and disadvantages of each digester design (Rittmann, 1982b). 6 OBJECTIVES in. OR.TF.CTIVES The general objectives of this study were to describe the effects of key operating variables on the performance of anaerobic digester designs and to compare the performance between these digesters. This should initiate a basis for the match-up of digester design with wastewater characteristics. The specific objectives were: 1. to incorporate the experimental findings, as reported in the literature, into an improved anaerobic digester design and mode of operation, and 2. to compare, via laboratory scale experimentation, the methane generation and wastewater treatment efficiency of each of the digester designs over a range of influent concentrations and liquid throughput rates for the most promising anaerobic digester designs. 7 8 LITERATURE REVIEW TV. LITERATI IRF. R SMISkV A. RTOCTTRMTCAT, BASICS Anaerobic digestion is effected by the integrated action of a mixed population of microorganisms which results in the conversion of water soluble organic feedstocks to methane and carbon dioxide. In order to exploit the process of methane fonnation, some understanding of the mechanisms involved and the factors affecting these mechanisms is required. The decomposition of organic matter and subsequent formation of methane gas by microorganisms is indigenous to nature. The environmental conditions under which methane is produced biologically are found in the organic sediments of most freshwater and marine environments, in soil, as well as in the gastro-intestinal tract of humans and animals, particularly in the rumen of herbivores and the enlarged caecum of certain non-ruminants (Mah, 1982). In all of these habitats, specific ecological constraints accompany the presence of fermentable organic compounds leading to the biological formation of methane (Mah, 1983 and Hashimoto et al., 1979). The fermentative bacteria are composed of a complex mix of bacterial species. Most are obligate anaerobes but some facultative anaerobes may be present in some environments. The methanogens are a unique group of bacteria which are composed of many different species that have quite different cell shapes and structures. They require very strict anaerobic conditions for growth. Their nutritional requirements are simple, using ammonia, sulphide, and carbon dioxide as the main nitrogen, sulphur and carbon sources respectively. Virtually all of the methane formers use carbon dioxide and hydrogen gas for growth. Methane formation does not occur in the presence of oxygen or where other electron acceptors such as sulphate or nitrate are available. As a group, the methanogens use a narrow range of substrates as energy sources. Methanogenic precursors such as acetate, methanol, methyl-, 9 10 dimethyl-, or trimemylamine are required, as is a pH range between 6.0 and 8.0, and the absence of light (Mah, 1982). These physical and chemical selective pressures enrich for a specialized microbial population collaboratively engaged in obtaining the most energy possible. A large variety of primary and secondary raw materials can be bioconverted anaerobically to methane. Substrates of interest in methane fermentation research differ greatly in their chemical composition. Wide variations in biodegradability have been observed (Chandler et cd., 1980). According to Young (1983), wastes which contain significant amounts of carbohydrate produce a higher yield and accumulation of solids than wastes containing predominantly proteins or complex organic acids. Carbohydrates are usually in the form of insoluble polysaccharides which have to be broken down into mono- and disaccharides before the bacteria can digest them (Cowley and Wase, 1981). Protein contributes little in terms of gas production, whereas carbohydrates and lipids are the 'fuel' supplies of the bacteria which determine the biogas output It is generally accepted that anaerobic digestion can be described as a three step process (Bhatia et al., 1981). The overall process is a symbiotic effort of diverse kinds of bacteria. These three steps occur simultaneously rather than in sequence as described. Visualizing the process may be aided by referring to Figure 1, a Simplified Process of Anaerobic Digestion. The first group, the fermentative bacteria, hydrolyse the principal substrates such as polysaccharides, proteins, and lipids to produce volatile fatty acids, alcohol, carbon dioxide, hydrogen gas, ammonia, and sulphide. Hydrolysis may be the rate-limiting step for the degradation of a concentrated high suspended solids wastewater (Olthof and Oleskiewicz, 1978). The second metabolic group, called the H 2 producing acetogenic bacteria, produce acetate, C0 2, and H 2 from the longer chained fatty acids (propionate and butyrate) generated in the first stage. In the third step, the methane forming bacteria consume acetate, C0 2 and H 2 produced in the first two stages to produce CH4 and C02. 11 FIGURE 1: Simplified Process of Anaerobic Digestion Acetate V Organic Feedstock Carbohydrate Protein Lipids \ / HYDROLYSIS (Fermentative Bacteria) \ / Fatty Acids DEHYDROGENATION (Acetogenic Bacteria) ACETATE DECARBOXYLATION (Methanogenic Bacteria) •> H 2 + C0 2 REDUCTIVE METHANE FORMATION (Metha nogenic Bacteria) V CH, + C0 o 4 2 V CH. + H-0 4 2 12 The methane formers produce methane by two distinct mechanisms (Kelly and Switzenbaum, 1983). Acetate is the chief precursor of methane where the methyl group of acetate is converted directly to methane and the carboxyl group to carbon dioxide. Approximately 70% of the methane formed originates from acetate: CHjCOOH — > CH« + C0 2 The remaining 30% of the methane formed is through the reduction of carbon dioxide: C0 2 + 4H2 —> CH4 + 2H20 In the digestion of soluble compounds, the rate-limiting step is methanogenesis of fatty acids. Anaerobic treatment requires a unique microbial balance of the robust and fast growing acid forming bacteria and the sensitive slow growing methane producing bacteria (Koybayashi et al., 1983 and Kroeker et al., 1979). A bottleneck arises from the low specific growth rate of the methanogenic bacteria. The malfunctioning of anaerobic digestion, commonly known as "souring", has occurred due to an imbalance between acid production and methane production. It has represented a serious problem in digester management (Melchior et al., 1982). Securing process stability through encouraging a balance between the microbial mix is central to the practical implementation of anaerobic wastewater treatment B. DIGESTER PERFORMANCE Special attention must be devoted to the gas production rate and to the organic content of the discharge, as these define the efficacy and the ecological impact of the digester. Digester efficiency is related to the two key operating variables: the HRT and the influent wastewater concentration. The volumetric loading rate combines the effects of those two independent variables. Volumetric loading of a digester is defined as (Sutton et al., 1982): VL = QSo/V = So/t 13 where, VL = volumetric loading (M/V t), Q = wastewater flow rate (V/t), So = wastewater substrate concentration (M/V), V = digester volume (V), and t = HRT (t). According to Jewell (1981), such a measurement facilitates a direct comparison of the required digester volume between digesters and therefore of the relative digester efficiencies. The use of the organic loading rate is problematic however, because it combines the the effects of the HRT and influent concentration, each of which exerts different influences on the digester performance. This shortcoming is in direct conflict with this set of experiments. For this reason, the organic loading rate was not used as a parameter of process performance. The anaerobic digester designs can be evaluated in terms of: 1. the biodegradation efficiency — substrate removal as a function of loading, 2. the process stability, and 3. the methane productivity. The efficiency of digestion can be described as either the extent of organic matter destroyed, measured as the percent substrate removed, or by the methane production rate. It is defined as (Lawrence and McCarty, 1969): E = ((S-s)/S)(100) where, E = efficiency of waste treatment (percent), S = influent concentration (mass/volume), and s = effluent concentration (mass/volume). 14 C. INDICATORS OF DTOKSTRR PERFORMANCE A single convenient and accurate measure of just what constitutes the actual substrate for anaerobic bacteria has not yet been found. The ability to apply models to anaerobic digesters has been limited because of the inability to pinpoint the significant physical and biochemical factors within the digester (Young, 1983). Measures of substrate include: volatile solids, total chemical oxygen demand, soluble chemical oxygen demand and total volatile fatty acids. What these parameters do measure and what their limitations are pertinent to anaerobic digestion is briefly considered. 1. VOLATILE SOLIDS The variable most commonly used as an estimate of substrate for anaerobic digestion is that of volatile solids concentration. The volatile solids concentration is a rough approximation of the mass of organic matter present in the solid fraction of the wastewater. Volatile solids include both soluble and suspended solids. The volatile solids fraction which is the most amenable to degradation is the soluble portion, followed by the fine particulates of very high surface area to volume ratios (Lo et al., 1983). Suspended solids aren't readily broken down in anaerobic systems, especially at the relatively short HRTs dictated by the economics of digester vessel sizing (Hashimoto and Chen, 1980). The limitations of simple analytical methods for volatile solids determination present themselves when considering the testing procedure itself. A number of compounds of particular importance in anaerobic digestion are volatile below the 105° C determination — volatile fatty acids, alcohols, and ammonium salts (Melchior et al., 1982). A significant variation in the methane yield per gram of substrate destroyed indicates that biodegradable volatile solids differ greatly in their energy content The parameter total volatile solids alone is inadequate to describe actively digesting biomass (Chandler et al., 1980). 15 2. TOTAT. COD The chemical oxygen demand is also used as an indicator of the substrate concentration. The total COD (TCOD) is a measure of the oxygen equivalent of organic matter that is susceptible to oxidation by a strong chemical oxidant It includes some compounds such as cellulose which are not a part of the immediate biochemical load on the oxygen of the receiving waters. TCOD also contains suspended solids, largely unavailable to the microorganisms as food. Alone, it is a poor indicator of substrate. 3. SOLUBLE COD Soluble COD eliminates the majority of the suspended particles through filtration prior to chemical oxidation. While SCOD is a measure of the soluble substrate which is more readily available for microbial breakdown, it also measures the solubilization of some of the suspended solids which occurs during the course of digestion. The extent that this solubilization has an effect is likely a function of the digester solids retention time. Larger residence times would allow an increased degree of solubilization of the suspended particulates. Hence, the usefulness of the parameter for the interpretation of the efficiency of feed conversion is limited to the shorter SRTs. The argument remains however as to the inappropriateness of strong chemical oxidation imitating the less than severe biochemical reactions of anaerobic digestion. 4. TOTAT. VOLATILE FATTY A ("TPS TVFAs can be used as an indicator of the available substrate. Total volatile fatty acids (TVFAs) include acetic, propionic and butyric acids — important intermediates in anaerobic wastewater degradation. Some acids are preferentially utilized more rapidly than others, however. Andrews and Pearson (1965) have shown that the rate of metabolism of the individual TVFAs varies in the order: acetic, butyric, propionic. Simply quoting a TVFA value without an indication of its constituents in terms of acetic, propionic and butyric acids 16 conveys incomplete information since the system's response varies with the different TVFA components. Not only are TVFAs degraded to simple organic acids and ultimately to CH4, C 0 2 and water, but the more complex carbohydrates present in the feedstock are themselves broken down to TVFAs. Measuring an aggregate reduction without accounting for its constituents and their relative shift, as well as the difficulty in easily predicting the additive effect of the feedstock degradation, means that TVFAs alone are not an useful indicator of substrate loading. This "snapshot effect" is analogous to the errors encountered with the interpretation of SCOD. From the above argument, it is apparent that no single and simple measurement is able to project much insight into the actual substrate loading. The conclusion in this study was to use all four measurements -volatile solids, total chemical oxygen demand, soluble chemical oxygen demand and total volatile fatty acids while recognizing the limitations of the individual indicators themselves. DESIGNING AN ANAEROBIC DIGESTER V. DESIGNING AN ANAEROBIC DIGESTER "Everything should be made as simple as possible, but not simpler." Albert Einstein According to Feilden (1983), anaerobic digestion is, from the process engineer's point of view, a very simple process. There is only one reaction stream. Horton and Hawkes (1979a) emphasized that while recognized as a low technology field, this does not mean that the design of anaerobic digesters is simple. The bioreactor is the heart of any biochemical process. Its fundamental function is to provide a controlled environment in order to achieve optimal reaction rates — for anaerobic digestion, biogas production or wastewater treatment efficiency. The feedstock must be brought into intimate contact with the bacteria for a sufficient period in order to allow the reactions to occur while keeping the bacteria as active as possible and avoiding losing them with the liquid stream. For essentially soluble wastewaters, these constraints have been addressed using two broad approaches (Callander and Barford, 1983): 1. increase the density of bacterial populations in the digesters, and 2. increase the bacterial activity by selecting optimal digester operating conditions and by ensuring an adequate supply of nutrients. The problem then, involved in digester design, is to select the digester design appropriate to each specific wastewater application and to determine the optimal operating conditions in order to meet the treatment objectives. Even at present, process design and operation is mainly based upon empirical knowledge, acquired in pilot plant experiments and in full scale operations (van den Beld, 1984). Difficulty arises because much of the current knowledge concerning plant design and operation is based on information gained from municipal sewage works. Limitations of the applicability of results of past experimentation from both the laboratory and full scale operations should be recognized because of their inability to provide a quantitative basis for design (Cowley and Wase, 1981). 18 19 Lawrence and McCarty (1969) have conjectured that optimum performance was seldom achieved because of the high degree of empiricism which has prevailed in digester design and operation. The design emphasis so far has focused on preventing the accumulation of floating or settled solids which cause operational problems such as the blockage of pipes, loss of effective digester capacity, and poor contacting of influent wastewaters with active biomass. There are two fundamental approaches to the design of anaerobic digesters (Morris et al., 1975). One is based on the kinetics of anaerobic digestion and on the theory of continuous culture of microorganisms. Knowledge of the reaction rates, the rate-limiting steps, the growth rates, the contacting pattern, and the mechanism of substrate breakdown are important to find out what a digester is able to do. The other approach is more empirical. It uses organic or hydraulic loading rates as the basis for digester design. This is the common method used in present agricultural waste management practice. The influent feed concentration of volatile solids, the effluent quality, the detention time and the operating temperature are the major design factors which determine the maximum methane production rate. Lo et al. (1984a) prescribed two stategies, not mutually exclusive, for the design of anaerobic digesters using the empirical orientation. These are: 1. to increase the rate of reaction by treating more waste in order to produce more biogas per unit volume of digester and thereby minimize the capital cost by keeping the digester small, or 2. to decrease both the operating and capital costs of the process by employing mechanically and operationally simple digesters. The limited income potential of the methane produced plus the minor credits for reduced wastewater discharge limit the acceptable capital investment for large scale digesters and requires that considerable attempts be made to keep digesters small and simple. Any increase in the complexity of the hardware has to be offset by the benefits of smaller 20 equipment in achieving the same conversion efficiency. Doyle et al. (1983) reported that about two thirds of the capital cost and one half of the operating costs are related to the cost of the digesters; therefore, the key economic parameter is the volumetric productivity. Maximizing the volumetric productivity leads to minimizing the capital investment. There seems to exist a power law relationship between size and cost Hashimoto and Chen, (1980), suggested that for digester vessels larger than 100 m\ the capital cost increases with volume to the 0.7 power. A cost optimizing exercise is necessary to determine whether for example the cost of a larger digester to enable a longer retention time is more than offset by the extra gas that can be produced. Biomass retention is expressed by the solids retention time (SRT) in the digester. It is a measure of the average residence time of microorganisms in the digester. Long SRTs can be achieved by reducing the loading rate or by retaining the solids while at the same time removing the liquid. Retention of the biomass is important because anaerobic bacteria, especially the methane formers, grow very slowly. The retention time of the sludge must be large, so that the appropriate relative populations of microbes in the mixed culture can develop during the adaptation period and can remain in the digester vessel during the subsequent pseudo-steady state periods of operation. Control of the SRT is essential for process stability. Higher flow rates and organic loadings would otherwise make the biomass more difficult to manage, thus making the process more sensitive and subject to failure (Jewell, 1981). Biological solids retention was shown by Speece and Parkin (1983) to be the key parameter in determining response to shock and toxic loadings. Long SRTs maximize potential for acclimation, and, in general, minimize the severity of response to toxicity. Observations of Anderson et al. (1982) have indicated that the possible organic loading is proportional to the number of the active microorganisms in the system. McKinney (1983) reasoned that more microbes mean more rapid contact and more substrate metabolized. High activity is due solely to the high biomass concentrations achieved; the 21 specific activities of each process is the same: 1.8kg TCOD/kg/day (Fynn and Whitmore, 1984). The available food limits the microbial growth and helps to determine the steady state population of active microbes in the system. Higher concentrations of biomass means that higher volumetric reaction rates, lower space requirements, and potentially lower capital costs can be achieved in the "advanced" or retained biomass digester designs (Walker et al., 1981). Extensive studies by van den Berg and Lentz (1980) have demonstrated that the advanced digesters exhibit higher performance, better conversion and greater reliability over the first generation designs. Especially under high organic loading rates and short HRTs, methane production was several times that found in conventional digesters. Active biomass retention permits HRT reduction from the 10 to 20 day processing times characteristic of conventional and anaerobic contact digesters to periods ranging from several days to hours. Even highly diluted wastewaters can be treated with high efficiencies provided the SRT is carefully controlled. Processes are required in which biomass retention can be controlled independent of the wastewater flow rate. The HRT should be minimized, increasing the hydraulic throughput, to reduce required digester volume and thereby improve system economy, and the SRT should be maximized to improve the volumetric removal rates and to prevent microorganism washout (Colleran et al., 1982 and Kennedy and Droste, 1983). Ideally, it is desirable to load digesters as high as possible for the shortest period of time consistent with targetted methane yields or wastewater treatment efficiency. Loading at a high concentrations reduces the required digester volumes at the same overall residence time, cuts down the preheat and digester temperature maintenance loads, and reduces water requirements, affecting storage and disposal. One effect of the increased organic load is to generate more biomass and, if retained, a corresponding increase in the digester processing capacity (Barnes et al., 1983). 22 Aivasidis (1983) and Genung et al. (1980) have considered the range of mechanisms for biomass retention in anaerobic digesters. The different methods include: 1. sedimentation using either flocculation, as in the anaerobic contact process, or pelletization, as in the upflow sludge bed digester, 2. immobilization employing support surfaces for microbial growth, as in the fixed film, fluidized or expanded bed and anaerobic filter, 3. filtration, as in the anaerobic filter, sludge bed and expanded bed, 4. flotation of sludge and foam, as in the upflow digesters with subsurface effluent withdrawal, and 5. centrifugation and recycling, as in the anaerobic contact process. Regardless of the means of microbial retention selected, the outcome must satisfy these general constraints (Bull et al., 1983, Cooney, 1983, and Hill, 1982b): 1. maximize gas yields and/or wastewater treatment efficiency on a volumetric or time basis, 2. minimize energy costs for gas compression, mixing, temperature control, 3. provide for efficient mass transfer by developing a large interfacial area, 4. ensure a steady biomass holdup to yield steady process performance, but at the same time minimize the excess sludge yield, 5. design for a low and uniform shear field in order to achieve uniform biomass retention with no dead volume or sedimentation or blockage, 6. allow for continuous operation with a narrow residence time distribution, 7. minimize the capital and operating costs by keeping the required controls, hardware and operator time low, and 8. ensure applicability for a wide range and variability of wastewater characteristics and processing rates. The differences in performance between the various digester configurations can be associated with the differences in residence time distribution and the concentration profile of 23 the digester contents. Factors affecting digester performance include (Lettinga et al., 1983a): 1. the feed characteristics: available substrate in terms of volatile fatty acid content, solids content, biodegradability, COD concentration, pH, nutrient content, and temperature. 2. the design of the process: a. the feed inlet system, b. the feeding frequency, c. the height:area ratio, and d. the extent of mixing. 3. the form of bacterial retention: stationary films, floes, or suspended particles. A. FORM OF BACTF.RTAT. RFTKNTTON Retained biomass in the form of biofilms or floes are the "centerpiece" of advanced digester design (Rittmann and McCarty, 1980). The retained biomass digesters are founded on the tendency of bacteria, especially the methanogens, for attachment to solid surfaces. In stationary fixed film digesters, bacteria attach themselves to the fixed inert supports, while in sludge bed digesters they attach themselves to each other or to suspended material to form conglomerates. In order to improve settling and retention, the active biomass can be immobilized on heavy particles and retained within the digester under the hydrodynamic conditions necessary for bed fluidization. The performance characteristics of the digesters are dominated by the form in which the microbial hold-up occurs (Atkinson and Knights, 1975). A biofilm is a layer-like aggregation of microorganisms attached to a submerged solid surface. In many natural aquatic systems, especially those with a high specific area and low nutrient concentrations, biofilms constitute 90 to 99.9% of the bacteria (Gibbons and Denton, 1981). Biofilms are found in or on streambeds, groundwater aquifers, lake benthos, water pipes, ship hulls, piers, and aquatic plants and animals. The film provides an extended surface for mass transfer and serves to cleanse the waters of organic materials (Fowler, 1981). Wastewater treatment is an enhancement of this natural process. 