Treatment of winery wastewater in a conventional

Water Science & Technology Vol 56 No 2 pp 79–87 Q IWA Publishing 2007

municipal activated sludge process: five years of
experience
D. Bolzonella*, M. Zanette**, P. Battistoni*** and F. Cecchi**
*Department of Wine Science, Technology and Marketing, University of Verona, I-37100 San Floriano,
Verona, Italy (E-mail: david.bolzonella@univr.it)
**Department of Science and Technology, University of Verona, Strada Le Grazie, 15. I-37134 Verona, Italy
(E-mail: franco.cecchi@univr.it)
***Institute of Hydraulics, Marche Polytechnical University, via Brecce Bianche. I-60100 Ancona, Italy
(E-mail: idrotre@mta01.unian.it)

Abstract A full-scale wastewater treatment plant where municipal and winery wastewaters were co-treated
was studied for five years. The experimental results showed that suspended solids, COD, nitrogen and
phosphorous were effectively removed both during the treatment of municipal wastewater and the co-
treatment of municipal and winery wastewater. The sludge production increase from 4 tons to 5.5 tons per
day during the harvesting and wine making period. In any case the specific sludge production was
0.2 kgMLVSS per kgCODremoved despite the organic loading increasing. About 70% of the COD was
removed through respiration. Also the energy demand increased from 6,000 to 7,000 kWh per day. The
estimated costs for the treatment of the winery wastewater was 0.2–0.3 Euros per m3 of treated
wastewater. With reference to the process efficiency, the nitrogen removal was just 20%. The co-treatment
of municipal and winery wastewater in conventional activated sludge processes can be a feasible solution
for the treatment of these streams at relatively low costs.
Keywords Activated sludge model; activated sludge process; energy consumption; winery wastewaters

Introduction
The worldwide production of wine in 2003 was nearly 267 million hl, an increase of
9 million hl (3.5%) compared with 2002. Europe, with a total production of 187 million
hl, still leads the rank with 70% of the production: France, Italy and Spain showed a
similar production of some 42 –46 million hl. Americas posted a total production of 45.8
million hl with the major increase in Chile (19% with respect to 2002). According to the
figures presented, it is evident that wine represents a very important market for several
countries. One of the drawback of wine making and bottling is the production of huge
amounts of wastewater, generally some 1.2–1.4 litres per litre of wine produced, up to 4
litres (Vlyssides et al., 2005). The treatment and disposal of winery wastewater still rep-
resents a controversial issue because of its particular characteristics: a high organic load
associated with a high biodegradability. About 80% of total COD is soluble (mainly etha-
nol), the ratio BOD5/COD ranges between 0.5 and 0.6 and the ratio BOD5:TKN:P is
100:1:0.25 (Canler et al., 1998; Andreottola et al., 2005; Vlyssides et al., 2005). Accord-
ing to Andreottola et al. (2005) some 85% of the COD is soluble and mainly readily bio-
degradable (RBCOD) because of the presence of compounds like ethanol and methanol
(see Table 1); then, also several sugars and organic acids are present (Colin et al., 2005).
Furthermore, these wastewaters are characterised by a noticeable seasonal variability in
terms of flow and pollution load.
doi: 10.2166/wst.2007.475 79
 Table 1 Characteristics of some Italian winery wastewaters in mg/l (Andreottola et al., 2005)

Total COD Soluble COD TSS NH4-N Organic N PO4-P

6,800 –7,100 5,600 –5,800 690 – 722 21– 27 22 –30 6–8

Because of this composition, winery wastewaters are generally treated by means of

biological processes (aerobic, anaerobic or combined). Sometimes these are associated
D. Bolzonella et al.

with chemical –physical pre- or post-treatments (Colin et al., 2005; de Heredia et al.,
2005). An interesting option for the treatment of these wastewaters is the possibility of
co-treating municipal and winery wastewaters in conventional wastewater treatment
plants (WWTP; Fumi et al., 1995; Chudoba and Pujol, 1996; Beck et al., 2005; Bruccu-
leri et al., 2005). This technique enables existing structures to be exploited without build-
ing new plants; furthermore, these plants may easily face the over-loading observed
during harvesting and wine making or bottling periods.
This paper is the prosecution of the one presented by Brucculeri et al. (2005) and
deals with the results and observations of four years of treatment of mixed municipal
and winery wastewater in a conventional activated sludge process. At the end, a possible
process modification to improve the plant efficiency is proposed.

