Nitrifikasyon-denitrifikasyon kinetiğinin deneysel karakterizasyonu

Abstract

Atıksu arıtma süreci incelendiğinde, yalan geçmişe kadar yürütülen biyolojik arıtma çalışmalarının daha çok organik karbon giderimine yönelik olduğu, azotlu bileşiklerin giderilmesi ile pek fazla ilgilenilmediği görülmektedir. Azotlu bileşiklerin alıcı ortamlarda sebep oldukları zararlı etkiler nedeniyle azot giderimine ancak son yıllarda Önem verilmeye başlanmış, karbon-azot gideriminin birlikte gerçekleştirildiği biyolojik nitrifikasyon-denitrifikasyon sistemlerinin en ekonomik çözüm olarak ortaya çıktığı görülmüştür. Bu sistemlerin yaygın olarak kullanımı nitrifikasyon-denitrifikasyon mekanizmalarının kinetik esaslarının ve sistem bileşenlerinin ortaya konmasını gerektirmektedir. Bu çalışmada nitrifikasyon-denitrifikasyon mekanizmalarının kinetik esaslarının araştırılması, ototrof ve heterotrof organizmaların maksimum çoğalma hızlarının belirlenmesi ve sistem bileşenlerinin ortaya konması hedeflenmiştir. Bu hedef doğrultusunda istanbul evsel atıksuyu, sentetik atıksu ve endüstriyel atıksular (et, süt, şekerleme endüstrileri) ile deneysel çalışmalar yürütülmüştür. Birinci bölümde genel olarak azotlu bileşiklerin alıcı ortamlarda yarattığı problemler üzerinde durulmuş, çalışmanın amaç ve kapsamı tanımlanmıştır. İkinci bölümde nitrifikasyon ve denitrifikasyon mekanizmalarının esasları detaylı olarak ele alınmış, stokiometrik ve kinetik ifadeleri ortaya konmuştur. Geniş bir literatür araştırması ile bu konuda yapılmış olan çalışmalar hakkında bilgi verilmiştir. Üçüncü bölümde ototrofik ve heterotrofik mikroorganizmaların maksimum spesifik çoğalma hızlarının belirlenmesi için kullanılan yöntemlerin kritiği yapılmış ve bu mevcut yöntemlerin deney ve değerlendirme aşamasında uygulanacak bazı değişiklikler ile geliştirilmesi önerilmiştir. Dördüncü bölümde ototrof ve heterotrof organizmalar ile yürütülen deneysel çalışmanın düzeni ve analiz yöntemleri açıklanmıştır. Beşinci bölümde ototrof ve heterotrof organizmalar ile çeşitli atıksularda yürütülen deneylerin sonuçlan detaylı olarak verilmiştir. Altıncı bölümde deney sonuçlan atıksu bazında değerlendirilmiş, literatürde mevcut olan veriler ile karşılaştırılmıştır. Yedinci bölümde sonuçlar ve öneriler tartışılmıştır.