24 Flocculation of microorganisms is the aggregation of a number of single cells into floes when suspended in a liquid medium with subsequent flotation or sedimentation (Esser and Kues, 1983). Synonyms include clumping, agglutination, coagulation, adhesion, association and agglomeration. Flocculation may be used to separate microbes easily from their culture fluid The mixed microbial cultures associated with wastewater treatment have excellent adhesion characteristics and readily form continuous layers of immobilized biomass on commonly available materials (Atkinson, 1981). Studies conducted by Eighmy et al. (1983), have demonstrated that different microbial species are preferentially selected for incorporation in the biofilm depending on the availability of substrate. Static films enable a biomass accumulation of up to 10 times the biomass density over the suspended growth systems (Jewell et al., 1981a). These attached cells allow for the higher liquid velocities at increased volumetric loadings and for some control of film thickness while still delivering effluent of good quality. Methanogenic acetate converters are the slowest growing bacteria in the system but the retention of these bacteria is a major factor in the high rate performance of retained biomass digesters. Neither the general mechanism of cell aggregation nor its environmental dependence are fully understood. Flocculation depends on the physiological age of the cells. It occurs as soon as the energy sources are exhausted during the stationary phase and is optimal during the endogenous growth phase. Flocculation depends on the concentration of cells in suspension. It is retarded under conditions of strong agitation leading to turbulence and thus to the occurrence of shearing forces. The presence of calcium, magnesium and sodium ions appears to stimulate the flocculation of bacteria (Lettinga et al., 1983a). The film support material affects both the rate of film development and ultimate maximum performance (van den Berg and Kennedy, 1981a). The rate of film development may be related to the ease with which bacteria are trapped and held on the support surface. A superior support material has a rough, porous surface which offers many attachment sites 25 to microorganisms (Murray and van den Berg, 1981). Many cavities and indentations on the surface provide local anchorage sites sheltered from fluid shear stresses. Other factors influencing the degree of biomass retention include (Gibbons and Denton, 1981): the wastewater contacting efficiency, the hydraulic load and fluid regime, film density, specific surface area, depth of packing, temperature, influent concentration, ionic strength, pH and time. Microbial films are to a degree self regulating (Fowler, 1981). Biofilms can be developed to a depth limited by hydraulic shear, organism decay, and of course, the digester volume. The steady state biofilm is the mass of biological growth that can be supported by a constant substrate concentration and loading rate. Binding forces between methanogens and the surface are weak and may impose a limit on the potential for process intensification in the fixed or fluidized bed digesters (Fynn and Whitmore, 1984). A common design strategy is to maximize the area available for contact either on a plane or particle surface. Even though microbes may be loaded to a high level on support structures, this does not demonstrate their utility. A very thin film lacks the maximum numbers and hence activity for the available surface. Beyond the substrate penetration depth however, the film becomes less active per unit biomass volume. Such a substrate limitation gives rise to endogenous respiration, and leads to the sloughing off of thick layers. Sloughing is not dependent upon the biofilm surface topography but is more likely to be a function of the viscoelastic nature of the slime film and the velocity of the solution across it (Gibbons and Denton, 1981). Besides the substrate diffusion limitations, the build-up of large microbial masses in the course of treating high concentration wastewaters presents problems of eventual clogging. 26 B. FFFD TNT .FT SYSTEM A fundamental difference exists in the form of the retained biomass between the upflow and downflow digesters. The downflow configuration ensures that suspended solids, including any suspended bacteria, are not retained in the digester. Upflow operation results in a combination of a fixed film and sludge bed digester resulting in elevated biomass concentrations and higher rates of methane production (Kennedy and van den Berg, 1982a). Readily degraded solids can be handled by all digesters. Since inert suspended solids cannot be easily separated from the active biomass, the easiest mode of operation is to remove the inert suspended solids prior to feeding into the digester, or, much less desireably, to allow the digester to accumulate a maximum possible solids concentration and then allow the excess solids to be discharged. More recalcitrant feeds with high particulate solids and relatively low nutrient concentrations can be more effectively dealt with in digesters, such as the upflow digester, which promote long SRTs. The retained biomass digester designs may pose operational problems during digestion of wastes containing high levels of hard to degrade suspended solids. They may decrease digester efficiency by plugging or clogging the sludge blanket or filter matrix. Accumulation of inorganics such as grit and refractory materials from the feedstock also causes considerable problems (Fannin et al., 1983). They can occupy volume within the digester, rendering it ineffective, and they block pumps and conduits, causing damage and plant downtime. The upflow designs are expected to give increased retention and conversion of suspended solids due to greater retention of both solids and active biomass (Callander and Barford, 1983 and Barnes et al., 1983). The downflow mode of operation is more trouble free than the upflow operation. Solids which are difficult to digest cannot be handled easily in upflow digesters due to accumulation and eventual clogging. With the downflow mode of operation, plugging is avoided (van den Berg and Kennedy, 1983). 27 The withdrawal of undigested solids, while lowering the risk of digester disruption, also reduces the effluent quality. According to van den Berg and Kennedy (1982b), the major advantages of the downflow mode of operation are the influent dispersion at the point of maximum turbulence from the rising biogas, and the withdrawal of undigested solids with the effluent While designing immediate dispersion of the feed caused by emitting biogas is of advantage if inhibitory substances in the feed pose a threat to digester operation, such dispersion in downflow operation presents a drawback in precluding plug flow operation. C. FRRDTNO FRFOITFNCY Although batch and semi-batch operation have the advantage of requiring less daily attention, the microorganisms in a batch fermenter are exposed to an ever-changing environmental history. In a typical batch fermentation, the microbial growth pattern exhibits different phases — lag, logarithmic, decline, and resting phases (Aiba et al., 1965). A small number of batch digesters operating in series approximates continuous operation (Levenspiel, 1979). In continuous operation, the change of composition is shifted to a space coordinate as opposed to a time co-ordinate in batch operations. Continuous processes are favoured for large scale operations due to (Doyle et al., 1983): 1. diminished labour costs and less down time, 2. amenability to automatic control, 3. greater constancy in digester conditions resulting in a more consistent gas output and effluent quality, and 4. the steady load on services required by the process results in smaller components — heaters, mixers, pumps, as well as the digester itself. van den Berg and Lentz (1980a) predicted that the application of anaerobic digesters on a larger scale would depend on continuous rather than periodic operation. Continuous processes often have higher volumetric productivities than batch processes (Cooney, 1983). 28 Biological waste treatment systems perform best when loaded continuously with relatively dilute wastes. As a result, these treatment systems are either oversized or require large equalizing tanks in many industrial applications where wastes are produced intermittently (van den Berg and Kennedy, 1982a). The present lack of continuous process applications may be attributed to the technical difficulties encountered. Intermittent loading amounts to short term hydraulic and organic overloading (Kennedy and van den Berg, 1981b). For a given hydraulic retention time, the extent of this periodic overloading is decreased as the loading frequency increases. The desired loading frequency will reflect the capacity of the digester to withstand or even benefit from the periodic hydraulic and organic overloading and its ability to consume the substrate efficiently. Digester stability increases as the frequency of loadings increase. Large oscillations in pH, in the rate of methane production, and in the carbon dioxide composition of the digester gas can be reduced substantially by loading the digester several times a day (Graef and Andrews, 1974). Asinari Di San Marzano et al. (1981) observed that semi-continuous loading appeared to be more reliable than continuous loading. However, they cautioned that no more than 20% of the digestion mixed liquor should be replaced at one time. Intermittent loading of a stationary fixed-film digester allowed the same or higher loading rates and rates of methane production as continuous loading (van den Berg and Kennedy, 1982a). For the same hydraulic loading rates, twice daily addition of wastes increased the proportion of methane formed as compared to once only feeding (van den Berg and Kennedy, 1982a). The maximum daily amount of waste added was 20% higher during two versus one daily loadings without adverse effects on subsequent digester performance (van den Berg and Kennedy, 1982a). Bhatia et al. (1981) and van den Berg and Kennedy (1982a) reported that not only were there no stability problems with drastic changes, but the system seemed to operate better after a complete cycle of step-up and step-down of inputs. They conjectured that it might be useful to operate the digester in a programmed transient condition for better conversion efficiency. 29 For a small digester, loading has to be intermittent due to downward limitations on commercially available pump sizes. Consequently, effluent withdrawal must also be intermittent. This restriction results in a sequencing operation, analogous to the sequencing batch reactor employed in aerobic processes. Short intervals of turbulence at the influent port, followed by a react phase and settling interval and subsequent supernatant withdrawal may take advantage of the positive aspects of both continuous and intermittent feeding (Mills, 1979). The settling period must be compatible with the substrate utilization rates. D . A S P F C T R A T I O The geometric arrangement of bioreactors, particularly as it affects the liquid flow pattern, contributes significantly to performance (Atkinson, 1974). Atkinson claimed that the potential for the exploitation of the interaction between digester configuration and the reaction is probably greater with biological systems than is currently achieved with chemical systems. Biological wastewater treatment is carried out in the decreasing growth phase and the reaction rate is generally controlled by substrate concentration (Busch, 1984). Below the microbial saturation concentration, that is, for dilute and moderate strength feedstocks, the reaction rate decreases as conversion increases and substrate is consumed if the reaction rate is concentration dependent The use of large bacterial populations to increase the overall reaction rates further removes the reaction rates from the maximum growth rate situation by lowering the food to microorganism ratio. Anaerobic digester operation has been centered around maintaining those conditions which favour the slow growing methane formers. Acid formers however, differ in their environmental optima from the methane formers (Jech and Brautigam, 1983). Separation of the groups of bacteria has been examined by numerous investigators in order to optimize the conditions for bacterial growth and substrate utilization for both. Anaerobic digestion, however, is not a two phase production of acid formers and methane formers, but a multistage process involving many different bacteria, each of which which have a specific 30 function in the step-wise degradation of substrate (Wilson, 1981). The challenge of anaerobic digester selection is to accomodate those conditions favouring the optimal reaction rates into a simple reaction vessel. Reactor designs from which the designer has a choice include (Levenspiel, 1979): 1. batch, where the fluid composition is uniform everywhere in the reactor, but the composition changes with time, 2. mixed flow, where fluid is uniformly mixed, resulting in uniform composition within the reactor and at the exit, and 3. plug flow, where the fluid passes through the reactor with no mixing of earlier and later entering fluid, as if the fluid moves along in single file with no overtaking. The continuously stirred tank reactor (CSTR) is the conventional mixing vessel. It is generally a cylinder with a 1:1 aspect ratio -the least cost practical vessel and that with the smallest surface area per unit volume (Mills, 1979). Because the CSTR concentration is homogeneous throughout the reactor and equal to the effluent concentration, the average reaction rate is set by the effluent concentration. A single stage CSTR can therefore only compromise the reactor conditions for the groups of bacteria. In order to achieve the same degree of conversion, a CSTR must be larger than a plug flow reactor or a batch reactor (Levenspiel, 1979). The aggregate volume of a series of CSTRs for the same degree of conversion is less than that for a single CSTR, but it is still greater than that of a plug flow reactor or a batch reactor. A plug flow reactor is approximated by a long, narrow tube, a packed bed, or a number of CSTRs linked in series in order to prevent back mixing (Klass, 1980). Since the columns used are long cylinders with bed heights many times greater than the bed diameter, it is reasonable to assume a relatively flat liquid velocity profile exists, provided the liquid flow is laminar and the liquid distribution is uniform (Ngian and Martin, 1980). A greater than 10:1 length to diameter ratio, the aspect ratio, is a necessary but not always sufficient condition for plug flow to occur. As well, longitudinal dispersion must be minimal or absent 31 In a plug flow reactor, concentration gradients develop along the length of the reactor. The average reaction rate has a value somewhere between the high initial rate and the low final rate (Busch, 1984). For all reactions where the rate increases with concentration, plug flow always requires a smaller volume than mixed flow (Levenspiel, 1979). The use of plug flow encourages the microorganisms to perform at an increased metabolic activity due to the development of essentially separate environments. Operation under plug flow conditions exposes the films at a particular location to an wastewater environment generated by the films upstream. With mixed cultures, the local species are determined by the local environment Since the substrate is depleted as it passes through the digester, a nutrient and hence bacterial profile develops along the axis in the direction of flow. Consequently, microbes engaged in the acid forming stages should predominate at the feed inlet, and methanogens at the feed exit This has been observed experimentally. Anderson et al. (1982) expected varying distributions of the organisms to develop when levels of acetate were high at the bottom and low at the top of the filter. Tesch et al., (1983) also reported this occurrence at high retention times. This nutrient profile is reflected in the process of start-up where Hilton et al., (1983) observed that visible colonization began at the bottom of the filter, but near the end of the experimentation, nearly all of the packing was colonized. During steady state operation, a profile of soluble COD and volatile acids concentration versus filter height developed in the anaerobic filter. The major fraction of the waste was removed in the lower part of the filter, particularly with higher strength wastes (Young and McCarty, 1969). If we can learn from natural processes, it might be useful to observe that the digestive tract of animals functions along a plug flow mode. This mode of operation combines increased reaction rates and hence increased treatment efficiency with enhanced effluent quality as compared to the more commonly used CSTR digesters. The more predictable residence times, as opposed to the short circuiting that occurs in completely mixed systems, means that wastewater reduction and pathogen destruction can be more 32 reliably predicted. E MIXING The degree of mixing influences the nature of a digester. Two extreme modes of digester operation are the plug flow reactor with no mixing, and the completely stirred tank reactor (CSTR). In the past, the CSTR was chiefly used. Mixing can aid the digester's performance in a number of ways. The key parameter which controls the rate of digestion is the degree of contact between the microorganisms and the substrate (Wilson, 1981). Mixing accelerates the rates of heat and mass transfer, and therefore promotes the rate of biochemical reaction. It maintains uniform temperatures and concentrations, and concentrations of inhibitory intermediates and end products are dispersed to the lowest possible levels. It inhibits particle settling, reducing stratification and suspending active bacteria. Efficient mixing keeps scum formation down to a minimum. Scum can otherwise accumulate and occupy digester volume as well as impede the evolution of biogas from the liquor surface. Mixing assists with the physical breakdown of coarser particles into smaller particles, and makes the substrate more available to microorganisms. Care must be taken, in the design of digesters, to ensure that shear, even locally, is kept to a minimum in order to avoid break up of bacterial floes or films followed by washout (Tramper et al., 1984 and Stafford, 1982). If mixing is not efficient, pockets of substrate will proceed at different stages of digestion, and may be affected by different pH and temperature zones. An incompletely mixed digester has a dead zone which is inactive biologically. The effect of this dead zone is to reduce the retention time of the active volume of the digester. The effect of incomplete mixing is minor at long detention times, a situation typical of most over-sized digesters. The energy required for mixing is a function of the solids concentration in the digester, the liquor viscosity, the digester configuration and the nature of the mixing equipment itself. The available options used to achieve mixing include mechanical mixing, 33 gas recycle or liquid recycle. Mechanical agitators such as augers, impellers, bottom scrapers or rotating disks suffer from the disadvantage of requiring gas seals and bearings. Methane, as a member of the paraffin family, breaks down greases in conventional seals and bearings even though the gas pressure may be low (Mills, 1979). Gas mixing, by compressing gas obtained from the top of the vessel and releasing it at the bottom, requires energy, but the energy requirement for the same degree of mixing is less for gas than it is for liquid recycle due to the buoyancy of the gas (Mills, 1979). Gas mixing should require less maintenance than conventional mechanical systems. The rate at which gas is recirculated is generally proportional to the digester area and the pressure required is proportional to the depth at which the pressurized gas is released. Reintroducing biogas into the digester helps to ensure that maximum stripping of carbon dioxide from the biogas occurs, thus enriching the biogas in methane content and supplying additional buffer capacity to the digester liquor. While biogas enrichment through carbon dioxide stripping in the digester liquor may appear to be attractive, the effects on the digester liquor and hence on the process biochemistry are worthy of consideration. In the course of investigations concerning end-product inhibition of the biogas, Hansson and Molin (1981) found that carbon dioxide was inhibitory (under 1 bar C02) and that methane was only slightly inhibitory. Effluent recirculation causes the applied waste load (the raw waste load plus the recycled load) to increase, the net HRT to decrease, and the influent waste strength to be reduced. Effluent recirculation is reportedly beneficial when treating very high strength wastes (COD greater than 10,000 mg/L), wastes containing toxic materials, or wastes with a low buffer capacity (Rittrnann, 1982b). The usefulness of a recycle capability should increase as the feed biodegradability decreases, since the increased solids retention time allows for greater solids breakdown, and as the variability of the incoming feed increases, since recycle 'smoothes' feed fluctuations. Although recycle of liquid or gas lowers the potential process efficiency by making the reactor more completely mixed, it can be of practical value in 34 controlling clogging (Callander and Barford, 1983). Recycling can be used to control clogging because it distributes substrate and hence the biofilm more evenly, and because it increases the flow velocity throughout the reactor. Using liquid recycle, transfer from the plug flow to the completely mixed regime can be accomplished. This allows for the possibilities of employing both mixing regimes in a single unit, depending on the changing conditions of the influent waste stream (Oleszkiewicz, 1981). Callander and Barford (1983) advocated rmmnuzing the extent of liquid recycle, since for a given loading, approaching plug flow results in higher waste conversion and reduced process energy requirements. As well, there is evidence that the metabolic activity of bacteria improves if they can operate in a more or less quiescent environment (Lettinga, 1981). Minimizing or altogether eliminating mechanical mixing of the digester should be pursued. In practice this is accomplished in digesters of a more or less plug flow design. ANAEROBIC DIGESTER DESIGNS VT. ANAFROBTC DIGESTER DESIGNS Advanced digester designs differ in the way in which biomass is retained. The method of microbial retention affects the design of wastes which can be treated, the method of operation, the loading rate and the conversion efficiency. Maintenance of active biomass within the digester independent of wastewater flow requires the development of different anaerobic digestion strategies and configurations. Many solutions have been proposed, of which the anaerobic filter, the upfiow anaerobic sludge bed, and the expanded bed are perhaps the most promising (Frostell, 1981, Switzenbaum, 1983, and Wilson, 1981). Based on this and the assessments of other investigators, these are the digester designs selected for study in this research. The differences between the high rate digesters chosen for study can be generalized in terms of both packing density and mobility. The expanded bed, sludge bed and upfiow anaerobic filter each use microbial retention, but in a different way. Each of the digester designs is considered in turn. A. TJPFT.OW ANAEROBIC? FITTER The anaerobic filter consists of a vertical filter bed filled with an inert support material. Anaerobic microorganisms accumulate on and in the void spaces between the packing and in particles and clumps, so that the waste comes into contact with a large active biological mass as is passes up through the filter. A very high biomass concentration may be retained in the digester, with up to 100,000 mg/1 being reported (Anderson et al., 1982). The movement of the upward flowing liquid and the rolling action of the rising gas bubbles causes the sludge particles to take on a granular shape. These granules can become as large as 0.31 cm (0.125 in) and settle readily. Essential features of the anaerobic filter include (Lettinga et al., 1980a): 1. a distributor in the bottom of the column, 2. a media support structure, 3. inert packing material, 36 37 4. an effluent draw-off, and 5. optional features which include: liquid and gas recycle, liquid backwash capabilities, and a sedimentation zone above and/or below the packing material. These features are illustrated in Figure 2. No extra space or design features are required for gas separation or settling of suspended material. The packing itself serves to separate the gas