Comportare in Proces Bioreactor.pdf 4 10

8/13/2019 Comportare in Proces Bioreactor.pdf 4 10

1/12

REV. CHIM. (Bucharest) 61Nr. 4 2010 http://www.revistadechimie.ro 407

Time Scale Behaviour of a Biological Wastewater Treatment Facility

MIRELA STOENICA, VASILE LAVRIC1*

University Politehnica of Bucharest, Chemical Engineering Department, 1-7 Polizu, 011061, Bucharest, Romania

The effects of recycle and purge ratios (as responsible for the two of the main time scales of the wastewatertreatment facility: the residence time of the biological reactors anoxic and aerobic, and the doubling timeof the microorganisms accountable for the biodegradation of substrates), on the performance of a biological

treatment process for wastewater are investigated in this study. A modified form of the ASM3 model withtwo-step nitrification-denitrification (ASM3-2N) [1] was used for the biological process. Three operatingtime windows with data collected for two weeks: dry weather, stormy weather and rainy weatherwere used to test the system at its largest time scale (http://www.benchmarkwwtp.org). The system understudy has two bioreactors in series, followed by a separator, a recycle (through which values the time scaleof the bioreactors are modified) and a purge, responsible for the time scale of the activated sludge. Recycleratios from 0.1 to 0.9 were used and purge ratios from 0.001 to 0.01. The time scale of the bioreactorsovercomes the time scale of the biological process, since the latter is, to some extent, determined by theformer. This means that the recycle and purge ratios are not independent so, when they are used as commandvariables to control the process, they should be treated accordingly.

Keywords: recycle ratio, purge ratio, time scale, wastewater treatment process, bioreactor

The characteristics of domestic wastewater are highlyvariable and complex in nature, due to the diversity of thepollutants. The main objective at any biological treatmentfacility is the reduction or, better elimination of itscontamination. Shock loads, high carbon load, presenceof toxic and inhibitory organics, solvents and inorganiccompounds, variability of the wastewater on both flow andcomposition diminish the treatment efficiency. The latterdepends on the type of the reactors used, their configurationin the treatment system and the associated operatingconditions adopted along with the nature andcharacteristics of wastewater being treated.

The biological treatment of wastewater depends uponthe complex, interdependent bacterial community whosestructure and composition is highly sensitive to suddenchanges in substrate composition, pH and temperatureand to certain toxic or inhibitory compounds. Basically,bacteria involved in the wastewater treatment belong totwo main groups: heterotrophs, whose main target is thecarbon based substrates expressed as carbon oxygendemand (COD), and autotrophs, targeting the nitrogenbased substrates. The bacteria have small growth rates ingeneral, but smaller for the autotrophs which also havepoor yields, meaning slow recovery after toxic shocks oroverloading; this requires increased biomass retention in

the bioreactors. As in all biological mediated processes,the performance of treatment facility is always a functionof the existing number of viable and active cells and awell-balanced biocenosis [2].

The heterotrophic organisms are responsible for CODdegradation under both aerobic and anoxic conditions.Under anoxic operating environment some heterotrophicbacteria are stimulated into utilizing nitrates and nitrites(resulted from nitrification) as final electron acceptors forcellular respiration in place of oxygen (complementaryprocess called denitrification) [3,4]. Zhao [5,6] showedthat heterotro-phic ammonia oxidation occurs when theC/N ratio is low and the organic carbon loading of acti-vated sludge is high.

Nitrification, the sequential transformation ofammonium to nitrate via nitrite, is typically the result of the

* email: [email protected]

activity of two distinct groups of autotrophic bacteria, i.e.,the ammonia oxidizing bacteria (AOB) and nitrite oxidizingbacteria (NOB). As oxidation of ammonia to nitrate is atwo step reaction in series, nitrite appears only when theammonia oxidation rate is higher than the nitrite oxidationrate. The decrease of the dissolved oxygen (DO)concentration causes the reduction of both oxidation rates.Nevertheless, NOB are more affected since their value ofthe oxygen affinity constant is higher than that of AOB.This is the reason for the observed nitrite accumulation forlower DO values [7,8]. AOB become dominant nitrifyingbacteria and NOB are inhibited or washout. The NOB areable to utilize both nitrite and organic carbon as energysources for growth. The influence of organic carbon on thenitrite removal performance and the community structureof NOB and heterotrophic bacteria is complex [9]. Manystudies presented the main factors affecting the nitritebuildup [8,10,11], such as higher temperature, short solidresidence time (SRT) [11], low DO concentration [12-14]and high free ammonia concentration [15]. Dissimilaritiesinto nitrifying populations are related to niche differentiationconcerning ammonium concentrations, system operation,and characteristics of wastewater. Moreover, thesecommunities maintain nitrifying capabilities whenconditions change such as sudden increases in ammonium

concentrations [16].In most cases, the mass of nitrifying bacteria is verysmall compared to the total organic sludge mass due tothe relatively low ammonium influent concentration andits low yield coefficient [17]; nitrification is generally a rate-limiting step in a biological nitrogen removal process. Amajor difficulty is to maintain adequate levels of nitrifiersin the aeration bioreactors. This is achieved using either aninternal separation device, like membranes, or a classicalexternal one, sedimentation followed by recycling.