Wastewaters are treated to a great extent mainly by biological wastewater treatment systems before being discharged in receiving waters. The system most commonly used is the activated sludge. Until 1970's these systems were designed and operated only for the removal of organic carbon. The observations and experience indicate that removal of organic matter is not enough to solve all the problems encountered in maintaining the quality of receiving waters for specific benefical uses. Most of these problems, such as eutrophication, excess depletion of oxygen, fish toxicity, etc., are readily associated with particular forms of nitrogen. Nitrogenous materials may enter the aquatic environment from either natural or man-caused sources. Natural sources of nitrogenous materials include precipation, dustfall, nonurban runoff and biological fixation. Man-caused sources may be listed as runoff from urban areas, subsurface drainage from agricultural lands, municipal and industrial wastewaters. Various forms of nitrogen present in a wastewater discharge can be undesirable in receiving waters for several reasons: as free ammonia, it is toxic to fish and many other aquatic organisms; as ammonium ion or ammonia, it is an oxygen-consuming component which will deplete the dissolved oxygen in receiving waters; in all forms, nitrogen will be available as a nutrient to aquatic plants and consequently contribute to eutrophication; as the nitrate ion, it is a potential public health hazard in water consumed by infants. On the basis of scientific evidence provided, nitrogen control has become increasingly important in water quality management. This new approach also triggered efforts to explore the merit and the potential of biological treatment, especially the activated sludge process, as a biochemical tool to secure the necessary conversion and removal of nitrogen forms. Nitrogen naturally exists in various compounds with a valence ranging from -3 to +5. Transformations of the nitrogen forms resulting in valence changes are associated with metabolic activities of different types of organisms. Oxidation of ammonia first to nitrite (N02~) and then to nitrate (N03~) is called nitrification, which is carried out by the autotrophic species Nitrosomonas and Nitrobacter, respectively. Conversely, the reduction of nitrate to molecular nitrogen by heterotrophic microorganism species is named denitrification. The effluent guidelines and standarts to protect the receiving waters promote nitrification- denitrification as a very feasible treatment process capable of ensuring the necessary level of conversion and removal of nitrogen forms. xvii For design and operation of a nitrification-denitrification system achieving simultaneous carbon and nitrogen removal, different mathematical models are used. Development of these models requires a complicated conceptual approach. It involves identification of all carbonaceous and nitrogenous components, correct assessment of the stoichiometric relationships between those components, and definition of the rate expressions for all the aerobic and anoxic processes. Within the framework of this study, the most critical design parameters in nitrification-denitrification process are experimentally surveyed along with the conventional characterization and COD fractionation for i) Istanbul domestic sewage, ii) synthetic waste, iii) a meat processing plant effluent, iv) a dairy effluent, v) a confectionary effluent, vi) different combinations of domestic-synthetic and domestic-industrial wastewaters. The critical design parameters considered include maximum spesific growth rate for autotrophic and heterotrophic biomass, correction factors for anoxic conditions and endogenous respiration rate. Maximum Specific Growth Rate for Autotrophic Biomass The maximum specific growth rate for autotrophic biomass, jxA is the most critical parameter in the modelling and design of nitrification systems, as it plays a dominant role on the magnitude of the washout sludge age for nitrifiers. The value of this kinetic coefficient is very much dependent on wastewater characteristics; therefore, it should be determined specifically for the wastewater of interest. Activated sludge reactors designed for nitrification utilize a mixed culture of heterotrophic and autotrophic biomass. For this reason, experimental techniques developed for pure cultures and relying on the evaluation of autotrophic growth by direct measurements of nitrifiers cannot be used for such systems. For suspended growth biological nutrient removal systems, the generally adopted approach is to determine (xA by monitoring the concentration of oxidized nitrogen, S^ in batch reactors, mainly because S^ is the only parameter solely related to autotrophic growth and batch systems offer a simpler interpretation of the reaction kinetics for experimental evaluation. Based upon the growth kinetics of nitrifying biomass, the following expression is applicable for the kinetic