anaerobic treatment of winery wastewater in fixed bed reactors
TRANSCRIPT
ORIGINAL PAPER
Anaerobic treatment of winery wastewater in fixed bed reactors
Rangaraj Ganesh • Rajagopal Rajinikanth •
Joseph V. Thanikal • Ramamoorty Alwar Ramanujam •
Michel Torrijos
Received: 28 May 2009 / Accepted: 8 October 2009 / Published online: 30 October 2009
� Springer-Verlag 2009
Abstract The treatment of winery wastewater in three
upflow anaerobic fixed-bed reactors (S9, S30 and S40)
with low density floating supports of varying size and
specific surface area was investigated. A maximum OLR
of 42 g/l day with 80 ± 0.5% removal efficiency was
attained in S9, which had supports with the highest specific
surface area. It was found that the efficiency of the reactors
increased with decrease in size and increase in specific
surface area of the support media. Total biomass accu-
mulation in the reactors was also found to vary as a
function of specific surface area and size of the support
medium. The Stover–Kincannon kinetic model predicted
satisfactorily the performance of the reactors. The maxi-
mum removal rate constant (Umax) was 161.3, 99.0 and
77.5 g/l day and the saturation value constant (KB) was
162.0, 99.5 and 78.0 g/l day for S9, S30 and S40,
respectively. Due to their higher biomass retention poten-
tial, the supports used in this study offer great promise as
media in anaerobic fixed bed reactors. Anaerobic fixed-bed
reactors with these supports can be applied as high-rate
systems for the treatment of large volumes of wastewaters
typically containing readily biodegradable organics, such
as the winery wastewater.
Keywords Anaerobic fixed bed reactor � Floatingsupports � Specific surface area � Winery wastewater �
Biomass attachment � Kinetic model
Introduction
Anaerobic treatment of industrial wastewater has become a
viable technique thanks to the development of a range of
high-rate reactors based on such technology as anaerobic
filters, UASB, fluidized or expanded beds [1]. High-rate
anaerobic reactors offer the advantage of high-load systems
requiring much less volume and space. Such advantages
are of interest to those industries which produce large
amounts of highly concentrated wastewater, notably the
food, paper, and pulp industries [2]. High-performance
anaerobic treatment in fixed-bed reactors has been applied
very successfully to wastewater from agribusiness indus-
tries that use agricultural products containing typically high
concentrations of organic substrates that are readily
degraded by anaerobic bacteria [3, 4].
Winery wastewater is a classic example of such agri-
based waste. Wine production is one of the foremost agri-
industries in Mediterranean countries, and it has also
acquired importance in a large number of countries in other
parts of the world (e.g. Australia, Chile, United States,
R. Ganesh � R. Rajinikanth � M. Torrijos (&)
Laboratoire de Biotechnologie de l’Environnement, INRA,
UR50, Avenue des Etangs, 11100 Narbonne, France
e-mail: [email protected]
R. Ganesh
e-mail: [email protected]
R. Ganesh � R. A. Ramanujam
Department of Environmental Technology, Central Leather
Research Institute (CLRI), Chennai 600020, India
R. A. Ramanujam
e-mail: [email protected]
R. Rajinikanth
Indian Institute of Technology, Roorkee, Uttaranchal, India
e-mail: [email protected]
J. V. Thanikal
Department of Civil Engineering,
Kumaraguru College of Technology, Coimbatore 641006, India
e-mail: [email protected]
123
Bioprocess Biosyst Eng (2010) 33:619–628
DOI 10.1007/s00449-009-0387-9
South Africa and China), with increasing impact on the
economy of these countries. The high amount of effluent
produced, along with the seasonal nature of wine produc-
tion, raises specific problems for the treatment process.
These relate to the volume and composition of the waste-
water produced and consequently treatment plants must be
versatile in relation to the loading regime and at the same
time be able to cope with a succession of start-ups and
closedowns, including periods of inactivity [5, 6]. Thus, in
winery wastewater treatment, involving effluent with high
COD but low nitrogen and phosphorus content, anaerobic
digestion offers advantages over aerobic treatment. Win-
eries have seasonal activity producing wastewater mainly
during harvesting and at the time of winemaking. The
wastewater can be treated as it is produced or can be stored
for treatment over several months. Anaerobic digesters
have the advantage of re-starting rapidly after a shutdown
[7].
Several treatment alternatives for winery wastewater
have been proposed by many authors based on both aerobic
and anaerobic processes. Based on aerobic systems are
conventional activated sludge systems [8, 9], SBR systems
[10, 11], jet-loop reactors [6, 12], aerobic biofilm systems
such as RBC [13] and more recently advanced treatment
systems such as moving bed biofilm reactors [14], fixed
bed biofilm reactors (FBBR) [15] and membrane bioreactor
(MBR) systems [16]. Based on anaerobic systems are
anaerobic filters [17], UASB [18, 19], hybrid reactors [20]
and anaerobic SBR [21].
