-
2745 IWA Publishing 2012 Water Science 8, Technology | 66.12 |
2012
Estimates of methane ioss and energy recovery potentiaiin
anaerobic reactors treating domestic wastewaterL. C. S. Lobato, C.
A. L. Chernicharo and C. L. Souza
ABSTRACT
This work aimed at developing a mathematical model that could
estimate more precisely the fractionof chemical oxygen demand (COD)
recovered as methane in the biogas and which,
effectively,represented the potential for energy recovery in upflow
anaerobic sludge blanket (UASB) reactorstreating domestic
wastewater. The model sought to include all routes of conversion
and losses inthe reactor, including the portion of COD used for the
reduction of sulfates and the loss of methane inthe residual gas
and dissolved in the effluent. Results from the production of
biogas in small- andlarge-scale UASB reactors were used to validate
the model. The results showed that the modelallowed a more
realistic estimate of biogas production and of its energy
potential.Key words | anaerobic reactors, biogas, COD balance,
domestic wastewater, energy recovery, UASB
reactor
L. c. s. Lobato (corresponding author)C. A. L. ChernicharoC. L.
SouzaFederal University of Minas Gerais.B.
Horizonte.BrazilE-mail:/s/tofot)ato@ya/joo.co/n.r
INTRODUCTION
The mass balance of chemical oxygen demand (COD) toestimate the
recovery of methane and energy in anaerobicreactors usually does
not consider the portion of CODused in sulfate reduction, nor the
portions lost as dissolvedmethane in the effluent or emitted to the
atmosphere. It isknown that the portion of COD used in sulfate
reductionis small due to the low concentration of sulfates in
domesticwastewater, usually in the range of 20-100 mg SO^^
L"^(Singh & Viraraghavan 1998; Metcalf & Eddy 2003), but
itis still important to consider it in the models that estimatethe
production of methane. On the other hand, in relationto the COD
converted to methane, a significant portioncould be dissolved in
tbe liquid pbase and be lost witb tbefinal effluent (Agrawal et al
1997; Hartley & Lant 2006;Souza & Cbernicbaro 2on).
Furthermore, methane losscan also occur due to emissions on the
settler surface ofupflow anaerobic sludge blanket (UASB) reactors.
Measure-ments taken by Souza & Cbernicbaro (2on) indicated
tbat,of all the methane produced in UASB reactors treating
dom-estic wastewater, tbe portion dissolved in tbe effluent
variedfrom 36 to 40%, wbile tbe portion emitted on tbe surface
ofthe settlers was in the order of 4%, constituting the wastegas.
To explain these huge losses of dissolved methane.Hartley &
Lant (2006) developed the hypothesis that themethane dissolved in
tbe effluent of different types ofanaerobic reactors could be
supersaturated in relation to
tbe saturation calculated according to Henry's law. There-fore,
the measurements taken by Souza & Chernicharo(2on) in anaerobic
reactors treating domestic (low concen-tration) wastewater and
operated at a very low organicloading rate (around 1.5 kg COD m"^
d"^) are in accord-ance with tbis hypothesis.
The loss of methane dissolved in tbe effluent or in thewaste gas
not only represents a loss of potential energy butalso contributes
to the emission of greenhouse gases. Pierotti(2007) reports a mass
balance that considers tbe portion ofCOD converted to metbane to be
divided into methane inthe biogas and metbane dissolved in tbe
effluent, in percen-tages from 20 to 25% of tbe influent COD, for
both portions.The same mass balance sbows tbe percentage of 40-50%
fortbe COD effluent and 10% for tbe COD that is convertedinto
sludge. This mass balance takes into account the por-tion of COD
converted to metbane and its division;however, it does not include
tbe portion tbat is due tosulfidogenesis.