description of such systems: where SNO0 is the initial Nox concentration at the beginning of the test, X^ is the concentration of autotrophic biomass initially added to the reactor, YA is autotrophic yield coefficient and bA is autotrophic decay coefficient. Since X^, YA and bA xvni cannot be separately determined during the experiment, it may be convenient to use an experimental setup with sufficiently low values of S^ and X^ so that the related terms may be neglected (Antoniou et al., 1990). In this case, the above expression may be expressed in the following linear form: In S" = In ^^- XA + (fi.A - bA) t (2) YA VA ~bA The value of £iA - bA is then obtained as the slope of this linear function; jxA is then estimated with a reasonable assumption for bA. However, a careful evaluation of this method shows that it is not generally mathematically justifiable to accept the simplifying assumptions leading to the above procedure, because (i) even for very low Sjjoo and X^ values, the initial concentration of oxidized nitrogen and the term including XA0 will not be equal and cancel out, and (ii) neglecting SNO0 as an initial condition will lead to the same values for the slope of the linearized expression at different temperatures. Within the framework of this study, the proposed new approach defines a "curve fitting" procedure in order to minimize the errors in the determination of jxA with the same experimental data. For this new procedure, expression (1) is arranged to yield: AA0 &A By setting, Y. û. - b.... k = -± LA ± (4) XA0 ftl the following logarithmic expression is obtained; k K5» " Ssoo) * + 1] = (A, - bA) t (5) In this expression k is a constant for a given experimental setup. The value of jxA-bA with the highest correlation coefficient is then computed using a search technique known as the Fibonacci search with different values for k. Experiments were conducted in batch reactors seeded with an initial biomass concentration of around 50 mg SS/1 from a mixed culture of the fill and draw unit operated at a sludge age of 10 days. The proposed method is successfully tested on Istanbul domestic sewage and on a synthetic substrate to depict possible inhibition effects. Similar studies are also carried out alone on samples from a meat processing plant effluent and on different combinations with domestic sewage. Characteristics Associated with Respirometric Measurements In this study, respirometric procedures, based on OUR (Oxygen Uptake Rate) and NUR (Nitrate Uptake Rate) measurements in batch reactors were adapted to assess xix the maximum specific growth rate of heterotrophic biomass under aerobic and anoxic conditions. The data of the same test were also used to calculate the readily biodegradable COD fraction. The most important point in NUR experiments is to measure the reduced electrons from nitrate to nitrogen gas, as an accurate representation of the amount of oxidized substrate. Any accumulation of nitrite means, that the second step of the denitrification process is stopped somehow and electrons used up in this step (3/5 electrons) can not be further transferred. To determine the real electron transfer rates under anoxic conditions the nitrate uptake rates should be corrected as AN03~-N- 0.6 AN02"-N. Comparing the results of the OUR method mentioned above with another procedure, which gives £H independent from XH, also based on OUR measurement the active fraction of biomass was calculated. The NUR measurements were also used to calculate denitrification rates. Endogenous respiration rate was determined with OUR measurements. Maximum Specific Growth Rate for Heterotrophic Biomass A respirometric method for the evaluation of (% has been developed by Ekama and Marais (1986). It is the procedure involving an aerated batch reactor where a preselected volume of wastewater, V^, is mixed with a preselected volume of mixed liquor, V^, having a total biomass concentration of XT'. Neglecting the endogenous respiration, the initial level of OUR observed in the test, OURj [mg 02/l.h] is proportional to u,H as indicated by the following expression: £" = J*_ OUR. (V"*+Vww) 24 hid (6) where yield coefficient YH is expressed as [mg VSS/mg COD]. The same procedure is also applied to an anoxic batch reactor to determine the specific growth rate of denitrifiers. The only difference is the electron acceptor, nitrate nitrogen. Nitrate respiration measurements are carried out in a closed reactor. 2.86 y" (V^+VwJ ÇLm = £ NUR, - * - T 24 hid (7) 1 *h L XTV, ml Both expressions include the active fraction of the mixed liquor, fa, which must be previously known. Some procedures are described for the estimation of this fraction. On the other hand Kappeler and Gujer (1992) developed another aerobic procedure independent of XH which involves a batch test with centrifuged wastewater and a very small amount of biomass corresponding to an initial COD/VSS ratio of 4. A linear relationship is defined between the logarithms of relative OUR (OUR/OURq) and time, with the slope equal to jxH-bH. xx (£" - bH) t =ln [2HEl] (8) \rH "j l0URQ In this study the endogenous respiration is neglected to make both procedures similar to each other. a t = In [-^] (9) Upon equalization of both expressions (6 and 9) the active fraction can be obtained as fallows; f = _?