Bacteria concentration increases with both residencetime and sludge recycling. Heterotrophs are responsiblefor the decomposition of dead cells and extracellularpolymeric substances (EPS) [18].When the ecosystem issupplied with organic carbon, the heterotrophs competewith the nitrifiers for ammonia and oxygen and are always


2/12

REV. CHIM. (Bucharest) 61Nr.4 2010http://www.revistadechimie.ro408

the winner due to their faster growth rate. Consequently,the nitrification efficiency is reduced in most cases [19].

The recycle ratio,,defined as the ratio between recycleand feeding flows (Q

rs/Qin), is a very important parameter,

which controls substantially the overall performance of thebiological treatment for any substrate, whatever thepending kinetics [20-23]. It affects the microbial ecology,hydraulic regime and characteristics of the operatingsystem by directly influencing the biological system. The

effects of recycle on the whole process are difficult tospecify beforehand and can be attributed to different timescales (growing characteristics) of the implied speciesapart from the operational factors. Recycle of treatedwastewater which mixes with the influent wastewaterprovides a method of balancing the hydraulic loading withthe carbon loading. A partial effluent recycle increases thereactor performance too. The net effect could be anincrease in the efficiency of substrate removal [24].

The purge ratio, , defined as the ratio between purgeand feeding flows (Q

p/Qin), plays a key role in shaping dead

cells and overall mass accumulation in the system.Removing a certain amount of sludge doesnt allow thebuildup of the particulate materials of any kind , thuslowering the concentration of dead cells and inert solids.

The mathematical modelVarious mathematical models for biological wastewater

treatment processes have been developed by th eInternational Association on Water Quality (IAWQ), manybelonging to the class of activated sludge models (ASM).The first, ASM1, considered as reference for design andassessment of advanced control strategies, includes thecarbon oxidation (organic matter consumption) by bothheterotrophs and autotrophs, nitrification (ammoniaoxidation to nitrite and subsequently to nitrate) anddenitrification (nitrate transformation till nitrogen gas) [25].

ASM3 is an extended improved version of ASM1,including the storage of organic substrates as a newprocess. The lysis process is replaced by an endogenousrespiration process [26]. Complete nitrogen removalinvolves nitrification and denitrification. Nitrite and nitrateproduced from nitrification are reduced to gaseous nitrogenby denitrifiers. For ASM3 models nitrification is a single-step process, considering the nitrite as a short-livedintermediate which is easily oxidized to nitrate. However,there are situations where considerable nitriteconcentrations may build-up in the system, which requiresa two-step nitrification for separate modelingof nitrite andnitrate fate. Such a model is ASM3-2N [1], which will beused subsequently in the overall model of the biological

treatment facility. The autotrophic biomass is split intoammonium-oxidizers (Nitrosomonas - AOBs), for the firststep of nitrification, and nitrite-oxidizers (Nitrobacter -NOBs), for the second step:

Nitrification first step (ammonia oxidation)

(1)

Nitrification second step (nitrite oxidation - ammoniabehaves as an inhibitor)

(2)

Considering nitrite into ASM3-2N also requires theseparate knowledge of the kinetics of ammonia- and nitrite-oxidizing bacteria together with the decay rates ofautotrophs, in order to reliably modeling nitrification [27].

The slowly biodegradable substrate in wastewater isfirst hydrolyzed to readily biodegradable substrate by theheterotrophs. The biomass can use it for simultaneousstorage and growth. When the readily biodegradablesubstrate is depleted (as low as the half saturationconcentration for the primary growth [28], the degradation(secondary growth) of the storage polymers takes place[29]. The intracellular storage polymers are taken intoconsideration, because these components are of a high

importance to microbial metabolism, effluent quality, andprocess dynamics [30].Denitrification is done by the facultative anaerobe

biomass (heterotrophs) that can remove organic carbonthrough anoxic respiration on nitrite or nitrate, which bothserve as electron acceptors in the absence of dissolvedoxygen.

Anoxic growth (denitrification) on nitrate:

(3)

Anoxic growth (denitrification) on nitrite:

(4)

The process generally occurs in two steps: the biologicaldegradation of pollutants, taking place in a series of twobiological reactors (the volume of the anoxic tank is 2000m3 while the volume of the aerobic tank is 4000 m 3),followed by the separation of the activated sludge in asettling unit (or clarifier), in which the solids settle to thebottom of the unit [31]. Activated sludge, along with mixedliquor, is recycled from the bottom of the clarifier into theanoxic bioreactor. The sketch of the system is given infigure 1, together with the pending notations.