from the biomass and to provide quiescent areas for settling of flocculated biomass (van den Berg and Kennedy, 1983). Anaerobic filters are generally operated as single pass plug flow digesters, taking advantage of the high driving force of the undiluted waste stream. The hydraulic head required to push the waste through the filter is low at all loadings, including operation at low HRTs, except during periods of high solids accumulatioa The biological solids remain in the filter and are not readily lost in the effluent (Young and McCarty, 1969). Solids do not readily become attached to the surfaces but lay loosely in the interstitial spaces. Although the upflowing gas bubbles lift some sludge particles, these separate when the sludge hits the overlying packing. Most particles settle back to their original position and the gas bubbles continue to rise. Eventually, some of the flocculated particles are carried up through the filter and are lost in the effluent. The proper selection of packing material will depend to a large extent on wastewater characteristics, especially the concentration of particulates. According to Kennedy and van den Berg (1981b), a suitable configuration of the digester's support surfaces enables the fixed film digester to handle large amounts of suspended solids, even if they are indigestible, without these solids accumulating or plugging the digester. Jennings et al. (1976) have suggested that the packing should be selected to have a high surface area to volume ratio to provide a large surface area for film attachment, while having a sufficient void volume to prevent the digester from plugging. In order to prevent clogging, a lower limit on the solid packing size and an upper limit on loading, especially the loading of suspended solids, may be necessary. A significant fraction of the retained biomass is not attached but is rather 38 FIGURE 2: Schematic of the Anaerobic Filter Process GAS RELEASE <-WATER RESERVOIR GAS COLLECTOR" WATER, GAS •> SAMPLE GAS EFFLUENT EFFLUENT SAMPLE EFFLUENT PUMP FEED TANK IGESTER RASCHIG RINGS ALUMINUM FOIL WRAP EATING PAD POLYSTYRENE FOAM INSULATION DISPERSION PLATE FEED PUMP SAMPLE FEED EFFLUENT TANK 39 entrapped in the void spaces of the packing. A greater porosity allows for greater biomass entrainment by presenting more working volume per digester volume, and by reducing interstitial flow velocities for a given loading (Dahab and Young, 1981 and Hudson et al., 1978). Other desireable packing material qualities include (Marshall and Timbers, 1982): 1. an irregular surface suitable for bacterial attachment, 2. support material composition able to buffer or supply trace nutrients, 3. a flow channel adequate to prevent the blocking of pores with solids accumulation or microbial growth after prolonged periods of operation, 4. ease of fashioning supports suitable for both laboratory and commercial scale applications, 5. available in lengths compatible with full scale digester height, 6. lightweight, requiring little in the way of structural support, and 7. commercial availability and ease of installatioa Materials used have included: rocks, gravel, plastics, polyvinyl chloride, needle punched polyester and other synthetic materials, wood, carbon, potter's clay and draintile clay, each in various configurations. Performance is more related to the manner in which wastes flow through the media matrix than to the media surface area itself. Tests by Young (1983) have shown that fluid flow initially follows an almost ideal plug flow pattern (also Young, 1968 and Rittmann, 1982b). As biological solids accumulate and as evolved gases induce mixing at elevated organic and hydraulic loadings, the hydraulic regime nears completely mixed conditions. Random packing such as loose fill Pall rings, Raschig rings OT perforated spheres reduce the opportunity for wastewater to come into contact with the biomass as compared to the modular media. With sustained biomass accumulation, waste flows around rather than through the packing and reduces the contact time within the digester. By using channels instead of random packing, potential problems of plugging and channelling could be largely 40 avoided, although channel design packing allows for vigorous agitation and a consequent wasting of bacteria as the biogas is produced. Anaerobic filters, operated in a quasi-plug flow mode, have valuable advantages over complete-mix, suspended growth reactors. They promote faster recovery from toxicant exposure through rapid toxicant washout and favour acclimation through bacterial retention according to Speece and Parkin (1983). I^ ttinga et al. (1980a) seem to disagree however. They considered the lack of effective mixing of the anaerobic filter to render it more sensitive to shock loadings, low influent pH, and toxic compounds than the other well-mixed processes. Their disagreement appears to stem from the degree of toxicity considered. Plug flow is superior to completely mixed operations if the shock load stresses but doesn't kill the microorganisms by washing out the shock through controlled residence times. On the other hand, high shock loads capable of causing permanent damage to the microbial system could be dispersed under completely mixed conditions. The desireable flow regime is determined by the strength and duration of the influent shock load. Even a filter with relatively consistent influent concentrations and operating conditions can exhibit a widely fluctuating output (Howell and Atkinson, 1976). The sloughing of large areas of attached biomass can lead to plugging. This can result in short circuiting and loss of pore volume, both of which reduce treatment efficiency by reducing the effective HRT and SRT. The anaerobic filter was developed for soluble, low strength wastes. It cannot treat all designs of wastes satisfactorily. Only small amounts of degradable suspended solids can be accepted without problems of plugging. The anaerobic filter does have a tendency to plug and channel with even moderate suspended solids loads, increasing the pressure drop and decreasing its long term effectiveness and perhaps leading to complete shutdown (Young, 1985). Channelling is a problem particularly for lower flow rates. Filters should be operated continuously to help counter this problem (Atkinson et al., 1980). 41 Packing offers a number of advantages. It reduces the likelihood of washout, allows for simple operation, obviates mixing and settling, and enables no recycle requirement of effluent or solids. With bacteria adhering to a surface, the digester height is not as limited as it appears to be in suspended growth digesters. Due to less efficient substrate biomass contact caused by developed biomass or accumulated suspended solids blocking channels, loading rates as high as in the sludge bed cannot easily be reached with the anaerobic filter process. Compared with the suspended growth systems, the fixed film digester is limited in its film surface area to digester volume ratio. This places an upper limit on digester loading. Biomass levels should be controlled to avoid bed clogging as well as to maximize volumetric efficiency. Bull et al. (1984c) have indicated that such control is effected by allowing the natural mechanisms of adhesion and flocculation to take place in a low shear environment and then controlling the size of the resulting particle or film hydraulically. Accumulated suspended material such as sloughed biomass or indigestible suspended matter in the feed such as lignin, clay, or sand can be eliminated from the digester as a matter of routine operation by upflow recycling digester liquor or biogas, or by backwashing (van den Berg and Lentz, 1979a). The anaerobic filter is suitable for efficiently treating dissolved wastes at high organic and hydraulic loading rates (Lettinga et al., 1980a). Waste strengths varying from 4,400 mg TCOD/1 to 140,000 mg TCOD/1 could be added directly to the digesters without adverse effects (van den Berg et al., 1981b). The anaerobic filter, with the exceptionally long SRTs inherent in its operation, allows the retention of biomass over exceptionally long acclimation periods so that high waste utilization efficiencies are eventually possible (Speece and Parkin, 1983 and Young, 1983). The fixed film digester could be operated at shorter HRTs and higher loading rates than the conventional digester. Lo et al. (1984a) reported that as the HRT is reduced and consequently, the loading rate is increased and the difference in methane production rates and effluent quality between the fixed film and the contact digester increased sharply. Kobayashi et al. (1983) cautioned that the limits to the HRT and 42 organic volumetric loading rates depend upon the influent waste concentration, degradability, temperature, and the desired efficiency of organics removal or the rate of gas production. The fixed film digester exhibits a high degree of adaptibility to feed variations (Lo et al., 1984b). In spite of the long initial start-up period, the process is capable of rapid restart after long periods without feed, and offers good stability and reasonable recovery when stressed by shocks of organic loads, pH, or toxic materials (van den Berg and Lentz, 1978a). B. UPFLOW ANAEROBIC EXPANDED BFD Lettinga and Vinken (1980), in exploring various directions for anaerobic digester design, have indicated that a promising approach is to increase the bacterial population through maximizing the surface area available for attachment while nTinimizing the volume occupied by the media. Packed bed digesters such as anaerobic filters often experience problems of increased pressure drop due to the accumulation of biomass (Boening and Larsen, 1982). In order to prevent such plugging, relatively large voidage must be maintained. This limits the specific surface area and therefore the biomass concentration. Expanded bed digesters have overcome this problem by using low voidage, mobile, high surface area media. Expanded beds can considered to be a hybrid between the fixed film and the free floating digester systems. The expanded bed process consists of inert sand-sized particles contained in a vertical column which expand as a result of the upward direction of liquid or gas flow. Once at or beyond the minimum point of fluidization, the media particles are individually and hydraulically supported. A desireable fluidization velocity is one which barely suspends the largest particle but does not wash out the smallest one (Sutton et al., 1982). Like the anaerobic filter, the components of an anaerobic expanded bed digester consist of: 1. a wastewater distributor, 43 2. a media support structure, 3. media, 4. headspace, 5. effluent draw-off, and 6. gas or liquid recycle capabilities. These features are illustrated in Figure 3. A settling device to aid in separating the biomass covered inert particles from the effluent wasting is also sometimes incorporated. The inert particles provide a support surface for the growth of microorganisms. Each particle eventually becomes covered with biofilm and the vast available support surface afforded by the media results in biomass concentrations of an order of magnitude greater than what is achieved in suspended growth systems (Shieh, 1980). Thus, for a comparable treatment efficiency, the required digester volume should be greatly reduced. The problem of cell washout which aggravates continuously operated digesters is answered by maintaining the superficial velocity below the settling velocity of the solid particles (Andrews, 1982), or by operating in a sequencing mode, with short periods of settling preceeding effluent withdrawal. Washout of the solids can be controlled by adjusting the degree of bed expansion through changing the recycle ratio. Surface areas on the order of 300mVm3 of digester volume (Hickey and Owens, 1981) have been reached. Jeris et al., (1977) have seemed to contradict these low figures, reporting the surface area to be 3300 mVm3 of digester volume — an order of magnitude difference. Jewell et al., (1981a) have expressed the effective biomass concentrations in the digester to be between 20 and 30 kg volatile solids per m3 of digester (empty volume). The bed porosity or the concentration of particles in a fluidized bed at a given set of operating conditions is a function of: the expanded bed height, the amount of support media used, the particle fluidization velocity, particle size, shape, density and biofilm thickness, fluid viscosity and density, and the hydraulic characteristics of the system (Shieh et al., 1981b). 44 FIGURE 3: Schematic of the Anaerobic Expanded Bed Digester GAS RELEASE <r WATER RESERVOIR GAS COLLECTORV WATER. •» SAMPLE GAS EFFLUENT GAS RECYCLE EFFLUENT SAMPLE EFFLUENT PUMP FEED TANK .DIGESTER .EXPANDED MEDIA LLUMINUM FOIL WRAP .HEATING PAD POLYSTYRENE 'FOAM INSULATION -DISPERSION PLATE t/FEED PUMP i SAMPLE FEED GAS RECYCLE PUMP EFFLUENT TANK 45 In order to fluidize a bed, the shear stress must be at least great enough to balance the net negative buoyancy of the media. The extent of bed expansion is strongly dependent on the geometry of the bed, methods of gas injection, and the presence of a retaining grid or internal structures (Muroyama and Fan, 1985). The selection of particles requiring low flow velocities and accompanying low shear stresses is advantageous for biofilm formation. In contrast to stirred tank systems, gas mixing requires no moving parts and the only power requirement comes from the compressor system which delivers gas to the sparging system. The gas bubbles are responsible for the induced turbulent liquid mixing and accompanying increased mass transfer. Fluidization of the particles using gas provides less buoyancy for the particles in the bed than liquid recycle would provide. The main advantages of gas mixing as opposed to mechanical mixing or liquid recirculation are the low shear and low energy requirements, and the simplicity of construction since seals are not required around rotating shafts (Lee and Buckley, 1981) and the gas is already separated from both the support particles and the digester liquor. The main variables available for manipulation by the designer are the amount, design, size and density of the support particles (Andrews, 1982). There are tradeoffs between the size and the density of support particles and the stability of operation (Switzenbaum, 1983). An increase in the support particle size would result in a decrease of the particle surface area per unit bed volume. This would require the biofilm thickness to increase in order to yield the optimal biofilm volume. The selection of smaller particles would provide for a greater surface area to volume ratio and thus give a greater surface area for attached microbes. There are practical limits to the carrier particle size reduction, however. Excessive bed expansion associated with the use of small media would cause a decrease in the biomass concentration and thus in the efficiency of the digester (Shieh, 1980). On the other hand, biomass concentration would decrease with increasing media size when a certain media size is exceeded. This would result in a decreasing conversion rate. Larger media requires an increased superficial velocity in order to attain a given bed 46 expansion. The use of particles with a diameter below 0.1 mm is recommended (Jewell, 1981 and Shieh et al., 1981). Maxham and Wakamiya (1980) and Kargi and Park (1982) have claimed that the ability of particles to concentrate biomass is a function of the volume of biomass to particle ratio and hence the overall particle specific gravity, not particle size. As the microbial film forms on the particles, the overall density of the biofilm-coated particle starts to decrease and the bed expansion becomes higher. The biofilm wet density has been estimated as 1.14 g/cm3 (Maxham and Wakamiya, 1980). The higher the specific gravity of the biofilm covered particle, the greater is the biomass concentrating effect Any such advantage must be weighed against the higher energy requirements and other operational problems for the fluidization of heavier media. The use of a heavy carrier material makes it difficult to obtain the desired amounts of attached biomass. Lighter particles can be fluidized at lower upflow velocities which reduces the recycle rate required for fluidizatioa The density of the particles should be as small as possible but still enable easy separation from the digester liquor. No definitive statement can be made about the best particle density because different densities give different advantages. In terms of the ease of particle control, the best density is in the range of 1.06 to 1.1 g/cm3 (Andrews and Trapasso, 1985). In this range, film growth has little effect on the particle settling velocity until the biomass fraction becomes very large. A range of film support materials have been used: aluminum oxide pellets, ion exchange resin, diatomite, expanded clay, activated carbon, crushed brick, anthracite coal, poly vinyl chloride, and anion and cation resin particles, each with diameters between 0.5 and 1 mm. Sand has been most frequently used as a microbial support material -since it is cheap, widely available and robust (Maxham and Wakamiya, 1980). Natural particles such as sand or coal are the cheapest but the particles are neither spherical nor of uniform size. Spherical particles have advantages associated with ease of fluidization, and the extensive knowledge of their fluidization characteristics. Sand does not seem to be an ideal carrier 47 material although it is attractive due to its low cost (Atkinson et al., 1981). Fluidized bed digesters can achieve superior performance to complete mix and fixed bed digesters because the biofilm is evenly distributed throughout the digester while the liquid regime may retain some plug flow characteristics (Rittmann, 1982b). By distributing the biofilm throughout the digester, the fluidized bed digester utilizes the biomass and digester volume effectively. In theory, the solid material is completely mixed, but in practice, due to the wide range of particle sizes introduced into the digester and the low bed expansions used, this is not always the case. This may lead to the situation whereby the two main groups of bacteria, the acids formers and the methane formers, exhibit different activities throughout the digester, effecting a form of phase separation (Bull et al., 1984b). Lower flow through velocities resulting in reduced bed expansion is the major operational difference between the expanded and fluidized beds. In practice it has been found that fluidization is not necessarily a requirement and that modest relative particle movement is often sufficient (Atkinson et al., 1980). Under conditions of considerable upfiow velocities, biofilm accumulation and the subsequent increase in bed expansion appeared to be reduced due to hydraulic shear (Jewell et al., 1981a and Hall and Jovanovic, 1982). In order to avoid the high flow rates required to maintain the particles in a fluid state (greater than 100% increase in bed volume), upfiow velocities in an expanded bed are limited to achieve 10 to 20% bed expansion. An increased recycle rate would cause the digester to diverge from any plug flow characteristics and approach a complete mix system. This would result in increased effluent concentrations (Rittmann, 1982b). The high effluent suspended solids appear to be due to the completely mixed nature of the digester, encouraging the flotation of solids (Bull et al., 1984a). Maintaining a high upfiow velocity via liquid or gas recycle might be costly in terms of net energy efficiency, may require a larger digester volume to achieve an equivalent HRT in order to accommodate bed expansion, or may require a high recycle ratio. All of these aspects are related to the size, shape and density of the biofilm encapsulated particles and to 48 the scale of operation and the characteristics of the wastewater treated. Pumping energy requirements could be too high for the marginal methane yields derived from low strength waste treatment Two factors appear to be responsible for the efficiency and stability of the expanded bed in treating low wastewater concentrations: 1. the development of a dense and concentrated biomass with concentrations as high as 30 kg/m3, and perhaps 100 kg/m3, and 2. the entrapment and filtration of Fme particles reducing effluent suspended solids concentrations. Both of the above factors require low velocities within the expanded bed compared to those required to fluidize heavy particles. Anaerobic processes which operate in an expanded or fluidized bed mode may become feasible for processing very dilute wastes at adverse environmental conditions (Jech and Brautigam, 1983). The anaerobic fluidized bed system provides a compact high intensity treatment option, particularly for the pretreatment of high strength wastes. Disadvantages of the system include: cost of the media, energy cost and process complexity required for expansion of the media, a time-consuming start-up, and the maintenance requirements for the bed expansion system. The expanded bed has shown to be effective for handling both dilute and concentrated waste streams, capable of achieving high organic removal efficiencies at low temperatures (10° C, 20° C) treating low strength wastes (COD less than 600 mg/L) and at high loading rates (up to 8 kg TCOD/m3/day) (Jewell and Morris, 1981 and Switzenbaum and Jewell, 1980). Factors that contribute to the effectiveness of the expanded bed include (Hsu, 1978 and Hickey and Owens, 1981): 1. the support particles allow the settling velocity, and, therefore, the bed porosity to be controlled independently of the substrate concentration, 49 2. the contact between the substrate and fine grained media is considerable, 3. the system accommodates particulates in the feed stream, 4. problems of channelling, plugging, and gas holdup encountered in packed beds are largely avoided, and 5. there exists excellent potential for scale-up. C. IJPFTDW ANAEROBIC SLUDGE BED Anaerobic sludge inherently has superior flocculating and settling characteristics over the more conventional suspended growth systems, provided the physical and chemical conditions for sludge flocculation remain favourable (Lettinga and Vinken, 1980). The UASB achieves internal retention without packing because the bacteria can be stuck to each other to form a bacterial floe Rocs are groupings of microorganisms with a characteristic size many orders of magnitude greater than that of a single microorganism. The sludge bed digester aims at removing some of the problems of the anaerobic filter while retaining its chief virtue: efficient retention of biological solids. The sludge bed digester combines the advantages of the high loading rates and low effluent suspended solids of the anaerobic filter process with the low construction costs and the wide acceptable range of influent suspended solids of the anaerobic contact process (Jech and Brautigam, 1983). High SRTs are achieved independent of the HRT and without the need for any support material. The sludge bed is designed to eUminate the need for costly media, reduce channelling and plugging problems and improve the resistance to changes in pH, temperature and organic and hydraulic loading. The support structures of the fixed film digesters occupy reactor volume. Increasing porosity by decreasing the support structure volume would lead to longer contact times and assist the wastewater conversion within the digester. Such a strategy is subject to the constraint that the fluid velocities do not exceed the bacteria particle settling rates. Hydraulic loading rates and fluid velocities must not exceed the bacteria particle settling rates. 