Treatment of slaughterhouse wastewaters using anaerobic filters

Environmental Technology, 2014Vol. 35, No. 3, 322–332, http://dx.doi.org/10.1080/09593330.2013.827729

Treatment of slaughterhouse wastewaters using anaerobic filters

Sandra Luz Martineza, Vincenzo Torrettab, Jèsus Vázquez Minguelac, Faustino Siñerizd, Massimo Rabonib,Sabrina Copellib, Elena Cristina Radae∗ and Marco Ragazzie

aFaculty of Agronomy and Agroindustry, National University of Santiago del Estero, Av. Belgrano 1912, Santiago del Estero 4200,Argentina; bDepartment of Science and High Technology, Insubria University of Varese, Via G.B. Vico, 46, Varese I-21100, Italy;

cDepartment of Rural Engineering, Politechnic of Madrid, Ciudad Universitaria, Madrid 28040, Spain; d National Council of Technicaland Scientific Research (PROIMI-CONICET), Av.Belgrano y pje. Caseros, San Miguel deTucumán 4000, Argentina; eDepartment of

Civil, Environmental and Mechanical Engineering, University of Trento, Via Mesiano 77, Trento I-38123, Italy

(Received 15 April 2013; accepted 16 July 2013 )

In this paper, a laboratory-scale experimentation allowed comparing the performances of two upflow anaerobic packed-bed filters filled with different packing materials and operating at mesophilic conditions (30◦C) for treating slaughterhousewastewaters. Methane production was experimentally evaluated considering different volumetric organic loading rates as wellas feeding overloading conditions. Although filter performances declined with loading rates higher than 6 kg CODin m−3 d−1,the chemical oxygen demand (COD) removal efficiency remained always above 60%. The experimental results allowed fordetermining kinetic parameters for bacterial growth rate and methane production, following Monod and Chen–Hashimotomodels, respectively. Results demonstrated that the reactors reached a cellular retention time significantly greater than thehydraulic retention time. The kinetic parameter values (Ks, μmax) revealed the low microorganisms’ affinity for the substrateand confirmed the moderate biodegradability of slaughterhouse wastewater. The kinetic analysis also allowed the comparisonof the filters performances with another anaerobic system and the assessment of the parameters useful for real-scale plantdesign. The system design, applied to a medium-sized Argentinean slaughterhouse, demonstrated to (i) be energeticallyself-sufficient and (ii) contribute to the plant’s water heating requirements.

Keywords: anaerobic digestion, anaerobic filter, kinetics, slaughterhouse, wastewater

IntroductionThe management of slaughterhouses waste and wastewa-ters is a very significant problem under the environmentaland economic point of view. In particular, slaughterhousewastewaters have a high organic content, with a chem-ical oxygen demand (COD) which spans between 1100and 20,000 mg L−1.[1] Anaerobic digestion is a widelyused solution for wet residues treatment [2,3] with theaim of water pollution reduction and energy recovery. Theprocess converts a large part of COD into biogas (com-posed by methane) or biohydrogen, thanks to its highremoval efficiency. The anaerobic process has been stud-ied over the years from many points of view.[4] As aresult, both the conventional [5] and unconventional aspects[6–8] of such a process are adequately known.[9,10] Inorder to obtain a good removal of organic matter dur-ing anaerobic digestion, it is necessary to properly selectthe system to be implemented. In this frame, the attachedgrowth reactors are systems where bacteria are attached toan inert support, developing a biofilm.[11] This kind ofreactor is widely used for the food industry wastewaterstreatment.[12–14] There are many configuration of attached

∗Corresponding author. Email: [email protected]

growth reactors [11]: fluidized bed reactors, anaerobicexpanded bed reactors and upflow packed-bed reactors(UPBR) or filters (UPBF).[15–18]

Saravanan and Sreekrishnan [19] argued that to opti-mize the design and scale break, mathematical models areneeded. In anaerobic reactors, performances are affectedby the hydrodynamics of the reactor (i.e. pattern flow), themass transfer in the biofilm and the kinetic effects, which arealso influenced by the high loading rates and the presenceof toxic compounds.[20,21] The methods for analysing themodels that describe the studied systems are mainly basedon the process phenomenology and the concentration gra-dient. The determination of the parameters describing thebehaviour of a system can be accomplished using empiricalfacts or applying mathematical models. For model applica-tion, empirical data provided by operating plants, laboratoryand pilot-plant experiences, collected using effluents ofsimilar characteristics to those presently under study, arenecessary. However, the assay conditions, the technologyused and the loading rate cannot always be identical, addinga degree of uncertainty in the evaluation of alternatives.[22]Moreover, if the purpose is the process control, the use

© 2013 Taylor & Francis

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

Environmental Technology 323

of focused models is desirable (i.e. models that take intoaccount only the methanogenic stage or at most two of theprocess stages, the acidogenic and the methanogenic ones).