Materials and methods

The WWTP studied is located in San Bonifacio, near Verona, northeast Italy. This area is
famous for the production of white wines Soave and Recioto which reaches some
500.000 hl per year. The design capacity of the plant, upgraded in 2001, is 60,000 people
equivalent (PE): the design flowrate is 4,650,000 m3/year of wastewaters,
250,000 m3/year from the wine-makers. Those treat some 65,000 tons/year of grapes.
Figure 1 shows the trends of flow and COD loading during the period 2001 –2004: it is
evident that the periods September –October can show peaks of COD loading up to
12,000 –16,000 kg/d (some 100,000 PE) while the flowrate remain similar.
Figure 2 shows a schematic of the WWTP. After pre-treatment (screening, grit and
oil/grease removal) the wastewater reaches the 5,400 m3 aerobic bioreactor. Treated

30000
Flowrate, m3/d
COD load, kg/d
25000

20000
Value

15000

10000

5000

0
07

04

03

31

30

27

24

22
/0

/0

/0

/1

/0

/0

/0

/0
8/

1/

6/

0/

3/

8/

1/

6/
20

20

20

20

20

20

20

20
01

02

02

02

03

03

04

04

Days

80 Figure 1 Trends of the flow rate and the COD load in 2001 –2004
 Raw
wastewater Effluent
1 2 3 4

Disposal 6 5

D. Bolzonella et al.
Figure 2 Plant schematic: 1: pre-treatment, 2: oxidation, 3: clarifier, 4: disinfection tank, sludge: 5:
thickening, 6: belt-filter press

water and sludge are separated in five settlers: three with a surface area of 1,038 m2 each
and two with a surface area of 770 m2 each.
Treated water is (eventually) disinfected and discharged. The sludge treatment process
consists of a thickening section followed by a belt-filter press and no stabilisation is
present. The plant was monitored for four years (from 2001 to 2004). The main character-
istics for the influent wastewater, the effluent and wasted sludge were determined in
terms of COD, suspended solids, nitrogen and phosphorous (all according to Standard
Methods), analysing both grab and average samples; furthermore, the operational
parameters of the biological process (hydraulic and solid retention time, sludge concen-
tration, flows) were also monitored.

Results and discussion

During the first three years studied the flow of treated wastewater was relatively constant
with an average flow of 15,000 m3/d with variations related to the rainy and dry periods
while in 2004 an increase in the flowrate due to new connections to the sewer system
was observed.
As for the COD load, that varied according to the harvesting periods. Figure 3 shows
the trend of the COD concentration during the harvesting and wine making period in
2003: it is clear how the COD concentration increased from an average of 250 mg/l up to
500 mg/l, with peaks of 1,000 mg/l, during the period from September to mid-November
and then decrease slowly to typical levels.
According to the COD load treated in years 2001 –2004 (Figure 4), the operation of
the WWTP can be divided in two different conditions:

1200 250

COD IN 200
COD OUT
COD out, mg/L
COD in, mg/L

800
150

100
400
50

0 0

/ 03 / 03 / 03 / 03 / 03 / 03 / 03 / 04 / 04 / 04 / 04
/6 /7 /8 /9 /10 /11 /12 /1 /2 /3 /4
15 15 14 13 13 12 12 11 10 11 10
Date

Figure 3 COD concentration during the harvesting and wine making period in 2003 –2004 81
 COD IN
6000 600
TKN IN
5000 500
Kg/d COD IN

4000 400

kg/d TKN IN
3000 300
D. Bolzonella et al.