The use of conventional activated sludge systems is
sometimes problematic due to bad settling properties of the
sludge caused by the development of filamentous bacteria
or the formation of dispersed flocs, resulting in an effluent
with a high solids concentration and turbidity [16]. Aerobic
biofilm systems such as FBBR are an alternative to the
conventional activated sludge plants as these systems offer
advantages such as low reactor volume requirements;
reduction of bulking problems; the absence of return flow
and backwashing due to the high void ratio of the filling
media and an easier management with respect to the con-
ventional activated sludge plants [15].
Compared to aerobic processes, anaerobic process do
not require oxygen and therefore are less energy intensive
and sludge production is much less making the process
simpler and cheaper [16, 22].
Anaerobic upflow reactor is a packed-bed reactor where
biomass can be retained or attached to packings. Because
of the retained biomass associated with the packings the
process is stable with respect to high organic loadings,
shock loadings or changes in pH and temperature [23]. The
different types of support material studied in fixed-bed
reactors include Raschig rings, pall rings, string-shaped
plastic media saddles, plastic cylinders, clay blocks [17,
23–26]. The characteristics of the support material used
determine biomass retention and, hence, the efficiency of
the treatment system [25]. The major factors influencing
bacterial attachment include roughness, porosity, surface
area, surface charge, hydrophily, surface energy and
chemical composition. Along with these, many other fea-
tures also influence biomass retention capability: micro-
crystals, macro- and micro-pores, fibers, ridges and degree
of smoothness of the support material [27–30].
The aim of the work presented here was to study the
treatment of winery wastewater using anaerobic fixed-bed
reactors with small floating supports. The influence of the
size and specific surface area of the supports on biomass
retention and reactor performance was investigated.
Kinetic model application was carried out for the predic-
tion of the performance of the reactors.
Materials and methods
Reactor details
Three upflow anaerobic fixed-bed reactors of similar
dimensions (1,000 mm height 9 125 mm diameter) were
used for the study. The reactors were made of Plexiglas
with an effective volume of 10 l. The reactors were
equipped with hot water jackets to maintain a mesophilic
temperature of 35 �C. The feed wastewater was pumped
into the bottom of the reactor by means of a peristaltic
pump. A perforated plate was placed at the bottom to
obtain uniform flow of feed across the reactor. The reactor
liquid was recirculated from the top to the bottom by means
of a recirculation pump at a rate of 10 l/h, with a liquid
upflow velocity between 0.8 and 0.9 m/h throughout the
experiment. Provision was made in the reactor, at the top
and above the support medium level, for collection of
effluent, which passed through a U-tube for separation of
gas. Temperature and pH were measured online with the
help of probes inserted through the top of the reactor. Each
reactor was filled with randomly distributed supports to
80% of reactor volume. The schematic diagram of the
reactor is shown in Fig. 1.
Characteristics of support media
Three small polyethylene floating carriers (S9, S30 and
S40) were used as media for biomass immobilization and
retention. The main characteristics of the supports are
shown in Table 1. The supports were cylindrical in shape
and baffled with compartments. These supports are suitable
thermoplastic materials as biological carriers and are
inexpensive, nontoxic and non-reactive in most biological
applications.
620 Bioprocess Biosyst Eng (2010) 33:619–628
123
Feed characteristics and reactor inoculum
To simulate winery wastewater, diluted wine with a COD
concentration of around 20 g/l was used as feed for the
experiments. The feed was made alkaline with the addition
of NaOH (0.5–1.0 ml 10 N NaOH/l feed), and was sup-
plemented with nutrients corresponding to a COD/N/P ratio
of 400/7/1 [9]. Due to the low nutrient content present in
winery wastewater, the addition of nitrogen and phosphate
sources is recommended for cellular growth in the
biological treatment process [14, 16, 31] to maintain a
minimum COD:Tot N:P ratio of 400:7:1. The main
characteristics of the feed are given in Table 2.
Sludge from a large-scale anaerobic reactor treating
distillery vinasse was used as inoculum for the reactors.
One liter of concentrated sludge with a volatile solids (VS)
content of 40 g/l was used to inoculate each reactor, thus
providing an initial VSS concentration of 4 g/l in the
reactors.
Analytical methods
Effluent samples were analyzed for alkalinity, TSS and
VSS according to the procedures given in Standard
Methods [32]. Total and Soluble COD was measured by
colorimetric method using Hach 0–1,500 mg/l vials.