50 Instead of using packing material to support and concentrate biological growth, the upflow sludge blanket operates entirely as a suspended growth system. The key to successful operation of the sludge bed system is to keep this sludge within the system. Because they have a density only slightly different from that of water, very little hydrodynamic drag is required to maintain the floes in suspension (Christensen et al., 1984). Attaining high biomass densities requires floe settling velocities higher than the upflow liquid velocity. This in turn requires large, heavy floes and low liquid velocities. Large dense particles allow high hydraulic loading resulting in increased volumetric productivities. The formation of large floes is in conflict with another requirement, that of high activity per unit volume. The lower activity of the largest particles could be the result of more inactive material in the core 'seeing' less substrate than the bacteria on the perimeter (Tramper et al., 1984). Three parts can be distinguished in the digester (Heertjes and van der Meer, 1978a): the sludge bed, the sludge blanket, and the settler. Sludge granules isolated from the sludge bed were grey-white, high in volatile suspended solids content (about 90%) and 0.5 to 1.0 mm in diameter with a settling velocity of about 0.5 m/minute (Hulshoff Pol et al., 1982 and Lettinga and Vinken, 1980). In digesters operated for several months at high capacity, sludge particles were as large as 3 to 7 mm; these could join together to bridge distances of several centimeters. The characteristics of the sludge blanket differ significantly from those of the sludge bed. Sludge blanket densities were very high containing 8 to 9 wt% suspended solids of which 45% were volatile (Maxham and Wakamiya, 1980). Different species and strains of microorganisms are known to produce very different floe sizes (Andrews, 1982). Lettinga and his co-workers have shown that the microorganisms themselves could function as a filtering medium. In order to facilitate this, the feed inlet system should introduce the influent wastewater as homogeneously as possible over the lower part of the digester (Frostell, 1981). Essential components of the sludge bed system include: 1. a distributor at the base of the column, and 2. optional features which include: provision for liquid and gas recycle, and a settling zone 51 from which to withdraw effluent These features can be seen by referring to Figure 4, a Schematic of the Anaerobic Sludge Bed Digester. Because of the fragile nature of the sludge, mechanical mixing appears to be detrimental to sludge retention within the digester (van der Meer, 1980). Instead of using mechanical mixing, the system is mixed by the upflowing feed and the rising gas bubbles produced. Start-up of the sludge bed is a delicate balance between the washout and retention of sludge which possesses good settling properties. The digester is initially seeded with digester sludge and then fed in the upflow mode. The upflow sludge bed digester itself selects for the best settling sludge particles. This characteristic carries with it both positive and negative implications. While the better settling sludge is retained and a more consistent digester performance is ultimately achieved, the sludge which is lost is that which is buoyed up by its own gas. Hence, the more active gas-producing strains are lost during start-up prior to flocculation. lettinga et al., (1980a) have prescribed the following directions for start-up: 1. A low food to microorganism ratio should be supplied. The initial sludge load should be below 0.1 to 0.2 kg TCOD/kg TS/day. 2. The loading rate of the digester should not be increased unless all of the TVFAs are degraded. 3. The environmental conditions for anaerobic bacteria should be favourable, including (Callander and Barford, 1983, and Lettinga and Vinken, 1980): a. conditions of low turbulence, b. the use of upflow feeding, and a homogeneous and gentle mode of mixing such as occurs with gas evolution while avoiding erosion of pellets, c the presence of certain cations such as calcium ions, and d. the absence of large concentrations of suspended solids. 52 FIGURE 4: Schematic of the Anaerobic Sludge Bed Digester GAS RELEASE <~ WATER RESERVOIR GAS COLLECTOR^ WATER GAS SAMPLE GAS EFFLUENT EFFLUENT SAMPLE EFFLUENT PUMP FEED TANK IGESTER SLUDGE BLANKET SLUDGE BED ALUMINUM FOIL WRAP HEATING PAD POLYSTYRENE FOAM INSULATION ISPERSION PLATE FEED PUMP SAMPLE FEED EFFLUENT TANK 53 Calcium ions exert a positive effect on the flocculation ability of the sludge, presumably mainly by improving the mechanical strength of the floes (Lettinga et al., 1980a, and Hulshoff Pol et al., 1982). The dynamics of the liquid flow and the sludge movements are interdependent and influence the performance of the process. Different densities of the influent and sludge influence the flow pattern of the liquid, so that bypassing of the sludge bed can occur (Heertjes and van der Meer, 1978a). Tracer studies have shown that liquid moves up through the granular sludge bed in a plug flow mode with some bypassing, is well mixed in the floe blanket region, and again moves in a plug flow mode through the settling region (Callander and Barford, 1983, and van der Meer and de Vletter, 1982, and Switzenbaum, 1983). Under these conditions the digester could be described as a slowly moving packed bed. In the bed, the sludge concentration is high and does not vary over a large range of process conditions. In the blanket, the sludge concentration is lower and is more dependent upon the process conditions (van der Meer and Heertjes, 1983). The sludge in the sludge blanket originates from the bed where it is buoyed up by rising gas bubbles. The distribution of the sludge between the bed and blanket depends on: 1. the properties of the sludge such as its settling velocity, particle size distribution, and density, 2. the organic load of the wastewater, 3. the linear upward fluid velocity, 4. the manner of influent distribution, 5. the digester configuration and mode of operation, 6. the suspended solids content of the digester liquid, and 7. the efficiency of gas separation prior to effluent withdrawal (Heertjes and van der Meer, 1978a and van den Berg et al., 1981a). 54 Ultimately, the maximum amount of sludge that can be retained in the digester is dictated by the organic loading rate applied (Lettinga et al., 1983). Increasing the loading will raise the gas production and consequently, the bed expansioa At low loading rates the sludge will also be markedly expanded, mainly as a result of occluded gas. The occluded gas occasionally causes parts of the bed to rise, inducing a pulse-like liberation of the entrapped gas and a local turnover through vertical mixing of the sludge bed (de Zeeuw and Lettinga, 1980). At increased loading and gas production rates, the sludge bed expanded to 80% of the digester volume (de Zeeuw and Lettinga, 1980). The improved agitation brought about by the increased gas production actually improves sludge thickening through reducing bridging of the floes and results in enhanced rates of sedimentation (Switzenbaum, 1983 and van der Meer, 1980). Present application of the sludge bed process has been focused on wastes with a large soluble COD fraction, such as sugar beet and potato processing wastewaters. The sludge bed is not an effective treatment system for those wastewaters with high suspended solids concentrations (Lettinga et al., 1983b). Potential feasibility has been demonstrated for the treatment of low strength industrial wastes. Digesters could operate with waste strengths between 2,000 and 10,000 mg TCOD/1 (Jewell et al., 1981a). Below 2,000 mg/1, the sludge tended to wash out at high hydraulic loading rates; at 10,000 mg/1 and higher, the mixing was insufficient to prevent locally high levels of volatile acids, causing failure, van den Berg and Lentz (1980a) indicated that sludge bed digesters could be operated at reasonably high loading rates (10,000 to 15,000 mg TCOD/l/day) with bean blanching wastes containing 3,000 to 10,000 mg TCOD/1. Settling was slower and more variable at high volatile solids loading rates. The biomass substrate concentration for sludge bed digesters should be diluted if necessary to less than about 0.3% VS (volatile solids) by using effluent recirculation if necessary (van den Berg and Lentz, 1980a). Restricted applicability of sludge bed digesters is due to (van den Berg and Kennedy, 1983, and Atkinson et al., 1980): 55 1. the flocculant strains which are required, 2. the dynamic but delicate balance needed between growth, attrition and elutriation, leading to a constant biomass hold-up, 3. the evolved gas causing floe breakup and flotation, and 4. the restriction to low superficial velocities because of the size and density of floes. Wash-out of discrete floe particles released from the sludge blanket can be nn'nimized by creating a quiescent zone within the digester, enabling the sludge particles to flocculate and settle. A major drawback is the difficulty in accumulating bacteria without retaining other solids. If these solids are not degraded then they will accumulate and cause problems of plugging and loss of active digester volume. Measures have to be taken to prevent an accumulation of suspended substrate solids in the lower part of the digester because solids may otherwise push out and replace the active biomass. Sludge beds are capable attaining of high rates of methane production and high rates of conversion for dilute and mostly soluble waste. Simple in construction and operation, the digester requires no mechanical parts. It does not require the expense and energy consumption of recycle pumps, centrifuges, or support media. The process accomodates fairly well to hydraulic and organic shock loads, temperature fluctuations and low influent pH values. The sludge bed digester has become established commercially for these reasons. CONCLUSIONS FROM THE LITERATURE VII- CONCLUSIONS FROM THE LITERATURE Just what comprises an improved design and mode of operation for anaerobic wastewater treatment can be pieced together from the findings reported in the literature. In order to facilitate reaching increased reaction rates, both the density and activity of the anaerobic bacteria must be increased. The strategies to do this are as follows: 1. employ a number of the following mechanisms for bacterial retention: sedimentation, flocculation, immobilization, filtration, flotation, or solids recycling, 2. employ upflow designs for increased biomass retentioa (this requires efficient removal of the inert and less readily biodegradable suspended solids prior to introduction into the reactor vessel), 3. employ a feeding system at or approaching continuous operation (subject to the constraints of the scale of operation), 4. employ a vessel configuration of at least 10:1 working height to diameter in order to accomodate the requirements for plug flow operation, 5. minimize the extent of mixing, using recycled biogas or effluent recycle on an intermittent basis. If the aspect ratio, feeding frequency and feed inlet system are appropriately selected, mixing will be advantageously used under a small set of circumstances, and 6. employ the following digester designs in this experimental work: a. the upflow anaerobic filter, b. the upflow anaerobic expanded bed, and c. the upflow anaerobic sludge bed based on the assessments of the investigators in this field. 57 MATERIALS AND METHOD vm. MATERIALS AND METHOD The experimental approach was to achieve operation approaching steady state at a set feed dilution and hydraulic loading rate and then to increase the loading stepwise for each digester design. Operating conditions were employed to demonstrate the effects of relatively high and low influent concentrations and long and short detention times. A. EXPERIMENTAL SET-TIP AND OPERATION The digesters were operated in the UBC Bio-Resource Engineering wastewater treatment laboratory. The digester designs examined were the upfiow anaerobic filter, the upfiow anaerobic expanded bed and the upfiow anaerobic sludge bed. Schematics illustrating the key components of each are found in Figures 2, 3 and 4. Each digester system was comprised of the following components: the digester itself, feed and effluent tanks, pumps and timer, and a gas collection system. B. ASPECT RATIO In selecting the actual diameter of the laboratory digesters a compromise was made. An extremely small diameter digester might be dominated by wall effects, while a greater diameter, for the same volume, would decrease the beneficial effects of digester height and an aspect ratio favourable to plug flow. A 10:1 aspect ratio for the working volume was selected in an effort to meet the plug flow criterion. Each digester was fabricated from translucent blue polyvinyl chloride (PVC) pipe, 5.08 cm ID x 94.5 cm (2.0 in x 37.2 in), and capped with PVC pipe fittings. These dimensions targetted for a working volume of one liter (approximately half full). 59 60 C. SUPPORT SURFACES 1. HXPANDFD RFJ) In order to facilitate comparison with the work of Switzenbaum and Jewell (1980), the pioneers of the anaerobic expanded bed research, approximately 250 g of support media was added to each expanded bed column. The media used was 60 Mesh Norton Alundun RR, purchased from Fisher Scientific The support medium was composed of aluminum oxide, a porous water insoluble inorganic biomass support The support medium was estimated to displace a liquid volume of 90 ml. The material was chosen for its uniformity. It does not represent a practical material for full scale operations because of its high cost The particles were not spherical — the irregular shape of the small support particles provided a very large surface area and thus facilitated the development of higher biomass concentrations. The particles had an apparent diameter of approximately 500 X 10 ~6 m a particle density of 2.79 g/cm3 and a bulk density of 0.6 g/cm3 (Switzenbaum and Jewell 1980). Poor initial expanded bed digester performance and considerable particle loss during the plugging problems necessitated particle reseeding. Because the aluminum oxide demonstrated unsatisfactory results, sand — readily available and widely used in the literature — was selected instead for reseeding. In an attempt to accelerate biofilm formation and a consequent decrease in the overall particle density, the sand was heated in the muffle furnace to approximately 550° C and dribbled, while still hot, into a container of the approximately 3% VS dairy cow manure feedstock. Qualitatively, the individual particles seemed lighter and easier to suspend in the liquid than before the induced fouling. The particles were then added to the expanded bed digesters. Approximately 340 g were added to each expanded bed, the upper limit which could remain in the digesters without loss to the effluent at the selected gas recycle level. Again, subsequent plugging problems in the feed 61 lines and minute amounts periodically found in the effluent tanks reduced the total particle loading to less than the 590 g total of the aluminum oxide and sand added to each. 2. ANAEROBIC FITTER Random loose fill Raschig rings were added to the columns to comprise the anaerobic filters. The rings were made of carbon and measured 1.27 cm OD x 1.27 cm x 0.16 cm wall (0.5 in OD x 0.5 in x 1/16 in wall). Each ring had a surface area of 9.42 cm2 (1.46 in2) and contributed approximately 2500 cm2 of surface area as measured by fluid displacement when packed into the digesters. The Raschig rings themselves occuppied approximately 185 ml of the digester volume. 3. SLUDGE BED The sludge bed digesters required no added support material. D. METHOD OF FEEDING A low speed 6-head 'Cole Parmer' peristaltic pump supplied influent to the base of the digesters. Feeding, wasting and gas recirculation for the expanded beds were controlled automatically by a 4-channel table-top 'ChronTrol' digital timer. The timing of the feed and effluent movements was co-ordinated to maximize the time interval between feeding and withdrawal in order to utilize gravity settling of the readily settleable solids. By varying the duration of pumping and the frequency of feeding (in intervals of one hour), a range of HRTs was targetted. The cycle modes of adding, mixing and withdrawing feed are listed in Table 1. The hydraulic retention times in days were calculated by using V*effl, the measured effluent volume corrected by the change in digester liquor height from the preceeding period, calculating an average digester working volume, Vd, and employing the following equation: TABLE 1: Cycle Modes of the Experiments HRT (days) 5 2.5 1.25 Cycle Length (hrs) 2 1 1 Start (hr:rain:sec) 0:00:00 0:00:00 0:00:00 Feed 0:00:00 0:00:30 0:00:00 0:00:30 0:00:00 0:01:00 React 0:00:30 1:58:00 0:00:30 0:58:00 0:01:00 0:58:00 Stop Gas Recycle (Expanded Bed) 1:49:00 1:59:00 0:49:00 0:59:00 0:49:00 0:59:00 Withdraw 1:58:00 1:58:30 0:58:00 0:58:30 0:58:00 0:59:00 Repeat Cycle 0:00:00 0:00:00 0:00:00 63 HRT = time/(V*effl/Vd) 'Time' in the above equation is the time interval for effluent collection. Loading was performed on a intermittent basis. The loading rate and loading frequency was limited by what the peristaltic pumps could deliver. Most variation in the feeding rates likely resulted from entrained gas in the feed lines and the differing resistances that the contents of the various digester designs presented to the feed pumping rates. In order to minimize these variations and still retain the beneficial effects of continuous feeding, a 30 second minimum duration of feeding was targetted. The duration of feeding in the minimum hourly increments (timer limitation) was then calculated. The fluid pumping rate was set at the smallest possible rate in order to minimize hydraulic disturbances in the digesters. The constraint on the bottom end of this rate was based on what seemed replicable from digester to digester. The feeding and wasting cycle was approximately 30 seconds each in duration, initiated every 2 hours for the nominal 5 day HRTs, 30 seconds every hour for the nominal 2.5 day HRTs, and 60 seconds every hour for the nominal 1.25 day HRTs. This feeding rate corresponded to flow rates in the digester of 0.56 and 0.83 ml/second. This timing of wasting, feeding and settling (and gas mixing for the expanded beds) resembled the sequencing operation of the aerobic sequencing batch reactors. This allowed a maximum period of settling to occur in order to better retain suspended bacteria and particulates. A plastic flow distribution plate was placed approximately 5 cm from the bottom of each column. The distribution plate distributed the incoming wastewater (and gas recycle in the expanded beds) over the cross section of the digester base prior to contact with the retained biomass, in its various forms, in the digesters. The aim of such a flow distribution was to minimize the upward jetting and destruction of the sludge beds by the influent and to provide for additional mixing of the feed. A fine stainless steel mesh screen was installed in the base of the expanded beds in order to retain the mobile particulate support media. The expanded bed digester feed was 64 pumped in above the particle retaining screen and distribution plate. This arrangement addressed an encountered difficulty. The previous set-up, with the influent entering the digesters below the retainer screen and plate, allowed the feed to back up the gas recycle lines in preference to entering the digester. This short circuiting presumably arose from biomass and particle accumulation at the base of the digesters and was exacerbated by a high impermeability of the retainer screen. The retainer screen mesh size was the largest possible while still supporting the media particles. The effluent withdrawal port was set at the 1.0 litre level but the digesters were operated at levels approximately 10 cm above this. The working volumes were 1200 ml at this level. This allowed for the level fluctuations which would arise from time to time without removing biogas from the headspace. Removing the digester liquor from below the surface retained a liquid seal between the digester headspace and the effluent peristaltic pump. It also enabled the foam and floating sludge particles to be retained within the digester. Short-circuiting of the feed and effluent was unlikely under this arrangement since the points of entry and exit of the wastestream were at opposite ends of the digester. E MTXTNO AND HFATTNG A peristaltic pump recycled the headspace gases in order to expand the support particles of the expanded beds. The gas recirculation rate was targetted to provide a minimum expansion of the column, and yet avoid washout of the bacteria coated support particles. Gas mixing for the expanded beds was ceased 10 minutes prior to effluent withdrawal in order to more easily retain the support particles by allowing them to partially settle away from the effluent port, and to retain the gas itself. The six digesters were aligned in a row and sandwiched on two sides by two electrical 'Solaray' Model 719 heating blankets. The heating pads were set to provide a consistent digester temperature of 33 ± 2°C. Temperature fluctuations were minimized by encasing the digesters and heaters in 3.8 cm (1.5 in) thick polystyrene sheets, reaching as 65 high as the effluent port of the digesters. The operating temperature was not a manipulated variable in this work. The increase over ambient temperatures was used to more clearly illustrate the differences between the digesters. Feilden (1983) similarly assumed that the effect of temperature applied equally to all digester designs. He reasoned that it is not a significant factor in the choice of the digester design, since, in principle, any digester can be run at any temperature. The entire arrangement was covered by an opaque nylon cloth in order to reduce any effects of sunlight F. F E E D S T O C K The mixed herd of Holstein and Ayreshire dairy cattle housed at the University of British Columbia Dairy Barn provided the manure feedstock for this study. Feed rations for the cattle consisted of four parts alfalfa cube, three parts grain pellet (14% protein), and two parts beet bulbs. No antibiotics were added to the feed. Whole manure less than 12 hours old was manually scraped off the concrete floor holding pens, mixed and diluted 2:1 with tap water, and passed through a solid liquid separator (Prater Industries VS1-13-1H Eccentro Set) equipped with a No. 10 mesh sloped vibrating screen to remove the coarse solids. The liquid filtrate was then trucked to the laboratory and analysed for total and volatile solids. Additional tap water dilutions were used as needed to correct the feed to approximately 3% VS. The filtrate was stored at 4 °C and diluted as necessary just prior to loading the room temperature (average temperature approximately 22° C) feed tanks. Feedstock was sampled and tested just prior to feed tank loading. The feed was loaded into tanks for intermittent withdrawal. The 4 L capacity feed tanks were loaded to a level to provide just enough feed for a 24 hour period if sampling was planned for, or greater to allow for operator time away from the laboratory. Effluent was collected in similar tanks, also at room temperature. Effluent collected for the purposes of chemical analysis was allowed to accumulate in the tanks for a maximum of 24 hours. 