The present paper shows the experimental results ofa laboratory-scale UPBF fed with typical slaughterhouseresiduals. Two types of packing materials were consid-ered. Experiments, carried out under mesophilic conditions,allowed to (i) compare the filters performances (in termsof methane production), (ii) study reactors behaviour dur-ing feedstock overloading conditions and (iii) determinethe parameters governing the process kinetics for biomassgrowth and methane production yield, useful for real-scaleUPBF design.

Materials and methodsLaboratory-scale reactorThe adopted UPBF were two polyvinyl chloride (PVC)made tubular reactors with a 4:1 height to diameter ratio(Figure 1). The filters were composed of a 0.3 L mixingchamber (Vmc) for flow homogenizing and a 4.0 L anaer-obic reactor (Vr) separated by a perforated plate (Figure 1– Section C-C), which also supports the packing material.Moreover, the plate prevented the formation of preferentialpaths and avoided dead spaces with consequent yield losses.The liquid outlet was placed on the upper side of the reac-tor, while the gas one was near the lid; the released gas wascollected in airtight test tubes. Such kind of design createsa behaviour similar to an upflow anaerobic sludge blanketreactor.[23] Continuous feeding with identical flow rate (Q)and feedstock characteristics for both the filters was per-formed using a peristaltic pump equipped with two heads.Such configuration ensures an identical organic loadingrate (OLR).

Figure 1. Drawing of the reactors.

Flanges were welded at both ends of the tubular body toenable rapid system unblocking and facilitate the access forattached or suspended biomass sampling. Upper and lowerlids were sealed using rubber gaskets to ensure watertight-ness. Eight sampling ports were uniformly distributed alongthe tube (Figure 1 – Section B-B).

For operating in mesophilic conditions, the filters werearranged in a water bath made out of a simple fibrecoated with expanded polystyrene. The inner walls werealso coated with tiles equipped with a temperature-controlsystem.[24]

Packing materialTwo packing materials were tested: 5/8 in. polypropy-lene Pall rings (in Reactor 1; Figure 2(a)) and expandedpolyurethane foam (in Reactor 2; Figure 2(b)). Pall ringswere cylinders with slotted walls and internal partitions orribs which allow the biofilm growth as well as uniform liq-uid and gas flow patterns.[25] The expanded polyurethanefoam was prepared in the form of supporting disks (diam-eter: 100 mm and thickness: 0.7 mm) and stripes (length:200 mm; width: 20 mm and thickness: 0.7 mm).

Packing materials were analysed by determining thevoid volume (V0) and density (ρpm). For this purpose, the

Figure 2. Packing material: 5/8 in. Pall rings and reactor (a);expanded polyurethane foam: supporting disks and stripes (b).

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

324 S.L. Martinez et al.

packing material was placed in a test tube of a known vol-ume (Vtube); packing material mass was measured (mpm).Water was added to the test tube until the material wascovered. The added volume of water (Vw) was measured.The calculations were performed using the expressionsdescribed below:

V0 =(

VW

Vtube

)100, (1)

ρpm = mpm

(Vtube − Vw). (2)

Tests were repeated five times for both the materials.Average results were:

Pall rings: V0 1 = 86% ρpm1 = 130.0 kg m−3

Expanded polyurethane foam: V0 2 = 89%

ρpm2 = 91.3 kg m−3.

Feedstock selection and inoculum preparationThe feedstock selection was randomized to ensure that thedigesters feeding have been subjected to the same varia-tions as the real case. The feedstock was a mix of liquidwaste (e.g. blood) and washing water (e.g. meat and roomscleaning) coming from a cattle slaughterhouse in Santiagodel Estero (Argentina). The wastewater was collected usingPVC containers, which were transported in conditionedthermal boxes to the laboratory where they were kept inrefrigeration chambers at 2 ± 1◦C for preservation beforethe use.

The inoculum preparation procedure consists in pour-ing 2 L of sludge into 3 L vial sealed up with airtightrubber caps. Remaining space of the vial was free to con-tain the generated biogas. Two glass tubes of differentlengths were inserted into the vial caps to, respectively,feed the vial and collect the produced gas. The sludge wasacclimated at 30◦C by feeding it with increasing concentra-tion ratios of slaughterhouse wastewater. The concentrationratios (expressed as dilution factors: 1:100, 1:80, 1:50, 1:20,1:10, 1:5, 1:2 and 1:1) were increased according to theinoculum response, evaluated as a function of the volume ofproduced biogas.[17] Fresh substrate additions, also calledrefill, were 10% of the inoculum volume. The same volumewas withdrawn from the clearer supernatant. After the fill-ing operation was completed, the vials were homogenizedfor ensuring the contact between the sludge and the freshsubstrate.