_ OURx K « v** 24 hid (10) I-?* û Xi V Correction Factors for Anoxic Conditions Two important parameters, T]g and %, reflect changes in the rates of microbial growth and substrate utilization when the system is switched from aerobic conditions to anoxic conditions. These values can be estimated by measuring OUR and NUR in two parallel tests. Tig is defined as a ratio of maximum specific growth conditions, whereas îih indicates the ratio of hydrolysis phases. ûT NUR,"",. T] = rSB. = 2.86 I (11) * A* OUR, NUR. Tj = 2.86 * (12) * OURh Readily Biodegradable COD In the aerobic batch test, the OUR initially measured is associated with both readily and slowly biodegradable substrate in the sewage sample and endogenous respiration. The initial OUR may stay constant during a certain time where the readily biodegradable substrate S^ is high enough to sustain maximum growth rate, with the selection of a suitable F/M ratio. After the consumption of S^, the OUR is expected to drop to a lower level. At this time OUR is correlated only to hydrolysed substrate and endogenous respiration. The readily biodegradable substrate S^, in the wastewater is calculated with this relationship: Sso = 7V A° (13> XXI where AO is the difference between total respiration and respiration due to hydrolysed substrate and endogenous metabolism. The same parameter may also be calculated from the corrected N03"-N profile in the anoxic reactor. In this test, the initial N03"-N utilization is faster, due to oxidation of readily biodegradable substrate. The amount of N consumed during this period, if corrected for the interference of the hydrolysed, substrate may be used to calculate Sso with the following relationship: 1 lH Endogenous Respiration Rate In a batch aerobic digester, under endogenous decay conditions containing no external substrate, OUR = IAS (l-fE)bHXH (15) and X = X e'*"»1 (16) AH AH0 e V / Substituing the value of XH above, In OUR = In [1.48(1 -fj bH Xm) - bH t (17) shows that the slope of a plot of In OUR versus time yields the value of 1%. Results of Istanbul Domestic Sewage The experimental survey program to characterize the Istanbul domestic sewage was carried out approximately 3 years including the experiments with autotrophs and heterotrophs. The results show that for raw sewage, average concentrations of 560, 64, 43 and 11 mg/1 can be associated with COD, TKN, NH4+-N and TP parameters respectively; the corresponding COD/N ratio is computed as 8.8. The maximum specific growth rate of nitrifiers was experimentally determined for 12 domestic sewage samples selected to represent different wastewater properties. The results show that jlA - bA levels at 20 °C were highly variable within a range of 0.24-0.52 d~\ with an average value of 0.38 d"\ No correlation was possible or justifiable between jxA-bA values and any of the conventional parameters characterizing domestic sewage samples. At 10 °C, which is the critical wastewater temperature in winter for the design of treatment systems, the average value of ftA-bA was observed to drop to 0.14 d"1, approximately one third the level associated with 20 °C and its variation for different sewage samples stayed within a narrower xxn range of 0.10-0.17 d"1. The results obtained also confirm the validity of this expression yielding an average value of 1.098 for the temperature coefficient, 6, quite in accordance with the range of 1.08-1.123 reported in the literature. Experiments which have been carried out with synthetic substrate parallel to domestic sewage reactors to evaluate the existence and the extent of inhibitors indicate similar levels of ftA-bA in the range of 0.25-0.52 d"1. These observations lead to the conclusion that inhibition, if present, is of no practical importance for Istanbul domestic wastewaters. It is determined that the maximum specific growth rate for heterotrophic growth rate [% of Istanbul domestic wastewater vary in the range of 2.7-6.5 1/day, with an average of 4.6 1/day. For anoxic conditions, values of around 0.88, consistently calculated for r\h are significantly higher than the level typically suggested for this parameter by the IAWPRC Task Group (Henze et al., 1987), but support the findings of Oles and Wilderer (1991) and Kristensen et al.