66 Tests were performed in order to assess the extent of feedstock degradation over this storage period. Because the feed tanks were not mixed, some settling of the particulates occurred. In order to minimize a possible error in the reported feed characteristics, the feed tanks were washed out one complete hydraulic retention time prior to and immediately after testing at each feed strength/HRT combination. Such a procedure was assumed to address the major shortcomings in this aspect of the experimental set-up. Three anaerobic digester designs plus duplicates were operated from August 1984 to September 1985. At start-up, the digesters were innoculated with effluent from the laboratory fixed film anaerobic digesters treating dairy cow manure at mesophilic temperatures. Due to alterations in the means for feeding and effluent withdrawal, the digesters were emptied and refilled with the same packing and liquor on two separate occassions. Acclimation of the digesters involved the adaptation of the bacteria as well as reworking the plumbing system to result in a workable process. Acclimation occurred in the 7 month period of August to February. Experimental results were recorded in the 7 month period of March to September. After an initial acclimation, the HRTs of the digesters were lowered from the nominal 5 days to 2.5 days to 1.25 day in a stepwise fashion. This procedure was repeated for dairy manure dilutions of 5 times, 2 times and undiluted. Undiluted manure had the following characteristics: 40,000 to 50,000 mg/1 TCOD (total chemical oxygen demand), 5,000 to 10,000 mg/1 SCOD (soluble chemical oxygen demand), 900 to 1300 mg/1 TVFA (total volatile fatty acids) and 30,000 to 35,000 mg/1 VS. Because the digester performances seemed to change over the course of the experimentation period, the feed concentration was stepped down from undiluted, to 2 times to 5 times dilutions as a mirror of the previous runs. 67 G. SAMPLING PROCTRDirRF A N D A N A L Y S T S A steady-state condition was assumed after three complete HRTs had passed through the digesters. Such an assumption is in agreement with a number of investigators who have conducted research in this area (Chakrabarty et al., 1981, Hamoda and van den Berg, 1983, Hickey and Owens, 1981, and Schraa and Jewell, 1983). Sampling at each combination of waste strength and HRT was performed only after this 'steady state' was achieved. The analytical tests (described below) were performed on at least three successive days for each waste strength/HRT combination. Additional testing days were added if considerable problems in controlling the HRT were encountered. Sampling could be performed at intervals not significantly less than one day due to the volumes of effluent required for complete sampling and analysis. This limitation applied to the nominal 5 day HRT and was followed for the larger throughput rates for experimental consistency. In both stepping up and down the HRT and feed strength, the route of process changes was selected in order to ntinimize the change in the organic loading rate. Due to the high frequency of sampling and consequently the demands placed on the laboratory technicians' time and on the drain on chemical reagents, as well as the limited capacity of the analytical equipment, the tests themselves were not performed in duplicate. Since the variability of the results was anticipated to lie more significantly in the biochemical processes than in the chemical testing, two digesters were operated and tested for each digester design. This assumption has proved to be a reasonable one as evidenced by a comparison of the precision of the tests with the results of the experimentation. Laboratory analyses conducted on both influent and effluent samples were: total solids (TS), volatile solids (VS), suspended solids (SS), total and soluble chemical oxygen demand (TCOD and SCOD), acetic, propionic and butyric acids (volatile fatty acids (VFAs)), total Kjeldahl nitrogen (TKN), ammonia nitrogen (NH3) and pH. Analyses for TS, VS, SS and TCOD were carried out according to according to the procedures outlined in The 68 Standard Methods For The Examination of Water and Wastewater (1975). SCOD samples were filtered using 15 cm Reeve Angel Folded Circles, Grade 802 with subsequent analyses according to Standard Methods. Volatile fatty acid samples were analysed on a Hewlett Packard 5890 gas chromatograph. Acetic, propionic, butyric and isobutyric acids were analysed on a 6' x 1/4" i.d. glass column filled with carbopack C/0.3% carbowax 20M/0.1% phosphoric acid. Total Kjeldahl nitrogen and ammonia nitrogen were assayed using a block digester and a Technicon Auto Analyzer II as described in the method of Schumann et al. (1973). pH was read from a 'Good' Digital Model 201 ATC pH meter. TKN, NH3, and suspended solids were measured once during each setting of feed dilution and HRT. Solids, chemical oxygen demands, volatile acids and pH analyses were performed during each sampling day. Headgas was collected in a fluid displacement system consisting of a graduated cylindrical collector connected to a water-filled reservoir. Gas production was read directly from the collector, and corrected to standard temperature (0 C) and pressure (760 mm Hg). Gas samples were collected in gas-tight sampling tubes and were analysed on a Hewlett Packard 5890 gas chromatograph. A stainless steel 6' x 1/8" i.d. column filled with porapack Q-80 mesh at 80 °C was used to determine the gas composition of the headspace. Gas was sampled at each setting of feed dilution and HRT. More frequent analyses were performed when possible during the loading conditions which promoted increased gas production. INTERPRETATION OF RESULTS IX. INTERPRETATION OF RESULTS A. PRECISION OF THE ANALYTICAL TESTS An estimation of the precision of the analytical tests was performed for the feed dilutions. Specifically for these tests, four samples at each dilution were prepared from fresh feedstock for the testing of each parameter reported in this experiment (except for gas composition). Precision estimates of the analytical tests are presented in Table 2. Each feed dilution was mixed thoroughly prior to loading the feed tanks and before each sample was taken. Each test was performed in duplicate or in triplicate depending on the time required to perform each analysis. The standard deviation and the standard deviation calculated as a percentage of the mean concentration are listed in Table 2 for each dilution for the parameters which measure the constituents in the digester liquor. Measurements of the biogas, including volume and methane and carbon dioxide content were not analysed for precision. Generally, the standard deviations for each of the parameters increased with increasing feed concentration. This likely occurred due to: 1. the difficulty in sampling small quantities of the feed laden with suspended particulates, and 2. the increased number of dilutions needed to carry out the chemical tests. The standard deviations quoted as percentages indicate that the relative error associated with each measurement generally decreased with increasing feed concentration. The standard deviations of the TCOD and SCOD analyses may be attributable to the errors in sampling and to the subsequent dilutions, up to 100 times for the most concentrated samples (Liao and Lo, 1985). The mean standard deviation for the TCOD and SCOD analysis of 6.3% compares favourably with that quoted in Standard Methods (1975). The errors in sampling and analysis of VS and SS were small, 1.6% and 2.8% respectively. No dilutions 70 TABLE 2: Precision of the Analytical Measurements Parameter D i l u t i o n Number of Number of Mean Standard Samples Measurements Concentration Deviatioi TCOD 5 4 8 6 270 433 mg/L 2 4 8 17 660 388 1 4 8 58 460 5 371 SCOD 5 4 8 2 430 194 mg/L 2 4 8 4 800 279 1 4 8 7 750 122 VS 5 4 12 4 470 115 mg/L 2 4 12 10 980 104 1 4 12 37 540 437 TVFA 5 4 12 300 32 mg/L 2 4 12 520 36 1 4 12 1 350 46 SS 5 4 12 3 810 121 mg/L 2 4 12 10 290 419 1 4 12 43 330 495 NH 5 4 12 57 0.73 mg/L 2 4 12 161 0.95 1 4 12 350. 2.30 TKN 5 4 12 340 4.5 mg/L 2 4 12 1 350 10.4 1 4 12 2 260 7.5 pH 5 4 8 7.53 0.05 . 2 4 8 7.41 0.05 1 4 8 7.27 0.01 72 were required for these samples and the sample volume itself was considerably larger than it was for the other measurements. The standard deviations associated with the TVFA measurements are likely due to the extreme sensitivity of the gas chromatograph to even minor differences in the the sample injection. The very large numerical values required to calculate the sample concentration from a known standard served to magnify any such error. The error in NH } , TKN and pH was also quite small, perhaps due to the automated nature of the testing procedures. B. HYPRAIJLIC RETFNTTON TIMES OF THE EXPERIMENTAL PFRTOD An evaluation of the digester performance over a range of conditions was carried out by varying the influent feed concentrations and hydraulic retention times. Even with the 7 month sampling period, time limitations and technical difficulties enabled only an approach to stability at each setting. It is accepted by some authors that a steady state is not reached before at least two or three mean hydraulic retention times under the same running conditions (Chakrabarty et al., 1981, Hamoda and van den Berg, 1983, Hickey and Owens, 1981, and Schraa and Jewell, 1983). In fact and in spite of what these investigators might argue, any true steady state conditions could not easily be reached with the drastic changes in the operating variables or with the slowly developing microbial populations. The results presented are a transient response of the digesters to a sequence of targetted HRTs and influent waste concentrations. The HRTs of the experimental period are shown in Figure 5. The hydraulic retention times were calculated on an empty volume basis. (The anaerobic filter and expanded bed digester were filled with 0.185 and 0.200 L of packing respectively. The volume occuppied by the sludge in the sludge bed digester was not measured.) Although this approach does not reflect the actual retention time, it facilitates comparisons to be made with processes such as the sludge bed digester which do not contain inert media (Atkinson, 1974). This is standard practice. 74 The targetted HRTs of the digesters were varied from 5 days to 2.5 days to 1.25 days for dilutions of 5 times, 2 times, and undiluted feed. In an attempt to assess the time effect on the experiments, the sequence was mirrored after the HRT cycles of undiluted feed back to targets of 2.5 and 5 days. In the course of stepping down the feed concentrations, changes in the organic loading rates were minimized by running the 2 times and 5 times dilutions in the reverse order of HRT, from 1.25 to 2.5 to 5 days. The data presented should not be interpreted as a measure of the steady state response of the digester designs but rather as a preliminary screening exercise in deternuning the relative performance between the digesters. It remains uncertain but expected that performance would improve over extended periods of operation at more stable conditions. C. PROBLEMS WITH THE EXPERIMENTAL A P P A R A T U S Several of the operating difficulties encountered were inherent in the design and scale of the experimental set-up. However, these inadequacies confronted each digester design approximately equally. Since it was the digester designs being compared, and each functioned within very comparable systems, the comparisons between the designs were assumed to be demonstrative of the relative performance between the digesters. 1. FEED DEGRADATION AND SFTTT.TNG Because the feed in the feed tanks remained at room temperature for as long as 24 hours prior to introduction into the digesters, it was subject to degradation outside of the digesters themselves, unaccounted for in the feed samples. The same difficulty was presented by the effluent tanks. The effluent samples taken were 24 hour composites, and, as such, were also subject to degradation during the collection period. In order to characterize the extent of feed degradation, additional experiments were performed. Undiluted feed and dilutions of 2 times and 5 times were added to the feed 75 tanks. Over a 24 hour period, feed at room temperature was pumped for 30 second periods at hourly intervals from the feed tanks (corresponding to the 2.5 day HRT) into the room temperature effluent tanks. The set-up was that of the regular experimental apparatus except the digesters themselves were bypassed. Sampling of the feeds, both initially and after the 24 hour composite was collected, was followed by performing the following chemical tests: TCOD, SCOD, VS, SS, TVFA, NH3, TKN and pH. The results of these tests are presented in Table 3. Significant degradation of the feed was generally observed for each of the substrate parameters. While the majority of the tests demonstrated statistically significant differences, only the TVFA degradation was of such magnitude to cause serious distortions in actual substrate addition or removal. This places some uncertainty regarding the value of using TVFA as an informative substrate parameter in this experimental work. Settling effects were expected to arise due to the unmixed nature of the feed tanks. The feed line rested on the bottom of the tank in order to not mimick a supernatant withdrawal process. Solids settling in the feed tanks, reflected in the TCOD, VS and SS measurements, was not of a sufficient magnitude to seriously distort the interpretation of these parameters. Settling of solids in the effluent tanks was of little consequence since the effluent was thoroughly mixed prior to sampling in both the treated and untreated cases. Generally, due to small quantities of active biomass already present in the feed and that which was sloughed off from within the digesters, wastewater treatments were exaggerated (biomass acting in the effluent tanks outside of the recorded HRT) and gas yields were conservative (degradation in feedtanks) compared to what might comprise an 'ideal' system where the wastewater for sampling purposes would not be subject to any changes in the feed or effluent tanks. This 24 hour period of feed degradation and settling was not a direct measure of the lowered feed concentrations that the digesters would receive. The actual digester influent would be better estimated by approximately 50% of this degradation since the feed, on PARAMETER FEED FEED ± S.D. Z REDUCTION SIGNIFICANCE % REDUCTION FOR SIGNIFICANCE DILUTION CONCENTRATION mg/L OVER 24 HRS X C.I. DIGESTER FEED % C.I. rag/L (% INCREASE) OF 24 HR REDUCTION) TCOD 5 X 6 273 593 10.7 95 5.4 no 2 X 17 662 581 8.7 95 4.4 95 none 58 461 2 903 3.5 no 1.7 no SCOD 5 X 2 432 218 0 no 0 no 2 X 4 802 170 0 no 0 no none 7 754 170 18.1 95 9.1 95 VS 5 X • 4 472 165 2.6 95 1.3 95 2 X 10 981 96 2.4 95 1.2 95 none 37 542 321 1.0 95 0.5 no TVFA 5 X 303 24 32.0 95 16.0 95 2 X 517 42 37.5 95 18.8 95 none 1 330 58 36.5 95 18.3 95 SS 5 X 3 814 132 3.8 95 1.9 no 2 X 10 287 695 9.3 95 4.7 90 none 43 329 566 3.8 95 1.9 95 NH3 TKN 5 X 56.9 0.75 8.5 95 4.2 95 2 X 160.7 1.2 8.0 95 4.0 95 none 350.3 1.2 4.0 95 3.7 95 5 X 341 4.5 10.9 90 5.5 no 2 X 1 350 10.4 2.9 no 1.5 no none 2 263 7.5 3.4 no 1.7 no PH 5 X 7.53 0.06 (0.5) no (0.25) no 2 X 7.41 0.07 (0.7) no (0.35) no none 7.27 0.01 2.1 95 1.1 95 77 average, would only be exposed for half of the 24 hour period in the feed tanks and the other half in the digesters themselves. This linear interpolation appears to be the most reasonable assumption in the absence of more detailed information. Degradation of the effluent should not be as significant as that of the feed due to the lowered concentration of nutrients in the effluents although some bacteria would have been present in the effluent especially at elevated loading rates. In both the influent and effluent, significant accumulations of bacteria were unlikely, given the slow growth rates of anaerobic bacteria and their stress from exposure to air. 2. DTOHSTER PIIJGGING PROBLEMS Generally, there are limitations in the use of bench scale data for full scale performance predictions. In this set of experiments, there arose problems in digester operation attributable to the scale of the laboratory set-up. Plugging problems would have been greatly overcome by the installation of larger conduits, since the size of both the suspended particulates and bacteria is independent of the scale of operation. While continuous feeding was considered to be a desireable operating mode by most researchers, the scale of the digesters and the peristaltic pumps' operating range rendered this unattainable. Intermittent operation was a compromise between what was considered ideal and what was possible. Continuous operation or intermittent operation conducted at a greater frequency and/or flow rate would have helped to overcome some of the plugging difficulties. Whenever plugging did occur, it was addressed by recycling the digester contents. The effects of this recycle were evident in lowering the effluent quality, presumably by changing the digesters from an approximation of plug flow to completely mixed. The response to plugging by liquid recycle altered the removal efficiency so abruptly that the tests performed during these periods did not reflect the normal operation. These results were not included for interpretation. 78 D. FF.FD CHARACTERISTICS The results presented are the results of the three digester designs, each with one replicate digester. A summary of the feed characteristics for the range of feed dilutions used in the experimental period is included in Table 4. Mean values of the TCOD, SCOD, VS, and TVFA concentrations are listed with their associated standard deviations. The standard deviations of the substate concentration measurements increased with increased substrate concentration. It should be noted that the dilutions of feedstock were not reflected in the same proportion in these four measures of substrate. The dilutions, instead of the 1:2:5 ratios targetted through tap water dilutions, were instead: TCOD: 1:2.1:6.3, SCOD: 1:1.3:3.6, VS: 1:2.1:6.0, and TVFA: 1:1.8:3.6. The two measurements which included settleable solids — TCOD and VS — were similar in their dilution ratios. SCOD and TVFA, the tests which precluded the suspended particulates, had dilution ratios similar to each other. This implies the effect of solids settling in preparing the feed dilutions themselves as well as in sampling for subsequent chemical analysis. The relationship between the substrate indicators of the feedstock is plotted in Figure 6. Mean values at each HRT and waste dilution of SCOD, TVFA (magnified ten times for illustrative purposes) and VS are plotted against TCOD. While some scatter is evident, the general relationships do appear to be linear. VS most closely followed TCOD. SCOD and TVFA were less sensitive to changes in TCOD. Because the relationships between TCOD and SCOD, TVFA and VS each appear to be linear, the relationship between any two of these variables must also be linear. Oscillations in the levels of intermediates and end-products of anaerobic digestion occur even at 'steady state' operation (Chakrabarty et al., 1981). The performance of anaerobic digestion in treating a consistent volume and feed strength normally varies with the fluctuating biomass and solids accumulations. Such wanderings were exacerbated in these experiments under the experimental regimen of frequent and relatively abrupt changes in both the HRT and influent concentration. TABLE 4: Summary of Feed Characteristics D i l u t i o n TCOD mg/L 5X 6 051 ± 1 2X 18 099 ± 5 IX 37 890 ± 8 SCOD mg/L 576 1 879 ± 941 126 5 348 ± 1 586 878 6 764 ± 2 155 VS mg/L 4 676 ± 1 263 13 518 ± 3 729 27 931 ± 5 092 TVFA mg/L 303 + 135 609 ± 181 1 094 ± 144 50000 40000 30000 E. o O 20000 H I O O O O H o o X A A A A y A 5000 10000 15000 20000 25000 SUBSTRATE CONCENTRATION [mg/L] L e g e n d o S C O P  X JVFA_X10 A vs 30000 81 Fluctuations occur in the microbial systems presumably due to shifts in the microbial population dynamics, inhibition effects and a lack of an optimum microbial environment in the digesters (Chakrabarty et al., 1981). The variation in the results can be explained in terms of: 1. the variability of the feedstock, 2. the difficulty in attaining consistent HRTs, 3. the step changes in both the HRT and the influent concentration, 4. the slowly changing nature of the retained biomass within the digesters, 5. the inadequacy of the individual parameters — TCOD, SCOD, VS and TVFA -in describing the actual substrate 6. the errors in representative sampling of solids-laden liquid wastes, and 7. the errors inherent in carrying out the analytical tests themselves. 1. TCOD AND SCOD The chemical oxygen demand of the feed, described in terms of both total and soluble COD is shown in Figure 7. The TCOD ranged from 3330 mg/1 for the 5 times dilution to 40,800 mg/1 for the most concentrated feed — a 12 fold difference. The feed SCOD varied from 1520 mg/1 to 8120 mg/1, a 5 fold change. The variability of the feedstock itself, in spite of the careful laboratory dilutions, is apparent Lo et al. (1984) ascribed these variations in the screened dairy manure feedstock to the variability in the characteristics of the manure, the duration and conditions of the manure on the floor prior to collection, the approximated pre-dilution and subsequent varying efficiency of liquid solid separation, and the duration of cold storage prior to feeding the digesters. Even within the targetted dilutions in these experiments the variations are considerable. The mean standard deviation over these experiments was 14% for the TCOD tests and 19% for the SCOD tests. On average, the feed TCOD was 4.4 times greater than the SCOD. This points to an important difference between the screened dairy manure treated in this study, and the C H E M I C A L O X Y G E N DEMAND [m g / L ] — NJ U* J*. (JI O O O O O O O O O O O O O O O O O O O O O I 1 1 I I I m o X TJ m Wi ° m o o o Q -< o • o HUB millMIIIISKSKS 11111111111111 k\\ \ \ \ \ \V^\\^^^^^ 11111111 IIIIIIIIIISSSSSSSSS^!^ o O O CD != ><n CD CD m o Q . o o o o psaj am jo aco siqnios pm? pnoi sranou 38 83 substrates used by many other researchers in anaerobic digestion. Here, approximately 80% of the influent TCOD was suspended, whereas other investigators have processed feedstocks of a considerably greater soluble nature. A considerable portion of these suspended solids were the partially digested grasses which were so fine as to pass through the liquid solid separator. 2. TOTAL. VOLATILE AND SUSPENDED SOT IDS Total, volatile and suspended solids are illustrated in Figure 8. The range of influent concentrations as measured by these parameters was considerable. Total solids levels were tested from 3900 to 38000 mg/1, volatile solids from 2890 to 28400 mg/1, and suspended solids from 1870 to 34000 mg/1. These spreads consituted at minimum a 10 fold change. The mean standard deviation for all three solids measurements approximated 22%. The total solids concentrations differed very little from the suspended solids levels. This leads one to conclude that the volatile solids themselves were predominantly in the suspended form. By difference, and of particular significance for assessing available substrate levels, the levels of soluble volatile solids were very small. Again the variability of the feed was noteworthy. A strong correlation between the total COD and the total, volatile and suspended solids was observed through inspection of Figure 6. 3. TOTAL VOLATILE FATTY ACIDS The influent volatile fatty acids, measured in terms of the acetic, propionic and butyric acid constituents, are presented in Figure 9. The influent TVFAs ranged between 140 mg/1 for the most dilute feedstock to 1160 for the most concentrated. The mean standard deviation over this range was 23%. In order to calculate the TVFA levels, the individual fatty acid constituents were factored and summed up as acetic acid equivalents. While the exact ratios between the TVFA constituents varied depending on the conditions of both the feed and the digesters, some general observations can be made. Acetic acid was the o 50000 -i 4 0 0 0 0 -cn 00 Q _ J O 00 3 0 0 0 0 -20000 10000 i 0 1. T 50 I Legend OID TOTAL SOLIDS EZa VOLATILE SOLIDS ^ SUSPENDED SOLIDS 100 EXPERIMENT 150 200 DAY [days] 250 2 o 3 a. s» a. a. o ro m n oo 3 o 1500-1 Q O < >-o > 1000 500 50 100 150 200 EXPERIMENT DAY [days] Legend om TVFA ACETIC ACID PROPIONIC ACID S BUTYRIC ACID 250 < o ST ct. > o 51 n n n oo 86 predominant form and on average comprised 65% of the total volatile fatty acids. Propionic acid levels followed the fluctuating acetic acid levels but at lower concentrations, on average 24% of the TVFA concentrations. The butyric acid concentrations quoted are the sum of butyric and iso-butyric acids. They represent 11% of the total on average. 4. TOTAT. KJELDAHL AND AMMONIA NTTROGFN The measured total Kjeldahl and ammonia nitrogen levels of the influent are shown in Figure 10. TKN, the total Kjeldahl nitrogen, is the sum of ammonia nitrogen and organic nitrogen. Both parameters followed the general trend demonstrated by the previous substrate indicators. By referring to the preceeding Figures, it can be seen that the ammonia nitrogen fraction, on average 27% (± 7%), varied with the feed concentration. At all levels, most of the nitrogen present in the feed was organic nitrogen. The TKN of the feed and digester effluents is shown in Figure 11. For each of the digesters, TKN concentrations of the effluents followed the trend of the influent, and within experimental error were conserved over the digestion process. Hashimoto and Chen (1979) and Lo et al. (1984) reported that all of the nitrogen contained in the feed was recovered in the effluent This is consistent with what was observed, allowing for the expected nitrogen removal through ammonia volatilization which occurred both in the feed and effluent tanks. The ammonia component of the total Kjeldahl nitrogen typically exhibited a general tendency to increase in the effluent The anaerobic filter and expanded bed showed mean ammonia concentration increases of 12% ± 3% and the sludge bed 22% ±6%. Ammonia is produced by the deamination of organic nitrogen-containing compounds, by the hydrolysis of urea, and by the reduction of nitrate anaerobically (Fannin et al., 1983). The ammonia content of the digester effluent is influenced to some extent by the pH. At pH above neutral, the ammonia volatilizes quite readily and the ammonia concentration in the digester liquor decreases. This appears to be a factor in the relatively low ammonia concentrations found in these experiments shown in Figure 12. Both the elevated pH values 1000 800 cn UJ O O 600 400 200-Legend DTfJJ TKN NH3 N 50 100 150 200 EXPERIMENT DAY [days] 250 FIGURE 12: NH 3-N of the Feed and Digester Effluents U. UJ TJ o 0 3 ° ^ CD H 00 & ° Q o 5 u, Oct z o <U H < a Q ^ -< ™* *-J U J J l D u Z X -I Ul < u in L D i l l o -m CN r o o IT) ~T" O O o o o o CM o o [-|/6iu] N300dl lN V INOIW 90 of the digester effluent and the opportunity afforded by the open feed and effluent tanks for ammonia volatilization might aid in explaining the ammonia levels recorded here. 5. £ J J The pH measurements taken over the experimental period are illustrated in Figure 13. The influent pH hovered around neutrality. It generally decreased as feed concentration increased, in keeping with the rise in TVFA. According to Li and Sutton (1983), and Anderson et al. (1982b), anaerobic digestion has a pH optimum of 6.7 to 7.4, and pH values below 6 or above 8 are restrictive. The range of permissible pH values quoted are generally based on observations of completely mixed digesters, those most commonly investigated. Differences in the hydraulic flow patterns — plug flow versus completely mixed—. would be reflected in different pH ranges of the effluent. In a plug flow reactor, the pH initially would decrease due to hydrolysis and TVFA formation, and subsequently increase in the direction of flow due to TVFA utilization by the methane formers (DeWalle and Chian, 1976). The pH of each of the digester effluents was higher than that of the influent, mean increases of 0.78, 0.94 and 0.77 for the sludge bed, expanded bed, and anaerobic filter respectively up the pH scale. Influent values ranged from 6.6 to 7.5. A comparison of Figure 13 with Figure 5 reveals that the effluent pH increased with longer HRTs. 6. SUSPENDED SOLIDS Suspended solids in the waste stream present various problems: 1. in materials handling, causing plugging and rendering the digester inoperable, 2. by significantly increasing the digester size in order to accomodate the extra volume required in the digester vessel, or 3. by sacrificing gas production and wastewater treatment efficacy by decreasing the effective HRT. 91 FIGURE 13: pH of the Feed and Digester Effluents 3 Q U . UJ "£ y <? e c 3 2 <° O Q ui COg z o _l < a. 3 Z X - I u 4 u vi bi D i l i o cs r o r o CD 00 92 More important than an actual removal rate of suspended solids is the capacity of the digester to perform uninterrupted due to plugging. Apparent removal rates could represent solids removal by settling or entrapment. Suspended solids of the feed and digester effluents is plotted on Figure 14. In all cases, the typical volume fraction of suspended solids in the feed was small, presenting no significant distortion of the measured HRT. Solids accumulation presented particular problems in the anaerobic filter especially. The extent of solids accumulation was not assessed. Effluent SS levels were lower than those of the feed, removals of 47% ± 11%, 53% ±10%, and 50% ±9% for the sludge bed, expanded bed and anaerobic filter respectively. This suspended solids removal is to be expected when considering the potential removal mechanisms at play. While biodegradation of suspended solids at these relatively short HRTs is likely to be insignificant, forces of filtration, flotation and sedimentation are bound to be important factors. The SS levels of the digesters appear to be similar except at the higher loading levels where the sludge bed effluent SS concentrations were higher. Here, reduced sedimentation and filtration due to increased channelling, sludge flotation from biogas generation at these increased loadings, and subsequent sludge and feed solids wasting in the effluent are the probable explanations (Bull et al., 1983). E. S U B S T R A T E R E M O V A L Curves describing the substrate removal efficiency and the methane production rate for each digester design are grouped according to their targetted HRTs. The standard deviations associated with the averaging of the HRTs are indicated with each plot Simple regression lines were drawn in an attempt to illustrate the differences between the digesters' performances. The high variability of the results precludes any rigid statistical interpretation of the regression lines themselves. The plotted values do not take into account the degradative changes in the feedstock prior to actual loading into the digester. While degradation tests were performed for each of 93 FIGURE 14: Suspended Solids of the Feed and Digester Effluents =1 o u. ui "O o m g O l a : y o <U H 3 Q Q " — ui Ul 1 2 < 0. 