Identification and modelling of aerobic hydrolysis in activated sludge systems, application of optimal experimental

Aktif çamur sistemlerinde aerobik hidroliz kinetiğinin tanımlanması, modellenmesi ve optimal deney tasarımı uygulaması

Abstract

Appropriate determination of accurate COD fractionation together with degradation kinetics of organic matter has a prime importance on the design and operation of activated sludge systems. In general, regarding the total biodegradable COD, raw domestic wastewaters approximately contain readily biodegradable and slowly biodegradable substrates at fractions around 30% and 70%, respectively. On the other hand, compared to domestic wastewaters, industrial wastewaters contain much higher fraction of slowly biodegradable substrate which directly influence the effluent quality of treatment plants. In this respect, the determination of slowly biodegradable matter as well as its degradation characteristics are crucial in terms of wastewater treatment plant design and operation. Hydrolysis process has already been known as the rate limiting step in organic carbon removal from industrial and domestic wastewaters. Considering the effluent quality, hydrolysis mechanism also plays a dominant role in delicate balance of electron donor/electron acceptor ratios in biological nutrient removal type activated sludge systems as an important carbon source. In addition to that, sludge production from activated sludge plant is also affected by the nature of slowly biodegradable matter. In parallel to the vast developments in activated sludge modeling, respirometry has always been effectively utilized as a convenient tool for influent wastewater characterization which can be regarded as a corner stone of activated sludge modeling. To date, numerious respirometric tests have been applied for gathering information on the stoichiometry and kinetics of biodegradable substrates in raw wastewaters. However, the proposed methodologies were mostly devoted to the estimation of growth associated parameters such as maximum growth rate, active fraction of biomass, heterotrophic yield and half saturation growth constant for heterotrophs etc. In this study, surface-saturation type hydrolysis kinetics was investigated based on short-term oxygen uptake rate measurements.An identifiability study were performed in order to find out best identifiable parameter groups from respirometric data with the aid of non-linear degradation model. Basically, the model has been constitued using the reactions of (i) aerobic heterotrophic growth. (ii) hydrolysis of particulate matter and (Hi) endogenous decay processes. Basically, the model has 7 parameters to be estimated from a single respirogram. The model identification procedure comprises the theoretical and practical identifiability studies. In theoretical identifiability study, identifiable combinations of model parameter that can be extracted from available data is studied for a certain model. Non-linear model under study was linearized with the aid of Taylor Series Expansion method. By neglecting the growth of heterotrophs under low initial F/M ratio, 6 parameter combinations were found to be theoretically identifiable from batch respirogram. On the other hand, the maximum rates governing the growth ($mu _H$) and hydrolysis ($k_h$) were found to be individually identifiable if considerable growth of heterotrophs are taking place during the course of the experiment. From the identifiability study, it was also found that all parameters combinations include the heterotrophic yield coefficient, $Y_H$. In addition, the parameter combinations containing hydrolysis parameters always include the initial active heterotrophic biomass, $X_{Ho}$ as a state variable. In practical identifiability, the identifiable parameter combinations of a selected model were estimated. In this study, the information of the experiments were simulated for different initial F/M (Food/Microorganism) ratio by comparing the amount of information as well as the correlation degree among the estimated parameters. The information contents of the experiments were evaluated on the basis of Optimal Experimental Design (OED) methodology. In this regard, the effects of initial conditions on information content of experiments was evaluated via comparing the scalar functions of Fisher Information Matrix (FIM). These scalar functions were selected as the D-Criterion and E- Criterion which summarize the information volume and the correlation degree among parameters, respectively. Finally, it was found that applying lower initial F/M ratio increases the information content of the experiment, on the other hand it also increases the correlation among the estimated parameters.

Özet

Atıksularda Kimyasal Oksijen İhtiyacı (KOİ) fraksiyonlarının ve bunlara ait giderim kinetiğinin belirlenebilmesi aktif çamur tasarımı ve işletilmesi açısından büyük önem taşımaktadır. Özellikle, yavaş ayrışan organik maddenin atıksularda yüksek miktarlarda olduğu bilinmektedir. Genelde, ayrışabilen KOİ fraksiyonlarını hızlı ve yavaş ayrışan KOİ olarak sınıflandırmak mümkündür. Ancak, atıksuyun tipi ve özelliğine bağlı olarak bu fraksiyonlar değişkenlik gösterdiği gibi çözünmüş, partiküler veya çökelebilir formlarda bulunabilmektedir. Hidroliz prosesi aktif çamur tesislerinin işletilmesinde çıkış suyu kalitesinin, fazla çamur oluşumu, oksijen ihtiyacı ve nutrient giderimi açısından büyük rol oynamaktadır. Özellikle deri, tekstil vb. endüstriyel atıksularında olduğu gibi çok yüksek miktarda yavaş ayrışan çözünmüş formdaki organik maddeye ait biyolojik giderim atıksu arıtma tesisi çıkış suyu kalitesi açısından önem taşımaktadır. Bu çalışma kapsamında, kesikli yürütülen respirometrik deneylerin kullanılması ile çoğalma ve hidroliz kinetiğine ait sistem tanımlanması yapılmıştır. Seçilen aktif çamur modeli çoğalma, hidroliz ve ölüm olmak üzere 3 proses ve 7 parametreden oluşmaktadır. Sistemin tanımlanması, teorik ve pratik sistem tanımlama adımlarını kapsamaktadır. Teorik sistem tanımlamada, Taylor Serileri ile lineerize edilen modelden respirometrik verileri kullanarak hangi parametre gruplarının elde edilebileceği bulunmuştur. Pratik sistem tanımlamada ise gerçek veriler kullanılarak, parametre tahminleri yapılmış ve güvenilirlik aralıkları belirlenmiştir. Optimal Deney Tasarımı simülasyonlarından, deneyin başlangıç F/M (Substrat/Mikroorganizma) oranının azaltılmasının deneyin güvenilirliğini arttıracağı ancak bunun yanında parametreler arasındaki korelasyonu da arttıracağı kanıtlanmıştır.