Two different types of sludge were tested. The firstone was extracted from the bottom sediments of a slaugh-terhouse wastewater tank that showed anaerobiosis signs(release of bubbles). Bacteria were found to be very sen-sitive to pH and feedstock. For this reason, a secondfermented sludge extracted from a wastewater treatmentplant was used for additional experimentation; the sludge

satisfied the expectations regarding rapid adaptation to theslaughterhouse effluent with good results in terms of biogasproduction.

Performed analysesFeedstock pH and solids concentration (divided into total,suspended and considering also fixed and volatile part)were analysed. pH and COD were measured in both thereactor influent and effluent. The reactor effluent characteri-zation was completed with alkalinity (Alk) and volatile fattyacids (VFAs) concentrations analysis. Daily biogas produc-tion (Vbg, STP m3 d−1) and methane content (%CH4) wereanalysed.

The following techniques were used:

• Solids concentrations and Alk were measured accord-ing to standard methods [26];

• pH was measured by the means of an Orion 520 pHmeter with an Orion 91-04 glass electrode;

• COD was determined by the means of a colorimet-ric method with concentrated reagents (1 N K2Cr2O7,H2SO4 and Ag2SO4 solutions); the method was basedon the semi-micro digestion in a reactor (Hach CODmodel 16,500) and a spectrophotometric technique(Hitachi U-2000); glucose and potassium biphthalatewere used to prepare the concentration–absorbanceresponse for COD patterns [26];

• VFAs were determined following a titration methoddescribed by Lahav and Morgan [27]

• Vbg was measured by the means of a volumetricmethod [28] and

• %CH4 was measured by taking a given volume ofproduced biogas with a perfectly lubricated syringeand treating it with a 50% w/w KOH solution [28];the determinations were verified afterward using agas chromatographic technique.

Experimental procedureIn order to analyse the influence of the packing materialon the reactors performances, constant COD concentra-tion in the feedstock was maintained between 12 and16 g COD L−1. When necessary, diluted blood was added.Five operating conditions were tested, varying flow rateand, as consequence, hydraulic retention time (HRT =VrQ−1) and (OLR = CODinHRT−1). Temperature was setto 30 ± 2◦C.

Each test was maintained for approximately one monthafter the reactors reached steady state conditions. Duringthe digesters start-up, the pH of the inflow was adjusted at5–5.5 in order to counterbalance the increase in pH relatedto NH3 generation.

Experiments were carried out considering four operat-ing parameters (CODin, HRT, OLR and packing materialtype, respectively; see Table 2) and monitoring Vbg, %CH4,

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

Environmental Technology 325

VFAout and CODout.[29] The COD removal efficiencies(RECOD) were computed for each operating condition.

Kinetic parameters determinationIn the preliminary tests performed at the laboratory scaleby the means of a tracer, the packed-bed anaerobic filterspresented flow patterns resembling the completely stirredtank reactors (CSTR) due to bubble motion within thesystem.[30] Therefore, the CSTR kinetic models can beapplied to UPBR.

Substrate degradation modelFor the stages that follow hydrolysis, substrate consumptionis well defined using the model proposed by Monod.[8]Therefore, the determination of the kinetic parameters forsubstrate degradation can be performed using the Monodequation in Lineweaver–Burk linearized form [19]

1μ

= Ks

μmax

1S

+ 1μmax

, (3)

where μ is the bacteria specific growth rate (d−1), Ks is thesaturation constant (mg COD L−1), μmax is the maximumspecific growth rate (d−1) and S is the substrate concentra-tion (mg COD L−1). With at least a pair of values of 1/μ

and 1/S and applying linear regression, kinetic parameters(Ks and μmax) can be obtained and compared with similarreference values of previous works [31] using the followingformulas:

μmax = q−1, (4)

KS = mμmax, (5)

where m and q are, respectively, the gradient and y-interceptof the linear equation resulting from Lineweaver–Burkplot. The parameter values intrinsically show the packingmaterial retention effects. But the comparison of differentsystems can be difficult.

Considering a reference system 0 with a similarhydraulic behaviour, the kinetic parameters of a system ican be expressed as

μmax i = μmax 0 Fi, (6)

KS i = KS 0Gi, (7)

where F and G are, respectively, dimensionless factors thatquantitatively describe filter effectiveness. Any system ismore efficient than the reference one when greater is F andlower is G. Therefore, Equation (3) becomes

1μ

= GiKS 0

Fiμmax 0

1S

+ 1Fiμmax 0

. (8)

Differences of F and G in the filters may indicate that thematerial influences not only the number of microorganisms

retained but also their type and kinetics. The effect of reten-tion is usually quantified from μmax, but F and G displaymore clearly the retention effect of, respectively, biomassand solids; the factors can be considered as a packing mate-rial feature and can be associated with the material densityand roughness. As consequence, Equation (8) has beenadopted to estimate the HRT required for a known substrateconcentration S, taking into account that HRT = μ−1.