3 X J u iU 1/1 u . [n/6ai] sanos cnaN3dsns 94 the parameters and over the range of feed dilutions, the intent of the degradation tests was to indicate the approximate magnitude of feed alteration rather than developing scaling factors with which to 'correct' the samples. 1. TCOD REMOVAL . The TCOD removal plotted against the TCOD addition is found in Figures 15 to 17 for the three digester designs. The TCOD removal, the initial TCOD level less the final measurement, appears to increase with increasing TCOD addition for each of the digesters and for the range of HRTs. This generally linear TCOD removal is in agreement with the fmdings of Frostell (1981) for both the sludge bed and anaerobic filter, and with van den Berg and Kennedy (1981b) for the fixed film reactor. The smaller volumetric processing rates, expressed as larger HRTs, exhibited the highest TCOD removal efficiencies, 47%, 45% and 53% removal efficiencies for the anaerobic filter, expanded bed and sludge bed respectively for the target 5 day HRT. Each of the digesters showed decreased removal efficiencies at increased HRTs. This was likely due to the retained microbial populations within the digester approaching a condition of saturation with respect to the substrate. Figure 18 illustrates the effect that HRT exerted on the TCOD removal efficiency for each of the digester designs. The sludge bed demonstrated the greatest decline in removal efficiency with decreasing HRT. This variability in performance was characteristic of the sludge bed digesters. The expanded bed digester was the most stable over the range of HRTs, but also demostrated the lowest removal efficiencies. The removals at the intermediate and lower HRTs, although generally lower than for the highest HRTs, were not in direct proportion to the decline in HRT. For the nominal 2.5 and 1.25 day HRTs, the removal efficiencies were: 51% and 35% for the anaerobic filter, 38% and 28% for the expanded bed, and 45% and 33% for the sludge bed. FIGURE 15: TCOD Removal vs TCOD Addition - - Anaerobic Filter [1/6LU] 1VAOH3CJ QOOl Q 25000 CD O CH Q o O 20000H 15000 H 10000 5000 H o o X X o o o O K ft o y O X V x x o o OA A A X 10000 20000 30000 40000 TCOD ADDITION [mg/L] Legend • UUN WU s <J ll D/>rS x uc *H HPT ^_J«_«/j 0^4 DMS * SSftS.tSJ.S.'.-A.t/r..?:?.?.1!!?. 50000 8 o 3 o > & c-. § $ CD CX td ro 97 FIGURE 17: TCOD Removal vs TCOD Addition ~ Sludge Bed 8.2 2, q , a l - I o o r ° fs [i/6ui] ivAowsa aooi FIGURE 18: Effect of HRT on TCOD Removal Efficiency y T J § C ce CU " CQ \ X l [%] A3N3I3LLG lVA0M3d QO01 99 Deviations from linearity increased with decreasing HRT and with increasing TCOD addition, as seen in Figure 17. Deviations from linearity were slightly larger for the expanded bed compared to the anaerobic filter and the sludge bed, with a standard deviation of approximately 11% for each of the HRT settings. The size of the data spread increased with increasing substrate concentration. This is perhaps reflective of the mixing regimes of the digesters. The expanded bed can be assumed to be well mixed independent of the hydraulic loading rate or feed concentration. The sludge bed however was essentially unmixed at low influent concentrations, but as the feed levels increased, biogas production correspondingly improved and likely induced some dispersion of the digester's sludge. The variability of TCOD removal for the anaerobic filter can be explained in terms of alternate filtration and sloughing of the biomass. At the 1.3 day HRT, periodic alleviation of column plugging required that the liquid contents of the anaerobic filter be recycled. This led to the condition of a completely mixed digester as opposed to the plug flow regime under more ideal circumstances. Any reasonable conjecture beyond the investigated substrate addition levels would be difficult, given the variability in the data. 2. SCOD REMOVAL Substrate removal, the final SCOD level less the initial concentration, measured as soluble chemical oxygen demand, is plotted against SCOD addition in Figures 19 to 21 for each of the three digester designs. SCOD removal increased with increased SCOD addition in an generally linear fashion at each of the three HRTs. As with TCOD removal, the SCOD removal was affected by the HRT, although not in direct proportion to the HRT. The SCOD removal efficiency was highest for the sludge bed digester, 85% for the 3.8 day HRT and 62% for each of the 2.2 and 1.2 day HRTs. No overall difference can be concluded for the SCOD removal efficiencies for the expanded bed and anaerobic filter, FIGURE 19: SCOD Removal vs SCOD Addition — Anaerobic Filter •fl/6ui] IVAOWBcl Q03S 10000 - i 8000 CD 6000 H __l o CY. 4000 H O o o 00 2000 H A A X X --x O X X X X o o o o o o X A A X A —I—: 2000 4000 6000 SCOD ADDITION [mg/L] i 8000 A A Legend e UUH HRT ; X UtAN^ HRt : A MCAN HRT > \}.A/7-1.1 QMS 0J M r s 0.2 DAIS 1 10000 C u C u W ro 102 FIGURE 21: SCOD Removal vs SCOD Addition — Sludge Bed 103 both averaging 57% over the range of HRTs tested. The SCOD removal efficiency decreased with decreasing HRT, in agreement with Dewalle and Chian (1976) and as would be expected with an increased food to microorganism ratio, but this decline in removal efficiency would appear to be of little consequence when considering the four-fold difference in HRT. 3. TVFA RF.MOVAL Total volatile acids removal versus TVFA addition is illustrated for each of the digester designs in Figures 22 to 24. TVFA removal increased with increasing TVFA addition, indicating a high and almost constant removal efficiency. The anaerobic filter and sludge bed demostrated superior performance compared to the expanded bed digester. The anaerobic filter degraded 97%, 86% and 68% of the influent TVFAs and the sludge bed removed 90%, 94% and 77% for the target 5, 2.5 and 1.25 day HRTs respectively. The expanded bed digesters were visibly less efficient. They removed 81%, 78% and 43% at these target HRTs. Little difference is visible between the two longest processing rates. The digesters responded to the smallest HRT with both a marked variability and decreased efficiency in TVFA removal, in contrast to the near linearity of the higher HRTs. This variability increased with increased TVFA addition. The performance of the expanded bed digester at the 1.3 day mean HRT was characterized by its extreme instability. According to Asinari Di San Marzano et al., (1981), more important than any given TVFA concentration is its rate of change. A sharp rise in the TVFA content indicates that something has happened to either retard the activity of methane forming bacteria or to stimulate the activity of the acid formers. This inability of the expanded bed digester at the 1.3 day HRT to efficiently degrade the TVFAs was perhaps due to digester overloading. Since each of the digester designs were loaded at approximately the same rate, a poorly developed microbial population retained in the expanded bed is indicated. 104 FIGURE 22: TVFA Removal vs TVFA Addition — Anaerobic Filter ci s ! =: « a [1/6LU] nvA0^N3d VJA1 105 F I G U R E 23: TVFA Removal vs TVFA Addition - - Expanded Bed Si 2 2 [1/601] 1VA0l*Gy V J A l FIGURE 24: TVFA Removal vs TVFA Addition - - Sludge Bed [1/6LU] lVAOH3cd VJA1 107 4. VS REMOVAL Volatile solids removal is plotted against volatile solids addition in Figures 25 to 27 for the three digester designs. A general trend of increased VS removal with increased VS addition was observed, although obscurred by the extreme variability in removal at all but the lowest hydraulic throughput rates. Increased VS addition led to a generally increased VS removal, but was marked by magnified fluctuations in the removal. A least squares fit of the data gave the following removal efficiencies: 55%, 49% and 46% for the anaerobic filter, 46%, 43% and 36% for the expanded bed, and 45%, 53% and 27% for the sludge bed at the targetted 5, 2.5 and 1.25 day HRTs. When the hydraulic loading was increased and influent concentration was maintained, the percent removal decreased and became more variable, suggesting that the percent removal was determined by the HRT (DeWalle and Chian, 1976). The anaerobic filter demonstrated the greatest stability in VS removal. This was likely due to the dual mechanisms of VS elimination from the effluent through both anaerobic degradation and physical filtration of the influent The expanded bed and sludge bed VS removals showed greater variability compared with the anaerobic filter. The sludge bed digester variability at increased organic loading rates was in reasonable agreement with ForsteT et al., (1982) and Heertjes and van der Meer (1978). The performance curves for the 2.4 and 1.3 day mean HRT are nearly parallel, with the longer HRT operation demonstrating a marginally superior level of performance. A summary of substrate removal efficiency for the four parameters is found in Table 5. A general interpretation of the digester performances for the various measures of substrate removal was facilitated by calculating a 'mean composite substrate removal efficiency'. This is the average of the substrate removal efficiencies measured in terms of TCOD, SCOD, VS and TVFA. From Table 5, it can be seen that both the anaerobic filter and the sludge bed demonstrated superior overall removal efficiencies, 67% ±18%, 58% ±14%, 51% ±15%, and 68% ±15%, 64% ±18% and 50% ±14% respectively for the target 5, 2.5 and 1.25 day FIGURE 25: VS Removal vs VS Addition — Anaerobic Filter [-|/6iu] IVAOHBcd SA 109 F I G U R E 26: VS Removal vs VS Addition — Expanded Bed — o V 51 i fc z j x l ?3I ii o o rO m [n /6o i ] nvAOwaa SA FIGURE 27: VS Removal vs VS Addition — Sludge Bed [-|/6LU] 1VAOM3CJ SA I l l HRTs. The expanded bed degraded substrate with reduced efficacy, having a mean substrate removal efficiency of 60% ± 19%, 53% ± 16% and 39% ± 13% for the target 5, 2.5 and 1.25 day HRTs. The high variabilities associated with each removal efficiency precludes definitive statements to be made concerning the statistically significant difference in performance of these digesters. Interpretation is difficult for a number of reasons: 1. the mean HRTs at the three target levels differ for the different digester designs, 2. the HRTs and the feed substrate concentration span a wide range over the experimental regimen, the HRTs varying four fold, and the organic loading rates ranging 25 fold for TCOD and VS, and 14 fold for SCOD and TVFA, and 3. the 'steady state' digester performance fluctuated at all HRT and feed level settings. F. BIOGAS PRODUCTION The average biogas production of the digester designs at each HRT and waste strength over the experimental period is depicted on the histogram in Figure 28. The biogas includes methane, carbon dioxide and the impurities present in the gas above the digester. Comparison with the levels of substrate concentration over the experimental period reveals that biogas production increased with decreasing HRT (see Figure 5) and with increasing feed concentration (see Figures 7 to 10). Because the histogram is not symmetrical over the roughly mirrored experimental regimen, some time effect is apparent Gradual biomass accumulation over the experimental period is the probable explanation for the increased biogas production levels at comparable loading rates with the prolonged digester operation. The expanded bed gave a stable but very low gas production over the entire experimental period. The anaerobic filter performance was markedly superior to that of the expanded bed. The sludge bed produced the greatest gas quantities but also demonstrated the greatest spread. DIGESTER HRT (DAYS) MEAN SUBSTRATE REMOVAL EFFICIENCY (%) MEAN COMPOSITE SUBSTRATE H TARGET ACTUAL TCOD SCOD VS TVFA REMOVAL EFFICIENCY (%) > ± S.D. ± S.D. ± S.D. ± S.D. ± S.D. £ Anaerobic 5 4.0 ± 1.0 47 ± 14 68 ± 19 55 ± 16 97 ± 25 67 ± 18 g» Filter g 3 Expanded 5 4.3 ± 1.1 45 ± 15 69 ± 23 46 ± 15 81 ± 23 60 ± 19 •§ Bed o Sludge 5 3.8 ± 0.7 53 ± 9 85 ± 13 45 ± 13 90 ±26 68 ± 15 Bed Sludge 2.5 2.2 ± 0.4 45 ± 12 62 ±22 53 ± 20 94 ±18 64 ± 18 Bed Anaerobic 1.25 1.3 ± 0.2 35 ±11 54 ± 15 46 ±18 68 ± 18 51 ± 15 Filter Expanded 1.25 1.3 ± 0.2 28 ± 9 48 ± 13 36 ± 15 43 ±15 39 ± 13 Bed Sludge 1.25 1.2 ± 0.2 32 ±10 62 ± 16 27 ±12 77 ± 18 50 ± 14 Bed o o c CT VI Pf Anaerobic 2.5 2.3 ± 0.7 51 ± 9 45 ± 17 49 ± 14 86 ± 15 58 ± 14 5*J CD 3 o < P3 Filter Expanded 2.5 2.4 ± 0.7 38 ±12 53 ± 19 43 ±13 78 ± 20 53 ± 16 Bed gj D. CD O o T J o I — CJ z> Q O CC Q_ CO < o o CO 1.5-1 H 0.5 H " T 0 50 100 150 200 EXPERIMENT DAY [days] Legend CD ANAEROBIC FILTER B EXPANDED BED E3 SLUDGE BED 250 114 G. BIOGAS CQMPQSmQN The mean methane and carbon dioxide concentrations in the biogas are illustrated on Figure 29. Gas composition is given as percent by volume on a water free basis. Since the experimental regime was symmetrical over time in stepping up and stepping down the feed concentrations, the trends of increasing methane concentration in the biogas over the experimental period support the conjecture of an increasing biomass concentration and/or activity. This is likely due to a gradual accumulation of the slow growing methane formers within the retained biomass. It would appear that the methane concentration was more strongly affected by the feed concentration than was the carbon dioxide content This broad observation is made by comparing Figure 26 with the previous Figures depicting substrate loading — Figures 7 to 10. The anaerobic filter and sludge bed digesters produced a consistent gas quality over the experimental period, in slight excess of 60% at all but the lowest loadings. The mean methane concentration of the expanded beds ranged from below 40% to 50%. Carbon dioxide concentrations were generally stable and below 20%. The ratio of methane to carbon dioxide ranged from approximately 2:1 for the expanded beds, to in excess of 3:1 for both the sludge beds and anaerobic filters. Gases other than methane and carbon dioxide were always present in the biogas samples analysed. Such findings, often obscurred in the literature by reporting gas compositions only as a methane/carbon dioxide split, have been noted by some investigators (lettinga et al. , 1983b). In addition to carbon dioxide, other contaminants commonly detected in the biogas are: nitrogen, ammonia, oxygen, water vapour, hydrogen, hydrogen sulphide, a range of mercaptans, and volatilized organic acids (Lettinga et al., 1983b). Apparently, their biogas always contained nitrogen, averaging approximately 30%. According to Lettinga et al. (1983b), the nitrogen in the biogas originated from the dissolved nitrogen in the influent solution; it is stripped from the liquid phase by the methane gas produced. Since the solubility of methane on a volume basis is about twice as high as 100 -i o oo O Q_ O o 00 < o o CQ 80-60-40-\ x— x 20- X— ; / . / a J-*a • r a a—— 50 100 150 200 EXPERIMENT DAY [days] 250 Legend O CH4 ANAEROBIC FILTER X CH4_EXPANDEQ_BED_ A CH4 SLUDGE BED m CQ2 ANAEROBIC FILTER 8 C02 EXj>AN0ED BED a C02 SLUDGE BED 116 nitrogen, the biogas is enriched with nitrogen not in proportion to what is produced in the digester (Lettinga et al., 1983b). Difficulties in flushing the air from the sampling tubes likely accounted for a considerable fraction of the nitrogen impurities in the gas in these experiments, especially at low gas production rates where the biogas was scarcely sufficient to flush out the air present in the sample tubes prior to sampling. Anaerobic methods are always subject to error from oxygen contamination, but the high oxygen consuming capacity of most wastewaters should prevent significant oxygen accumulation in the headgas. It is unlikely that any significant quantities of air entered through leaks since the system always operated at pressures slightly greater than atmospheric due to the fluid displacement method of gas collection. Very small volumes of air could have entered the digesters through the feed lines, however, as a result of the prior disconnecting of tubing in response to the periodic plugging problems. H. 'METHANE PRODUCTION Methane production is plotted against TCOD addition in Figures 30, 32 and 34, and against TCOD removal in Figures 31, 33 and 35 for the three digester designs. While a general methane production increase with both TCOD addition and removal is evident considerable fluctuations reduce any confidence in accurately predicting methane production for any of the three digesters from the performance data presented. Methane production generally increased with increased HRTs and influent waste concentrations for each of the digesters. This finding is consistent with the observations of Lo et al. (1984). According to a number of researchers, the rate of gas production is a straight function of TCOD loading rate and TCOD conversion (van den Berg and Kennedy, 1983, Singh et al., 1983, and Hudson et al., 1978). Other investigators seem to concur but in less definite terms, stating that methane production could be expected to vary proportionately with substrate removal or substrate addition below some maximum loading value (Frostell, FIGURE 30: CH 4 Production vs TCOD Addition — Anaerobic Filter 7 ? « 5 1 3 1 X X \ \ \ \ o c\ \ \ \ oo o \ \ \ \ \ \XfX< o o • o m O o • o o Tt-o o • o IT) K ) O i O O O CD E, o o o cs C J o != CM o Q ° O ° CJ o o o o V o o o m C N CO d d d CN d [XDP 1/T/HO I] NOIlOnaOdd 3NVH13W FIGURE 31: CrL. Production vs TCOD Removal ~ Anaerobic Filter [ X D P n / t / H O i] Nononaodd BNVHJJW FIGURE 32: CH* Production vs TCOD Addition — Expanded Bed 119 II 3 1 2! =1 «" 5| ; i ^ 'I ~, "g II I1 ii « *i * i *: S>J 1, 2: J> at 31 a: • x < T CN 00 o o o m O o ho o -a-o o • o m t o o o o 2 CD z & o o 2: CN O Lr 5 CN oo o o CN O O O m Q Q < Q O O o o o o o o o m [ X D P IAHO I] Nouonaodd 3 N V K G W FIGURE 33: CH 4 Production vs TCOD Removal ~ Expanded Bed g z i z o al 3 a: 0 X 4 \ \ Ls . E . x I \ x\ 4 O o r- O IT) CN O O h o o CM O ID O O O O O Ld at Q O o o o o IT) CM CO d UD d T f d CM d [Xop i / ^ H 3 1] NOIlOnaOdd 3NVH13W 121 FIGURE 34: CH* Production vs TCOD Addition -- Sludge Bed A XI X • al a [/Dp I/^HO l] NOIlOnaOcJd 3NVH13rN 1.2n O TJ H X o 0.8 H X X g r— o Q O Cr: Q_ Ld < 0.6 H 0.4 H 0.2 H x x -w o A / O A y 0 0 A X A x 4r< * x ° A X n< <9s T • 1 1 1 5000 10000 15000 20000 TCOD REMOVAL [mg/L] Legend x UC»H HRt »1.J_»A 04 o»rs 4 H.? itf.H?.'..*..W.!/.". .9:?.P.*!?. 25000 123 1981, and Hill, 1982). Such a performance was not displayed in these experiments however. Methane production appeared to be more strongly related to TCOD addition than to TCOD removal for the anaerobic filter and sludge bed digesters. No strong correlation of methane production with TCOD addition or removal for the expanded bed is obvious. Of the three digesters, the anaerobic filter delivered the most stable methane productivities for each HRT. The sludge bed performance was higher on average than the anaerobic filter but was also highly variable. The variability in methane productivity was extreme in the expanded bed digester. This was exacerbated by low biogas productivity precluding frequent or regular sampling intervals. Inspection of similar graphs for the soluble substrate indicators revealed no relationship between the SCOD or TVFA parameters, either in terms of substrate addition or removal, and methane production. Any gas production increase with increased loadings or removals as might be expected was masked by the large fluctuations in performance. The lack of a correlation might be explained in the following terms. Generally, elevated substrate additions resulted in increased gas production. Increased microbial activity was likely not confined to the methanogens alone, however. A balanced microbial activity would require hydrolytic and acetogenic activity as well, so while soluble substrate was consumed it was also produced. The usefulness of these transitory soluble intermediates as indicators of substrate removal is therefore quite limited. Measured against increasing volatile solids addition or removal, methane production generally increased. These performances are plotted in Figures 36 to 41. The anaerobic filter performance appeared to be the most stable of the digesters. The expanded bed demonstrated fluctuations large enough to obscure any trend. The sludge bed demonstrated its characteristic variability. Methane yield expressed in terms of L CH4/g VS added gives an indication of biodegradation efficiency. Theoretically, a reduction of one gram TCOD is equivalent to the 124 FIGURE 36: CH, Production vs VS Addition — Anaerobic Filter V 11 5 5 Z l X o> 2 2, « 31 Si X x X \ \ \ < \ oo o \ \ X, \ \ \ \ \ \ \ \ ON o o r O o N-o o o m ro o o o o ro o o o cn E, tno o o o CN Q Q < o o o !£2 CO o o o o o o o m CN co d " T -O* d CN d [/Dp I/^HQ -]] NOIlOnaOdd 3NVH13W 125 FIGURE 37: CH 4 Production vs VS Removal — Anaerobic Filter V Si « al V \ X \ \ V \ o o \ \ \ <I A o o r o IT) cs o o h o o CN r CN CO d d d ~T~ CN d o _ o CO o c O U J ct: oo > o o • o o o o o m [XDP l / ^ H O l] NOIlOnaOdd 3 N V H 1 3 W FIGURE 38: CH* Production vs VS Addition ~ Expanded Bed Si i! ol o tl 6' 6 <• i • ' " * l •*»: ••i ft. " i tt' •: •-•o c si S1 5: V D> 3 <: w a> 51 at 3: —I o X 4 X <3 T-o o r CM 00 d ID d [ X D P H/^HO i] Nouonaodd 3NVHJJH FIGURE 39: CH* Production vs VS Removal — Expanded Bed ¥ si s m I I X » al a X < < X CM CO d d d [Xpp I/^HO l] NOIlOnaOcJd 3NVH13W HGURE 40: CKL Production vs VS Addition — Sludge Bed s| I i = : d l o i » at a a: e x « r CN <1 X < X \ <3 < "... \ X X X , \ o \ i \ \ N \ \ No- o X \ N CO o o • o o T* o o • o m ro O O O o O O m r-o o ^ o O O 7— C M Q o Q § < in 00 > o o o o o o o m CM [Aop I /VHQ N] NOIlOnOOdd 3NVH13W 129 FIGURE 41: CH, Production vs VS Removal — Sludge Bed 11 1 £' «•: si Si ° 6-• ' ni I1 ij © 51 • < <\ 31 3: x < o r o U J CN! o - o v—1 CN x x < 15000 [mg/L \ \\ o o < 0 X x V < , \ \ X x X x / x y < .