Fig. 1. The sketch of the system, with the main notations

The state variables of the system (table 1) are dividedinto two categories; soluble components, whoseconcentrations are denoted by S, assumed to betransported by water, and particulate components, whoseconcentrations are denoted by X, assumed to be

associated with the activated sludge concentrated in thesettling tank. According to the ASM general philosophy,the unstructured COD is split into the required partitions[32]: 13% of inert soluble COD (SI), 22% of readilybiodegradable COD (SS), 11% of inert particulate COD (XI)and 54% of slowly degradable COD (XS).

The mathematical model is given by the overall andpartial mass balances together with the appropriateexpressions for the kinetic of different biological processesassociated with the degradation of the pollutants.

Assuming perfect mixing, the partial mass balance forthe first tank in series, the anoxic bioreactor, reads:

` (5)where,

- the vectors xin, xrs, xedenote the concentrations (kg/m3) in the influent, Qin, in the recycled sludge, Qrs, and in


3/12

REV. CHIM. (Bucharest) 61Nr. 4 2010 http://www.revistadechimie.ro 409

outlet, Qin+Qrs flows (m3/s) respectively. Their components

are:SO2, SS, SN

2, SNH

4, SNO

2, SNO

3, SI, SALK, XI, XS, XH,

XSTO, Xns, Xnb(table 1 for details);- v is the vector formed by the reaction rates of each

component, kg/m3. s

where cji is the stoichiometric

coefficient ofithcomponent injthreaction, according to thestoichiometric matrix from table 2, while j is the kineticrate ofjthreaction,kg/m3s, as given in t table 3.

- Vanis the volume of the anoxic bioreactor (m3).

The partial mass balance for the second tank in series,

the aerobic bioreactor, reads:(6)

In equation (6) the notations are the same as in theprevious mass balance equation, except for ar, which isthe residence time, Var/ (Qin+Qrs) of the aerobic bioreactor.In the aeration tank the mass balance for oxygen has anadditional term which describes the oxygen mass transferfrom the air bubbles fed through the diffusers, kg/m3 daybeside its consumption rate due to the biological process:

(7)

In equation (7), the notations are:KLa - oxygen mass transfer coefficient,d

-1

SO2sat- - the saturation concentration of oxygen in the

liquid phase, kg/m3SO2- the liquid oxygen concentration, kg/m

3.The influence of the recycling parameters, and ,

upon the concentrations of the soluble and insolublespecies in the inlet of the reactor x inRis expressed in thefollowing equations:

(8)

for the soluble items, and

(9)

for the insoluble species, respectively. In both cases, thedependency upon the system feed is the same, while thatupon the recycled concentrations differs because of theeffect induced by the presence of the separation unit.

Performance criteriaIn order to measure the fitness of the process to the

rapid change of the influent, two types of performancecriteria were envisaged: an averaging one the cumulativeconversion, taking into account the whole history of thesystem and an immediate one the instant conversion,

which measures the instantaneous behavior of the system,showing if, for this given particular conditions at this instantof time, a failure appears or not.

The cumulative conversion is defined as the ratiobetween the amount of species P transformed due to thebiological process from the beginning of the time windowtill present and the amount of species P fed into the systemduring the same interval:

(10)

In equation (10), Pin stands for the species concentrationat inlet, Pout stands for the species concentration at outlet,while Pp stands for the species concentration at purge (fig.1 for the sketch of the system).

Theinstant conversion is defined as the ratio betweenthe amount of species P transformed at the moment t andthe amount of species P fed into the system at the sametime:

(11)

It must be emphasized that the main difference betweenthe instant and the overall conversions is that the formermeasure the performance of the system regardless of theload still existing in it, while the latter takes this intoaccount. In other words, there could be cases in which theoverall conversion is higher than the instant one, meaningthat there is an ongoing accumulation of thePspecies in

Table 1

NOTATIONS USED FOR THE STATE VARIABLES DESCRIBINGTHE BEHAVIOUR OF THE WHOLE SYSTEM


4/12

REV. CHIM. (Bucharest) 61Nr.4 2010http://www.revistadechimie.ro410

the system; conversely, when the overall conversion issmaller than the instant one, the system performs betterthan it did beforehand, consuming not only thePspecies

fed into it, but also a part from thePspecies accumulatedso far in the biological reactor. When the system reachessteady state, both conversions should have the same values.

In order to cover the whole spectrum of species thissystem has to deal with, three components were pickedup to compute both criteria for: the readily biodegradablesubstrate (SS), the ammonia (NH4

+) and the slowlybiodegradable substrate (XS).

Results and discussionsThe benchmark input files referred to as dry weather,

stormy weather and rainy weather and the processsizing reported in the aforementioned benchmark study

(http://www.benchmarkwwtp.org) were used for theoperating and initial conditions for an operational windowof 14 days (fig. 2 for the variation of the incoming flowwhich has to be treated in the system). It should be notedthat the same variation appears to any inlet concentrations,

so the system is intrinsically in a dynamic state. Theaeration of aerobic tank is achieved using such mixingand inlet air flowrate so the value of KLais 15 h

-1, irrespective

of the inflow conditions. This value is high enough so thatthe process would be on the safe side with respect to theoxygen supply for all situations. It was assumed thatprocess developed at an average temperature of 20C.