Methane production modelAccording to the Chen–Hashimoto model,[32] the methaneproduction of a completely mixed anaerobic digester can becalculated using the following expression:

YCH4 i = YCH4 max i

[1 − ki

μmax i HRTi − 1 + ki

], (9)

where YCH4 and YCH4 max are, respectively, the reac-tor and maximum methane-specific production rate(m3 CH4 kg−1 CODin d−1) and k (dimensionless) is a kineticperformance parameter which depends on the inflow solidsconcentration.

It can be observed that YCH4 is zero when the HRT =μ−1

max. This finding means that μmax determines the min-imum HRT (HRTmin) below which fermentation is notpossible: HRTmin > μ−1

max. On the contrary, above a certainvalue of HRT the methane production performance is notpractically influenced.

The balance proposed by Chen and Hashimoto is thefollowing:

YCH4 i = YCH4 max i

(1 − Si

Sin i

), (10)

where Sin is the inlet substrate concentration and S isthe reactor concentrations, which can be expressed usingEquation (8) developed by the Monod equation for filterswhere biomass decay is assumed negligible (as demon-strated by [33,34])

Si = GiKS 0

HRTiFiμmax 0 − 1. (11)

The following relation for filters results from couplingEquation (11) to the Chen–Hashimoto model Equation (10):

YCH4 i = YCH4 max i

[1 − GiKS 0

Sin i(HRTiFiμmax 0 − 1)

]. (12)

The equation allows for the calculation of the maximumproduction of methane that can be obtained with the anaer-obic filters by applying kinetic parameters obtained fromthe Monod model.

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

326 S.L. Martinez et al.

Reactor modelThe concentrations of both the solids and the microorgan-isms in the digester are usually evaluated as the quotient oftheir total mass over the reactor volume, although most ofthem adhere to the packing material. Therefore, the solidand microorganism concentration in the effluent are lowerthan that into the reactor. The packing material effectivenesscan be established by determining the microorganism-retention factor (f , dimensionless; 0 < f ≤ 1) and thesolid-retention factor (g, dimensionless; 0 < g ≤ 1)

f = xout

xr, (13)

g = Sout

Sr, (14)

where x is the biomass concentration while the subscript rand out indicates, respectively, the situation into the reactorand in the effluent.

Hence, the reactor kinetic model is completely definedusing Equation (12), the equality HRTmin = μ−1

max and thefollowing equations:

θx = HRTf

, (15)

θS = HRTg

, (16)

where θx and θS are, respectively, the cellular and the solidsretention time. Equation (15) shows that, for UPBF (f < 1),the cellular retention time (CRT) is higher than that in acompletely mixed reactors. This finding means that UPBFallow to work with lower HRT, and hence with smallerreactors.

Results and discussionFilters performanceTable 1 shows the results of the influent characterization.The pH is slightly basic. A high percentage (87%) of thetotal solids (TS) is composed of volatile matter and about69% of the TS are suspended.

During the reactors start-up, it was observed that the pHincreased also over 8, with a lowering of biogas production,as suggested also by technical literature [35] which states anoptimal range at 6.6–7.4. The phenomenon is due to the high

Table 2. Main characteristics of the experimentations.

HRT CODin OLR (kg COD pHin pHoutReactor (d) (mg L−1) m−3 d−1) (−) (−)

1 1.90 14,683 7.726 5.1 7.971.95 15,800 8.102 5.5 8.202.12 15,196 7.165 5.2 7.982.50 14,130 5.640 5.3 7.983.00 12,000 4.000 5.5 7.95

2 1.90 14,683 7.726 5.1 8.011.95 15,800 8.102 5.5 7.942.12 15,196 7.165 5.2 7.982.50 14,130 5.640 5.3 7.973.00 12,000 4.000 5.5 8.01

proteins content of slaughterhouse wastewaters which cre-ates a high production of NH3 during the hydrolysis stage,which neutralizes the acidity generated by methanogenicbacteria. Therefore, the pH of the influent (Table 1) hadto be adjusted. The operating conditions are listed inTable 2.

The results were found consistent with the literatureregarding anaerobic filters,[13] which describes experi-ences with typical OLR of 16 kg COD m−3 d−1, HRTequal to about 0.5–4.0 d and influent concentrations of1–20 g CODin L−1.

The effluent pH was constant at about 8, while the alka-linity and the VFA content increased with the increase in theOLR (Figure 3). The high VFA concentration (Figure 3(b))demonstrated the partial activity of the anaerobic bacteria.