Volatile fatty acid (VFA) concentration was measured
using a gas chromatograph (GC-8000, Fisons instruments)
equipped with a flame ionization detector and an auto-
matic sampler (AS 800, Fisons instruments). The column
used was a semi-capillar Econocap FFAP (Alltech) col-
umn with 15 m length, 0.53 cm diameter and Phase ECTM
1000 film 1.2 lm. The temperature of the spitless injector
was 250 �C, the temperature of the detector was 275 �C.
The temperature increased from 80 to 120 �C in 3 min.
The carrier gas was nitrogen (25 kPa). The volume of
sample injected was 1 ll. The calibration was made from
a mixture of 6 acids (standard solution): acetic (C2),
propionic (C3), butyric (C4), isobutyric (iC4), valeric (C5)
and iso-valeric (iC5) acids at 1 g/l each. The calibration
range was 0.25–1 g/l by dilution of the standard solution.
The internal standard method (1 g of ethyl-2-butyric acid
in 1 l of water acidified with 50 ml of H3PO4) was used to
measure total VFA concentration by mixing 1/1 volume of
the internal standard solution and the sample or the stan-
dard solution. The margin for error of this measurement
was between 2 and 5% with a quantification threshold of
0.1 g/l.
Quantification of attached/entrapped solids
At the end of the experiment, the reactors were emptied to
quantify the amount of volatile solids inside the reactors.
Supports were removed by batches of 1 L (645, 23 and 11
numbers, respectively, for S9, S30 and S40) starting from
the top (near the outlet) to the bottom (above the sludge
bed). To estimate total solids, the supports were placed in
aluminum foil and oven dried for 24 h. To estimate the
volatile solid content, oven-dried solid samples were
scrapped from the supports and ignited at 550 �C for 2 h.
Fig. 1 Schematic diagram of the anaerobic fixed bed reactor
Table 1 Characteristics of supports
Media Dimension
(D) 9 H (mm)
Surface
area (m2/m3)
Density
(kg/m3)
S9 9 9 7 800 0.92
S30 30/35 9 29 320 0.94
S40 40/45 9 35 305 0.95
Table 2 Characteristics of feed winery wastewater
Parameter Value
pH 8–11
Total suspended solids (mg/l) 150–200
Volatile suspended solids (mg/l) 100–130
Total COD (mg/l) 18,000–21,000
Soluble COD (mg/l) 17,200–20,000
TOC (mg/l) 4,400–5,140
Bioprocess Biosyst Eng (2010) 33:619–628 621
123
Experimental design
The experimental protocol was designed to examine the
effect of increasing organic loading rates (OLR) on COD
removal efficiency in the three upflow anaerobic fixed-bed
reactors filled with the different polyethylene floating car-
riers, the aim was to attain the maximum loading rate while
maintaining 80% removal efficiency. The influent COD was
maintained constant and OLR was gradually increased over
time by decreasing the hydraulic retention time (HRT).
Results and discussion
Operation of the reactors
The same operational strategy was employed for all the
three reactors and Fig. 2 shows the example of the reactor
filled with S9.
Initially, the reactor with S9 was operated at a low OLR
of around 0.4 g/l day. During the first 2 weeks of operation
at this loading rate, the effluent COD remained quite high,
then decreased gradually from 2.7 to 0.3–0.4 g/l, corre-
sponding to a removal efficiency of 98%. This period of
2 weeks, considered as the acclimation phase, was,
however, quite rapid given the type of wastewater employed
(ethanol, which is readily biodegradable, is the primary
constituent of winery wastewater). The OLR was gradually
increased by 10–20% once a week, provided that COD
removal efficiency remained above 80%. An 80% removal
efficiency was considered as the threshold level, both for
safe operation and from an economical point of view. In
cases where the COD remained high after an OLR increase,
the OLR was temporarily decreased by 10–20% to ensure
lower COD at outlet with corresponding removal efficiency
higher than 80%. The main results for the three reactors with
the different supports are summarized in Table 3.
Biodegradability of winery effluent
Figures 2 and 3 show, respectively, the variations with
OLR for non VFA-COD and VFA-COD in the effluent at
the reactor outlet. At low OLR (below 6 g/l day), the non
VFA-COD at outlet (Fig. 3), which was at fairly constant
and very low values, can be considered as the non-biode-
gradable COD fraction of winery wastewater. The average
non-biodegradable COD in the treated effluent was
320 mg/L. The soluble COD removal rate of 98.4% shows
that winery wastewater is highly biodegradable, with
refractory soluble COD representing less than 1.6% of
Fig. 2 Typical operating
strategy of the reactors, example
of reactor with S9
Table 3 Comparison of reactor performance with the different supports used
Parameters S9 S30 S40
Maximum OLR attained, g COD/l day (for 80% COD removal efficiency) 42 27 22
Total reactor volatile solids, g 354 291 225
Attached/entrapped volatile solids, g 302 225 162
Suspended volatile solids, g 52 66 63
Specific biomass activity, g COD/g VSS day 1.19 0.93 0.98
Biomass yield, g VSS/g CODdes 0.056 0.053 0.060
Maximum removal rate constant (Umax), g COD/l day 161.3 99.0 77.5
Saturation value constant (KB), g COD/l day 162.0 99.5 78.0
622 Bioprocess Biosyst Eng (2010) 33:619–628
123
initial soluble COD. The residual COD concentrations
measured were close to those reported earlier [10, 21, 33].