2000 200

1000 100

0 0
1° 2° 3° 4° 5° 6°
(Harvest) (Civil) (Harvest) (Civil) (Harvest) (Civil)

Period

Figure 4 COD load and TKN load of the influent during the different periods

† Harvesting, wine making and bottling period: from September to February. In this
period we observed the highest organic pollution load due to harvesting (September)
and wine production (October – December) followed by the bottling process
(December –February). In this condition the typical load exceed 3,000 kg of COD per
day and can occasionally rise up to 15,000 kg per day.
† Normal period: from March to August. In this period the organic pollution load is due
to the normal load originating from the municipality and is generally lower than
3,000 kg per day (see Figure 4).
However, it is important to emphasise that the peaks of concentrations are limited to a
few days so the average load during the harvest periods is only 20 –30% greater than the
one observed during the rest of the year.
While the COD loads differed according to the operation, the TKN load was constant
and equivalent to some 450 kgN/d (see Figure 4). The same was true for TP which
showed a typical load of some 35 kgP/d. The variation of the COD load and the stability
of the TKN resulted in different COD/N ratios ranging from 5.6 to 7.7 for the normal
period and from 7.7 to 13 for the harvesting period. According to these data the typical
loading of the plant was some 30,000 PE for the ordinary period, when only municipal
wastewaters were treated, while it reached the design capacity of 60,000 PE (on average)
during the periods of harvesting and wine making. With specific reference to the oper-
ational conditions applied during the two periods in these years, these are reported in
Table 2.
According to data reported in Table 2, it is clear that the high organic loads of the
harvesting were treated with increasing biomass concentration in the reactor. This
increased from 4 to more than 5–6 kg per m3. As a consequence the two different periods

Table 2 Operational conditions, sludge production and observed yields

Period Type Flowrate CODIN/TKNIN MLSS YOBS F/M SRT Sludge

(m3/d) (g/L) (gVSS/gCOD) (gCOD/gVSSd) (d) (ton/d)

18 Harvest 13,095 13.1 8.4 0.12 0.07 34 5.3

28 Civil 13,835 5.6 4.3 0.21 0.10 28 3.6
38 Harvest 16,879 7.7 5.1 0.22 0.13 23 5.2
48 Civil 13,477 7.7 4.6 0.18 0.12 26 4.1
58 Harvest 14,540 9.7 5.1 0.22 0.12 19 6.2
68 Civil 18,790 6.5 3.9 0.23 0.12 24 3.9
82
presented similar F/M ratios, in terms of kgCOD/kgMLVSS per day. The production of
waste activated sludge during the harvesting periods was higher (40%) while SRT and
observed yield maintained similar values compared to the normal periods. However,
despite the high organic load the sludge production increased relatively little. Probably,
the higher biomass concentration mitigated this aspect shifting the metabolism of the acti-
vated sludge to the respiration pathway rather than growth (Low and Chase, 1999). This
is also confirmed by data related to COD utilisation (see below). The harvest periods

D. Bolzonella et al.
obviously presented larger sludge production and these determined an increase of some
1.7 ton per day of sludge to be disposed of, for a total cost of some 36,000 Euros per
year. This additional cost is completely attributed to the treatment of winery wastewaters.
It is then important to note that a stabilisation step of sludge is absent in the plant and its
presence could reduce these costs. Despite the high content of MLSS in the bioreactor
and the increased load to the clarifiers, those did not undergo to difficulties; in fact, the
sludge volume index (SVI) was the lowest level during the harvesting period owing to
the increased settability of sludge. This fact can be probably ascribed to the formation of
floc-forming bacteria due to the high availability of organic substrates; SVI reached
values less than 100 cc/g in the period September –February and rose to values over
150 cc/g for the rest of the year. Figure 5 shows that the SVI decreased for increasing
COD loads. This resulted in a decrease in the loss of suspended solids in the effluent
from 35 to 20 mg/l. As a result, the COD, TKN and TP also showed very low concen-
trations in the effluent during the harvesting and wine making periods being the sus-
pended solids perfectly removed from the effluent.