The experiments carried out with the lowest OLR (cor-responding to the highest HRT, equal to 2.5 and 3 d) showedproduction of about 0.12 m3 CH4 kg−1 CODin (Figure 4(a)).Generally, the filter filled with polyurethane foam demon-strated higher methane production. The result was duealso to the higher methane content in the produced biogas(Figure 4(b)), above all with higher OLR. Such increase isprobably due to the best performances in VFA degradation(Figure 3(b)).

In addition, the OLR responses in both of the reactorsexhibited overloading effects at about 6 kg CODin m−3 d−1,stressed by a drastic reduction in methane production anda slight decline in the COD removal (Figure 5). Both thereactors demonstrated good stability during the steady stateconditions period, with fluctuations of about 5% in CODoutand methane production.

Table 1. Characteristics of the feedstock (inlet).

TS (mg L−1) Suspended solids (mg L−1)

Total Volatile Fixed Total Volatile FixedT (◦C) pH (−) (TS) (TVS) (TFS) (TSS) (VSS) (FSS)

Mean 22 7.5 926 809 117 637 579 58Standard deviation 1.0 0.3 56 47 6 41 17 3Coefficient of variation (%) 4.5 4.5 6.0 5.8 5.5 6.5 3.0 4.5

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

Environmental Technology 327

Figure 3. Evolution of alkalinity (a) and VFAs (b) in the effluent at different OLRs for the two filters.

Kinetic parameters

Substrate degradation model

A comparison between the two UPBF and a reference anaer-obic CSTR working at similar operating conditions (tem-perature: 30◦C; HRT: 2/8 d; CODin: 9800 mg L−1; OLR:1.2/4.9 kg COD m−3 d−1) with slaughterhouse wastewaters[36] has been carried out in order to identify the effects oftwo packing materials on bacterial and solid retention withrespect to a standard anaerobic CSTR. Results were mod-elled by applying Equation (3). The Lineweaver–Burk plot(Figure 6) shows the experimental results.

The kinetic parameters (μmax and KS ) as well as thedimensionless factors (F and G) for both the filters wereobtained by applying Equations (3)–(5); results are listed inTable 3.

The results showed that the filters were more efficientthan the reference CSTR due to two simultaneous effects:bacterial and solids retention (F and G, respectively). Infact, F1 (Pall rings) was higher than that of F2 (expanded

polyurethane foam). The greater efficiency of Filter 2 canbe asserted because of the low value of G. The difference ofG, and consequently of K , between the filters indicates thatthe material influenced not only the number of microorgan-isms retained but perhaps also the type. The difference alsoreveals that a significant difference in the solids retentiontime existed.

Comparing the obtained parameters with those reportedin the literature,[17] it can be observed that the obtainedμmax are lower than those of slower methanogenic reac-tors. This finding enables to assert that the chosen packingmaterials allow the growth of attached bacteria and, there-fore, the cellular retention increases. For this reason, itwas possible to operate the reactors at higher dilution ratesor at lower HRT. Considering a slaughterhouse wastewa-ter with a concentration of 10 g COD L−1, both the filtersHRT are threefold lower than that of the analogous com-pletely mixed reactor. In synthesis, the results demonstratethat the reactors reached a CRT significantly greater thanthe HRT.

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

328 S.L. Martinez et al.

Figure 4. Methane specific production rate (a) and methane content (b) in biogas produced by the two filters.

Figure 5. COD removal efficiencies of the two tested biofilters.

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

Environmental Technology 329

Figure 6. Comparison between substrate degradation of the UPBfilters and a reference completely mixed anaerobic reactor [27]kinetic performances.

Table 3. UPB filter kinetic parameters and comparison with acompletely mixed anaerobic reactor.[27]

CSTR – Filter 2reference Filter 1 (polyurethane

Parameter unit case [30] (Pall rings) foam)

μmax i d−1 0.58 5.60 2.26KS i g L−1 9.05 50.89 12.80Fi – 1 9.65 3.89Gi – 1 5.62 1.41

Moreover, the elevated Ks values reveal the lowmicroorganisms’ affinity for the substrate and confirms themoderate biodegradability of slaughterhouses effluents.[37]

Kinetic parameters of biogas productionThe methane specific production rate results at differentfeeding rates (D = HRT−1) are shown in Figure 7. The max-imum production of methane was obtained by applying themodified Chen–Hashimoto Equation (15) on the two results

obtained in steady state conditions with the highest HRT;average results were:

YCH4max 1 = 0.162 m3 CH4 kg−1 CODin

for Pall ring filter,

YCH4max 2 = 0.158 m3 CH4 kg−1 CODin

for Expanded polyurethane foam.