(1992). An average value of T|g 0.59 is calculated. Readily biodegradable substrate Sso is calculated to vary in the range of 12-92 mg COD/1 with an average of 50.5 mg COD/1 on the basis of NUR test; the aerobic test yielded an average value of 50 mg COD/1 within a range of 21-86 mg CODA. The readily biodegradable fraction is around 9%. The denitrification rates in maximum growth, hydrolysis and endogenous decay phases are calculated 0.029, 0.014 and 0.010 mg N/mg VSS.h on the basis of active biomass, respectively. The endogenous decay rate is determined as 0.24 1/day. Additionally, experiments with domestic-synthetic waste mixtures were carried out to investigate the impact of S^ on OUR and NUR measurements. Syntetic waste as defined by Henze (1992) represents the readily biodegradable substrate in domestic sewage. This part of the study indicated that S^, externally added, could be recovered successfully with the electron uptake rate measurements. Results of the Meat Processing Plant Effluent The characterization program of the meat processing plant effluent was carried out for approximately 6 months including the experiments with autotrophs and heterotrophs. The results show that for raw sewage, average concentrations of 2130, 158 and 80 can be associated with COD, TKN and NH/-N parameters respectively; the corresponding COD/N ratio is computed as 13.5. Experimental assessment of ftA for this industrial wastewater was also realized with a monitoring program extended over three months. A point of interest in connection with this evaluation is the observation that the meat processing wastewater yields xxm markedly higher p,A- bA values as compared to domestic sewage. The average flA- bA level for the meat processing effluent is 0.59 d"1 at 20°C, 80 % higher than 0.32 d"1 calculated as the average value for domestic sewage at the same temperature. The same observation remains also valid for the experiments at lower temperatures: At 10 °C for example, average values of 0.28 d"1 and 0.13 d"1 have been found to characterize this coefficient for the meat processing waste and domestic sewage, respectively. Another significant observation relates to the experiments conducted on mixtures of sewage and meat processing wastes, where the meat processing waste appeared to control the rate of nitrification with a higher autotrophic activity. The maximum specific growth rate of heterotrophs is determined to be 3.8 1/day (average) within a range of 3.6-4.2 1/day. The t|g and T|h values are both calculated as 1.5, greater than 1, which is the maximum value given in literature. Sgo varies in the range of 304-416 mg COD/1 with an average value of 374 mg COD/1 on the basis of NUR test. The readily biodegradable fraction is around 16%. The denitrification rates in maximum growth, hydrolysis and endogenous decay phases are calculated 0.064, 0.027 and 0.016 mg N/mg VSS.h on the basis of active biomass, respectively. Results of the Dairy Industry Effluent The results show that the influent to biological process has average concentrations of 1745, 75 and 23 can be associated with COD, TKN and NH/-N parameters respectively; the corresponding COD/N ratio is computed as 23. The maximum specific growth rate of heterotrophs is calculated to be 3.1 1/day (average) within a range of 2.9-3.3 1/day. The T|g and T|h values are both determined greater than 1, as 1.02 and 2.25, respectively. Sgo is calculated in the range of 394-425 mg CODA with an average value of 406 mg CODA on the basis of NUR test. The readily biodegradable fraction is around 23%. The denitrification rates in maximum growth, hydrolysis and endogenous decay phases are calculated as 0.079, 0.039 and 0.028 mg N/mg VSS.h on the basis of active biomass, respectively. The endogenous decay rate is determined as 0.14 1/day. xxiv Results of the Confectionary Industry Effluent The wastewater taken from the influent of a biological system analized for conventional parameters. The results show that average concentrations of 3790 and 13 and 23 can be associated with COD and TKN parameters respectively, with a COD/N ratio of 292. The purpose of the NUR measurements in this part of the study is to outline the denitrification potential in a possible common treatment scheme where it is handled together with a domestic sewage or a wastewater with a high nutrient content. (Ih is calculated as 4.1 1/day, r\s and % values are determined 0.86 and 1.7 respectively. Sgo is calculated with an average value of 720 mg COD/1 on the basis of NUR test. The readily biodegradable fraction is around 19%. The denitrification rates in maximum growth, hydrolysis and endogenous decay phases are calculated to be 0.111, 0.074 and 0.006 mg N/mg VSS.h on the basis of active biomass, respectively. The endogenous decay rate is determined as 0.24 1/day. The experimental results summarized above provide the required, scientific background so far unavailable in the literature for a comprehensive evaluation of the behaviour of different types of wastewaters in biological treatment. It is also believed that these results may be interpreted for the clarification of the fundamental issues in the kinetic description of aerobic and anoxic processes. They also confirm the common understanding that growth characteristics of autotrophic and heterotrophic biomass are very much wastewater-specific and should be separately determined for each case.