Mass balances
With specific reference to the COD mass balance (Figure 6), the different operational
conditions during the harvest and the civil periods generated a similar distribution of the
COD mass balance with a prevalence of respiration processes (about 60– 70%) over
growth (about 23%) and with a COD in the effluent under 20%, except in the last period
when it increased to 25% due to the loss of suspended solids from the clarifier. It is
noticeable from Figure 6 that during all the periods (except the last) the percentage of
removal was about 80%. Finally, the COD concentration in the effluent averaged 50 mg/l,
except in period six which showed a removal of about 75%. With specific reference to
nitrogen, this was removed through biomass growth for 10% in almost all the periods
considered in this study (Figure 7). There are no evident differences between the two

200
180
160
140
SVI (cc/g)

120
100
80
60
40
20
0
0 0.25 0.5 0.75 1 1.25 1.5 1.75
OLR (KgCOD/m3d)

Figure 5 Profile of the SVI to OLR during 2001 –2002 83

 Respiration
80.0 Growth
Out
70.0
60.0
50.0
COD %

40.0
D. Bolzonella et al.

30.0
20.0
10.0
0.0
1° 2° 3° 4° 5° 6°
(Harvest) (Civil) (Harvest) (Civil) (Harvest) (Civil)

Period

Figure 6 COD partition of the mass balance in differing operating periods

periods: the percentage for nitrogen denitrified varied from 40– 55% and effluent nitrogen
varied from 37–48%, with TKN output near 10% of influent N. In terms of concen-
trations the effluent TN was some 15–20 mgN/l and nitrates were clearly predominant.
According to reported data, the process was not effectively controlled and the efficiency
in nitrogen removal was related to high or low organic loading.
Finally, total phosphorus in the effluent showed a constant concentration of
1 –2 mgP/l. The average removal efficiency was 50% without chemical addition.

Energy consumption
Further evidence for the difference between the harvesting and ordinary periods came
from data related to the energy consumption (see Figure 8). It is evident here, that energy
consumption increased during October and November, when the highest organic load was
observed. This was clearly due to an increased demand in oxygen requirements.
Compared to the rest of the year, the treatment of winery wastewaters determined an
increase of the energy consumption of 750 kWh per day (as an average) reaching values
of 475 Wh per m3 treated (equivalent to 120 Wh/PE). This value is similar to the highest
reported in the literature for the treatment of municipal wastewaters (Balmer and Matt-
son, 1994; Battistoni et al., 2003a, b).