In terms of maximum methane production, the obtainedresults with the model are slightly better using Pall ringsbut, considering the methane production obtained with real-istic operating conditions, expanded polyurethane foamhas better performances: for example, operating with aHRT = 3 d the YCH4 are 0.125 and 0.133 m3 CH4 kg−1

CODin for Pall rings and polyurethane foam, respectively.This is explained by the low Ks obtained by using expandedpolyurethane foam.

Case-study application of the kinetic model resultsConsidering a UPBF working with a certain HRT at 30◦C,the active reactor volume (Vr) can be simply defined byknowing the average daily slaughterhouse effluent flow rate(Qsl, m3 d−1):

Qsl = nh wh vww, (17)

where nh is the number of daily slaughtered cattle head(head d−1), wh is the average cattle head weight (kg head−1)and vww is the volume of wastewater produced for each unitweight of slaughtered head (m3 kg−1).

The methane volumetric production rate (rv , m3

CH4 m−3 reactor d−1) can be evaluated as follows:

rv = YCH4 OLR. (18)

The OLR is computed by using the classical formula, whileYCH4 is computed using Equation (12), and the previously

Figure 7. Methane specific production rate (YCH4 ) at different feeding rates (D).

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

330 S.L. Martinez et al.

Table 4. Estimation of methane production in a medium-sized Argentinean slaughterhouse.

Description unit Filter 1 Filter 2

Step 1: Reactor design and gross thermal energy production from biomethaneNumber of daily slaughtered cattle head, nh head d−1 100a

Average cattle head weight, wh kg head−1 350a

Volume of wastewater produced, vww m3 kg−1 0.005b

Effluent flow rate, Qsl m3 d−1 175HRT d 3.15Overall reactors volume, Vr m3 551.3OLR kg CODin m−3 d−1 4.6Methane volumetric production rate, rv m3 CH4 m−3 d−1 0.583 0.614Daily methane production, PM m3 CH4 d−1 321 338Gross thermal energy production, Ec

pot MJ d−1 10,069 10,611Step 2: Reactor heat energy assessment

Energy for wastewater heating, Eheat ww MJ d−1 5868Energy recovery from effluent, Erec out MJ d−1 −1174Heat losses from walls, El wall MJ −1 387Biogas heat losses, El bg MJ d−1 22.1 19.7Reactor heat energy requirements, Er MJ d−1 5104 5101Net thermal energy production, Enet prod MJ d−1 4965 5509Hot water production m3 d−1 19,770 21,935

% slaughterhouse requirement 41 46

a[32,36].b[37].cCalculated considering the lower heating value of CH4 equal to 50 MJ kg−1 [33] and boiler efficiency.

obtained kinetic parameter YCH4 max. Therefore, the pro-duced methane (PM) is obtained by multiplying rv and Vr .The application of such a method to a medium-sized Argen-tinean slaughterhouse [38] with a wastewater similar tothe previously analysed one for COD content (average ofthe values listed in Table 3) and temperature (Table 1) ispresented.

The proposed simplified plant is composed of threeUPBR unit (in order to facilitate maintenance operations),a boiler (with an efficiency of 80%) for heating (i) influ-ent at the reactor operating temperature (30◦C) and (ii)water for slaughterhouses services, such as meat treatmentand production, rooms cleaning and internal laundry (about0.48 m3 head−1 [38] at 80◦C in order to balance distributingnetwork heat loss).

The energy requirements take into account [39,40]:wastewater heating, heat recovery from effluents by themeans of a countercurrent heat exchanger (efficiency: 20%)and reactor heat loss from the walls (overall heat transfercoefficient: 0.83 W m−2◦C−1) as well as from the producedbiogas (composed of water vapour, methane and carbondioxide).

After an optimization procedure [41] based on the reac-tor space requirements and the methane production, anHRT = 3.15 d has been chosen. The results are listed inTable 4. For both the packing materials, the reactor heatenergy requirements (Er) are about 51% of Epot, the grossthermal energy production from methane combustion (47%for the influent heating and 4% for the heat losses). Theresidual thermal energy production can contribute to cover

about 41% of the factory hot water requirements (46% ifpolyurethane foam is used as packing material).

ConclusionsThe paper presents a contribution to the study of solutionsfor the management of slaughterhouse wastewater [42,43]by using upflow anaerobic packed-bed filters filled withtwo different packing materials (Pall rings and expandedpolyurethane foam). The results showed that the two filters,operating in mesophilic conditions (30◦C), were more effi-cient than a completely mixed system due to an increasedbacterial and solids retention.

Although filter performances declined with OLR higherthan 6 kg CODin m−3 d−1, the COD removal efficiencyremained always above 60%. In general, better results wereobtained with expanded polyurethane foam.

The application of kinetic models allowed defining theparameters and a procedure useful for real-scale upflowanaerobic packed-bed filters design.