Considering tbat the potential for production and recov-ery of
biogas in UASB reactors tbat treat domestic waste isconsidered low
(Noyola et al 2006), rarely resulting insome type of energy use
(the biogas is usually burned), thedevelopment of models that
permit more precise estimatesof the effective potential energy tbat
may be recovered, aswell as the emission factors (losses), becomes
important.
doi: 10.2166/wst.2012.514
-
2746 L. C. S. Lobato et al. Estimates of methane loss and energy
in anaerobic reactors Water Science & Technology | 66.12 |
2012
Thus, the COD mass balance models must incorporate allthe main
routes of conversion and of loss, that is: the portionused in
sulfate reduction; the portion converted into sludge,which may be
subdivided into the fraction remaining in thereactor and the
portion that is lost with the effluent; the dis-solved portion not
converted into methane and dischargedwith the effluent; the portion
converted into methane thatis recovered as biogas (which can be
used as energysource); as well as the portion converted into
methanethat escapes dissolved in the effluent or as waste
gas(losses). So, the aim of this study was to develop a
math-ematical model that better represented the mass balance ofCOD
and the potential for energy recovery in UASB reac-tors treating
domestic wastewater.
MATERIAL AND METHODS
The mathematical model for calculating the mass balance ofCOD
and the energy potential was structured with areduced number of
input data, with the goal of facilitatingits use in a broader way.
Conceptually, the model was struc-tured according to the COD
conversion routes and methaneflow in UASB reactors shown in Figure
1.
The model was developed considering three scenariosthat lead to
different methane recovery potential: (i) worstscenario; (ii)
typical scenario; (iii) best scenario. The worstscenario, in which
the energy potential is lower, involvessystems operating with more
diluted waste, higher sulfateconcentrations, lower COD removal
efficiencies, andhigher rates of methane loss. The best scenario
involves sys-tems operating with more concentrated waste, lower
sulfateconcentrations, higher COD removal efficiencies, and
lower
rates of methane loss. The typical scenario uses intermedi-ary
values for the input data.
Input data
The input data considered in the simulation are shown inTable 1.
The variability of the input data (Pop, QPC,QPCcoD and T) was
incorporated into the interpretationof the model results, through
Analysis of Uncertainty,which is based on a large number of
simulations (in thiscase, 250 simulations for each scenario),
making the so-called Monte Carlo simulation. For each run of the
model,a different set of values for the input data is chosen
forwhich uncertainty exists. The input data were generated
ran-domly following uniform distribution and within pre-established
ranges.
Fractions of the mass baiance of COD and of potentiairecovery of
CH4
Once the input data were defined, the portions of CODremoved
from the system, converted into sludge and con-sumed in sulfate
reduction, are firstly estimated. Based onthese portions, the total
COD converted into CH4 andthe consequent volumetric production are
calculated. Inorder to calculate the volume of CH4 actually
availablefor energy use, the model considers the losses of CH4
dis-solved in the effluent and in the gaseous phase with thewaste
gas, in addition to other eventual losses in the gas-eous phase
(e.g. leaks). Finally, deducting these losses,the potential
available energy is calculated. The equationsused to calculate all
the portions of the mass balance ofCOD and the potential for energy
recovery are shown inTable 2.
COD converted into CH4present in the biogas
* - . - . _ . i . J COD converted into CH4 and lost into the
atmosphere
Load of infiuent COD tnr -tf
%
w * -
i,I 1
-=HL
i
1 -
COD converted into CH4 and lost with the waste gas
COD converted into CH and lost dissoived in the effluent
COD not converted into CH4, and lost with the effluent
COD used by the BRS in sulfate rduction
COD converted into sludge
Figure 1 I COD conversion routes and methane fiow in UASB
reactors.
-
2747 L. C. S. Lobato e al. Estimates of methane loss and energy
in anaerobic reactors Water Science & Technology | 66.12 |
2012
Table 1 Input data considered in the model
scenario
Unit worst Typical BestReference
Contributing population (Pop)Per-capita wastewater
contribution
(QPC)Per-capita COD contribution
(QPCCOD)
Expected efficiency of COD removal
Dissolved CODCH4 lost with tbeeffluent (p^)
Percentage of CH4 in tbe biogas(CcH4)
inbab.
