Aktif çamur sistemlerinde substrat inhibisyonunun modellenmesi

Abstract

Bu çalışmada, endüstriyel atıksuların biyolojik arıtımında karşılaşılan temel problemlerden biri olan substrat inhibisyonu mekanizmasını açıklayan yeni bir matematik model geliştirilmiştir. Birinci bölümde, aktif çamur sisteminde substrat inhibisyonunun önemi vurgulanmış, öne sürülen modellerin bu olayı açıklamada yetersiz kalması nedeniyle yeni bir model geliştirilmesi gereği ortaya konarak çalışmanın amaç ve kapsamı açıklanmıştır. İkinci bölümde, aktif çamurda substrat giderim esasları ve kinetiği mevcut bilgiler dahilinde özetlenmiştir. Çoğalma kinetiğinin enzim mekanizması ile ilişkisi ve yaygın olarak kanlaşılan inhibisyon türleri hakkında genel bilgiler verilmiş, bu inhibisyon türleri için enzim mekanizmasından yola çıkarak oluşturulan modeller açıklanmıştır. Üçüncü bölümde, substrat inhibisyonunun tanımı yapılmış, substrat inhibisyonu ile ilgili gerek enzim mekanizması bazlı matematik modeller ile gerekse deney sonuçlarına göre en uygun eğriyi geçirme metodu uygulanarak oluşturulmuş ampirik modeller özetlenmiştir. Ayrıca, ürün oluşumu ve ürün inhibisyonu ile ilgili modeller ve çalışmalar da Özetlenmiştir. Substrat inhibisyonu ile ilgili yapılmış çeşitli deneysel veya teorik bazlı çalışmalar özetlenmiş ve değerlendirmeleri yapılmıştır. Dördüncü bölümde, aktif çamurda substrat inhibisyonunun ürün oluşum adımı ile kısıtlandığı kabulü ile oluşturulan yeni bir model tanımlanmıştır. Önerilen modelin iki aşamada teorik bazlı irdelemesi yapılmış, elde edilen sonuçlar özetlenmiştir. Beşinci bölümde, bu çalışmada önerilen modelin literatürdeki deneysel verilerden yararlanılarak kontrolü yapılmıştır. Bu kontrolde kullanılan literatür verilerinin seçim esasları ve kontrol yöntemi hakkında bilgi verilmiş, elde edilen sonuçlara göre modelin değerlendirmesi yapılmıştır. Altıncı bölümde, oluşturulan yeni modelin teorik irdeleme ve deneysel sonuçlarla karşılaştırılmasından yola çıkılarak varılan sonuçlar ve öneriler özetlenmiştir.