The design of an upflow anaerobic packed-bed reactorfor treating the medium-sized Argentinean slaughterhousewastewaters demonstrated that the system could reduceenergy requirements, therefore the carbon footprint of themeat production sector.

References[1] Fuchs W, Binder H, Mavrias G, Braun R. Anaerobic treat-

ment of wastewater with high organic content using a stirred

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

Environmental Technology 331

tank reactor coupled with a membrane filtration unit. WaterRes. 2003;37:902–908.

[2] Tokumoto H, Tanaka M. Novel anaerobic digestion inducedby bacterial components for value-added byproducts fromhigh-loading glycerol. Bioresource Technol. 2012;107:327–332.

[3] Allen E, Browne JD, Murphy JD. Evaluation ofthe biomethane yield from anaerobic co-digestion ofnitrogenous substrates. Environ Technol. 2013. Avail-able from: http://www.tandfonline.com/doi/full/10.1080/09593330.2013.806564

[4] Andreottola G, Ragazzi M, Foladori P, Villa R, LangoneM, Rada EC. The unit integrated approach for OFMSWtreatment. Sci Bull. 2012;74:19–26.

[5] Rada EC, Ragazzi M, Torretta V. Laboratory-scale anaerobicsequencing batch reactor for treatment of stillage from fruitdistillation. Water Sci Technol. 2013;67:1068–1074.

[6] Rada EC, Ragazzi M. Critical analysis of PCDD/F emis-sions from anaerobic digestion. Water Sci Technol. 2008;58:1721–1725.

[7] Torretta V, Conti F, Leonardi M, Ruggeri G. Energy recoveryfrom sludge and sustainable development: a tanzanian casestudy. Sustainability. 2012;4:2661–2672.

[8] Ylitervo P, Akinbomi J, Taherzadeh MJ. Membranebioreactors’ potential for ethanol and biogas produc-tion: a review. Environ Technol. 2013. Available from:http://www.tandfonline.com/doi/full/10.1080/09593330.2013.813559

[9] Braguglia CM, Gianico A, Mininni G. Comparison betweenozone and ultrasound disintegration on sludge anaerobicdigestion. J Environ Manage. 2012;95:S139–S143.

[10] Callegari A, Torretta V, Capodaglio AG. Preliminary trialapplication of biological desulfonation in anaerobic digestorsfrom pig farms. Environ Eng Manage J. 2013;12(4):815–819.

[11] Tchobanoglous G, Burton FL, Stensel HD. Wastewater engi-neering: treatment and reuse. New York: McGraw-Hill;2003.

[12] Chan YJ, Chong MF, Law CL, Hassell DG. A reviewon anaerobic-aerobic treatment of industrial and municipalwastewater. Chem Eng J. 2009;155:1–18.

[13] Debik E, Coskun T. Use of the Static Granular Bed Reactor(SGBR) with anaerobic sludge to treat poultry slaugh-terhouse wastewater and kinetic modeling. BioresourceTechnol. 2009;100:2777–2782.

[14] Tawfik A, El-Kamah H. Treatment of fruit-juice industrywastewater in a two-stage anaerobic hybrid (AH) reactor sys-tem followed by a sequencing batch reactor (SBR). EnvironTechnol. 2012;33:429–436.

[15] Acharya BK, Mohana S, Madamwar D. Anaerobic treatmentof distillery spent wash – a study on upflow anaerobic fixedfilm bioreactor. Bioresource Technol. 2008;99:4621–4626.

[16] Del Pozo R, Diez V. Integrated anaerobic-aerobic fixed-filmreactor for slaughterhouse wastewater treatment. Water Res.2005;39:1114–1122.

[17] López I, Borzacconi L. Modelling of slaughterhousesolid waste anaerobic digestion: determination of param-eters and continuous reactor simulation. Waste Manage.2010;30:1813–1821.

[18] Curio CH, Geissen S, Knodel J, Roeckel von Bennewitz M,Lillo EA. Modelling of integrated anoxic–anaerobic–aerobictreatment for salmon fishery wastewater in an upflow fixed-bed biofilm reactor. Environ Technol. 2012;33:607–622.

[19] Saravanan V, Sreekrishnan TR. Modelling anaerobic biofilmreactors – a review. J Environ Manage. 2006;81:1–18.

[20] Torretta V. PAHs in wastewater: removal efficiency in aconventional wastewater treatment plant and comparison

with model predictions. Environ Technol. 2012;33:851–855.

[21] Torretta V, Katsoyiannis A. Occurrence of polycyclic aro-matic hydrocarbons in sludges from different stages ofa wastewater treatment plant in Italy. Environ Technol.2013;34:937–943.

[22] Vázquez Minguela J. Biometanización de residuos orgánicosagroindustriales. Madrid, Spain: Universidad Politécnica deMadrid; 2010.

