121

[Japanese Journal of Water Treatment Biology Vol.35 No.2. 121-129 1999]

Start-up of UASB Reactors with Activated

and Granular Sludges for
the Treatment of Dairy Waste

RIHAB ARIF TARAWNEH and ZAFAR IQBAL BHATTI*

Institute of Environmental Science and Engineering, National University of Sciences and

Technology, Tamiz-ud-Din Road, Lalkurti, Rawalpindi, Pakist

ABSTRACT
Upflow anaerobic sludge blanket (UASB) process is one of the potential biological processes

for the treatment of dairy waste especially in the developing countries. One of the factors of

major concern in anaerobic systems is their long start-up periods. In this study, start-up of

UASB reactors with activated and methanogenic granular sludges was examined. The results

showed that volumetric loading rates that could be achieved at 80% COD removal efficiencies

were 3.7 and 8.5 g-COD/l•Ed while time taken to establish these rates were 60 and 30 days,

respectively for reactors started-up with activated and granular sludges. A solids loading rate of

1.25 to 1.5 g-COD/g-VSSd was found to be a guide value for UASB reactor seeded with ac-

tivated sludge. The results of this study indicated that the UASB reactors can successfully be

started-up with granular or activated sludges but use of granular sludge is advantageous.

Key words : UASB reactors, activated sludge, granular sludge, start-up, dairy waste

treatment of a variety of industrial effluents

INTRODUCTION
including dairy wastes 1-12).Its low-cost and
Use of industrially processed milk and its low-skill characteristics render it to be a
products is becoming more and more popular viable technology for reducing organic
in developing countries resulting in the pollution loads in developing countries.
growth of dairy industry. Like most of the One of the drawbacks of the anaerobic
agro-based industries, dairy industry systems is known to be their long start-up
generates strong wastewater characterized periods. However, this drawback has been
by high BOD and COD concentrations. overcome to a great extent in the UASB
Biological treatment remains the most systems because of the existence of granular
favorable option to control these organic sludge. Formation of methanogenic granular
loads. sludge is an important feature of upflow
Upflow anaerobic sludge blanket (UASB) anaerobic sludge blanket reactors (UASBRs).
process is now widely used as an The presence of granular sludge enhances
energy-conservative biotechnology for the the reactor's resistance against shock loads

*Corresponding author
122 Japanese J. Wat. Treat. Biol. Vol.35 No.2

and against mild dissolved oxygen

MATERIAL AND METHODS
concentrations in the influent. Seeding with
granular sludge also results in rapid start-up This study was carried out in two
of UASBRs. For these reasons, UASBRs are phases. In phase-I, a lab-scale UASBR
now frequently seeded with granular sludges (UASBR-1) was seeded with activated sludge
from other running UASBRs to shorten the whereas in phase-II the same reactor
start-up periods. However in circumstances (henceforth termed as UASBR-2) was seeded
where granular sludge is not available, the with mature and fully developed
UASBR must be seeded with activated methanogenic granular sludge. The duration
sludge or other suitable material. It is, of experimental monitoring was 60 and 35
therefore, important to investigate the days for phase-I and phase-II, respectively.
start-up response of UASBR when seeded UASB reactors The lab-scale UASBR had a
with material other than granular sludge. total volume of 6.85 l including a reactor
This study was undertaken with the portion of 5.5 l and a settling portion of
objective to examine the start-up of UASBRs 1.35 1. The reactor was made of transparent
seeded with activated and fully developed acrylic resin and was equipped with a water
granular sludges and to investigate if jacket to maintain constant temperature.
granular sludge developed on brewery waste The height of the reactor portion was 95 cm
could be used in the start-up of UASBR for with an internal diameter of 9 cm and had
the treatment of dairy waste. sampling ports (SPs) at the height of 3.5, 14,
30, 54 and 94 cm from the bottom. The
reactor was also equipped with a gas solids
separator (GSS). Fig. 1 shows schematic
diagram of the UASBR(s).

Fig. 1 Schematic diagram of the experimental UASB reactor.

 Start-up of UASB Reactors with Activated and Granular Sludges 123

Sources of seed sludges The activated sludge maintained in both the reactors. The reactors
used in UASBR-1 was obtained from a local were started-up in stepwise loading pattern
activated sludge plant treating municipal starting from an influent COD concentration
wastewater. A total of 28.4 g-VSS was of 600 and 500 mg/l; and hydraulic retention
provided in this reactor. In UASBR-2, two times (HRT) of 40 and 30 hours,
liters of fully developed granular sludge respectively for R-1 and R-2.
(69.8g-VSS/l)donated by Asahi Brewery Analysis COD concentration, effluent
Co. Ltd., Osaka from its full-scale UASB plant suspended solids (SS), and mixed liquor
treating brewery waste was added as seed volatile suspended solids (MLVSS) were
material. measured according to Standard Methods13).
Substrate and nutrients Synthetic substrate Gas was measured over tap water saturated
made from dilution of milk powder with tap with NaCL Methane content of the gas was
water was used in this study. No external taken as 75% based on literature7) and our
nutrients were added since all necessary previous work14).
nutrients were available in the milk powder
RESULTS
itself. For every 100 g, the milk powder