Between 6 and 15 g/l day, non-VFA-COD increased
very little (Fig. 3) and any increase in COD at outlet was
mainly linked to VFA accumulation (Fig. 4). For OLR
above 15 g/l day, VFA was always in the range of 2–3 g
COD/l and increase in COD at outlet were mainly linked to
a gradual increase in the non VFA-COD, indicating the
presence of other anaerobic intermediates or of non-acidi-
fied organic matter. This shows that above 15 g/l day, the
acidification of the organic matter started to deteriorate and
this phenomenon was more pronounced when the size of
the support was bigger.
Acetic acid was the major VFA component in all the
reactors but there was a slight build-up in propionic,
butyric and valeric acid concentrations as the OLR
increased. The acetic acid concentration in S9, S30 and
S40, corresponding to the maximum OLR of 42, 27 and
22 g/l day, were 1.67, 1.42 and 1.04 g/l, respectively.
Propionic, butyric and valeric acid concentrations remained
less than 0.5 g/l at these OLRs.
Reactor performance
Figure 5, which presents the changes in COD removal
efficiency related to the OLR, gives an overview of the
behavior of the three reactors.
The maximum OLRs attained while maintaining above
80% removal efficiency were 42, 27 and 22 g/l day for S9,
S30 and S40, respectively.
Furthermore, Fig. 5 indicates that, for a given loading
rate, removal efficiency was always higher and, thus, sol-
uble COD at outlet always lower for reactor S9 compared
to S30 and, similarly, for reactors S9 and S30 compared to
S40. These results show that quite a high OLR could be
applied to the reactors filled with the floating supports and
that the reactor with S9 performed better than that with S30
which was, in turn, better than the reactor with S40. This
Fig. 3 Variation of Non VFA-
COD at outlet with OLR
Fig. 4 Variation of VFA-COD
at outlet with OLR
Bioprocess Biosyst Eng (2010) 33:619–628 623
123
indicates that the performance of reactor was enhanced by
a decrease in size and an increase in the specific surface
area of the supports used.
Table 4 shows a general comparison of the perfor-
mances of the anaerobic fixed-bed reactors used in this
study with various other reactor configurations used for the
treatment of winery wastewater. These configurations
studied were anaerobic filters [17], UASB [18, 19], a
hybrid system consisting of a UASB and an anaerobic filter
[20] and an anaerobic SBR [21]. It can be seen that
anaerobic fixed-bed reactors offer enhanced performance
in the treatment of winery wastewater in terms of high
OLR and COD removal efficiency. String-shaped plastic
media were used by Yu et al. [17] for biomass immobili-
zation in an anaerobic filter reactor and the results reported
were similar to those in the present study, with a maximum
loading rate of 37.7 kg/m3 day with 82% COD removal
efficiency. The performances of the other reactor configu-
rations were much lower. Andreottola et al. [20] used a
PVC-type filling material (Flocor R) with a specific surface
area of 230 m2/m3 in the hybrid system (a single reactor
with a combination of UASB and anaerobic filter). The
average COD loading rate attained was 10 kg/m3 day, with
93% COD removal. The comparison of the results of this
study with the data from the literature clearly shows that
the supports have high capability of biomass retention
which facilitated the operation of the reactors at high
OLRs. Higher OLR implies that for a constant volume of
effluent to be treated, the reactor volume required is sig-
nificantly reduced thus contributing to substantial savings
in land area and economics.