Sulfate concentration in the influent kg SO4 m(Cso,)
Efficiency of sulfate reduction {Eso) "/oOperational temperature
of tbe C
reactor (T)CODcH4 lost as waste gas (pw) %Otber CODcH4 losses
(e.g. biogas %
leaks) (po)kg m"^
0/0
1,000-1,000,000
0.12-0.22 von Sperling & Chernicharo (2005)
0.09-0.11 von Sperling & Chernicharo (2005)
60 65 70 von Sperling & Chernicharo (2005)0.08 0.06 0.04
Singb&Viraraghavan (1998); Metcalf& Eddy (2003);
Gloria et al (2008)80 75 70 Souza (2010)20-30 von Sperling &
Cbernicbaro (2005)
7.5 5.0 2.5 Souza & Cbernicbaro (2on)
7.5 5.0 2.5 Souza & Cbernicharo (2on)
0.025 0.020 0.015 Souza & Cbernicharo (2on)
70 75 80 von Sperling & Cbernicbaro (2005)
Validation of the mathematical modei
Following the simulations, the validation of the model
wascarried out based on measured biogas production and com-position
in small- and large-scale UASB reactors, accordingto the main
characteristics shown in Table 3, and thedescriptions given
below.
RESULTS AND DISCUSSION
Simulations
Figure 2 shows the average values obtained from the simu-lations
performed, in which the contribution of eachportion of the mass
balance of COD may be observed forthe three simulations analyzed
(worst, typical and best).
The analysis of Figure 2 supports the followingcomments:
13-15% of the COD applied to the system was convertedinto
biomass. In relation to simulations of sulfate concen-tration in
the influent (varying between 40 and 80 mg L~^),the percentage COD
use in sulfate reduction variedbetween 3 and 7%.
For the simulated rates of methane loss in the effiuent(varying
between 15 and 25mgL"^), 11-17% (average)of the COD load applied to
the system was convertedinto non-recovered methane in the biogas,
but lost dis-solved in the effluent.The portion of influent COD
effectively converted into CH4present in the biogas, which
represents the effective poten-tial of energy recovery, varied from
19% in the worstscenario to 39% in the best scenario. As it
involves domesticwaste, in which the relationship COD/SO4" is high,
metha-nogenesis (the sum of the portions of COD converted intoCH4)
is greater in relation to suldogenesis (39-52% con-trasted to 3-7%
of the influent COD, respectively).In the worst scenario, of all
the COD converted into CH4,on average 39%, only 19% refers to the
portion of CH4collected in the three-phase separator and available
foruse. This represents a loss of about 50% of energy poten-tial.
In the best scenario, an average of 52% of COD isconverted into
CH4, with 39% related to the portion ofCH4 available for use, which
characterizes only a 25%loss of energy potential. Thus, for a given
concentrationof influent COD and removal efficiency in the
reactor,the loss of dissolved methane in the effluent becomesan
extremely important factor in the energy balance ofthe system.