contained: protein 25.7 g, lactose 37.4 g, Start-up of UASBR with activated sludge

butterfat 28 g, minerals 5.7 g, calcium 0.93 g UASBR-1 was initiated with a COD

and phosphorous 0.75 g.1M NaHCO3 was concentration of 600 mg/l and HRT of 40 h

added to the influent in amounts as needed that corresponded to a volumetric loading

to maintain a pH around 7.0. rate(V:LR)of 0.36 g-COD/l・d. VLR was then

Operational conditions A temperature of 35 increased stepwise upto 3.7 g-COD/l・d.

℃ was tried to be maintained in the reactors Most of the time COD removal efficiency

but due to operational difficulties the actual remained above 85%as illustrated in Fig.2

temperature remained in the range of which shows the time course of COD

25-35℃ for UASBR-1 and 30-35℃ for removal efficiencies during the experimental

UASBR-2. A pH of closely around 7.0 was period. At day 42, the COD removal

Fig. 2 Time course of influent and effluent COD concentrations and COD removal efficiency for
UASBR-1.
124 Japanese J. Wat. Treat. Biol. Vol.35 No.2

efficiency dropped suddenly due to during the start-up period.

temperature control failure whereafter it Small methane gas bubbles were observed
increased above 80% as the temperature to be rising in the reactor from the start of
increased. Relationship between VLR and its the experiments. Because of low gas
removal rate is shown in Fig.3 which production rates, significantamounts of gas
indicates the limit of the reactor during the could not be collecteduntilday 30. After day
period of operation. Relationship between 30,1-21iters of gas were collected daily,
COD removal efficiency and solids loading however, the gas conversion rate (l-
rate (SLR) is illustrated in Fig.4. COD CH4/g-CODremoved)was low indicating that the
removal efficiency decreased as the SLR population of methanogenic bacteria had not
increased reaching 80% at a SLR of 1.5 grown significantly and that facultative
g-COD/g-VSS・d. This Fig. suggests that a microorganisms may have been competing
F/Mratio of 1.25-1.5 can be applied to a for the substrate with methanogenic bacteria.
UASB reactor seeded with activated sludge On average, methane gas conversion rate

Fig. 3 Relationship between volumetric loading and removal rates for UASBR-1.

Fig. 4 Relationship between COD removal efficiency and solids loading rate for UASBR-1.
 Start-up of UASB Reactors with Activated and Granular Sludges 125

waS 0.1k-CH4/g-CODremoved・Effluent suspended The VLR was started from 0.5g-COD/k・d

solids remained below 200 mg/l until day 42 and increased to 10 g-COD/l・d. The initial

indicating a fairly clear effluent. However, influent COD concentration was 500 mg/

its concentration increased to about 500 mg/ l increasing to 9500 mg/l. High COD removal

l during the last part of experiments when efficiencies greater than 85% were achieved

VLR was relatively high, i.e.,3.7 g-COD/l・d. as shown in Fig.5. Relationship between

The observed yield(Yobs)was found to be volumetric loading and removal rates is

0.051g-VSS/g-CODremoved which is very close shown in Fig.6. This Fig. indicates that a

to Yobs calculated for methanolic waste volumetric loading rate of about 8.5

treatment in a UASB reactor15). g-COD/l・d could be applied to the reactor

Start-up of UASBR with granular sludge with a COD removal efficiency of 80%. VLR

After phase-I study, the reactor was was also plotted against effluent COD

emptied and seeded with granular sludge. concentration as shown in Fig.7. The trend

Fig. 5 Time course of influent and effluent COD concentrations and COD removal efficiency for
UASBR-2.

Fig. 6 Relationship between volumetric loading and removal rates for UASBR-2.
126 Japanese J. Wat. Treat. Biol. Vol.35 No.2

Fig. 7 Relationship between volumetric loading rate and effluent COD concentration for UASBR-2.:

observed in this Fig. is a traditional one. the very first day of operation. Time course

High effluent COD concentration values at a of gas production and gas conversion of COD

V:LR of 6 g-COD/l・d or more do not reflect removed is shown in Fig.8. A maximum gas

the incapability of system, rather they are production rate of 24.6 l/d was observed

indicative of the fact that influent COD whereas average gas conversion rate was

concentrations were correspondingly very calculated to be 0.35 l-CH4/g-CODremoved・

high;still achieving about 90%COD removal. The granular seed sludge was of brownish

Nominal wash-out of biomass took place color having a size range of 0.54-3.2 mm and

from the reactor. Effluent SS remained below asettling velocity of 0.75-3.43 cm/s. Ash