Comparison with aerobic treatment systems
A general comparison of the performance of different aer-
obic systems for the treatment of winery wastewater is
shown in Table 5. For the conventional activated sludge
systems [8, 9] and SBR systems [10] operated with an OLR
less than 1 kg COD/m3 day, COD removal efficiency of
above 90% was achieved. Recently several advanced aer-
obic treatment systems were developed such as jet-loop
reactors [12], moving bed biofilm reactors [14], FBBR [15]
and MBR systems [16]. These systems are also capable of
achieving COD removal efficiency above 90%, but at
Fig. 5 Soluble COD removal
efficiency with respect to OLR
Table 4 Different anaerobic reactor configurations studied for winery wastewater
References Reactor type Media OLR (kg/m3 day) % COD removal
This study Anaerobic fixed bed S9 42 80 ± 0.5
This study Anaerobic fixed bed S30 27 80 ± 0.2
This study Anaerobic fixed bed S40 22 80 ± 0.3
Yu et al. [17] Anaerobic filter String shaped plastic media 37.7 82
Muller [18] UASB Activated microbial pellets 10 84.4
Keyser et al. [19] UASB Granular sludge 5.1 86
Granular sludge ? Enterobacter 6.3 90
Andreottola et al. [20] Anaerobic hybrid (UASB ? filter) Flocor R 10 93
Ruiz et al. [21] Anaerobic SBR Anaerobic sludge 8.6 98
624 Bioprocess Biosyst Eng (2010) 33:619–628
123
relatively higher loading rates. Anaerobic fixed bed reactor
with the floating supports used in the present study allows
operation at high OLRs (22–42 kg COD/m3 day) with a
removal efficiency of 80%, thus contributing to lower
reactor volume and land area requirements. Anaerobic fixed
bed reactors are advantageous than conventional aerobic
systems as they are less energy intensive and produces less
excess sludge. However, an aerobic post-treatment is
required to make the effluent fit for final disposal.
Suspended solids at outlet
Effluent total suspended solids (TSS) fluctuated with
changes in the OLR and ranged between 600 and
1,700 mg/l, though no significant difference between the
reactors was recorded. The average effluent TSS for the
entire study period was 839 ± 425 mg/l for S9, 859 ± 337
for S30 and 958 ± 438 for S40, respectively.
Quantification of suspended and attached/entrapped
solids
At the end of the experiment, the supports were removed
from the reactors in batches of 1 L and the volatile solid
concentration was measured in the sludge bed and in each
of the 1 L support fraction.
The supports used were able to retain a considerable
quantity of solids. The average attached/entrapped volatile
solids per support were 0.052 ± 0.024, 1.09 ± 0.28 and
1.65 ± 0.33 g VS/support for S9, S30 and S40,
respectively. Table 3 shows that the total quantity of
attached/entrapped solids in the reactors was, in decreasing
order, S9[S30[ S40 (301, 225 and 162 g, respectively)
which indicates that biomass attachment increased with a
decrease in the size of the supports coupled with an increase
in the specific surface area. The quantity of volatile solids in
suspension in the liquid phase (Table 3) was fairly similar
in the three reactors (52–66 g of VS) and represented,
respectively, for S9, S30 and S40, 15, 23 and 28% of total
VS in the reactors. Finally, thanks to the supports, the total
quantity of volatile solids retained in each reactor increased
significantly in comparison to conventional CSTRs: values
for the 10 l reactors S9, S30 and S40 were, respectively,
355, 290 and 225 g of volatile solids.
The specific mass loading rates or specific biomass
activity, that is to say the organic load in terms of g COD/day
divided by the quantity of volatile solids in the reactor, were,
for S9, S30 and S40, respectively, 1.19, 0.93 and 0.98 g
COD/g VSS day. This specific activity is close to 0.90 g
COD/g VSS day reported byMosquera-Corral et al. [34] for
the anaerobic flocculent sludge used in their study. These
values are quite high as a specific load between 1 and 1.5 g
COD/g VSS day is generally considered to be the upper
limit for the stable operation of an anaerobic reactor [2].
The VS yields for the three reactors were quite close
with 0.056, 0.053 and 0.060 g VS/g COD destroyed, for
S9, S30 and S40, respectively. The VS yield value obtained
is in good agreement with those of Borja et al. [35] and
Tatara et al. [36] where values reported by the authors were
0.06 and 0.0585 g VS/g COD removed, respectively.
Table 5 Comparison with aerobic reactor configurations studied for the treatment of winery wastewater
References Reactor type OLR,
kg COD/m3 day
% COD
removal
Salient features
Fumi et al. [8] Long term-activated
sludge
0.14–0.4
kg BOD/m3 day
98 Efficient than conventional activated sludge process
Very high retention time required
Brucculeri et al. [9] Activated sludge 0.8 90 High solids retention time (48 days) resulted
in low excess sludge production
Torrijos and Moletta. [10] Aerobic SBR 0.7 93 Highly flexible to seasonal loads
Suitable for small wineries
Petruccioli et al. [12] Jet-loop activated sludge 5.9 90 Sludge settling problems
Petruccioli et al. [12] High rate aerobic 8.8 92.2 Reduction in reactor space requirement
Andreottola et al. [14] Aerobic sequencing
batch biofilm
8.8 86–99 High organic load removal
Online control of biodegradation process possible
No sludge recycle required
Andreottola et al. [15] Aerobic fixed bed biofilm
(two-stage)
1.57 91 Simple management, good settling sludge without
bulking problems, no backwashing required.
Rapid start-up of reactor requires previously
colonized filling media.