The simulation, based upon the aforementionedmathematical model, was implemented in MATLAB usingode15s integration routine for stiff differential equations.

In the system with recycling of mixed liquorwe chooseas control parameters therecycle fraction,, and the purgefraction, . The performance (cumulative and instant) ofsystem for readily biodegradable substrates (SS), slowlybiodegradable substrates (XS) and ammonium (NH4

+)were analyzed for the lowest, the middle and highest values

of recycle and purge fractions, that is the recycle fraction,, was 0.1, 0.5 and 1.0, while for the purge fraction, , was0.001, 0.005 and 0.01. The first parameter influences thebioreactors time scales, lowering it from the original valuesof V

anox/Q

in and V

aero/Q

in to the actual values of V

anox/Q

in/

Tabel 2

STOICHIOMETRIC MATRIX OF THE ASM3-2N MODEL [1]

V = variable; P = process


5/12


6/12

REV. CHIM. (Bucharest) 61Nr.4 2010http://www.revistadechimie.ro

recycling and purge flows will be, continuously, adapted tothe actual values the question of how the system wouldrespond if the recycle and the purge would have been fixedwill be addressed in another study. Another simplifyingassumption is the biological process shifts instantaneouslyaccording to the external substrate concentrations, noadaptation lag being taken into consideration.

According to figure 2, the sys tem is forced into adynamic state; the crucial problem at hand is could thesystem fail? and, if so, when and for what combinationof operating parameters? In subsidiary, there is another

question: does the system reach a steady state for theaverage values of the inflow conditions?We consider the following time averaging formula as a

convenient relationship:

(12)

Applying equation (12) for the dry weather conditions(since the other two are exceptions, they cannot beconsidered as candidates to establish the averageconditions), the calculated common values for the

considered period are: Qin=18445 m3

/day, CODin=398.72g/m3andNH4in

=31.556 g/m3. With this constant input andfor =0.5 and =0.005 (chosen as the middle values ofthe proposed ranges), the system reached a steady statecharacterized by the following conversions: of the readilybiodegradable substrates (SS) XSS = 0.999 of the slowlybiodegradable substrates (XS) XXS=0.065 and of theammonium (NH4

+)XSNH4=0.983. The rest of the steadystate values are presented in table 4.

The cumulative and instant performance criteria of thesystem converged rapidly to the same value, characteristicfor the steady state. Due to the cascade system adoptedfor the two reactors (anoxic and aerobic), the values of theparameters corresponding to steady state do not differ asmuch as expected, although for the most sensitive thedifferences are conclusive (table 4). While the systemperforms excellent for the readily biodegradable substratesand the ammonium, consuming them almost entirely, the

conversion of the slowly biodegradable substrates is verylow. This means that a large part is disposed to theenvironment.

An interesting development would be to study thebehaviour of a slightly changed system, with a settler andrecycling after each reactor, which would permit to theactivated sludge from each reactor to adapt itself better toits particular operating conditions, but this is left for anotherstudy.

In figure 3, we present the time behaviour of the readily

biodegradable substrates for the three weather conditionsunder three different pairs of parameters and: middle,lowest and highest. The system is in a confined dynamic,meaning that its performance varies according to the inletchanges between some finite boundaries; both conversionsare high, slightly under the pseudo-steady stateperformance obtained with the averaged inlet data (fig.3A, dry case). The ascending trend of the conversions isdue to the slowly increase in heterotrophs concentration,which increases the total amount of readily biodegradablesubstrate processed. This has as effect higher instantconversions than cumulative ones for the last part of thetime window. For the rainy case, there is a significant dropin performance starting from the eighth day, due to the

increase in the influent, which increases the recycled flowand the purge flow. This affects negatively the process,reducing the time scale of the biological reactor. More, theincrease in both throughput and purge flows diminishesthe concentration of the biomass in the bioreactors, thusdecreasing the amount of processed substrates. After theshock ends, the instant conversion regains its behavior (fig.3A, dry and rainy cases after the tenth day), although witha slight delay. The same is valid for the cumulativeconversion, but due to the substrate accumulations in thesystem, it has lower values for the same time window.The stormy case is characterized by two sudden increasesof inlet flows, much larger than in the rainy case, but for

shorter periods (fig. 3A, stormy case). Consequently, thesystem behaves badly during these periods, worse that inthe previous case, but it recovers faster, due to the lessaccumulation of substrates in the biological reactors thistime.