-
2748 L. C. S. Lobato et al. \ Estimates of methane loss and
energy in anaerobic reactors Water Science & Technology | 66.12
| 2012
Table 2 I Equations for calculating the portions of the mass
balance of COD and energy recovery potential
Portions Equations Notes
Estimate of meaninfluent fiowrate
an = Pop X QPC
Estimate of daily COD CODremoved = Pop x QPCCODmass removed
fromthe system
= CODremoved xEstimate of daily CODmass used by the i'coD
=biomass
Estimate of sulfate load COSO4 convened = ^mean X Cso, X
so,converted intosulfide
Estimate of daily COD CODso, = COso, convertemass used in
sulfate S^" + 2O2 => SO4"reduction (32g) + (64g) ^ (96g)
Estimate of daily COD CODCH, = gmass converted into _ CODm, x K
x (275-F T)
QCH, - px/methane production
Estimate of available PEavaUabIe-CH4 = QN-actual-CH, X CH4energy
potential
mcan = mean influent flowrate (m^ d '); Pop =population
(inhab.); QPC = per-cupte wastewatercontribution (m'' inhab "^
d"')
oved = daily COD mass removed from thesystem (kg COD d^'); Pop =
population (inhab.);QPCCOD = per-capita COD contribution (kgCOD
inhab"' d"'); BCOD = efficiency of CODremoval (%)
CODsiudge = daily COD mass converted intobiomass (kg CODsiudge
d"'); ycoD = sludge yield,as COD (kg COD.iudge kg COD I^noved); Y=
sludgeyield, as TVS (kg TVS kg COD^ e^ nioved);TVS-COD = conversion
factor (1 kg TVS =1.42 kg CODsiudge); TVS = total volatile
solids
COSO4 ^j = load of SO4 converted into sulfide(kg SO4 d~'); Cso,
= average influent SO4concentration (kg SO4 m"''); 30, = efficiency
ofsulfate reduction (%)
CODso^ = COD used by the BRS for sulfatereduction (kg CODgo,
d"^); KCOD-SO^ = CODconsumed in sulfate reduction (0.667 kg
CODSO4
CODcH4 = daily COD mass converted into methane(kg CODcH4 d"^);
QCH4 = theoretical volumetricproduction of methane (m'' d"'); i? =
gas constant(0.08206 atm L m o r ' K"'); T= operationaltemperature
of the reactor ( C); P = atmosphericpressure (1 atm); KCOD = COD of
one mole ofCH4 (0.064 kg CODcH4 mor ' )
Qw-cH4 = methane loss as waste gas (m^ d"'); p =percentage of
methane in the gaseous phase lostas waste gas (%); QO-CH4 = other
methane lossesin the gaseous phase (m^ d"'); po = percentage
ofmethane in the gaseous phase considered as otherlosses (%);
QL-CH4 = loss of dissolved methane inthe liquid effluent (m^ d"');
pt = concentration ofdissolved methane in the liquid effluent (kg
m"');/cH4 = conversion factor of methane mass intoCOD mass (4 kg
COD kg CHj')
Qactuai-cH4 = actual producton of methaneavailable for energy
recovery (m'' d"')
PEactuai-CH4 = available energy potential (MJ d~');QN-actuai-cH4
= normalized methane production(Nm' d"'); CH4 = calorific energy of
methane(35.9 MJ Nm-')
The mass balance carried out by Souza (2010) throughmeasurements
in reactors, in both pilot- and demon-stration-scale, indicated the
following ranges in relation tothe global COD: soluble in the
effluent (14-24%), sludgein the effluent (10-20%), sludge retained
in the reactor
(8-10%), methane in the biogas (24-30%), dissolvedmethane
(16-18%), and sulfate reduction (4.5-5%). It isobserved that these
values are close to the ones obtainedfrom the simulations using the
mathematical modeldeveloped.
-
2749 L. C. S. Lobato et al. \ Estimates of methane ioss and
energy in anaerobic reactors Water Science & Technology | 66.12
| 2012
Table 3 Main characteristics of the UASB reactors used to
validate the model
Characteristic
LocationPopulation (inhab.)Mean influent owrate (L s"^)Number of
unitsDimensions (m)Useful depth (m)Useful volume (m'^ )
Pilot-scale
CePTS*80.021D = 0.34.00.36
Demo-scale
CePTS*1400.321D = 2.04.514.0
Full-scale
Ona - Belo Horizonte - Brazil330,0006602438.4x6.44.52,211.9
Full-scale
Laboreaux - Itabira - Brazil30,00070821.7x6.24.51,210.9
(*) Centre for Research and Training on Sanitation - Belo
Horizonte - Brazil.
(a)
COD usedin Sulfatereduction
7%
(b)COD
conveftedlntosludge ' ' ^
COO notconerted intoCH4, and lost COD used in ^^
witiithe sulfate \ " ieffluent reductionm 5%
OttierC0DCH4
losses3%
(c)COD
converted intoconverted into COD used in f^ngCH4, and lost
sulfate 1504
with the reductioneffluent 3%
35%
COD notconverted intoCH4, and lost
30%
Figure 2 I Result of the simulations of COD mass balance in UASB
reactors treating domestic waste, in relation to the infiuent COD
for the three scenarios: (a) worst; (b) typical; (c) best.