* • X o ~ CO > < • < \ o \ _ o o < < If) * * § r w :< * » » x \ I 1 1 CN — CO " d [ X D P HAHO l] > i i tf> • * CN d d o Nononaoyd 3N.VH.L3W r ° o 130 production of 0.35 litres CH4 (STP) (McCarty, 1974), although values quoted in the literature range from 0.09 to 0.4 (Kleinstreuer and Poweigha, 1982). van den Berg and Kennedy (1981b) indicated that most measured methane generation rates were approximately one third of the theoretical value at best A summary of the mean methane yield for the digesters at their mean HRTs is presented in Table 6. Lo et al. (1985) reported a comparable methane yields for screened dairy manure of 3% VS using a fixed film digester. I. SUMMARY The large standard deviations associated with both the substrate removal efficiency and the methane yield make any statistical interpretation of the results quite difficult although statements about apparent general trends can be put forward. Generally, increased influent waste strength led to reduced substrate removal efficiencies and increased fluctuations in overall performance. At the nominal 5 day HRT (3.8 ±0.7 days), the sludge bed digester demonstrated the highest and most stable removals in terms of TCOD and SCOD — 53% ± 9% and 85% ±13% respectively. The anaerobic filter at this HRT (4.0 ±1.0 days) was marginally superior to the sludge bed in terms of VS and TVFA removal, 55% ±16% and 97% ±25% respectively compared to 45% ±13% and 90% ±26% for the sludge bed. The higher removal efficiencies of soluble substrate (compared to the TCOD and VS removals which contain both soluble and suspended substrate) indicate that even at these relatively long HRTs, the suspended component of the feed was not significantly degraded. The calculated mean composite removal efficiency indicates that there was no discernable difference between the anaerobic filter and the sludge bed. For all three digesters, the associated variability of the removal efficiencies leads one to conclude that statistically, there was no significant difference between the digester designs. The methane yield of 0.095 L CH«/g VS added of the sludge bed digester was an order of magnitude superior to the expanded bed and considerably higher than that of the 131 TABLE 6: Summary of Methane Productivity Digester Feed Influent TVFA HRT (Days) Methane Production Type D i l u t i o n mg/1 ± S.D. Target Actual ± S.D. 1 CH4/1 day Anaerobic 5 X 303 + 135 5 4.1 + 0.4 .051 F i l t e r (±44%) 2.5 2.2 + 0.7 .202 1.25 1.2 ± 0.2 .198 ± .06 Expanded 5 4.8 + 0.9 Bed 2.5 2.3 + 0.4 .016 1.25 1.3 + 0.1 .038 ± .02 Sludge 5 4.1 + 0.5 .022 ± .00-Bed 2.5 2.1 0.3 .180 1.25 1.3 + 0.2 .270 ±. .12 Anaerobic 2 X 609 + 181 5 4.5 + 0.6 .178 ± .06 F i l t e r (±29%) 2.5 2.5 + 0.6 .191 ± .05 1.25 1.2 + 0.3 .288 ± .15 Expanded 5 4.0 + 1.2 .030 ± .02 Bed 2.5 2.4 + 0.4 .060 ± .04 1.25 1.2 + 0.2 .150 ± .06 Sludge 5 3.8 + 0.8 .541 ± .12 Bed 2.5 2.2 + 0.4 .296 ± .12 1.25 1.3 + 0.2 .400 ± .16 Anaerobic None 1 094 + 144 5 4.1 1.3 .334 ± .16 F i l t e r (±13%) 2.5 1.9 + 0.7 .849 ± .08 1.25 1.2 + 0.1 .508 ± .09 Expanded 5 4.1 + 1.1 .042 ± .01 Bed 2.5 2.5 + 0.3 .195 ± .07 1.25 1.5 + 0.1 .203 ± .07 Sludge 5 3.3 + 0.8 .244 ± .06 Bed 2.5 2.4 + 0.4 .724 ± .16 1.25 1.1 + 0.1 1.000 ± .13 132 anaerobic filter (0.052 L O V g VS added). At the nominal 2.5 day HRT, the sludge bed provided the largest treatment efficiencies in terms of mean composite substrate removal efficiency, 64% ±18%. The anaerobic filter and the sludge bed demonstrated comparable methane yields at these HRTs, 0.044 ± 0.023 L CH«/g VS added at the mean 2.3 day HRT of the anaerobic filter and 0.040 ± 0.022 L CH4/g VS added at the mean 2.2 day HRT of the sludge bed digester. At the lowest HRT, nominally 1.25 days, the sludge bed removal efficiency remained superior over the other digester designs in terms of soluble substrate. The sludge bed removed 62% ± 16% SCOD and 77% ± 18% TVFA at this targetted retention time. The anaerobic filter delivered the highest effluent quality in terms of suspended solids removal, 35% ± 11% TCOD removal efficiency and 46% ± 18% VS removal efficiency. A lowering of the HRTs led to reduced mean substrate removal efficiencies, although not in proportion to the change in HRT. The performance of the expanded bed digester dropped off significandy with declining HRT. The mean composite substrate removal efficiency declined 22% over the drop from the highest to the lowest HRT for the expanded bed and 18% and 16% for the sludge bed and the anaerobic filter respectively. The changes in treatment efficacy were greatest for those parameters which included suspended solids. This observation indicates that the HRT decrease led to increased solids suspension and wasting. This might have occurred due to the increase in hydraulic mixing and the increased biogas production at increased organic loading rates. The anaerobic filter appeared to be relatively stable against changes in the HRT. The stationary support surfaces likely served a function in addition to that of bacterial retention. They also physically filtered the wastewater and consequently produced a more consistent effluent quality until suspended solids accumulated to critical levels. Digester plugging, especially at the highest influent concentrations and lowest HRTs, caused considerable problems. Increased processing rates afforded insufficient time for the suspended solids to be degraded, and consequently, they accumulated and caused severe disruptions in 133 digester operation. The expanded bed was the most affected by the decline in HRT when the soluble solids are taken into account Increasing the throughput rates from 4.3 days to 1.3 days caused the substrate removal efficiencies to decline 21% and 38% for SCOD and TVFA removal respectively. There was difficulty in either removing the available feedstock by digestion, or in producing additional soluble substrate through solubilizing the suspended component Low methane yields and a raised effluent pH preclude the latter explanation. This implies that there were in fact some problems in digesting the readily degradable substrate. A deficiency in the bacterial numbers and overall activity is indicated. Stress on the system may have arisen from both physical and chemical sources. The hydraulic shear required to support the particles could also have retarded the biofilm formation (Hall and Javanovic, 1982). Hansson and Molin (1981) have discovered some end-product inhibition which occurs through gas mixing. The sludge bed was most affected by a change in HRT when the suspended solids indicators are taken into account Dropping from the 3.8 to 1.2 day HRT, the sludge bed treatment efficiency declined 26% and 20% for TCOD and VS removals respectively. The corresponding decline in performance for the soluble indicators was 9% and 4% for the SCOD and TVFA removals. The discrepancy between the removal efficiencies of the suspended and soluble indicators suggests that although capable of degrading the feed biochemically, difficulties were encountered in retaining the suspended solids. The source of these suspended solids originated from the feed and as a part of the microbial mass in the digester. This difficulty of solids retention was reflected in the widely fluctuating methane production of the sludge beds. CONCLUSIONS FROM THE EXPERIMENTS X. CONCLUSIONS FROM THE EXPERIMENTS Based on an extensive review of the literature and on this experimental work, the major results are summarized as follows: 1. Anaerobic digester design and mode of operation can profit from the experimental findings as reported in the literature, specifically: a. use a number of mechanisms in order to favour microbial retention, b. use efficient liquid solid separation prior to feeding, c. use upflow operation, d. use a feeding frequency at or approaching continuous operation, e. use a vessel configuration which accomodates plug flow, and f. use gas or liquid recycle and minimize the extent of mixing. 2. The upflow anaerobic filter, the upflow anaerobic expanded bed and the upflow anaerobic sludge bed digesters demonstrated good stability and reasonable treatment efficiency over the range of hydraulic retention times and influent waste strengths performed in these experiments. 3. The anaerobic filter performance delivered the most stable effluent quality of the digester designs provided solids accumulation was small enough to preclude digester plugging. It was less affected by the decline in HRT than the other digesters. The mean TCOD removal efficiency was 47% ± 14% for the mean 4.0 day HRT, 51% ± 9% for the mean 2.3 day HRT and 35% ±11% for the mean 1.3 day HRT. This stable performance was reflected in the methane yields, 0.052 ± 0.024 L CH4/g VS added for the mean 4.0 day HRT, 0.044 ± 0.023 L CH,/g VS added for the mean 2.3 day HRT and 0.025 ± 0.013 L CH«/g VS added for the mean 1.3 day HRT. The biogas contained 62.1% ± 5.9% methane and 17.1% ± 5.6% carbon dioxide. 4. The expanded bed demonstrated substrate removal efficiencies over the range of experimentation which were comparable with the corresponding performance of the other digester designs. The mean TCOD removals for the three targetted HRTs were as 135 136 follows: 45% + 15% at the mean 4.3 day HRT, 38% ± 12% at the mean 2.4 day HRT and 28% ± 9% at the mean 1.3 day HRT. This substrate reduction was not translated into biogas yields however. The average biogas composition was 43.6% ± 5.3% methane and 9.0% ± 4.9% carbon dioxide. Low methane production levels were observed throughout the experimental period. The cause of such low gas productivities was not determined although the effects of shear and end-product inhibition arising from the bed expansion through gas recycle are suspected based on the findings of the literature. 5. The sludge bed digesters delivered superior substrate removal efficiencies and the largest methane quantities compared to the two other digester designs. For the range of influent concentrations, the mean TCOD removals were: 53% ± 9% at the mean 3.8 day HRT, 45% ± 12% at the mean 2.2 day HRT and 32% ± 10% at the mean 1.2 day HRT. The sludge bed also demonstrated the greatest variability in process performance. The mean methane yields were as follows: 0.095 ± 0.038 L CH4/g VS added at the mean 3.8 day HRT, 0.040 ± 0.022 L CH /^g VS added at the mean 2.2 day HRT and 0.037 ± 0.020 L CH«/g VS added at the mean 1.2 day HRT. Biogas content averaged 61.1% ± 6.5% methane and 18.9% ± 6.2% carbon dioxide over the entire experimental period. It is thought that the fluctuations in methane productivity occurred due to a biogas induced flotation of the sludge with subsequent wasting in the effluent. Such difficulties are widely reported in the literature. RECOMMENDATIONS XT. RECOMMENDATIONS Based on the practical experience gained from operation and trouble-shooting in the laboratory and on the assimilated information from the literature, a number of recommendations are presented. These are pertinent to the design, operation and research into retained biomass digester designs for anaerobic wastewater treatment, They are as follows: 1. Longer periods of time than are typically quoted in the literature are required for a closer approximation to steady state operation at each combination of waste strength and hydraulic retention time. 2. The testing for soluble COD, TKN and NH3 is consumptive in terms of operator time and chemical reagents. These parameters contribute little information about the process itself and the usefulness of their continued testing is questionable. They may be used profitably for the characterization of the feedstock itself however, in order to facilitate comparisons with other wastewaters. 3. The variability in both the feedstock and the hydraulic retention times caused considerable difficulties in assessing the process performance. The feedstock variability could be better controlled by diluting to a particular specification following more detailed laboratory assays. The hydraulic retention time fluctuations could be addressed on several fronts. An increased digester volume would decrease the variability relative to the total digester volume by feeding wastewater for longer periods and at higher rates. A larger volume would increase the likelihood of rinding a suitable level control apparatus. Feed tank mixing, and the use of cooled and covered feed and effluent tanks would assist in retarding the deterioration of the influent and effluent samples outside of the digesters themselves. 4. Plugging caused difficulties in operation, leading to added time required for a re-establishment of a pseudo steady state. The use of larger digester vessels, and 138 139 shorter, larger diameter and straighter lines for both feeding and effluent withdrawal could help to address the plugging problem. The installation of a tapered vessel base would decrease the dead volume where solids would otherwise accumulate, increase the flow velocity at the point of feed entry and therefore decrease the incidences of plugging. Greater efforts should be made to preclude the introduction of suspended solids into the digesters. A secondary screening process would make a significant contribution in this area. 5. Upflow operation with intermittent or continuous feeding would seem to have important advantages over the more commonly practiced downflow daily feeding. The high aspect ratio similarily may have benefits over the more conventional digester configuration. These changes should be incorporated into existing operations where feasible. 6. In the anaerobic filter, the small randomly placed packing material led to plugging problems. Instead, vertically arranged channels would allow the passage of small quantities of suspended particles through the digester. A provision to retard the rising biogas at regular height intervals would be required. 7. In the expanded bed, use of support particles having only a marginally negative buoyancy in the wastewater would nunimize the energy expenditure and hydraulic shear which is required for the expansion of heavier particles. Continuous feeding or intermittent feeding at short time intervals might eliminate the need for gas mixing altogether. 8. The sludge bed demonstrated the highest but most variable performance. In order to achieve a semblance of process stability, some provision is needed to separate the gas-floated sludge from the effluent. Some investigators have suggested incorporating a settling chamber into the digester system. Instead, adding an element in the upper part of the digester to serve as both a gas solid separator and a bacterial support surface might be a simpler innovation. In this way, the digester volume would be better utilized 140 due to increased and more evenly distributed bacterial populations. Such an arrangement is a merging of the sludge bed with the anaerobic filter, a design gaining interest among researchers in this area. 9. Based on the observed relative stability of the experimentation over a 25 fold change in TCOD and VS loading rates and a 14 fold change in SCOD and TVFA loading rates, further investigations at increased feed strengths and decreased retention times appear to be warranted. Increasing the wastewater concentrations through decreasing the dilutions and increasing the throughput rate together will permit the use of smaller and hence less costly digester vessels. R E F E R E N C E S Aiba, S., Humphrey, A.E. and Mills, N.F., 1965, Biochemical Engineering, Academic Press, New York, Chapters 4, 5, 7. Aivasidis, A., 1983, Different Methods of Biomass Retention in Continuous Anaerobic Digestion, In: Biotech '83, Proceedings of the International Conference on the Commercial Applications and Implications of Biotechnology, Online Conferences Ltd., Middlesex, U.K., 1061-1072. American Public Health Association, 1975, Standard Methods for the Examination of Water and Wastewater, 14th Edition, APHA, AWWA, WPCF, Washington, D.C. Anderson, G.K., Donnelly, T. and McKeown, K.J., 1982, Application of Anaerobic Packed-Bed Reactors to Industrial Wastewater Treatment, Proceedings of the 37th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 651-660. Anderson, G.K., Donnelly, T. and McKeown, K.J., 1982b, Identification and Control of Inhibition in the Anaerobic Treatment of Industrial Wastewaters, Process Biochemistry, July/August, 28-32. Andrews, G.F., 1982, Fluidized-Bed Fermenters: A Steady-State Analysis, Biotechnology and Bioengineering, Vol. XXIV, 2013-2030. Andrews, G. and Trapasso, R., 1985, The Optimal Design of Fluidized Bed Bioreactors, Journal WPCF, Vol. 57, No.2, 143-150. Andrews, J.F. and Pearson, E.A., 1965, Kinetics and Characterization of Volatile Acid Production in Anaerobic Fermentation Processes, International Journal Air Water Pollution, Vol. 9„ 439-461. Archer, D.B., 1983, The Microbiological Basis of Process Control in Methanogenic Fermentation of Soluble Wastes, Enzyme Microbial Technology, Vol. 5, 162-170. Asinari Di San Marzano, CM., Binot, R., Bol, T., Fripiat, J.L, Hutschemakers, J., Melchior, J.L,, Perez, I., Naveau, H., and Nyns, E.J., 1981, Volatile Fatty Acids, An Important State Parameter For The Control of the Reliability and the Productivities of Methane Anaerobic Digestions, Biomass, Vol. 1, 47-59. Atkinson, B., 1981, Immobilized Biomass - a Basis for Process Development in Wastewater Treatment, In: Cooper, P.F. and Atkinson, B. (editors), Biological Fluidized Bed Treatment of Water and Wastewater, Ellis Horwood Ltd., Chichester, England, 22-34. Atkinson, B., 1974, Biochemical Reactors, Pion Limited, London, 129-152. Atkinson, B., Black, G.M. and Pinches, A., 1981, The Characteristics of Solid Supports and Biomass Support Particles When Used in Fluidised Beds, In: Cooper, P.F. and Atkinson, B. (editors), Biological Fluidised Bed Treatment of Water and Wastewater, Ellis Horwood Ltd., Chichester, England, 75-106. Atkinson, B., Black, G.M. and Pinches, A., 1980, Process Intensification Using Cell Support 141 142 Systems, Process Biochemistry, May, 24-32. Atkinson, B. and Knights, AJ., 1975, Microbial Film Fermenters: Their Present and Future Applications, Biotechnology and Bioengineering, Vol. XVII, 1245-1267. Barnes, D., Bliss, P.J., Grauer, B., Kuo, EM., Robins, K. and Maclean, G., 1983, Influence of Organic Shock Loads on the Performance of an Anaerobic Fluidized Bed System, Proceedings of the 38th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 715-723. Bhatia, D., Vieth, W.R. and Venkatasubramanian, K., 1981, Transient and Steady-State Performance of Anaerobic Bio-Reactors for Methane Production from Organic Wastes, Proceedings of the 2nd World Congress of Chemical Engineering, Montreal, Vol. 1, 306-314. Boening, P.H. and Larsen, V.F., 1982, Anaerobic Fluidized Bed Whey Treatment, Proceedings of the 37th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 295-304. Brauer, H., 1982, Development and Improvement of Bioreactors, The Chemical Engineer, June 1982, 224-229. Brown, D.E, 1982, Industrial-Scale Operation of Microbial Processes, Journal of Chem. Tech. and Biotech., 34-46. Buckland, B.C., 1984, The Translation of Scale in Fermentation Processes: The Impact of Computer Process Control, Bio/Technology, October, 875-883. Buhr, H.O. and Andrews, J.F., 1977, The Thermophilic Anaerobic Digestion Process, Water Research, Vol. 11, 129-143. Buijs, C, Heertjes, P.M. and van der Meer, R.R., 1982, Distribution and Behaviour of Sludge in Upflow Reactors for Anaerobic Treatment of Wastewater, Biotechnology and Bioengineering, Vol. XXIV, 1975-1989. Bull, M.A., Sterritt, R.M. and Lester, J.N., 1984, An Evaluation of Single- and Separated-Phase Anaerobic Industrial Wastewater Treatment in Fluidized Bed Reactors, Biotechnology and Bioengineering, Vol. XXVI, 1054-1065. Bull, M.A., Sterritt, R.M. and Lester, J.N., 1984b, The Distribution of Bacterial Activity in an Anaerobic Fluidized Bed Reactor, Water Research, Vol. 18, No. 8, 1017-1020. Bull, M.A., Sterritt, R.M. and Lester, J.N., 1984c, Developments in Anaerobic Treatment of High Strength Industrial Waste Waters, Chem. Eng. Res. Des., Vol. 62, July, 203-213. Bull, M.A., Sterritt, R.M. and Lester, J.N., 1983, Response of The Anaerobic Fluidized Bed Reactor to Transient Changes in Process Parameters, Water Research, Vol. 17, No. 11, 1563-1568. Bull, M.A., Sterritt, R.M. and Lester, J.N., 1983b, An Evaluation of Four Start-Up Regimes for Anaerobic Fluidized Bed Reactors, Biotechnology Letters, Vol. 5, No. 5, 333-338. Busch, A.W., 1984, Chemical Reactor Design Theory and Biological Treatment of Industrial Wastes ~ Is There a Gap?, Journal WPCF, Vol. 56, No. 3, 215-218. 143 Callander, IJ. and Barford, J.P., 1984. Improved Anaerobic Digestion of Pig Manure Using a Tower Fermenter, Agricultural Wastes 11, 1-24. Callander, I.J. and Barford, J.P., 1983, Recent Advances in Anaerobic Digestion Technology, Process Biochemistry, August, 24-30. Chakrabarty, T., Scharer, J.M. and Reilly, P.M., 1981, Modelling of Oscillating Methane Generation in Anaerobic Digestion, Proceedings of the 2nd World Congress of Chemical Engineering, Montreal, Vol. 1, 325-328. Chandler, J.A., Jewell, W.J., Gossett, J.M., van Soest, PJ. and Robertson, J.B., 1980, Predicting Methane Fermentation Biodegradability, Biotechnology and Bioengineering Symposium, No. 10, 93-107. Characklis, W.G., 1981, Fouling Biofilm Development: A Process Analysis, Biotechnology and Bioengineering, Vol. XXTTI, 1923-1960. Christensen, D.R., Gerick, J.A. and Eblen, J.E, 1984, Design and Operation of an Upfiow Anaerobic Sludge Blanket Reactor, Journal WPCF, Vol. 56, No. 9, 1059-1062. Colin, F„ Ferrero, G.L, Gerletti, M., Hobson, P., L'Hermite, P., Naveau, H.P. and Nyns, RJ., 1983, Proposal for the Definition of Parameters and Analytical Measurements Applicable to the Anaerobic Digestion Process, Agricultural Wastes, Vol. 7, 183-193. Colleran, E , Barry, M., Wilkie, A. and Newell, P.J., 1982, Anaerobic Digestion of Agricultural Wastes Using the Upfiow Anaerobic Filter Design, Process Biochemistry, March/ April, 12-17. Cooney, C L , 1983, Bioreactors: Design and Operation, Science, Vol. XXX, 11 February, 728-733. Cowley, I.D. and Wase, D.A.J., 1981, Anaerobic Digestion of Farm Wastes; a Review - Part 1, Process Biochemistry, Aug./SepL, 28-33. Cowley, I.D. and Wase, D.A.J., 1981b, Anaerobic Digestion of Farm Wastes; a Review -Part 2, Process Biochemistry, OcL/Nov., 16-19. Dahab, M.F. and Young, J.C, 1981, Energy Recovery from Alcohol Stallage Using Anaerobic Filters, Biotechnology and Bioengineering Symposium No. 11, 381-397. de Vocht, M., van Meenen, P., van Assche, P. and Verstraete, W., 1983, Influence of Sedimentation and Adhesion on Selection of Methanogenic Associations, Process Biochemistry, December, 31-34. de Zeeuw, W. and Lettinga, G., 1980, Acclimation of Digested Sewage Sludge During Start-up of an Upfiow Anaerobic Sludge Blanket (UASB) Reactor, Proceedings of the 35th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 39-47. DeWalle, F.B. and Chian, ES.K., 1976, Kinetics of Substrate Removal in a Completely Mixed Anaerobic Filter, Biotechnology and Bioengineering, Vol. XVTJJ, 1275-1295. Doyle, O.P., O'Malley, J.D., Clausen, EC. and Gaddy, J.L, 1983, Kinetic Improvements in the Production of Methane From Cellulosic Residues, Symposium Papers, Energy from Biomass and Wastes VII, January 24-28, Lake Buena Vista, Florida, Institute of Gas 144 Technology, 545-559. Eighmy, T.T., Maratea, D. and Bishop, P.L, 1983, Electron Microscopic Exarnination of Wastewater Biofilm Formation and Structural Components, Applied and Environmental Microbiology, Vol. 45, No. 6, 1921-1931. Ellwood, D.C, Keevil, C.W., Marsh, P.D., Brown, CM. and Wardell, J.N., 1982, Surface-Associated Growth, Phil. Trans. R. Soc. Lond. B, Vol. 297, 517-532. Esser, K. and Kues, U., 1983, Flocculation and its Implication for Biotechnology, Process Biochemistry, December, 21-23. Fannin, K.F., Srivastava, V.J., Chynoweth, D.P. and Bird, K.T., 1983, Effects of the Interaction Between Biomass Composition and Reactor Design on Anaerobic Digestion Process Performance, Symposium Papers, Energy From Biomass and Wastes VD., January 24-28, 1983, Lake Buena Vista, Florida, Institute of Gas Technology, 563-583. Fannin, K.F., Srivasta, V.J. and Chynoweth, D.P., 1982, Unconventional Anaerobic Digester Designs for Improving Methane Yields from Sea Kelp, Symposium Papers, Energy From Biomass and Wastes VI, January 25-29, 1982, Lake Buena Vista, Florida, Institute of Gas Technology, 373-396. Feilden, N.E.H., 1983, The Theory and Practice of Anaerobic Digestion Reactor Design, Process Biochemistry, October, 34-37. Feilden, N.E.H., 1981, A Note on the Temperature for Maximum Net Gas Production in an Anaerobic Digester System, Agricultural Wastes, 75-79. Forster, C.F., Rockey, J.S., Wase, D.A.J. and Godwin, S.J., 1982, Mixing Characteristics of Fixed Film Anaerobic Reactors, Biotechnology Letters, Vol. 4, No. 12, 799-804. Fowler, H.W., 1981, Microbial Adhesion: It's Significance in Biochemical Engineering and a Method for Evaluation of Adhesion, Proceedings of the 2nd World Congress of Chemical Engineering, Vol. 1, Montreal, 215-218. Frostell, B., 1982, Anaerobic Fluidized Bed Experimentation with a Molasses Waste Water, Process Biochemistry, November/December, 37-40. Frostell, B., 1981, Anaerobic Treatment in a Sludge Bed System Compared with a Filter System, Journal WPCF, Vol. 53, No. 2, 216-222. Fynn, G.H. and Whitmore, T.N., 1984, Retention of