Artiga et al. [16] Membrane bioreactor
(MBR)
2.2 97 Flexible and stable to high seasonal loads
Problems with membrane modules lifetime
and maintenance
Bioprocess Biosyst Eng (2010) 33:619–628 625
123
The higher OLR of 42 g/l day attained in S9 can thus be
attributed to the greater volatile solids present in the reactor
as well as to its higher biomass activity. In all three reactors,
attached volatile solids were distributed almost identically
throughout the reactor height. Indeed, the average volatile
solids were 43.2 ± 2.3, 26.4 ± 4.9, and 21.8 ± 1.2 g VS/l,
indicating that there was no solids gradient.
Photographs of the biocovered supports at the end of the
experiments are shown in Fig. 6. Two to three biocovered
supports were subjected to a normal tap water pressure jet
to evaluate the level of fixation of solids on the supports. It
was found that most of the solids were easily removed and
very little remained attached to the surface of the supports.
Thus, it was found that solids retention in all the three types
of support was based on entrapment rather than on actual
biofilm formation. Similar observations were made by
Henze and Harremoes [24], Weiland [3] and Alkalay et al.
[29] using other supports. They concluded that biofilm
thickness is of limited significance in fixed-bed reactors
since it makes up only a small part of the total percentage
of biomass whereas the biomass suspended/entrapped in
the gaps is of major importance: biomass builds up as
particles suspended in the spaces, forming the largest part
and thus contributing considerably to overall activity.
Kinetic model application
The most widely-used kinetic models for anaerobic filters
include the Monod model [37] and the Stover–Kincannon
model [38]. The major difference between these two
models is the use of the total loading rate concept in the
Stover–Kincannon model [39]. In the present study, the
Stover–Kincannon model was applied to the reactors for
the evaluation of kinetic constants and for predicting the
performance of the reactors.
In the modified Stover–Kincannon model, the substrate
utilization rate is expressed as a function of the OLR by
monomolecular kinetics, as follows:
dS=dtð Þ�1¼ V=Q Si � Seð Þ ¼ KBV=UmaxQ Sið Þ þ 1=Umaxð Þ
ð1Þ
where dS/dt, substrate removal rate (g/l day); Umax, maxi-
mum utilization rate constant (g/l day); KB, saturation
value constant (g/l day); V, volume of the reactor; Q, flow
rate (l/d); Si, influent substrate concentration (g/l); and Se,
effluent substrate concentration(g/l).
The plot between the inverse of the OLR V/(Q Si) and
the inverse of the organic removal rate V/[Q(Si - Se)]
yields a straight line, with 1/Umax as the intercept and
Fig. 6 Photographs of the biocovered supports (S9, S30, S40)
Fig. 7 Stover–Kincannon
model application to the
reactors
626 Bioprocess Biosyst Eng (2010) 33:619–628
123
KB/Umax as the slope (Fig. 7). For S9, S30 and S40,
respectively, the maximum removal rate constant (Umax)
was 161.3, 99.0 and 77.5 g/l day and the saturation value
constant (KB) 162.0, 99.5 and 78.0 g/l day (Table 3).
The modified Stover–Kincannon model was applied to
fixed-bed reactors treating different substrates. For exam-
ple, the values of Umax and KB, respectively, were: for
soybean wastewater, 83.3 and 85.5 g/l day [39]; for syn-
thetic wastewater, 83.3 and 186.23 g/l day [40]; and for
textile wastewater 31.69 and 45.37 g/l day [41].
The Eq. 2 shown below, which is the reorganized form
of Eq. 1 together with the values of Umax and KB obtained
from Fig. 7, is used to predict the effluent substrate
concentration.
Se ¼ Si � UmaxSi½ = KB þ Q Si=Vð Þð � ð2Þ
Fig. 8 shows the experimental Se values and the predicted
values using Eq. 2 for the three reactors. The high corre-
lation coefficient of 0.99 obtained indicates that the mod-
ified Stover–Kincannon model can be satisfactorily used
for the design of anaerobic fixed-reactors treating winery
wastewaters.
Conclusion
In this study, the potential for the use of upflow anaerobic
fixed-bed reactors with low-density floating carriers as
media for the treatment of winery wastewater was inves-
tigated and the following conclusions were drawn:
1. Anaerobic fixed-bed reactors with small floating
supports offer great promise as high-rate systems for
the treatment of high COD wastewater which typically
contains readily biodegradable organics, such as
winery wastewater. A maximum OLR of 42 g/l day
at 80% removal efficiency was attained in the reactor
with the supports of smallest size and highest specific
surface area.
2. Volatile solid retention, specific biomass activity and
the maximum OLR were found to increase as support
size decreased and its specific surface area increased.
3. The supports favored solids entrapment with very little
or no biofilm formation.