Fig. 2. The inflow characteristics for the benchmark, corresponding to the dry, rainy and stormy weather

412


7/12

REV. CHIM. (Bucharest) 61Nr. 4 2010 http://www.revistadechimie.ro

Working at very low recycling and purge ratios affectsslightly the behavior of the system, decreasing the overallperformance. But, since the heterotrophs concentration ishigh enough, the conversion varies between almost thesame values (fig. 3B, dry case see the overallperformance, compared to the previous case). The shocksinduced by the weather change (rainy or stormy fig. 3B)

is a little bit higher, but still the conversions are very good,

Table 4

THE STEADY STATE VALUES OF THE STATEVARIABLES FOR THE TWO BIOREACTORSCORRESPONDING TO THE PSEUDO-INLET

CONDITIONS RESULTED FROM AVERAGING THEDRY WEATHER BENCHMARK INPUT

Fig. 3AB. Performance of readilybiodegradable substrates, expressed ascumulative and instant conversions, for

different time scales and weather conditionsA) base case of =0.5 and =0.005; B) lowest

operating parameters values of =0.1 and=0.001;

413


8/12


Fig. 3C. Performance of readilybiodegradable substrates, expressed

as cumulative and instant conversions,for different time scales and weather

conditions C) highest operating

parameters values of =1.0and =0.01

meaning that the system performs well even with these

values of and .Increasing both ratios to their maximum values worsensthe behaviour of the system, but not to a great extent (fig.3C, dry, rainy and stormy profiles, compared to the basecase profiles). Working at high values of the recycle ratiosdecreases the time scale of the bioreactor significantly,especially when shocks arrive (see the overall and instantconversions for the rainy and stormy periods fig. 3C). Onthe other hand, higher values for the purge ratio decreasethe concentration of the living cells, which are withdrawnfrom the system through the purge. This way, the processedamount of soluble substrates diminishes although theirconcentrations remain the same (the purge fraction affects

only the concentration of the suspended particles, likemicroorganisms or any other solids which are present inthe system).

A similar behaviour of the system could be seen withrespect to the slowly biodegradable substrate consumption,as depicted in figure 4, where the profiles of its conversionare presented, for all three aforementioned scenarios. Evenfor the virtual steady state conditions expressed as theaveraged values for the inlet variables, the system in itsactual configuration well behaved , ensuring a highconversion, of 0.98. Actually, this means that almost allthe slowly biodegradable substrate entering the system ,defined as the sum of fragments of heterotrophic biomassresulting from decay by endogenous respiration, and slowlyhydrolysible substrate, was transformed, a small amountwas removed by purge. For any of the confined dynamicsscenarios (dry, rainy or stormy weather), the cumulative

conversion has positive values for the whole time window.Nonetheless, working at the lowest possible values for and seems beneficial, the cumulative conversion of the

Fig. 4A. Performance of slowly biodegradable substrates, expressed as cumulative and instant conversions,for different time scales and weather conditions A) base case of=0.5 and =0.005

414


9/12


Fig. 4BC. Performance of slowlybiodegradable substrates, expressed

as cumulative and instantconversions, for different timescales and weather conditions

B) lowest operating parametersvalues of =0.1 and =0.001;C) highest operating parameters

values of =1.0 and =0.01

slowly biodegradable substrate has most values, 0.99. Atthese low values of the operating parameters, the time

scale of the bioreactors is the highest and so is theconcentration of the microorganisms responsible for theconsumption of the slowly biodegradable substrate. Thus,the microorganisms, which are in a higher concentration,have a longer time to process this substrate, whichincreases its conversion (fig. 4B). For the highest valuesof the operating parameters and (fig. 4C), the systembehaves worst; the time scale of the bioreactorsdecreases dramatically diminishing the time themicroorganisms act upon the substrate, and so does theirconcentration. Less substrate processed is a consequenceof a smaller amount of bacteria, despite the slight increasein the process rate due to the raise in the concentration ofthis substrate.

The behavior of the system, with respect to theammonium, is presented in figure 5 for all the operatingconditions at hand (A dry weather, B rainy weatherand C stormy weather). The system behaves worse than

for the previous substrates (readly biodegradable substrateand slowly biodegradable substrate), although the confined

dynamics has as inferior margin 0.75 which could be seenacceptable for such daily fluctuations of the inlet flow andconcentrations. Although the steady state conversion ofthe system subject to the averaged inlet conditions is ratherclose to the conversion of the readily biodegradablesubstrates, the amplitude of the variations are higher sincethe class of the microorganisms responsible orammonium conversion is different, namely AOBs. Thesemicroorganisms have a much lower concentration in thesystem, so a comparable percentage of their variationcompared to the heterotrophs would imply a morepronounced effect upon the quantity of the processedsubstrate than for heterotrophs. That is why keeping a highconcentration of AOBs and a longer residence time (lowestvalues forand) is beneficial (fig. 5B), while the highestvalues for these operating parameters have detrimentaleffects (fig. 5C). The changes in the weather conditionsonly increase the lower spikes for the instant conversion(fig. 5, the first three days period corresponding to the rainy

415


10/12


Fig. 5. Performance of ammonium

substrate, expressed as cumulative andinstant conversions, for different timescales and weather conditions A) basecase of =0.5 and =0.005; B) lowest

operating parameters values of =0.1 and=0.001; C) highest operating parameters

values of =1.0 and =0.01

and stormy weather respectively), consequently affectingthe cumulative conversion too.

