气升式生物反应器在废水处理中的应用

Airlift Bioreactors: Application to Wastewater Treatment

Merchuk J.C.1 and Shechter, R.2

1 2

Dept. of Chemical Engineering, Ben-Gurion University of the Negev, Beer-Sheva, Israel AqWise – Wise Water Technologies Ltd. P.O.B. 8698, Netanya 42504, IsraelAbstract:

After an introductory review, this paper presents a three-phase reactor fluiddynamic model that is used in the development of a generic process for upgradingexisting wastewater treatment plants (WWT). The aim is to incorporate nitrificationand de-nitrification without additional reactor volume. The core of the process consistsof an attached growth airlift reactor (AGAR) in which a biomass carrier is used.Mixing is provided in the aerobic stages by partitions that create an airlift effect. Thedesign procedure for large airlift elements in an aeration basin containing floating solidcarriers is based on classic airlift calculation procedures adapted to this specific case.The combination of the organized flow pattern and the low density of the media enablehydraulic prevention of carrier migration between different process stages, eliminatingthe need of screens. Process performance in a demonstration plant, expressed as airliftliquid velocity, confirms the predicted values within a reasonable range.Keywords:

Airlift reactors; wastewater treatment; mathematical model; airlift fluid dynamics

New reports on use of ALRs for WWT:Since one of the recognized characteristics ofairlift reactors is the potential for scaling up andthe relatively low power consumption for agitationand oxygenation, it is only natural that manyprocesses related to wastewater treatment use thistype of reactors.

Jinet al. (2002) used an airlift reactor in acomprehensive pilot plant system for starchprocessing wastewater reclamation. The starch wasutilized byAspergilius orizae. Simultaneously to a95% COD, 93% BOD and 98% suspended solidsremoval, an important production of -amylase(~50 EU/ml) was obtained (Jinet al. 1998). Aninteresting point in this paper is the dependence offungal morphology on ALR fluid dynamics. In thistype of processes, morphology of the fungalbiomass is extremely important. Free mycelialgrowth (wild growth) increases strongly theviscosity, limiting the oxygen transfer rate fromthe gas to the culture. The solution that has beenalmost universally adopted for this problem, whichattains to citric and other organic acids,antibiotics, etc., is to find the conditions underwhich the biomass takes the form of fungal pellets.The advantage in gas-liquid transfer rate becauseof the decrease in viscosity usually overweighs theadded resistance stemming from the itraparticlediffusion of oxygen. But the formation of pellets inoptimal size and compactness is a very complexmatter.

Lazarovaet al. (1997) studied experimentallythe fluid dynamics and the performance forwastewater treatment of an split-vessel airlift witha rectangular section. They studied carefully theinfluence of suspended solids on gas holdup bothin the riser and the downcomer, as well as theinfluence of the ration of riser to downcomer crosssectional areas on liquid velocity. They comparethe experimentally measured velocities fordifferent reactor heights without proposing anycorrelation. The main aspect stressed by theresearchers is the capacity for nitrificationobserved in various stages.

Many applications of ALR have been reported inprocesses where the point of interest forresearchers interested in ALR and BC is simplythat the process, which can take place in aconventional stirred tank, can be run using an ALRas well, with the consequent savings in energyrequirements, etc. For example, the use ofAspergilius niger for textile wastewater (biologicaldiscoloration) was reported by Assadi andJahangiri (2001). Camposet al. used an ALR in acombined (microfiltration and biological)treatment of oilfield wastewater treatment. Theyobtained satisfactory results in TOC and CODreduction in a continuous process, using an ALRwith suspended Polystyrene particles and studyinghydraulic retention times from 12 to 48 hours.Both the above-mentioned studies where carriedout in small-scale reactors.

Loh and Liu (2001) used an external loopfluidized bed airlift bioreactor for treatment of highstrength phenolic wastewater. They used theincrease in gas holdup that they get closing a valvein the downcomer and restricting liquid circulationto control in this way the oxygen transfer. Therange of variation in their device goes in fact fromholdup in an airlift with unrestricted circulation tothe holdup in a bubble column, for similardiameters and gas superficial velocity. Obviously,the case of a completely closed valve implies thatthe downcomer volume is not contributing to theprocess.

Bakkeret al, (1996) immobilized their biomassinside -carageenan gel beads, and studied acascade of two ALRs (small scale) to study theoxidation of nitrite to nitrate byNitrobacter agilis.This is an important step in the nitrificationprocess (i.e. the oxidation of ammonia to nitratevia nitrite, usually followed by a deitrificationstage with reduction of nitrate to N2). They foundadvantages in the use of two bioreactors in series,and attributed it to the kinetics of the process (non-competitive substrate and product inhibition).Because the density of the beads was close tounity, there was no problem in fluidizing of thebeads in spite of the small scale.