Figure 3 shows the results from the simulations and
therespective adjusted lines for the production of biogas and
forthe energy recovery potential in UASB reactors treatingdomestic
wastewater. The simulations considered the vari-ation in the
contributing population from 0 to 1 millioninhabitants
(corresponding to influent flowrates from 0 to2.5 m'' s"^). The
input data for the model are shown in
Table 1, for the three scenarios considered (worst, typicaland
best), while the determination coefficients of theadjusted lines
are presented in Table 4.
A wide range of biogas production and potential forenergy
generation may be obtained, depending on the inputvariables.
Considering, for example, a wastewater flowrateof 2,000 L s \ the
expected biogas production and potential
(b)
Typicali Best
0 500 1,000 1,500 2,000
Wastewater flowrate (L.s-' )2,500
ergy
c
'SI r
gene
n(M
J.la
len
l
s.
400,000 -
300,000
200,000 -
i 00,000
0"Wonl Typical
500 1,000 1,500 2,000 2,500
Wastewater flow rate (L.s' ' )
Figure 3 I Expected ranges of biogas production (a) and
potentiai generation of energy (b) in UASB reactors treating
domestic wastewater.
-
2750 L. C. S. Lobato et al. \ Estimates of methane loss and
energy in anaerobic reactors Water Science & Technology | 66.12
{ 2012
Tabie 4 I Regression equations and determination coefficients of
the adjusted data
Scenario
Best
Typical
Worst
Linear regression equation"
y =
y =
8.95X
7.52*
5.16
MJd H s ^
y = 235.92X
y = 185.46
y = 118.86
Determinationcoefficient (/?')
0.83
0.75
0.64
"Regression equations were obtained for a set of simulated data
during a certain run of themodel.
of energy recovery may vary from 10,000 to 18,000 m^ d ^and from
240,000 to 480,000 MJ d'\ respectively.
Based on the simulations performed, tbe following uni-tary
relationships were also obtained for methane, biogasand energy
production in UASB reactors treating typicallydomestic wastewater
(Table 5).
The UASB reactors show an estimated volumetric biogasproduction
of 14 L.inbab"^ d"^ (mean for tbe typical scenario);tbis production
is lower tban tbat found in tbe sludge digesters.In the best
scenario, tbe mean value of tbe volumetric biogasproduction was
17L.inhab"^ d^^ The potential for energyrecovery in tbe UASB
reactors varied from 1.5 to 2.9 MJ perm^ of treated wastewater,
depending on the characteristics oftbe influent wastewater and tbe
efficiency of tbe system.
Finally, it is important to mention tbat tbe range ofmean
metbane yield predicted by tbe model (0.113-0.196 Nm^ CH4 kg
CODremoved - Table 5) is in close agree-ment witb tbe expected
range reported by Noyola et al(1988), of 0.08-0.18 Nm^ kg
CODremoved-'.
reactors used to validate tbe model (Table 3), were plottedon
tbe same grapb sbowing tbe tendency lines for the resultsobtained
in tbe simulations (Figures 4-7).
The mean values obtained for tbe pilot-scale UASB reac-tor were
0.12 m^ d"' for biogas production (Figure 4(a)), and2.7 MJd"' for
the energy recovery potential (Figure 4(b)),considering the mean
influent flowrate of 0.02 L s"^ For thedemonstration-scale UASB
reactor, considering the meaninfluent flowrate of 0.32 L s "\ 2.1
m^ d"' and 55.8 MJ d"'were obtained for biogas production (Figure 5
(a)) andenergy recovery potential (Figure 5(b)), respectively.