Analyzing the profiles presented in figures 3-5, weobserve that except the three days period when the

weather changed in the rainy and stormy periods, it is ratherdifficult to seize the differences between them for the lastthree days, even if it is obvious that they exist. Measuringthe distance between the perturbed trajectory of a process

416


11/12


and the nominal one is not trivial, since the process couldbehave differently immediately after the perturbationoccurred or after a rather long period. To quantitativelycharacterize the departure of the state trajectory given byany kind of perturbations from the original, non-perturbedone, we will use an intensive measure, derived from thenotion of Shannon entropy, the generated entropy of aperturbed trajectory, as defined in [33,34]. Since in thepresent analysis the accent is made more upon theexistence and magnitude of the variations and less uponthe direction of them, this measure is a reliable measuringtool. The sensitivity of the bioreactor to some parametershould imply larger generated entropies for some if not allthe state variables, for small changes in the former. Whenmore than one parameter has to be analyzedsimultaneously, care should be taken to adequatelydecouple their mutual influence upon the process.

Applying the relationship presented in [33,34], weobtained the values of the generated entropy of theperturbed trajectory for the aforementioned cases, whichare summarized in Table 5a and b. Analyzing the figuresfrom table 5, we observe that we can express in aquantitative manner the observations made so far for theinfluence of the operating parameters and upon theprocess, coupled with weather changes. According to thetype of substrate under scrutiny, loweringandcould bebeneficial or detrimental. For example, the generatedentropy for the cumulative conversion of the readilybiodegradable substrates, XSS, is high for the highest valuesof the operating parameters and low for the lowest values,irrespective of the weather change (table 5a, S column,for both Dry-Rainy and Dry-Stormy cases). This means thatthe system has the tendency of amplifying any departurefrom the normal confined dynamics. The same is true forthe ammonium substrate, although the figures are oneorder of magnitude lower, which is in agreement with the

Table 5a

THE GENERATED ENTROPIES OF THEPERTURBED TRAJECTORIES, ASRESULTED FROM THE WEATHER

CHANGE, FOR THE SAME VALUES OF THEOPERATING PARAMETERS AND (C

STANDS FOR CUMULATIVE

Table 5bTHE GENERATED ENTROPIES OF THE PERTURBED TRAJECTORIES,

AS RESULTED FROM THE OPERATING PARAMETERS AND CHANGE, FOR THE SAME WEATHER CONDITIONS (C STANDS FORCUMULATIVE CONVERSION, I STANDS FOR INSTANT CONVERSION) concentrations of the responsible species (table 5a, N

column, for both Dry-Rainy and Dry-Stormy cases togetherwith table 4, HET and AOB concentrations). The highesteffect of weather change, although in the opposite direction,is for the slowly biodegradable substrates, XXS, where thegenerated entropy is almost double than for the readilybiodegradable substrates (table 5a, X column).

The influence of the operating parameters change uponthe process is affected by the weather conditions, as can

be seen from table 5b. For the dry period, which isconsidered as the base case in this study, as the systemworks at higher values of the operating parameters, it willbe more affected by any of their change (table 5b) for thereasons aforementioned, proper to each type of substrate.

ConclusionsThe study at hand presents the time scale behaviour of

a wastewater treatment facility subject to environmentalstress, such as weather changes. Two of the time scalesare addressed here, the residence time of the biologicalreactors (aerobic and anoxic) which is mainly influencedby the recycling ratio , and the doubling time of the

biological process, corroborated with the concentration ofthe microorganisms, which is mainly influenced by thepurge fraction . In order to quantitatively measure theperformance of the system, two types of conversions wereused, a cumulative one, as introduced by equation (10),and an instant one, equation (11). The former, being ofintegral type, stands for the average processingperformance of the system, considering its time-historyas captured by the time window under scrutiny, while thelatter gives the actual performance, with respect to theinlet and outlet. While the former is valuable for describingthe long term performances, the latter shows when and towhat extent the system fails punctually and harms theenvironment. To quantify the time scale influences upon

the system, generated entropy of a perturbed trajectoryconcept was used, not only showing that lower time-scalesare detrimental for slowly degradable substrates, but alsoto what extent.