4. Kinetic models were applied to the reactors to evaluate
removal rate constants. The modified Stover–Kincan-
non model gave accurate predictions about reactor
performance and thus should be of effective use when
applied to the design of anaerobic fixed-bed reactors
treating winery wastewater.
5. Due to their higher biomass retention potential, the
supports used in this study offer great promise as
media in anaerobic fixed bed reactors. Anaerobic
fixed-bed reactors with these supports can be applied
as high-rate systems for the treatment of large volumes
of wastewaters typically containing readily biodegrad-
able organics, such as the winery wastewater.
6. Anaerobic fixed bed reactor with the floating supports
used in the present study allows operation at high OLRs
(22–42 kg COD/m3 day) with a removal efficiency of
80%, thus contributing to lower reactor volume and
land area requirements. Anaerobic fixed bed reactors
are advantageous than conventional aerobic systems as
they are less energy intensive and produces less excess
sludge. However, an aerobic post-treatment is required
to make the effluent fit for final disposal.
Acknowledgments The authors wish to express their gratitude to
the French Embassy in India for funding and supporting the program.
Fig. 8 Experimental and
predicted effluent COD for the
reactors
Bioprocess Biosyst Eng (2010) 33:619–628 627
123
References
1. Perez M, Romero LI, Sales D (1999) Anaerobic thermophilic
fluidized bed treatment of industrial wastewater: effect of F:M
relationship. Chemosphere 38:3443–3461
2. Jordening HJ, Buchholz K (2005) High-rate anaerobic waste
water treatment. In: Jordening HJ, Winter J (eds) Environmental
Biotechnology: concepts and applications. Wiley-VCH Verlag
GmbH & Co. KGaA, Weinheim
3. Weiland P, Wulfert K (1986) Anaerobic treatment of stillage
using different pilot scale fixed bed reactors. Biotechnol Bioeng
27:800–809
4. Di Berardino S, Bersi R, Converti A, Rovatti M (1997) Starting-
up an anaerobic hybrid filter for the fermentation of wastewater
from food industry. Bioprocess Eng 16:65–70
5. Sales D, Valcarcel MJ, Perez L, Martinez de la Ossa E (1987)
Activated sludge treatment of wine-distillery wastewaters.
J Chem Tech Biotechnol 40:85–99
6. Eusebio A, Petruccioli M, Lageiro M, Federici F, Cardoso Duarte
JC (2004) Microbial characterization of activated sludge in
jet-loop bioreactors treating winery wastewaters. J Ind Microbiol
Biotechnol 31:29–34
7. Moletta R (2005) Winery and distillery wastewater treatment by
anaerobic digestion. Wat Sci Technol 51:137–144
8. Fumi MD, Parodi G, Parodi E, Silva A, Marchetti R (1995)
Optimization of long-term activated sludge treatment of winery
wastewater. Bioresour Technol 52:45–51
9. Brucculeri M, Bolzonella D, Battistoni P, Cecchi F (2005)
Treatment of mixed municipal and winery wastewaters in a
conventional activated sludge process: a case study. Wat Sci
Technol 51:89–98
10. Torrijos M, Moletta R (1997) Winery wastewater depollution by
sequencing batch reactor. Wat Sci Technol 35:249–257
11. Houbron E, Torrijos M, Moletta R (1998) Application du procede
SBR aux effluentes viticoles: resultats de trois anne‘e de suivi. In:
Proceedings 2nd congress international sur le traitment des
effluents vinicoles, Bordeaux, 5–7 May
12. Petruccioli M, Cardoso J, Eusebio A, Federici F (2002) Aerobic
treatment of winery wastewater using a jet-loop activated sludge
reactor. Process Biochem 37:821–829
13. Malandra L, Wolfaardt G, Zietsman A, Viljoen-Bloom M (2003)
Microbiology of a biological contactor for winery wastewater
treatment. Wat Res 37:4125–4134
14. Andreottola G, Foladori P, Ragazzi M, Villa R (2002) Treatment
of winery wastewater in a sequencing batch biofilm reactor. Wat
Sci Technol 45:347–354
15. Andreottola G, Foladori P, Nardelli P, Denicolo A (2005)
Treatment of winery wastewater in a full scale fixed bed biofilm
reactor. Wat Sci Technol 51:71–79
16. Artiga P, Carballa M, Garrido JM, Mendez R (2007) Treatment
of winery wastewaters in a membrane submerged bioreactor. Wat
Sci Technol 56:63–69
17. Yu HQ, Zhao QB, Tang Y (2006) Anaerobic treatment of winery
wastewater using laboratory-scale multi- and single-fed filters at
ambient temperatures. Process Biochem 41:2477–2481
18. Muller D (1998) Treatment of winery wastewater using UASB
approaches: capability and efficiency. In: Proceedings of the
second international conference on wineries wastewater treat-
ment, Bordeaux France
19. Keyser M, Witthuhn RC, Ronquest LC, Britz TJ (2003) Treat-
ment of winery effluent with upflow anaerobic sludge blanket
(UASB)–granular sludges enriched with Enterobacter sakazakii.