Denitrification
60 Out N-NO3
Out TKN
50 Growth

40
N%

30

20

10

0
1° 2° 3° 4° 5° 6°
(Harvest) (Civil) (Harvest) (Civil) (Harvest) (Civil)
Period

84 Figure 7 Nitrogen partition of the mass balance in differing operating periods

 10000

8000

6000
KW/d

4000

D. Bolzonella et al.
2000

C ER

R
Y

R
Y

ST
H

IL

LY
AY
AR
AR

BE

BE
N

BE
C

JU

B
M

JU
AR

AP

EM

EM
U

TO
U

BR
N

AU
M

VE
JA

PT

EC
FE

O
SE

D
Months N

Figure 8 Typical energy consumption during a year

Possible process improvement

The application of the alternated cycle process (ACP) could be the right solution to
improve performances of the plant in terms of nutrient removal and to reduce the energy
consumption due to the optimisation of the cycles length (Battistoni et al., 2003a, b).
The comparison between the actual situation and the application of the ACP is car-
ried out through the simulation of the process by means of the activated sludge model
considering the alternation of 1 hour anoxic and 1 hour aerobic cycles. The character-
istics of the treated wastewaters were determined according to the simplified method-
ology reported in Koch et al. (2000) the soluble COD during the harvesting period was
67% of the total influent (Brucculeri et al., 2005; Stricker and Racault, 2005). Figure 9
shows the typical distribution of the influent COD for the municipal or civil period and
the mixed winery wastewaters. Table 3 compares the results determined by the process
simulation, considering the treatment of municipal and mixed wastewaters in the con-
ventional and the alternated cycles processes at the same operational conditions of tem-
perature and SRT: the result is the alternated cycles process allows for the optimisation
by biological nitrification with attendant recovery of oxygen and lowering of nitrates as
low as 6 mgN/l, while oxygen consumption would be lowered from 292 to 225 kg/h O2.
This can be roughly considered as a reduction of some 23% of the energy consumption.

Biomass;
x NBCOD;
Biomass; 5,7
x NBCOD; 8,6
2,2
x COD; 5,2
21,6

x COD; s COD;
s COD; 48,6
s NB COD; 25,7 s NB COD;
66,7
4,3 11,4
Winery Municipal

Figure 9 COD partition in winery and municipal wastewaters (x and s are the particulate and soluble
fractions, respectively) 85
 Table 3 Predicted effluent characteristics of an alternated cycle process (ACP)

Parameter Municipal Winery Winery ACP

SRT (days) 2.5 2.5 2.5

MLSS (g/l) 7.1 8.9 8.9
Ammonia (mgN/l) 0.5 0.4 0.6
Nitrates (mgN/l) 19.0 18.0 5.9
Oxygen (kg/h) 115 292 225
D. Bolzonella et al.

Economic considerations
Considering the data reported it is possible to calculate the extra costs determined by the
winery wastewater treatment. As for energy consumption, the winery wastewater deter-
mined an increase for energy demand of some 750 kWh per day and an extra cost of
some e 25,000 per year. Considering sludge disposal we observed an extra production of
some 1.5 tons of sludge per day corresponding to some e 36,000 per year of extra costs
for sludge disposal. According to these figures, the specific cost for the treatment of
winery wastewater at San Bonifacio WWTP was e 0.25 per m3 of winery wastewater
treated in terms of pure operating costs for energy and sludge disposal.
It is evident how the treatment of winery wastewater in a conventional plant can be
considered a good solution for this problem: in fact, it is possible to avoid costs for new
plants and facilities as well as personnel costs. Furthermore, considering the specific case
study presented here, costs could be also lowered considering the application of the alter-
nated cycles process and the introduction of a stabilisation process to reduce the content
of the volatile solid fraction in waste activated sludge and reduce costs for disposal (the
digester is already present in the WWTP).

Conclusions
The possibility of co-treating municipal and winery wastewater in a full-scale treatment
plant was studied. The main conclusions deriving from the experimentation can be
resumed as follows:
† The oxidation process was able to handle the high COD peaks (up to 15,000 kg/d)
during harvesting and wine making period, from September to December. The high
solids retention times (SRTs) applied to the system, always greater than 20 days,
determined low production of excess sludge (Yobs ,0.25 gVSS/gCOD). The winery
and municipal wastewaters induced a similar growth of the activated sludge, but
during harvesting period the respiration rate was higher.
† Energy consumptions increased during the harvest period from 6,000 to 7,000 kWh
per day. It was calculated that the sum of energy costs and sludge disposal accounted
for e 0.25 per m3 of winery wastewater treated.
† The possible application of the alternated cycles process (ACP) could lead to a perfect
control of nutrients emissions. The simulation of the process emphasised that the
application of the ACP can lead to a nitrogen concentration in the effluent of
6– 7 mgN/l (now is 20 mgN/l), while the oxygen consumption could decrease from
292 to 225 kg oxygen per hour, a 23% reduction in energy consumption.
† According to the results shown, the co-treatment of winery together with municipal
wastewater in conventional WWTPs seems to be a feasible and reliable solution and
municipal WWTPs with unused capacity can effectively and economically treat
86 winery wastewater.
Acknowledgements
The authors thank the Consorzio Le Valli, San Bonifacio (Verona), Pierluigi Dodi, Dario
Masenelli, Giuliano Fontanesi and Antonio Fornasiero and SAGIDEP SpA for their valu-
able help.