NotationsASM - Activated Sludge ModelAOB - ammonia oxidizing bacteriabA,NOX - Anoxic endogenous respiration rate of XA (d

-1)bH,NOX-Anoxic endogenous respiration rate of XH(d

-1)bH,O2 - Aerobic endogenous respiration rate of XH (d

-1)bnb,NO3 - Aerobic endogenous respiration rate of Xnb (d

-1)bnb,O2 - Aerobic endogenous respiration rate of Xnb (d

-1)bnsNO2 - Aerobic endogenous respiration rate of Xns (d

-1)

bns,U2 - Aerobic endogenous respiration rate of Xns (d -1)bSTO,NOX - Anoxic respiration rate of XSTO (d

-1)bSTO,O2- Aerobic respiration rate of XSTO (d

-1)COD - Chemical Oxygen Demand (g O2/m

3)C/N - Carbon to Nitrogen ratio

417


12/12


DO - dissolved oxygenEPS - extracellular polymeric substancesf

SI - Production of SI in hydrolysis

fXI

- Production of XI in endogenous respirationHET heterotrophic bacteriai

NSS - % amount of N in SS (g-N/g-COD)

iNXI

- % amount of N in XI (g-N/g-COD)i

NBM - % amount of N in biomass (g-N/g-COD)

iNSI

- % amount of N in SI (g-N/g-COD)i

NXS - % amount of N in XS (g-N/g-COD)

KA,ALK - Saturation constant for alkalinity for autotrophs (mol HCO3/m3)KALK - Saturation constant for alkalinity for XH(mol HCO3/ m

3)KA,NH4- Saturation constant for SNH4for autotrophs (g N / m

3)KA,NOX- Saturation constant for SNOX for autotrophs (g NOx / m

3)KA,NO2 - Saturation constant for SNO2 for autotrophs (g NO2/ m

3)KA,NO3- Saturation constant for SNO3 for autotrophs (g NO3/ m

3)KA,O2- Saturation constant for SO2 for autotrophs (g O 2/ m

3)K1,NE4 - Ammonia inhibition of nitrite oxidation (g NH4/ m

3)KH- Hydrolysis rate constantKX- Hydrolysis saturation constantK

La - Oxygen mass transfer coefficient (h-1)

KNH4

- Saturation constant for SNH4(g N / m3)

KNO2

- Saturation constant for SNO2of Xns (g NO2/ m3)

KNO3

- Saturation constant for SNO3of Xnb (g NO3/ m3)

KNOX - Saturation constant for SNOX(g NO3/ m3)K

O2 - Saturation constant for SO2(g O2/ m

3)K

S - Saturation constant for SS(g COD SS/ m

3)K

STO - Saturation constant for XSTO(g COD XSTO/ g COD XH)

kSTO

- Storage rate constant (d-1)NOB - nitrite oxidizing bacteriaQin feeding flowQrs - recycle flowSRT - short solid residence timeS

ALK - Alkalinity [mol HCO3/m

3]S

I - Soluble inert organics (g COD /m3)

SN2g

- Dinitrogen released by denitrification (g N / m3)S

NH4 - Ammonium (g N / m3)

SNOX

- Total oxidized nitrogen in the ASM3 model (g N / m3)

SNO2 - Nitrite nitrogen (g N / m3)S

NO3 - Nitrate nitrogen (g N / m3)

SO2

- Dissolved oxygen (g O2 / m3)S

S - Readily biodegradable substrates (g COD / m3)

Vaero

volume of aerobic reactorV

anoxvolume of anoxic reactor

XH

- Heterotrophic biomass (g COD / m3)X

I - Inert particulate organics (g COD / m3)

Xnb

- Nitrite-oxidizing autotrophs (g COD / m3)X

ns - Ammonia-oxidizing autotrophs (g COD / m3)

XS - Slowly biodegradable substrates (g COD / m

3)X

STO - Organics stored by heterotrophs (g COD / m3)

YA - Aerobic yield of XA(g COD XA/ g N NOX)

YH,NOX

- Anoxic yield of heterotrophic biomass (g COD XH/ g COD XSTO)

YH,O2 - Aerobic yield of heterotrophic biomass (g COD XH/ g COD XSTO)Ynb

- Aerobic yield of Xnb (g COD Xnb / g N NO3)Y

ns - Aerobic yield of Xns (g COD Xns / g N NO2)

YSTO,O2

- Aerobic yield of stored product for SS (g COD XSTO/ g COD SS)Y

STO,NOX - Anoxic yield of stored product for SS (g COD XSTO/ g COD SS)

- recycle ratio - purge ratio

NOX - Anoxic reduction factor

A - Autotrophic max. growth rate of XA(d

-1)

H - Heterotrophic max. growth rate of XH(d

-1)

nb - Max. growth rate of Xnb (d-1)

ns

- Max. growth rate of Xns (d-1)

Ac knowle dg em en ts ; This work was supported at University

Politehnica of Bucharest by UEFISCSU project No 175/1.10 .2007 -Complex behavior of mixed microbial populations induced by timescales and segregation. Case study: wastewater biological treatmentprocess.