During the last two decades much effort have been devoted to activated sludge modelling which almost totally based on Monod kinetics. Research has been focused on COD fractions and their expressions in a rather complicated forms which in turn require the assessment of at least several new kinetic parameters or coefficients representing growth conditions. However, the basic assumption is always that the Monod kinetic is valid. On the other hand, for the present the major environmental pollution control issue is the treatment of industrial wastewaters to meet the discharge standards which are becoming increasingly stringent. The problem with the activated sludge treatment of industrial wastewaters is the type of the substrate which frequently degraded in a way different from the Monod kinetic. There are several reasons for this behaviour however one of the most important phenomenon involved is substrate inhibition. Effect of substrate inhibition on the treatability of many industrial wastewaters is significant and the models employed to account for the substrate inhibition effect are numerous and diversified. Except for the Haldane model which is the most commonly used mathematical model to express substrate inhibition effect, no physical or empirical bases or a systematic for the use of these models have been developed. The models developed so far were mostly concentrated on either specific substrate or specific condition for a given substrate. As the importance of toxic organic material control increases, the prediction of the behavior of such substances in activated sludge which is the most common method applied to wastewaters have been becoming important. Therefore, a need for a basic approach to the substrate inhibition modeling demonstrated itself very strongly. Several substrate inhibition kinetics were proposed in the literature for understanding the behavior of especially toxic pollutant causing the inhibition effect on microorganisms. The models in the literature can be grouped in two categories. The first one was the Haldane model and the second were mostly empirical models derived through enzyme kinetics or purely experimental models. All of these models have been investigated using several toxic substances such as phenol, PCP, etc. Although these models can be fitted with the experimental data in most cases applying best-fit method, none of them explains the mechanism of the substrate XVI inhibition and they do not provide a general basis to explain for all experimental observations.. In this study, the substrate inhibition phenomenon in the activated sludge process has been throughly investigated and a new model explaining the most important aspect of the inhibition was proposed. The model has been analytically analyzed and experimentally tested. In the first chapter, the scope and objective of the study were emphasized. In the second chapter, substrate removal mechanism in the activated sludge process and types of inhibition phenomenon have been defined and basic equations were stated. Chapter third was entirely devoted to the review of the literature related to substrate inhibition. A detailed literature survey and analyses were made. All the proposed models and their experimental results have been reviewed and criticized. Conclusion of this chapter was that most of the models did not represent the physical meaning of the substrate inhibition mechanism. The Haldane model which was the most commonly applied model had its weakness due to the general response of the system as well as its inability to predict the substrate concentration at which the growth ceases. In the fourth chapter, starting from the conclusions of third chapter and considering the mechanism of substrate inhibition, a new model has been derived. The model was based on the fact that substrate inhibition was caused by a kind of suppression effect of product(s) an related to enzyme kinetics were employed to make a simulation for the activated sludge kinetics. The model equation was defined as follows: M MmmSKs+S + krP Mmsx" PJ_S (1) k, in which: S P urns Ks ump Kp ki substrate concentration, (M/L3) product(s) concentration, (M/L3) maximum specific growth on substrate, (1/T) half saturation concentration for substrate, (M/L3) maximum specific growth on product(s), (1/T) half saturation concentration for product(s), (M/L3) reverse rate constant of product xvn . For almost all the data set a satisfactory fit to the model was obtained. The general tendency of the model which was lower growth rates at low substrate concentrations did not caused greater deviation from the experimental data than those of Haldane and Monod model. Therefore the accuracy of the model at lower concentrations was also found to be quite realistic. The kinetical constants determined for each data set were fairly close to the ones determined for the Haldane model and this fit especially quite obvious for urns and Ks.. Experimental evaluation also indicated that the general behavior of the function expressed in the theoretical treatment of the model was in accord with the experimental fit and sensitivity of especially ump, Kp, ki, X was quite important in determining the inhibitor behavior of the substrate.. The most striking future of the experimental evaluation was that the proposed model exhibited almost exact fit after critical substrate concentration and for most of the data set experimental values were followed the theoretical model curve. For same of the experiments which provided very low growth rates at high substrate concentrations also demonstrated the functions realistic behavior as the growth rate approach to zero. Some recommendations for advance studies may be done were given. XX . For almost all the data set a satisfactory fit to the model was obtained. The general tendency of the model which was lower growth rates at low substrate concentrations did not caused greater deviation from the experimental data than those of Haldane and Monod model. Therefore the accuracy of the model at lower concentrations was also found to be quite realistic. The kinetical constants determined for each data set were fairly close to the ones determined for the Haldane model and this fit especially quite obvious for urns and Ks.. Experimental evaluation also indicated that the general behavior of the function expressed in the theoretical treatment of the model was in accord with the experimental fit and sensitivity of especially ump, Kp, ki, X was quite important in determining the inhibitor behavior of the substrate.. The most striking future of the experimental evaluation was that the proposed model exhibited almost exact fit after critical substrate concentration and for most of the data set experimental values were followed the theoretical model curve. For same of the experiments which provided very low growth rates at high substrate concentrations also demonstrated the functions realistic behavior as the growth rate approach to zero. Some recommendations for advance studies may be done were given. XX The model was analytically analyzed. The analyses showed that general response and character of the equation (1) fitted the generally observed substrate inhibition behavior. On the other hand, model predicts the substrate concentration at which the growth ceases. The prediction of the model especially beginning from the point where the substrate inhibition began to demonstrate itself was significantly better than other models proposed so far. Numerical analyses of the model were also made. A wide range of the kinetic constants representing experimental values found for phenol has been selected. The evaluations were made in parallel with the Haldane model to make a comparison between the models as well as to express the advantages of proposed model. The analyses were made in two steps. In the first step, the critical substrate concentration at which inhibition began was taken the same both of the proposed model and Haldane model. The analyses showed that the determining variable was the product concentration whose effect dependent mostly on kinetic parameters of ump and Kp for the cases where ums and ump were close enough the two models could be parallel provided that higher Kp values were applied. On the other hand, as the ums and ump were apart from one another other constants gained importance in providing a parallelism between the two models. If critic values of substrate concentration for proposed model was accepted to deviate about 10, 20, 30 % than that of the Haldane model, the fit of the kinetical parameters between two model were found to be at even lower values of critical substrate concentrations. Decreasing inhibition constant values provided a better match between the models. In the second step, the effect of um on the behavior of model was investigated, um was the most effective parameter in determining the character of the function. Analytical study has shown that as ums values increased the model results deviated from the Haldane model and for the reverse, better matches could be obtained. The sensitivity of the deviation was of course dependent on other kinetical parameters. In the fifth chapter, experimental evaluation of the model was made. Experimental data were obtained from the literature in such a way that a wide spectrum of experimental data in terms of both microbial culture and substrate type and concentrations could be accounted for. Phenol was the most common substrate investigated in the literature however in addition to phenolic substances such as PCP, OCP, 2,4-DCP, 2,4-DNP were also evaluated. The concentration range for phenol were 500-1000 mg/1 and for phenolics were 10-300 mg/1. Both batch and continuous cultures were evaluated. The method for the evaluation was in the main a curve fitting method since the proposed model could not be linearized. The curve fitting procedure was non-linear least squares method which was solved using MATCAD program. 13 data set of which 8 were for the phenol were used for the experimental evaluation.