Biogas production in tbe UASB reactors of LaboreauxWWTP was in
tbe order of 390 m^ d"^ (Figure 6(a)), result-ing in an energy
recovery potential of around 11,000 MJ d"^(Figure 6(b)). For tbe
UASB reactors of Ona WWTP, tbemean observed values were 3,900 m^d~^
for biogas pro-duction (Figure 7(a)) and 105,000MJd"' for
energyrecovery potential (Figure 7(b)).
The results for biogas production and tbe resultantpotential for
energy recovery in both the pilot- and demo-scale UASB reactors
(Figures 4 and 5), obtained using exper-imental data, were observed
to be witbin tbe simulatedranges (between worst and best
scenarios), with no dataobserved below tbe worst scenario line. For
tbe full-scalereactors (Figures 6 and 7), most of the results for
biogas pro-duction and energy recovery potential were verified to
beconcentrated between tbe simulated ranges. However,some data from
botb full-scale plants were situated belowtbe worst scenario
line.
Validation of the mathematical model
Tbe predicted results of biogas production and the
corre-sponding potential for energy recovery, for tbe four UASB
Model adjustment to the measured data
To better evaluate tbe model adjustment to the measureddata of
biogas production, tbese data were grouped in
Table 5 I Unitary relationships for the production of methane,
biogas and energy production In UASB reactors treating domestic
wastewater
unitary relationsiiip
Unitary metbaneyield
Unitary biogas yield
Unitary energypotential
unit
NLCH4 inbab"' day"'NLCH4 m"^wastewater
NLCH4kgCOD-eUved
NLbiogas inbab"' day"'NLbiogas m"^wastewater
NLbiogas kg CODremoved
MJ m^'^wastewater
MJ kg CODr'movedMJ Nm"^biogas
MJ inbab"' year"'
Worst scenario
iVIaximun'
9.9
81.7
154.1
14.1
116.7220.1
2.95.5
25.1
129.5
1 Minimum
3.616.766.0
5.2
23.8
94.3
0.6
2.4
25.1
47.7
Mean
6.8
42.2
113.4
9.860.3
162.0
1.5
4.1
25.1
89.7
Typicai scenario
iVlaximum
13.3
103.7
185.8
17.7
138.3247.8
3.7
6.7
26.9
173.8
1 Minimum
7.4
34.8124.2
9.946.4
165.6
1.2
4.5
26.9
96.8
iVIean
10.2
64.2
158.3
13.685.6
211.1
2.3
5.7
26.9
133.8
Best scenario
iVlaximum
16.7
134.6
219.1
20.8
168.3
273.9
4.8
7.9
28.7
218.4
1 Minimum
11.1
51.8
173.9
13.9
64.8
217.4
1.96.2
28.7
145.7
Mean
13.7
81.3
196.0
17.1
101.6
245.0
2.9
7.0
28.7
179.3
-
2751 L, C. S. Lobato et ai. \ Estimates of methane loss and
energy in anaerobic reactors Water Science & Teciinoiogy |
66.12 | 2012
1 0.02 0.03* measured data (pilot sea le)
Wastewater flowrate (L.s"')
I 0.02 0.03measured da ta (pilolseale)
Wastewater fiowrate (L.s-')
Figure 4 I Validation of the modei using the monitoring data
from the pilot-scale UASB reactor: (a) biogas production; (b)
potential ot energy recovery.
.1 0.2 0.3
* measured data (demonstration scale)
Wastewater flowrate fL.s-')
0.1 0.2 0.3 measured data (demonstration seak)Wastewater
flowrate {L.s'' )
Figure 5 I Validation of the model using the monitoring data
from the demonstration-scale UASB reactor: (a) biogas production;
(b) potential of energy recovery.