References1. IACOPOZZI, I.; INNOCENTI, V.; MARSILI-LIBELLI, S.; GIUSTI, E.Environmental Modelling & Software, 22(6),2007,p. 8472. MERKEL, W.; MANZ, W.; SZEWZYK, U.; KRAUTH, K. Water Research,33(10),1999,p. 23923. MOUSSA, M.S.; HOOIJMANS, C.M.; LUBBERDING, H.J.; GIJZEN, H.J.;

VAN LOOSDRECHT, M.C.M. Water Research, 39(20),2005,p. 50804. CHUANG, S.-H.; OUYANG, C.-F. Water Research, 34(8),2000,p. 22835. ZHAO, H.W.; MAVINIC, D.S.; OLDHAM, W.K.; KOCH, F.A. WaterResearch, 33(4),1999,p. 961

6. HU, J.; LI, D.; LIU, Q.; TAO, Y.; HE, X.; WANG, X.; LI, X.; GAO, P.Journal of Environmental Sciences, 21(3),2009,p. 3877. CAMPOS, J.L.; GARRIDO, J.M.; MOSQUERA-CORRAL, A.; MENDEZ,R. Biochemical Engineering Journal, 35(1),2007,p. 878. PLSZ, B.G.; JOBBGY, A.; GRADY, C.P.L. Water Research,37(4),2003,p. 8539. JIE, H.; DAPING, L. Current Microbiology, 57(4),2008,p. 28710. KIM, D.-J.; LEE, D.-I.; KELLER, J. Bioresource Technology,97(3),2006,p. 45911. KIM, J.-H.; GUO, X.; PARK, H.-S. Process Biochemistry, 43(2),2008,p. 15412. KIM, D.-J.; AHN, D.; LEE, D.-I. Korean Journal of ChemicalEngineering, 22(1),2005,p. 8513. CANZIANI, R.; EMONDI, V.; GARAVAGLIA, M.; MALPEI, F.; PASINETTI,

E.; BUTTIGLIERI, G. Journal of Membrane Science, 286(1-2),2006,p. 20214. XUE, Y.; YANG, F.; LIU, S.; FU, Z. Bioresource Technology,100(3),2009,p. 105515. PAMBRUN, V.; PAUL, E.; SPRANDIO, M. Chemical Engineeringand Processing: Process Intensification, 47(3),2008,p. 32316. WHANG, L.-M.; CHIEN, I.C.; YUAN, S.-L.; WU, Y.-J. Chemosphere,75(2),2009,p. 23417. CERVANTES, F.J.; PAVLOSTATHIS, S.G.; VAN HAANDEL, A.C.

Advanced biological treatment processes for industrial wastewaters:principles and applications, 2006.18. GAO, M.; YANG, M.; LI, H.; YANG, Q.; ZHANG, Y. Journal ofBiotechnology, 108(3),2004,p. 26519. MICHAUD, L.; BLANCHETON, J.P.; BRUNI, V.; PIEDRAHITA, R.

Aquacultural Engineering, 34(3),2006,p. 22420. WANG, H.; LI, Q.; WANG, Y.; HE, N.; SUN, D. Journal of HazardousMaterials, 168(1),2009,p. 20321. RIVAS, J.; GIMENO, O., BELTRN, F. Chemosphere, 74(6),2009,p. 85422. RAMAKRISHNAN, A., GUPTA, S.K. Bioresource Technology,99(9),2008,p. 374523. HAO, R.-X.; ZHOU, Y.-W.; CHENG, S.-Y.; LI, J.-B.; ZHAO, M.; CHEN,

X.; YAO, N. Chemosphere, 70(5),2008,p. 78324. YOOCHATCHAVAL, W.; OHASHI, A.; HARADA, H.; YAMAGUCHI, T.;SYUTSUBO, K. International Journal of Environmental Research,2(3),2008,p. 23125. TOMITA, R.K.; PARK, S.W.; SOTOMAYOR, O.A.Z. ChemicalEngineering Journal, 90(3),2002,p. 283

26. GUJER, W.; HENZE, M.; MINO, T.; VAN LOOSDRECHT, M. WaterScience and Technology, 39(1),1999,p. 18327. MANSER, R.; GUJER, W.; SIEGRIST, H. Water Research, 40(12),2006,p. 241628. SIRIPONG, S.; RITTMANN, B.E. Water Research, 41(5),2007,p. 111029. NI, B.-J.; YU, H.-Q. Chemical Engineering Journal, 140(1-3),2008,p. 10130. NI, B.-J., YU, H.-Q. Chemical Engineering Science, 63(10),2008,p. 278531. ILIE, M.; ROBESCU, D.N.; GHITA, G. Rev. Chim. (Bucharest), 60,no. 5, 2009,p. 52932. HENZE, M.; GUJER, W.; MINO, T. Activated sludge models ASM1,

ASM2, ASM2d and ASM3. IWA Scientific and Technical Report no. 9 ,

200033. LAVRIC, V.; OFITERU, I.D.; WOINAROSCHY, A. Biotechnology andApplied Biochemistr y, 41(2005,p. 8934.OFITERU, I.D.; LAVRIC, V.; WOINAROSCHY, A. Chemical EngineeringTransactions, 3(3),2003,p. 1845Manuscript received: 17.07.2009

418

Documents

Comportare in Proces Bioreactor.pdf 4 10