Methanogens in Colonized Reticulated Polyurethane Foam Biomass Support Particle, Biotechnology Letters, Vol. 6, No. 2, 81-86. Genung, R.K., Million, D.L., Hancher, C.W. and Pitt, W.W. Jr., 1978, Pilot Plant Demonstration of an Anaerobic Fixed-Film Bioreactor for Wastewater Treatment, Biotechnology and Bioengineering Symp. No. 8, 329-344. Genung, R.K., Pitt, W.W. Jr., Davis, G.M. and Koon, J.H., 1980, Energy Conservation and Scale-up Studies for a Wastewater Treatment System Based on a Fixed-Film Anaerobic Bioreactor, Biotechnology and Bioengineering Symposium No. 10, 295-316. Ghosh, S., 1981, Net Energy Production in Anaerobic Digestion, Symposium Papers, Energy from Biomass and Wastes V, Lake Buena Vista, Florida, January 26—30, Institute of Gas 145 Technology, Chicago, 253-273. Ghosh, S., 1981b, Energy Production Efficiencies in Anaerobic Gasification Processes, Process Biochemistry, June/July, 2-7. Ghose, T.K, and Bhadra, A., 1981, Maximization of Energy Recovery in Biomethanation Process, Part-1: Use of Cow Dung as Substrate in Multireactor System, Process Biochemistry, October/November, 23-25. Gibbons, D.B. and Denton, P.H., 1981, Surface Roughness and Biofouling, Proceedings of the 2nd World Congress of Chemical Engineering, Montreal, Vol. 1, 430-437. Graef, S.P. and Andrews, J.F., 1974, Stability and Control of Anaerobic Digestion, Journal WPCF, Vol. 46, No. 4, 666-683. Guiot, S.R. and van den Berg, L., 1984a, Performance and Biomass Retention of an Upflow Anaerobic Reactor Combining a Sludge Blanket and a Filter, Biotechnology Letters, Vol. 6, No. 3, 161-164. Hall, ER. and Jovanovic, M., 1982, Anaerobic Treatment of Thermal Sludge Conditioning Liquor With Fixed-Film and Suspended-Growth Processes, Proceedings of the 37th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbor Science, 719-728. Hall, ER., Jovanovic, M. and Pejic, M., 1982, Pilot Studies of Methane Production in Fixed-Film and Sludge Blanket Anaerobic Reactors, Proceedings of the Fourth Bioenergy R&D Seminar, March 29-31, Winnipeg, Manitoba, 475- 480. Hall, ER. and Melcer, H., 1984, Energy Recovery from Wastewater by High Rate Anaerobic Treatment, Water and Pollution Control, March/April, 20-25. Hamoda, M.F. and van den Berg, L., 1983, Effect of Settling on Performance of the Upflow Anaerobic Sludge Bed Reactors, Water Research, yet to be published Hansson, G. and Molin, N., 1981, End Product Inhibition in Methane Fermentations: Effects of Carbon Dioxide and Methane on Methanogenic Bacteria Utilizing Acetate, European Journal of Applied Microbial Biotechnology, Vol. 13, 236-241. Hansson, G. and Molin, N., 1981, End Product Inhibition in Methane Fermentations: Effects of Carbon Dioxide on Fermentative and Acetogenic Bacteria, European Journal of Applied Microbial Biotechnology, Vol. 13, 242-247. Hashimoto, A.G. and Chen, Y.R., 1980, Economic (Optimization of Anaerobic Fermenter Designs, Paper presented at the Fourth International Symposium on Livestock Wastes, Amarillo, Texas, April 15-17. Hashimoto, A.G., Chen, Y.R., Varel, V.H. and Prior, R.L, 1979, Anaerobic Fermentation of Animal Manure, Paper presented at the Summer Meeting of the ASAE and CSAE June 24-27, Winnipeg, Manitoba, ASAE Paper No. 79-4066. Hawkes, D.L and Horton, R., 1981, Anaerobic Digestion Design Fundamentals Part III, Process Biochemistry, Feb./March, 10-12. Heertjes, P.M. and van der Meer, R.R., 1978a, Dynamics of Liquid Flow in an Up-Flow Reactor Used for Anaerobic Treatment of Wastewater, Biotechnology and Bioengineering, 146 Vol. XX, 1577-1594. Henze, M. and Harremoes, P., 1982, Anaerobic Treatment of Wastewater in Fixed Film Reactors, Review Paper, IAWPR Specialized Seminar, June 16-18, Copenhagen, Denmark, 1093-1098. Hermanowicz, S.W. and Ganczarczyk, J.J., 1983, Some Ruidization Characteristics of Biological Beds, Biotechnology and Bioengineering, Vol. XXV, 1321-1330. Heyes, R.H. and Hall, RJ., 1981, Anaerobic Digestion Modelling - The Role of H2, Biotechnology Letters, Vol. 3, No. 8, 431-436. Hickey, R.F. and Owens, R.W., 1981, Methane Generation from High-Strength Industrial Wastes with the Anaerobic Biological Fluidized Bed, Biotechnology and Bioengineering Symp. No. 11, 399-413. Hill, D.T., 1982, Design of Digestion Systems for Maximum Methane Production, Transactions of the ASAE, 226-230. Hill, D.T., 1982b, Optimum Operational Design Criteria for Anaerobic Digestion of Animal Manure, Transactions of the ASAE, 1029-1032. Hilton, M.G., Turner, RJ., Powell, G.E and Archer, D.B., 1983, Methanogenesis from Acetate by a Mesophilic Anaerobic Filter, Process Biochemistry, December, 2-5. Horton, R. and Hawkes, D.L, 1979a, Anaerobic Digester Design Fundamentals: Part II, Process Biochemistry, September, 12-16. Howell, J.A. and Atkinson, B., 1976, Sloughing of Microbial Film in Trickling Filters, Water Research, Vol. 10, 307-315. Hsu, H.W., 1978, Characteristics of Tapered Fluidized Reactors: Two-Phase Systems, Biotechnology and Bioengineering Symp. No. 8, 1-11. Hudson, J.W., Pohland, EG. and Pendergrass, R.P., 1978, Anaerobic Packed Column Treatment of Shellfish Processing Wastewaters, Proceedings of the 33rd Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 560-574. Hulshoff Pol, L , Dolfing, J., de Zeeuw, W. and Lettinga, G., 1982, Cultivation of a Well Adapted Pelletized Methanogenic Sludge, Biotechnology Letters, Vol. 4, No. 5, 329-332. Jech, D.D. and Brautigam, J.A., 1983, Bioconversion of Industrial Residues to Methane Gas, In: Fuel Gas Developments, Wise, D. L (editor), CRC Press Inc., Boca Raton, Florida. 133-157. Jennings, P.A., Snoeyink, V.L and Chian, ES.K., 1976, Theoretical Model for a Submerged Biological Filter, Biotechnology and Bioengineering, Vol. XVIII, 1249-1273. Jeris, J.S., Owens, R.W., Hickey, R. and Rood, F., 1977, Biological Ruidized-Bed Treatment for BOD and Nitrogen Removal, Journal WPCF, May 1977, 816-783. Jewell, W.J., 1981, Development of the Attached Microbial Film Expanded-Bed Process for Aerobic and Anaerobic Waste Treatment, In: Cooper,P.F. and Atkinson, B. (editors), Biological Ruidised Bed Treatment of Water and Wastewater, Hlis Horwood Ltd., Chichester, England, 251-271. 147 Jewell, WJ., 1980, Future Trends in Digester Design, In: Stafford, D.A., Wheatley, BJ. and Hughes, D.E, (editors), 1980, Anaerobic Digestion, Applied Science Publishers, London, 467-489. Jewell, W.J. and Morris, J.W., 1981, Influence of Varying Temperature, Flowrate and Substrate Concentration on the Anaerobic Attached-Film Expanded-Bed Process, Proceedings of the 36th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 655-664. Jewell, W.J., Switzenbaum, M.S. and Morris, J.W., 1981a, Municipal Wastewater Treatment with the Anaerobic Attached Microbial Film Expanded Bed Process, Journal WPCF, Vol. 53, No. 4, 482-490. Kargi, F. and Park, J.K., 1982, Optimal Biofilm Thickness for Fluidised-Bed Biofilm Reactors, Journal of Chem. Tech. and Biotech., 744-748. Kelly, CR. and Switzenbaum, M.S., 1983, Temperature and Nutrient Effects on the Anaerobic Expanded Bed Treating a High Strength Waste, Proceedings of the 38th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 591-601. Kennedy, K.J. and Droste, R.L., 1983, The Effect of Influent Concentration on the Startup of Anaerobic Downflow Stationary Fixed Film (DSFF) Reactors, Proceedings of the 38th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 615-626. Kennedy, K.J. and van den Berg, L., 1982a, Stability and Performance of Anaerobic Fixed Film Reactors During Hydraulic Overloading at 10-35 C, Water Research, Vol. 16, 1391-1398. Kennedy, K.J. and van den Berg, L , 1982b, Effect of Height on the Performance of Anaerobic Downflow Stationary Fixed Film (DSFF) Reactors Treating Bean Blanching Waste, Proceedings of the 37th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 71-76. Kennedy, KJ. and van den Berg, L , 1982c, Continuous vs Slug Loading of Downflow Stationary Fixed Film Reactors Digesting Piggery Waste, Biotechnology Letters, Vol. 4, No.2, 137-142. Kennedy, K.J. and van den Berg, L., 1981, Effects of Temperature and Overloading on the Performance of Anaerobic Fixed-Film Reactors, Proceedings of the 36th Industrial Waste Confernce at Purdue University, Lafayette Indiana, Ann Arbour Science, 678-685. Kennedy, K.J. and van den Berg, L., 1981b, Use of Fixed Film Reactors for Production of Methane from Waste, Bioenergy R&D, 1981 NRCC No. 19239. Klass, D.L., 1980, Anaerobic Digestion for Methane Production - A Status Report, Proceedings, Bio-Energy '80, World Congress and Exposition, April 21-24, Atlanta, Georgia, 143-149. Kleinstreuer, C. and Poweigha, T., 1982, Dynamic Simulator for Anaerobic Digestion Processes, Biotechnology and Bioengineering, Vol. XXIV, 1941-1951. Kobayashi, H.A., Stenstrom, M.K. and Mah, R.A., 1983, Treatment of Low Strength 148 Domestic Wastewater Using the Anaerobic Filter, Water Research, Vol. 17, No. 8, 903-909. Kroeker, EJ., Schulte, D.D., Sparling, A.B. and Lapp, H.M., 1979, Anaerobic Treatment Process Stability, Journal WPCF, Vol. 51, No. 4, 718-727. La Motta, EJ., 1976, External Mass Transfer in a Biological Film Reactor, Biotechnology and Bioengineering, Vol. XVJJI, 1359-1370. Lavagno, E , Ravetto, P. and Ruggeri, B., 1983, An Analytical Model to Study the Performance of an Anaerobic Digester, Agricultural Wastes, Vol. 5, 37-50. Lawrence, A.L and McCarty, P.L., 1969, Kinetics of Methane Fermentation in Anaerobic Treatment, Journal WPCF, Vol. 41, No. 2, Part 2, R1-R17. Lee, J.C. and Buckley, P.S., 1981, Fluid Mechanics and Aeration Characteristics of Fluidised Beds, In: Cooper, P.F. and Atkinson, B., (editors), Biological Fluidised Bed Treatment of Water and Wastewater, Ellis Horwood Ltd., Chichester, England, 62-74. Lettinga, G., 1981, Anaerobic Digestion for Energy Saving and Production, In: Energy From Biomass, 1st EC. Conference, Palz, W., Chartier, P. and Hall, D.O. (editors), Applied Science Publishers Ltd., London, 264-278. Uttinga, G., Hulshoff Pol, L.W., Wiegant W., de Zeeuw, W., Hobma, S.W., Grin, P., Roersma, R., Sayed, S. and van Velsen, A.F.M., 1983a, Upfiow Sludge Blanket Processes, Proceedings, Third International Symposium on Anaerobic Digestion, August 14-19, Boston Mass., Evans & Faulkner Inc., 139-158. Lettinga, G., Roersma, R. and Grin, P., 1983b, Anaerobic Treatment of Raw Domestic Sewage at Ambient Temperatures Using a Granular Bed UASB Reactor, Biotechnology and Bioengineering, Vol. XXV, 1701-1723. Lettinga, G., van Velsen, A.F.M., Hobma, S.W., de Zeeuw, W. and Klapwijk, A., 1980a, Use of the Upfiow Sludge Blanket (USB) Reactor Concept for Biological Wastewater Treatment, Especially for Anaerobic Treatment, Biotechnology and Bioengineering, Vol. XXII, 699-734. Lettinga, G. and Vinken, J.N., 1980, Feasibility of the Upfiow Anaerobic Sludge Blanket (UASB) Process for the Treatment of Low-Strength Wastes, Proceedings of the 35th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 625-634. Levenspiel, O., 1979, The Chemical Reactor Omnibook, Oregon State University Bookstores, Inc. Li, A, and Sutton, P.M., 1983, Determination of Alkalinity Requirements for the Anaerobic Treatment Process, Proceedings of the 38th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 603-613. Liao, P.H. and Lo, K.V., 1985, Methane Production Using Whole and Screened Dairy Manure in Conventional and Fixed-Film Reactors, Biotechnology and Bioengineering, Vol. XXVTJ, 266-272. Lo, K.V., Liao, P.H. and Bulley, N.R., 1983, Biogas Production From Dairy Manure and its Filtrate, Canadian Agricultural Engineering, Vol. 25, No. 1, 59-61. 149 Lo, K.V., Chen, W.Y. and Liao, P.H., 1986, Mesophilic Digestion of Screened Dairy Manure Using Anaerobic Rotating Biological Contact Reactor, Biomass, Vol. 9, 81-92. Lo, K.V., Liao, P.H., Bulley, N.R. and Chieng, S.T., 1984a, A Comparison of Biogas Production from Dairy Manure Filtrate Using Conventional and Fixed-Film Reactors, Canadian Agricultural Engineering, Vol. 26, No. 1, 73-78. Lo, K.V., Whitehead, A.J., Liao, P.H. and Bulley, N.R., 1984b, Methane Production from Screened Dairy Manure Using a Fixed-Film Reactor, Agricultural Wastes, Vol. 9, 175-188. Mah, R.A., 1983, Interactions of Methanogens and Non-Methanogens in Microbial Ecosystems, Proceedings of the Third International Symposium on Anaerobic Digestion, 14-19 August, Boston, Mass., Evans & Falkner Inc., 13-22. Mah, R.A., 1982, Methanogenesis and Methanogenic Partnerships, Phil. Trans. R. Soc Lond. B., Vol. 297, 599-616. Margaritis, A. and Wallace, J.B., 1984, Novel Bioreactor Systems and Their Applications, Bio/Technology, May 1984, 447-453. Marshall, D. and Timbers, G.E., 1982, Development and Testing of a Prototype Fixed Film Anaerobic Digester, Paper presented at the 1982 Winter Meeting of the ASAE Paper No. 82-6519. Maxham, J.V. and Wakamiya, W., 1980, Innovative Biological Wastewater Treatment Technologies Applied to the Biomass Gasification Wastewater, Proceedings of the 35th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 80-94. Mclnerney, M.J. and Bryant, M.P., 1981, Review of Methane Fermentation Fundamentals, In: Fuel Gas Production From Biomass Vol. 1, Wise, D. (editor), CRC Press, Boca Raton, Florida, 19-46. McKinney, R.E., 1983, Anaerobic Systems vs Aerobic Systems, Proceedings of the Third International Symposium on Anaerobic Digestion, 14-19 August, Boston, Mass., Evans & Falkner Inc., 493-499. Melchior, J.L., Binot, R., Perez, LA., Naveau, H. and Nyns, E.J., 1982, Biomethanation: Its Future Development and the Influence of the Physiology of Methanogenesis, Journal of Chem. Tech. and Biotech., 189-197. Messing, M.S., 1982, Immobilized Microbes and a High-Rate, Continuous Anaerobic Waste Processor, Symposium Papers, Energy From Biomass and Wastes VI, Lake Buena Vista, Florida, January 25-29, Institue of Gas Technology, Chicago, 425-438. Mills, P.J., 1979, Minimisation of Energy Input Requirements of an Anaerobic Digester, Agricultural Wastes, Vol. 1, 57-66. Morris, J.W. and Jewell, W.J., 1981, Organic Particulate Removal with the Anaerobic Attached-Film Expanded-Bed Process, Proceedings of the 36th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 621-630. Morris, G.R., Jewell, W.J. and Casler, G.L., 1975, Alternative Animal Waste Anaerobic Fermentation Designs and Their Costs, Proceedings of the Agricultural Waste 150 Management Conference at Cornell University, Ithaca, New York, 317-335. Mueller, J.A. and Mancini, J.L, 1975, Anaerobic Filter — Kinetics and Application, Proceedings of the 30th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbor Science, 423-447. Murk, J.S., Frieling, J.L, Tortorici, LD. and Dietrich, C C , 1980, Use of Bench-Scale Digesters to Evaluate Full-Scale Digester Performance, Journal WPCF, Vol. 52, No.ll, 2709-2716. Muroyama, K. and Fan, LS., 1985, Fundamentals of Gas-Liquid-Solid Fluidization, American Institute of Chemical Engineering, Vol. 31, No.1, 1-33. Murray, W.D. and van den Berg, L , 1981, Effects of Support Material on the Development of Microbial Fixed Films Converting Acetic Acid to Methane, Journal of Applied Bacteriology, Vol. 51, 257-265, NRCC No. 19493. Ngian, K.F. and Martin, W.R.B., 1980, Biologically Active Fluidized Beds: Mechanistic Considerations, Biotechnology and Bioengineering, Vol. XXII, 1007-1014. Oleszkiewicz, J.A., 1981, Aerobic and Anaerobic Biofiltration of Agricultural Effluents, Agricultural Wastes, Vol. 3, 285-296. Oleszkiewicz, J.A. and Koziarski, 1982, Anaerobic Wastewater Treatment in Flowthrough Digesters Without Sludge Recycle, Agricultural Wastes, Vol. 4, 365-386. Oleszkiewicz, J.A. and Koziarski, S., 1982b, Low Temperature Anaerobic Biofiltration in Upfiow Reactors, Journal WPCF, Vol. 54, No. 11, 1465-1471. Olthof, M. and Oleskiewicz, J., 1982, Anaerobic Treatment of Industrial Wastewaters, Chemical Engineering, Nov. 15, 121-126. Orcutt, J.C, Carpenter, B.H. and Click, C.N., 1973, Fluidized Bed aarification in Wastewater Treatment, Journal WPCF, Vol. 45, No. 9, 1908-1927. Pirt, S.J., 1978, Aerobic and Anaerobic Microbial Digestion in Waste Reclamation, J. Appl. Chem. Biotechnol., Vol. 28, 232-236. Ripley, L E and Boyle, W.G, 1983, Anaerobic Digestion Models: Implications for the Design Engineer, Proceedings of the Third International Symposium on Anaerobic Digestion, 14-19 August, Boston, Mass., Evans & Falkner, 451-463. Rittmann, B E , 1982, The Effect of Shear Stress on Biofilm Loss Rate, Biotechnology and Bioengineering, Vol. XXIV, 501-506. Rittmann, B.E, 1982b, Comparative Performance of Biofilm Reactor Types, Biotechnology and Bioengineering, Vol. XXIV, 1341-1370. Rimnann, B.E and McCarty, P.L, 1980, Model of Steady-State-Biofilm Kinetics, Biotechnology and Bioengineering, Vol. XXII, 2343-2357. Robinson, R.W., Akin, D.E., Nordstedt, R.A., Thomas, M.V. and Aldrich, H.C., 1984, Light and Electron Microscopic Examinations of Methane-Producing Biofilms From Anaerobic Fixed-Bed Reactors, Applied and Environmental Microbiology, Vol. 48., No. 1, 127-136. 151 Rodrigues, A,, Grasmick, A. and FJmaleh, S., 1983, Modelling Biofilm Reactors, The Chemical Engineering Journal, Vol. 27, B39-B48. Schraa, G. and Jewell, W.J., 1984, High Rate Conversions of Soluble Organics with a Thermophilic Anaerobic Attached Film Expanded Bed, Journal WPCF, Vol. 56, No. 3, 226-232. Schumann, G.E, Stanley, M.A. and Knudsen, D., 1973, Automated Total Nitrogen Analysis of Soil and Plant Samples, Soil Science Society American Proceedings, Vol. 37, 480-481. Shelton, D.R. and Tiedje, J.M., 1984, General Method for Deterrriining Anaerobic Biodegradation Potential, Applied and Environmental Microbiology, April, 850-857. Shieh, W.K.,1980, Suggested Kinetic Model for the Fluidized-Bed Biofilm Reactor, Biotechnology and Bioengineering, Vol. XXII, 667-676. Shieh, W.K. and Chen, C.Y., 1984, Biomass Hold-Up Correlations for a Fluidised Bed Biofilm Reactor, Chem, Eng. Res. Des., Vol. 62, March, 133-136. Shieh, W.K., Mulcahy, LT. and LaMotta, EJ., 1981, Unified Design Basis for the Fluidized Bed Biofilm Reactor, Proceedings of the 2nd World Congress of Chemical Engineering, Montreal, Vol. 1, 165-181. Shieh, W.K., Sutton, P.M. and Kos, P., 1981, Predicting Reactor Biomass Concentration in a Fluidized-Bed System, Journal WPCF, Vol. 53, No. 11, 1574-1584. Singh, R., Jain, M.K. and Tauro, P., 1983, Mathematical Modelling of Substrate Utilization and Biogas Formation From Cattle Waste, Water Research, Vol. 17, No. 3, 349-351. Speece, R.E and Parkin, G.F., 1983, The Response of Methane Bacteria to Toxicity, Proceedings of the Third International Symposium on Anaerobic Digestion, 14-19 August, Boston, Mass., Evans & Falkner Inc., 23-35. Stafford, D.A., 1982, The Effects of Mixing and Volatile Fatty Acid Concentrations on Anaerobic Digester Performance, Biomass, Vol. 2, 43-55. Stuckey, D.C, 1983, Biogas in Developing Countries: A Critical Appraisal, Proceedings of the Third International Symposium on Anaerobic Digestion, 14-19 August, Boston, Mass., Evans & Falkner Inc, 253-269. Sutton, P.M. and Evans, R.R., 1983, Anaerobic System Designs For Efficient Treatment of Industrial Wastewaters, Proceedings of the Third International Symposium on Anaerobic Digestion, 14-19 August, Boston, Mass., Evans & Falkner Inc. Sutton, P.M. and Li, A., 1981, Anitron and Oxitron Systems: High-Rate Anaerobic and Aerobic Biological Treatment Systems for Industry, Proceedings of the 36th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 665-677. Sutton, P.M., Li, A., Evans, R.R. and Korchin, S.R., 1982, Dorr-Oliver's Fixed-Film, Suspended Growth Anaerobic Systems for Industrial Wastewater Treatment and Energy Recovery, Proceedings of the 37th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 667-676. Switzenbaum, M.S., 1983, Anaerobic Fixed Film Wastewater Treatment, Enzyme Microb. 152 Technol., Vol. 5, July, 242-250. Switzenbaum, M.S. and Danskin, S.C, 1981, Anaerobic Expanded Bed Treatment of Whey, Proceedings of the 36th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 414-424. Switzenbaum, M.S. and Jewell, W.J., 1980, Anaerobic Attached-Film Expanded-Bed Reactor Treatment, Journal WPCF, Vol. 52, No. 7, 1953-1965. Tesch, W., Schneider, K. and Bachofen, R., 1983, Performance of an Anaerobic Filter for Purification and Methane Production from Waste Water of a Sugar Refinery, Process Biochemistry, January/February, 34-37. Tramper, J., van Groenestijn, J.W.,Luyben, K.ChAM and Hulshoff Pol,LW., 1984, Some Physical and Kinetic Properties of Granular Anaerobic Sludge, In: Innovations in Biotechnology, Houwink, EH. and van der Meer, R.R., (editors), Proceedings of a Poster Symposium, Delft, The Netherlands, November 22, 1983. Trulear, M.G. and Characklis, W.G., 1982, Dynamics of Biofilm Processes, Journal WPCF, Vol. 54, No. 9, 1288-1301. van den Beld, H., 1984, Environmental Biotechnology, In: Houwink, EH. and van den Meer, R.R., (editors), Innovations in Biotechnology, Proceedings of a Poster Symposium, Delft, The Netherlands, November 22, 1983. van den Berg, L., Hamoda, M.F. and Kennedy, K.J., 1981a Performance of Upfiow Anaerobic Sludge Bed Reactors, Proceedings of the Third Bioenergy R&D Seminar, March 24-25, 1981, Ottawa, Energy Project Office, NRC, 109-112. van den Berg, L and Kennedy, K.J., 1983, Comparison of Advanced Anerobic Reactors, Proceedings of the Third International Symposium on Anaerobic Digestion, August 14-19, Boston, Mass., Evans and Faulkner, 71-89. van den Berg, L. and Kennedy, K.J., 1982a, Comparison of Intermittent and Continuous Loading of Stationary Fixed-Film Reactors for Methane Production from Wastes, J. Chem. Tech. Biotechnol., Vol. 32, 427-432. van den Berg, L, and Kennedy, K.J., 1982b, Factors Affecting Methane Production Rates of Downflow Stationary Fixed Film Reactors, Proceedings of the Fourth Bioenergy R&D Seminar, March 29-31, Winnipeg, Manitoba, 457-460. van den Berg, L and Kennedy, KJ., 1981a, Support Materials for Stationary Fixed Film Reactors for High Rate Methanogenic Fermentations, Biotechnology Letters, Vol. 3, No. 4, 165-170. NRCC No. 19161. van den Berg, L. and Kennedy, K.J., 1981b, Potential Use of Anaerobic Processes for Industrial Waste Treatment, Paper presented at the Seminar on Anaerobic Waste Water Treatment and Energy Recovery, November 3,4, Pittsburg, Pennsylvania. NRCC No. 19727. van den Berg, L Kennedy, K.J. and Hamoda, M.F., 1981b, Effect of Type of Waste on Performance of Anaerobic Fixed Film and Upfiow Sludge Bed Reactors Paper presented at the 1981 Industrial Waste Conference at Purdue University, Lafayette, Indiana, NRCC Number 19315. 153 van den Berg, L and Lentz, CP., 1980a, Effect of Digester Configuration, Waste Composition and Inoculum on Rates of Production of Methane From Wastes, Paper presented at the Bioenergy R&D seminar, March 25-27, Ottawa NRC Number 18478. van den Berg, L and Lentz, CP., 1980b, Effects of Film Area-to-Volume Ratio, Film Support, Height and Direction of Flow on Performance of Methanogenic Fixed Film Reactors, Paper presented to the U.S. Department of the Environment Workshop/Seminar on Anaerobic Filters, Howey-in-the-Hills, Florida, 9-10 January 1980, NRCC No. 17998, 1-10. van den Berg, L and Lentz, CP., 1979a, Comparison Between Up- and Downflow Anaerobic Fixed Film Reactors of Varying Surface-to-Volume Ratios For the Treatment of Bean Blanching Waste, Proceedings of the 34th Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 319-325. van den Berg, L. and Lentz, CP., 1978a, Factors Affecting Sedimentation in the Anaerobic Contact Fermentation Using Food Processing Waste, Paper presented at the 33rd Industrial Waste Conference at Purdue University, Lafayette, Indiana, Ann Arbour Science, 185-193. van der Meer, R.R., 1980, Anaerobic Treatment of Wastewater Containing Fatty Acids in Upflow Reactors, Water Research, Vol. 14, 389-394. van der Meer, RR. and Heertjes, P.M., 1983, Mathematical Description of Anaerobic Treatment of Wastewater in Upflow Reactors, Biotechnology and Bioengineering, Vol. XXV, 2531-2556. Anaerobic Digester, Agricultural Wastes, Vol. 5, 37-50. Walker, J.F. Jr., Handier, C.W., Genung, R.K., Patton, B.D. and Kowalchuk, M. 1981, Biological Treatment of Nitrate Wastewater Using Fluidized-Bed Bioreactors, Biotechnology and Bioengineering Symp. No. 11, 415-427. Wastewater Engineering: Treatment Disposal Reuse, Metcalf & Eddy, Second Edition, McGraw-Hill Book Co., 1979. Wilson, L. E., 1981, Digestion Design Concepts, In: Fuel Gas Production From Biomass, Vol. II, Wise, D., (editor), CRC Press, Boca Raton, Florida, 63-109. Woods, J.L and O'Callaghan, J.R.,1975, A Theoretical Analysis of Microorganism Cultivation in Cyclic Batch and Recirculating Plug Flow Processes, Biotechnology and Bioengineering, Vol. XVTI, 779-784. Young, J.C, 1983, The Anaerobic Filter — Past, Present and Future, Proceedings of the Third International Symposium on Anaerobic Digestion, 14-19 August 1983, Boston, Mass., 91-106. Young, J.C, 1985, Operating Problems With Wastewater Filters, Journal WPCF, Vol. 57, No. 1, 22-29. Young, J.C. and McCarty, P.L., 1969, The Anaerobic Filter for Waste Treatment, Journal WPCF, Vol. 41, No. 5, Part 2, R160-R173. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

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.831.1-0097233/manifest

Comment

Related Items