20 40 60 80 measured data (Laboreaux WWTP)Wastewater flowrate
(L.S"')
20 40 60 80 meastjreddata (Labotcaux WWTP)Wastewater flowrate
(L.s-')
Figure 6 I Validation of the model using the monitoring data
from the Laboreaux WWTP: (a) biogas production; (b) potentiai of
energy recovery.
two sets of results, as follows: (i) measured results of
biogasproduction in the pilot- and demo-scale reactors (Figure
8(a));and (ii) measured results of biogas production in the
full-
scale plants (Figure 8(b)). It can be seem from Figure 8(a)that
the linear adjustment of the biogas production datafrom the pilot-
and demo-scale UASB reactors was very
-
2752 L. C. S. Lobato et al. | Estimates of methane loss and
energy in anaerobic reactors Water Science & Technology | 66.12
| 2012
(a) (b)7.000
6.000 5,000g3 4,0001 3.000 a.
a 2.000
1.000
0
CQ
.
/y
- /y
A
fi
200 400 600 800 measured dala (Ona WWTP)
Wastewater flowrate (L.s-')
1,000 )0 400 600 800 measured data (Ona WWTP)
Wastewater flowrate (L.S"')
1,000
Figure 7 I validation of ttie model using the monitoring data
from the Ona WWTP; (a) biogas production; (b) potential of energy
recovery.
y=5.70xR"=0.81
measured data (pilot and dei|- - linear adjustment of measured
data
Wastewater flowrate (L . s ' )
250 500 750 measured data (WWTP Laboreaun and Ona)
linear adjustment of the measureddataWaslewater flowrate
(L.s"')
Figure 8 I Model adjustment to the measured biogas production
data; (a) pilot- and demonstration-scale UASB reactors; (b)
full-scale UASB reactors.
close to tbe tendency line predicted by the model in
typicalscenario, confirming great adberence to the model.
Tbedetermination coefficient {R^) for this set of data was0.92. On
the other hand, the linear adjustment of thedata from the two
full-scale plants stayed close to the ten-dency line of the worst
scenario (Figure 8(b)). The R^coefficient for this set of data was
0.81.
The greater adherence of the results from the pilot-and
demonstration-scale reactors to the model may beexplained by the
fewer variations in the concentration ofthe infiuent wastewater,
and also by the greater precisionof the biogas meters used in the
CePTS. In the case of thetwo full-scale plants, the excessive
dilution of the influentwastewater is a recurring problem, owing to
the contri-butions of rainwater and fiood on the river banks,which
result in the reduction of biogas production.Moreover, the lack of
calibration of biogas metersduring some periods may have prompted
erroneousmeasurements.
It is worth mentioning that the model was developedand validated
for the COD balance in UASB reactors treat-ing domestic (low
concentration) wastewater. Its use forother situations should
therefore considrer the review ofthe input data presented in Table
1.
CONCLUSION
The mathematical model developed enabled better represen-tation
of the mass balance of COD and of the potential forenergy recovery
in UASB reactors treating domestic waste-water. The results of the
simulations performed showedthat the model enables a more realistic
estimate of theamount of biogas that can be recovered from the
interiorof the three-phase separators, which effectively
representthe portion available for energy recovery.
The incorporation into the model of the losses ofmethane
dissolved in the effiuent and in the gaseous phase.
-
2753 L. C. S. Lobato et al. \ Estimates of methane loss and
energy in anaerobic reactors Water Science & Technology | 66.12
| 2012
as well as the portion of COD used for sulfate reduction, maybe
considered an advance as the available models usuallyoverestimate
biogas production and the potential for energyrecovery in anaerobic
reactors used in the treatment of dom-estic wastewater. The results
of the simulations indicate thatsignificant portions of the
influent COD may not be recov-ered as methane in the biogas,
depending primarily on theloss of methane dissolved in the effluent
and the concen-tration of sulfate in the influent. In worst
scenario (seeTable 2), only 19% of the influent COD was recovered
asmethane in the biogas. In the best scenario, the percentageof
methane recovered in the biogas reached 39% of the influ-ent COD.
Of all the COD converted into methane, theportion recovered in the
biogas varied from 49 to 75%,depending on the losses mentioned
above.
When all COD fractions are considered in the mass bal-ance, as
well as the possible losses in the liquid and gasphases, the values
obtained for the theoretical amount ofmethane available for energy
recovery are much closer tothe actual values measured in the field.