[23] Del Pozo R, Diez V, Salazar G, Espinosa JJ. The influenceof influent distribution and blood content of slaughterhousewastewater on the performance of an anaerobic fixed-filmreactor. J Chem Technol Biotechnol. 2006;81:282–288.

[24] Massé DI, Masse L. Characterization of wastewater from hogslaughterhouses in Eastern Canada and evaluation of their in-plant wastewater treatment systems. Canadian Biosyst Eng.2000;42:139–146.

[25] Rivera A, Guerrero B, Nieves G. Tratamiento de efluentesde la producción de leche de soya en un filtro anaerobio aflujo pistón. Rev Int Contam Ambient. 2004;20:117–128.

[26] Clesceri LS, Greenberg AE, Eaton AD. Standard methodsfor the examination of water and wastewater. Washington,DC: APHA American Public Health Association; 1998.

[27] Lahav O, Morgan BE. Titration methodologies for monitor-ing of anaerobic digestion in developing countries – a review.J Chem Technol Biotechnol. 2004;79:1331–1341.

[28] Valcke D, Verstraete W. A practical method to estimatethe acetoclastic methanogenic biomass in anaerobic sludges.J Water Pollut Control Federation. 1983;55:1191–1195.

[29] Huang JS, Jin CG. Deep-biofilm kinetics of substrate utiliza-tion in anaerobic filters. Water Res. 1997;31:2309–2317.

[30] Soroa S, Gomez J, Ayesa E, Garcia-Heras JL. Mathemat-ical modelling of the anaerobic hybrid reactor. Water SciTechnol. 2006;54:63–71.

[31] Vázquez Minguela J. Estudio de los parámetros de labiometanización de aguas residuales de mataderos y suaplicación al diseño de digestores. Madrid: UniversidadPolitécnica de Madrid; 1986.

[32] Kumar S. Biogas, InTech, 2012. Available from: http://www.intechopen.com/books/biogas

[33] Pérez J, Aldana G, Cárdenas C. Upflow anaerobic sludgeblanket reactor (UASB) performance through sludge ageload and kinetic coefficients. Rev Téc Ing Univ Zulia.2012;35–1:98–108.

[34] Coelho N, Capela I, Droste RL.: Application of ADM1 toa UASB treating complex wastewater in different feedingregimes. Proceedings of the WEFTEC; Dallas, TX: 2006,p. 7123–7135.

[35] Méndez-Romero DC, López-López A, Vallejo-Rodríguez R,León-Becerril E. Hydrodynamic and kinetic assessment ofan anaerobic fixed-bed reactor for slaughterhouse wastew-ater treatment. Chem Eng Process: Process Intensification.2011;50:273–280.

[36] Martínez SL. Diseño y construcción de un filtro anaeróbicopara el tratamiento de efluentes de matadero. Santiago delEstero, Argentina: Universidad Nacional de Santiago delEstero; 2005.

[37] López-López A, Vallejo-Rodriguez R, Méndez-Romero DC.Evaluation of a combined anaerobic and aerobic systemfor the treatment of slaughterhouse wastewater. EnvironTechnol. 2010;31:319–326.

[38] Food and Agriculture Organization of the United Nations(FAO-UN), Structure and functioning of medium-sizedslaughterhouses in Developing Countries [Original titlein Spanish: Estructura y funcionamiento de mataderosmedianos en países en desarrollo], FAO Animal Pro-duction and Health Paper, ISBN 92–5-303148-4, 1993.

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14

332 S.L. Martinez et al.

Available from: http://www.fao.org/docrep/004/T0566s/T0566S00.htm

[39] Hashimoto AG, Chen YR, Varel VH.: Anaerobic fermen-tation of beef cattle manure – Final report, 1981, U.S.Department of Energy.

[40] Maestri F, Copelli S, Rota R, Gigante L, Lunghi A,Cardillo P. Simple procedure for optimally scaling-up finechemical processes. I. Practical tools. Ind Eng Chem Res.2009;48(3):1307–1315.

[41] Maestri F, Copelli S, Rota R, Gigante L, Lunghi A, CardilloP. Simple procedure for optimal scale-up of fine chemicalprocesses. II. Nitration of 4-chlorobenzotrifluoride. Ind EngChem Res. 2009;48(3):1316–1324.

[42] C.d.E.d.C.A. ABC, Argentinean meat exporters consor-tium. Available from: http://www.abc-consorcio.com.ar

[43] Del Pozo R, Diez V, Beltrán S. Anaerobic pre-treatment of slaughterhouse wastewater using fixed-filmreactors. Bioresource Technol. 2000;71:143–149.

Dow

nloa

ded

by [

Uni

vers

ity o

f U

dine

] at

07:

59 2

2 Ja

nuar

y 20

14