While the basic characteristics of an airliftreactor indicate its fitness for aeration of largevolumes of wastewater, the problem of ammoniaremoval calls for special handling. Twoapproaches have been lately presentedincorporating the nitrification-denitrificationelement into a basic airlift arrangement. The firstone is the biofilm airlift suspension extensionreactor (van Benthum 1999-a, 1999-b), whichpresents the very compact design. Theconventional airlift with its three phases inenclosed into an additional vessel (extension) thatbecomes the anaerobic volume. Part of the liquidand suspended biofilm coated solids overflows theaerobic airlift core and enter the top of theextension, reentering at the bottom. The designallows the control of aerobic/anaerobic times forthe biofilm-coated particles suspended in thesystem in order to improve thenitrification/denitrification of wastewater. Theflow of liquid and suspended solids in theextension, which is the anaerobic volume, can becontrolled manipulating the overpressure in theheadspace of the reactor. A mathematical modelwas developed and used for the design of a pilotplant. The experimental results of gas and solidholdups concur satisfactorily with the modelpredictions.

A different approach to the integration ofnitrification/denitrification in a wastewatertreatment process is the one presented by Shechteret al. (2002) that follows.

The AGAR system:

The AGAR process is used for retrofittingwastewater treatment plants for nitrogen removaland capacity increase, by increasing the quantityand improving the quality of microorganisms inthe aeration basin of the biological treatment. Theprocess utilizes shaped plastic carriers forsupplying a high surface area on which a bio-filmdevelops. The biomass carriers are mixed in theaeration basin by the air required for supplying theoxygen demand of the aerobic process. Energycost considerations require that the air supply tothe system, one of the major operating costs, beminimized.

The mixing of a high concentration of floatingbiomass carriers in water requires arrangement ofan organized flow pattern, in order to avoidexcessive aeration expenses. Airlift hydraulicsprovides a well organized flow pattern whichensures a liquid velocity sufficiently high tofluidize the plastic carriers in every part of thebasin. Since space limitation does not allow toexplicit the detailed mathematical expression ofthe hydraulic model, which will be presentedelsewhere (Shechter and Merchuk, 2003), onlysome results obtained with the model are shownhere.

Simulations:

Figure 1 shows the profiles of gas holdup in theriser, solids holdup in the riser and downcomer,and the superficial velocities of liquid and solid inthe riser for a wide range of superficial gasvelocity at the inlet of the airlift reactor.

Figure 1: Influence of the gas superficial velocity in the riser(UGr) on the fluid-dynamic variables.

The figure shows that the gas holdup does notchange sharply, while the liquid velocity almostdoubles itself. The picture is completed by the

profile of solids holdup, which increases in theriser and decreases in the downcomer. The modelpredicts thus that most of the energy input isrelated to the movement of the suspended particles.Figure 2 shows the influence of the total solidsload in the system. Again, the variables

Figure 2: Influence of the solids total loading s on the fluid-dynamic variables. Legends as in Fig. 1

represented on the graph are gas holdup inthe riser, solids holdup in the riser anddowncomer, and the superficial velocities ofliquid and solid in the riser.

This graph is built for a constant gas inputcorresponding to a gas superficial velocityof 0.006 [m/s]. While the gas holdup doesnot change, the liquid velocity drops sharplyas the solids holdup both in the riser and thedowncomer increase.

Figure 3 shows the influence of thedensity of the suspended particles on thevariables of the process: also in this case thegraph is built for a constant gas inputcorresponding to a gas superficial velocity

of 0.006 [m/s].

Figure 3: Influence of the density of the suspended solids, S on the fluid-dynamic variables. Legend as in Fig. 1.

While the gas holdup in the riser and thevelocity of the solids in the riser remain almostconstant when the density of the solid changes in5% under and over the density of water, thesolids holdup in the riser and downcomer areaffected in opposite ways: while sr increases sd

decreases. The lines cross at particle densityequal to that of water. The difference betweenriser and downcomer changes thus from negativeto positive as s increases. This difference, inabsolute terms, is smaller at higher particledensity (approximately 0.03versus 0.074), andthe predicted increase in liquid velocity,therefore, concurs with the general conclusionsby Heijnenet al. (1997). Since in the AGARsystem the solids do not move from stage tostage, each stage is, with respect to solidparticles, a closed reactor. If the density of thesolid is larger than that of the liquid, its risingvelocity will be smaller and the holdup of solidsin the riser will be larger than in the downcomer,andvice versa.The presence of solids, therefore,will always diminish the driving force forcirculation, independently of their density, asshown clearly on Fig. 3.