This can be con-firmed by the validation of the mathematical model
usingthe results for biogas production and percentage of
CH4obtained in pilot, demo- and full-scale UASB reactors.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the support obtained fromthe
following institutions: Companhia de Saneamento deMinas Gerais -
COPASA; Conselho Nacional de Desenvol-vimento Cientfico e
Tecnolgico - CNPq; Fundaao deAmparo Pesquisa do Estado de Minas
Gerais - FAPEMIG;Sistema Autnomo de Agua e Esgoto de Itabira -
SAAEItabira.
REFERENCES
Agrawal, L. K., Harada, H. & Okui, H. 1997 Treatment of
dilutewastewater in a UASB reactor at a moderate
temperature:Performance aspects. Journal of Fermentation
andBioengineering 83 (2), 179-184.
Gloria, R. M., Souza, C. L., Silva, S. Q., Aquino, S. F.
&Chernicharo, C. A. L. 2008 Remoo de sulfeto de hidrogniode
efluente de reator UASB: Contribuio da camada deescuma e da tubulao
de descarte. In: IX Taller y SimposioDigestin Anaerobia, Anais,
Ilha de Pscoa.
Hartley, K. & Lant, P. 2006 Eliminating non-renewable
CO2emissions from sewage treatment: An anaerobic migratingbed
reactor pilot plant study. Biotechnology andBioengineering 95 (3),
384-398.
Metcalf & Eddy Inc. 2003 Wastewater Engineering: Treatment
andReuse. 4th edition, McGraw Hill, New York, 1818 pp.
Noyola, A., Capdeville, B. & Roques, H. 1988 Anaerobic
treatmentof domestic sewage with a rotating-stationary fixed
filmreactor. Water Research 12, 1585-1592.
Noyola, A., Morgan-Sagastume, J. M. & Lpez-Hernndez, J.
E.2006 Treatment of biogas produced in anaerobic reactors
fordomestic wastewater: Odor control and energy/resourcerecovery.
Reviews in Environmental Sciences andBiotechnology 5, 93-114.
Pierotti, S. M. 2007 Avaliao da partida do reator anaerobio
demanta de Iodo (UASB), em escala real, sob condieshidrulicas
desfavorveis (Evaluation of the Startup of a Full-scale Upflow
Anaerobic Sludge Blanket (UASB) Reactorunder Adverse Hydraulic
Conditions). Dissertao deMestrado, Mestrado em hidrulica e
saneamento. Escola deEngenharia da Universidade de Sao Paulo, 156
pp. (inPortuguese).
Singh, K. S. & Viraraghavan, T. 1998 Start-up and
operationof UASB reactors at 20 C for municipal
wastewatertreatment. Journal of Fermentation and Bioengineering85
(6), 609-614.
Souza, C. L. 2010 Estudo das rotas de formao, transporte
econsumo dos gases metano e sulfeto de hidrognioresultantes do
tratamento de esgoto domstico em reatoresUASB (Pathways of
formation, transportation andconsumption of methane and hydrogen
sulfide resulted fromdomestic wastewater treatment in UASB
reactors).Tese de Doutorado, Programa de Pos-Graduaao emsaneamento,
meio ambiente e recursos hfdricos. Escolade Engenharia,
Universidade Federal de Minas Grais(in Portuguese).
Souza, C. L. & Chernicharo, C. A. L. 2on Methane losses in
UASBreactors treating domestic wastewater: Mass balance andpossible
improvements. In: 10th Specialized Conference onSmall Water and
Wastewater Treatment Systems. IWASpecialist Conference. Venice,
Italy.
von Sperling, M. & Chernicharo, C. A. L. 2005
BiologicalWastewater Treatment in Warm Climate Regions.
IWAPublishing, London, 1452 pp.
First received 28 March 2012; accepted in revised form 19 July
2012
-
Copyright of Water Science & Technology is the property of
IWA Publishing and its content may not be copiedor emailed to
multiple sites or posted to a listserv without the copyright
holder's express written permission.However, users may print,
download, or email articles for individual use.