Biotechnol Lett 25:1893–1898
20. Andreottola G, Nardelli P, Nardin F (1998) Demonstration plant
experience of winery anaerobic treatment in a hybrid reactor. In:
Proceedings of the Second International conference on wineries
wastewater treatment, Bordeaux, France
21. Ruiz C, Torrijos M, Sousbie P, Lebrato-Martinez J, Moletta R,
Delgenes JP (2002) Treatment of winery wastewater by anaerobic
sequencing batch reactor. Wat Sci Technol 45:219–224
22. Rajeshwari KV, BalakrishnanM,Kansal A, Lata K, Kishores VVN
(2000) State of the art of anaerobic digestion technology for
industrial wastewater treatment. Renew Sust Energ Rev 4:135–156
23. Choi YS, Shin EB, Chang HN, Chung HK (1989) Start-up and
operation of anaerobic biofilters with packing alternatives. Bio-
process Eng 4:275–281
24. Henze M, Harremoes P (1983) Anaerobic treatment of waste-
water in fixed film reactors: a literature review. Wat Sci Technol
15:1–101
25. Young JC, Dahab MF (1983) Effect of media design n the per-
formance of fixed-bed reactors. Wat Sci Technol 15:369–383
26. Farhadian M, Borghei M, Umrania VU (2007) Treatment of beet
sugar wastewater by UAFB bioprocess. Biores Technol 98:3080–
3083
27. Harendranath CS, Anuja K, Singh A, Gunaseelan A, Satish K,
Lala K (1996) Immobilization in fixed film reactors: an ultra-
structural approach. Wat Sci Technol 33:7–15
28. Bouallagui H, Touhami Y, Ben Cheikh R, Hamdi M (2005)
Bioreactor performance in anaerobic digestion of fruit and veg-
etable wastes. Process Biochem 40:989–995
29. Alkalay D, Guerrero L, Chamy R, Schiappacasse M (1997)
Microbial adherence studies for anaerobic filters. Bioprocess Eng
16:311–314
30. Agamuthu P (1999) Specific biogas production and role of
packing medium in the treatment of rubber thread manufacturing
industry wastewater. Bioprocess Eng 21:151–155
31. Kalyuzhnyi S, Gladchenko M, Sklyar V, Kizimenko Y,
Shcherbakov S (2001) One and two stage upflow anaerobic
sludge bed reactor pre-treatment of winery wastewater. Appl
Biochem Biotechnol 90:107–123
32. APHA, AWWA, WEF (1995) Standard methods for the exami-
nation of water and wastewater, 19th edn. American Public
Health Association, Washington, DC
33. Canler JP, Alary G, Perret JM, Racault Y (1998) Aerobic biological
treatment ofwinery effluentusing tanks in series. In: 2nd international
specialized conference on winery wastewater. Bordeaux, France
34. Mosquera-Corral A, Belmar A, Decap J, Sossa K, Urrutia H,
Vidal G (2008) Anaerobic treatment of low-strength synthetic
TCF effluents and biomass adhesion in fixed-bed systems. Bio-
process Biosyst Eng 31:535–540
35. Borja R, Sanchez E, Martin A, Jimenez AM (1996) Kinetic
behaviour of waste tyre rubber as microorganism support in an
anaerobic digester treating cane molasses distillery slops. Bio-
process Eng 16:17–23
36. Tatara M, Yamazawa A, Ueno Y, Fukui H, Goto M, Sode K
(2005) High-rate thermophilic methane fermentation on short-
chain fatty acids in a down-flow anaerobic packed-bed reactor.
Bioprocess Biosyst Eng 27:105–113
37. Williamson K, McCarty PL (1976) A model of substrate utili-
zation by bacterial films. J Wat Poll Cont Fed 48:9–16
38. Stover EL, Kincannon DF (1982) Rotating biological contactor
scale-up and design. In: Proceedings of the first conference on
fixed-film biological processes. Kings Island, OH
39. Yu HQ, Wilson F, Tay JH (1998) Kinetic analysis of anaerobic
filter treating soybean waste. Wat Res 32:3341–3352
40. Buyukkamaci N, Filibeli A (2002) Determination of kinetic con-
stants of an anaerobic hybrid reactor. Process Biochem 38:73–79
41. Sandhya S, Swaminathan K (2006) Kinetic analysis of treatment
of textile wastewater in hybrid column upflow anaerobic fixed
bed reactor. Chem Eng J 122:87–92
628 Bioprocess Biosyst Eng (2010) 33:619–628
123