Figure 4 shows the prediction of the modelwith respect to the influence of the ratio Ar/Adon the variables considered, for a constantsuperficial gas velocity in the riser. As expected,the increase in Ar/Ad produces an increase in theholdup of solids in the downcomer and a

decrease in the liquid velocity in the riser.

Figure 4: Influence of the riser to downcomer area ratio on thefluid-dynamic variables. Legends as in Fig 1

An interesting point that was observed is thatthe model would not converge to a result if thegas input was not enough to produce the minimalliquid velocity required for recirculation of solidsin the downcomer. This observation was the basefor the construction of figure 5, that shows theminimal gas flow rate required to allow the propercirculation of solids in the airlift reactor at eachsolids loading (shown as: % fill=100* S/f, wheref is the solid volume fraction in a packed bed ofparticles), for four different densities of theparticles. The lines obtained for a constantparticle density are straight and the slopeincreases slightly as the particle densitydecreases.

210840

185

860

/h1603m 880m135,riaQ110900

856015

20

25

30

35

40

45% fill

I

Figure 5: Minimum driving air requirements for differentparticles fill ratios. The different particle densities areindicated in each line in units of [kg /m3].

It is worth noting the strong sensitivity of themodel to particle density. The heavier the particle(the closer to water density) the lower the gas flowrate required for the inverse fluidization of theparticles in the bed.

Conclusions:A mathematical model describingthe fluid dynamics of a split vessel, airlift unit ofrectangular section in the AGAR process waspresented. The model allows the prediction ofliquid velocity, gas holdup in the riser, and solidsholdup and velocity in both riser and downcomer.The model shows a coherent picture of thebehavior of the fluid dynamic system. It was usedto predict the minimal gas input rates required forinverse fluidization in the downcomer of thebiofilm-carrying solids.

ARiser cross-section area, given, m2ADown-comer cross-section area, m2AFree passage area for flow fromomdowncomer to riser, m2

fVolume of solids per volume of packed bed of solids, - Overall gas holdup rGas holdup in the riser dGas holdup in the down-comer SrSolids holdup in the riser Sdsolids holdup in the down-comer Liquid density, kg/m3

Solid density, kg/m3

Citations:

Assadi, M. M. and Jahangiri, M. R., Textilewastewater treatment byAspergilius nigerDesalination141, 1-6 (2001)

Bakker W. A. M., Kers, P., Beefting, H. H.,Tamper, J. and de Gooijer, C. D., NitriteConversion by ImmobilizedNitrobacter agilisin an airlift Loop Bioreactor Cascade: Effectsof combined Substrate and Product Inhibition,J Ferm. Bioeng.81, 390-393 (1996)

Campos,J. C., Borges, R. H. M., Olivera Filho,A. M., Nobrega, R. And Sant’ Anna Jr., G. L.,Oilfield Wastewater treatment by combinedmicro filtration and biological precess, WaterResearch36, 95-104 (2002)

Jin, B., van Leeuwen, H. J., Patel, B. and Yu,K. Utilization of starch processing wastewaterfor production of microbial biomass proteinand fungal -amylase byAspergilius orizae,Bioresource Technol.66:201-206 ( 1998).Lazarova, V., Meyniel, J., Duval, L. AndMenem, J.,A Novel circulating bed reactor,Hydrodynamics, mass transfer and nitrificationcapacity, Chem. Eng. Sci.52:3919-3927(1997)

Loh, K. and Liu, J., External loop inversedfluidized bed airlift bioreactor (EIFBAB)(2001) Chem. Eng. Sci.56:6171-6176 ( 2001)Shechter, R., Merchuk, J. C. and T. Ronen,Demonstration of an attached growth AirliftReactor for capacity Increase and Nitrification,WEFTEC 2002, Chicago, Sep.28-Oct 2 (2002)Shechter, R. and Merchuk, J. C., Modeling ofan Airlift Reactor with Floating Solids forWastewater Treatment, GLS-6, Vancouver,August (2003).

van Benthum, W.A.J., van den Hoogen,J.H.A., van der Lans, R.G.J.M., vanLoosdrecht, M.C.M. and Heijnen, J.J., Thebiofilm airlift suspension extension reactor.Part I: Design and two-phase hydrodynamicsChem. Eng. Sci. 54, 1909-1924 ( 1999-a)van Benthum, W.A.J., van der Lans, R.G.J.M.,van Loosdrecht, M.C.M. and Heijnen, J.J., Thebiofilm airlift suspension extension reactor.Part II: Three-phase hydrodynamics Chem.Eng. Sci. 54, 1909-1924 ( 1999-b)