References

D. Bolzonella et al.
Andreottola, G., Foladori, P., Nardelli, P. and Denicolo, A. (2005). Treatment of winery wastewater in a full-
scale fixed bed biofilm reactor. Water Sci. Technol., 51(1), 71– 79.
Balmer, P. and Mattsson, B. (1994). Wastewater treatment operation costs. Water Sci. Technol., 30(4), 7– 15.
Battistoni, P., Boccadoro, R., Bolzonella, D. and Marinelli, M. (2003a). An alternate oxic-anoxic process
automatically controlled: theory and practice in a real treatment plant network. Water Sci. Technol.,
48(11/12), 337 – 344.
Battistoni, P., de Angelis, A., Boccadoro, R. and Bolzonella, D. (2003b). An automatically controlled
alternate oxic-anoxic process. A feasible way to perform high nitrogen biological removal also during wet
weather periods. Ind. Eng. Chem. Res., 42(3), 509 –515.
Beck, C., Prades, G. and Sadowski, A.G. (2005). Activated sludge wastewater treatment plants optimisation
to face pollution overloads during grape harvest periods. Water Sci. Technol., 51(1), 81 –88.
Brucculeri, M., Bolzonella, D., Battistoni, P. and Cecchi, F. (2005). Treatment of mixed municipal and
winery wastewaters in a conventional activated sludge process: a case study. Water Sci. Technol., 51(1),
89 –98.
Canler, J.P., Perret, J.M. and Racault, Y. (1998). Traitment biologique aerobie par bassin en serie des
effluentes vinicoles. In Proceedings 2nd Congres International sur le Traitment des Effluents Vinicoles,
Bordeaux, May 5 – 7.
Chudoba, P. and Pujol, R. (1996). Activated sludge plant facing grape harvest period. A case of study. Water
Sci. Technol., 34(11), 25 –32.
Colin, T., Bories, A., Sire, Y. and Perrin, R. (2005). Treatment and valorisation of winery wastewater by a
new biophysical process (ECCF (R)). Water Sci. Technol., 51(1), 99 – 106.
Fumi, M.D., Parodi, G., Parodi, E., Silva, A. and Marchetti, R. (1995). Optimization of long-term activated
sludge treatment of winery wastewater. Bioresour. Technol., 52, 45 – 51.
Koch, G., Kühni, M., Gujer, W. and Siegrist, H. (2000). Calibration and validation of activated sludge model
No. 3 for swiss municipal wastewater. Water Res., 34, 3580– 3590.
Low, E. and Chase, H. (1999). Reducing production of excess biomass during wastewater treatment. Water
Res., 33(5), 1119– 1132.
Stricker, A.E. and Racault, Y. (2005). Application of activated sludge model No. 1 to biological treatment of
pure winery effluents: case studies. Water Sci. Technol., 51(1), 121 – 127.
Vlyssides, A.G., Barampouti, E.M. and Mai, S. (2005). Wastewater characteristics from Greek wineries and
distilleries. Water Sci. Technol., 51(1), 53 –60.
de Heredia, J.B., Torregrosa, J., Dominguez, J.R. and Partido, E. (2005). Degradation of wine distillery
wastewaters by the combination of aerobic biological treatment with chemical oxidation by Fenton’s
reagent. Water Sci. Technol., 51(1), 167 – 174.

87