100mg/ltill day 23 increasing to around 600 contents of the sludge was 9.34%. After 35

mg/l towards the end of the study period days of operation the dairy waste treating

when the loading rates were significantly granular sludge became grayish in color with

high. Methane gas production started from size in the range of 0.71-3.2 mm and a settling

Fig. 8 Time course of methane production and methane conversion rate for UASBR-2.
 Start-up of UASB Reactors with Activated and Granular Sludges 127

velocity of 0.65-3.6 cm/s. The ash content of granular sludge manifested better

this sludge was 8.92%. These results performance than the UASBR seeded with
indicated that there was no marked activated sludge. This can be seen from

difference in the physical characteristics of Table l which provides a summary of

the seed and dairy waste treating granular operational and performance data of the two

sludges during the start-up period. reactors. A VLR of 3.7 g-COD/l・d could be

achieved in UASBR-1 after 60 days of

DISCUSSION
operation as compared to 20-25 days in

Although a rational comparison of both the UASBR-2. This indicates that the start-up

reactors cannot be made in this study periods can be significantly reduced by

because of differences in their operational seeding with granular sludge. Moreover,

conditions, the results do provide useful better gas production rates were achieved

evidence of the start-up behavior of these with granular sludge. High removal

reactors. The UASBR started-up with efficiencies in UASBR-2 at an influent COD

Table 1 Summary of operational and performance data of UASB reactors seeded with activated
and granular sludges.
128 Japanese J. Wat. Treat. Biol. Vol.35 No.2

concentration of 7000 mg/l or greater

REFERENCES
(corresponding to a VLR of 5 g-COD/l・d or

greater) and significantly high gas production 1) LettingsG., HulshoffP. L. W., Koster I.
rates after 25 days of operational period W., Wiegant W. M., De Zeeuw W. J.,
suggest that the start-up period was over. Rinzema A., Grin P. C., Roersma R. E.
Thus a start-up period for the treatment of and Hobma S. W.: High-rate anaerobic
dairy waste in a UASBR seeded with waste treatmentusing the UASB reactor
granular sludge was ascertained to be about under a wide range of temperature
25 days. Start-up period of UASBR seeded conditions,Biotechnol.Gen. Eng. Rev.,
with activated sludge could not be concluded 2,253-284 (1984).
in this study, nevertheless, because of the 2) KosterI. W. and LettingsG.: Application
fact that an influent COD concentration of of the upflow anaerobic sludge
3600mg/l corresponding to a VLR of 3.7 bed (UASB) process for the treatment of
g・COD/l・d could be reduced by 80% and complex wastewaters at low
that clear gas production had started, it was temperatures, Biotechnol. Bioeng.,27,
estimated that the start-up period might have 1411-1417,(1985).
been near to completion. On this basis, an 3) Wiegant W. M., ClaassenJ. A. and Lettings
approximate start-up period for this case is G.: Thermophilic anaerobic digestionof
suggested to be about 60-70 days. This result high strength wastewaters, Biotechnol.
is comparable with start-up periods of 40 and Bioeng.,27,1374-1381 (1985).
100days for VFAs treatment in anaerobic 4) Wiegant W. M. and Lettings G.
filters(3000 mg/l VFA mixture;seeding Thermophilic anaerobic digestion of
with thick and light digested sewage sludges sugars in upflow anaerobic sludge
respectively) and 42-49 days for VFAs blanket reactors,Biotechnol.Bioeng.,27,
treatment in UASBRs (5g-COD/l・d of 1603-1607,(1985).
VLR)16). 5) Habets L. H. A. and KnelissenJ. H.
The seeding granular sludge developed on Applicationof the UASB reactor for
brewery waste was adapted to dairy waste. anaerobictreatment of paper and board
There was no reactor upset or sign of low mill effluent,Wat. Sci. Technol.,17,
COD removal efficiency. Thus, the UASBR 61-75 (1985).
was successfully and rapidly started-up with 6) Baily D. G., Tunik M. H., Friedman A.
seed granular sludge cultivated on a different A.and Rest G. B.: Anaerobic treatment
substrate. On the basis of the results of this of tannery waste, Proc.38th Ind. Waste
study, it can be concluded that UASBRs can Conf.,Purdue Univ.,673-682 (1984).
be rapidly started-up by seeding them with 7) Ozturk I.,Eroglu V., Ubay G. and Demir
granular sludge obtained from other UASBRs I.: Hybrid upflow anaerobic sludge
but in circumstances where granular sludge blanket reactor (HUASBR) treatment of
is not available, UASBRs can also be dairyeffluents,Proc.2ndIntl.Symposium
successfully started-up with activated on Waste Management Problems in
sludge. The results of this study also Agro-industries, Istanbul, Turkey,
indicated that treatment of dairy waste is 111-120 (1992).
highly feasible in UASB reactors. 8) Ozturk I.,Ubay G., Sakas S.,and Ubay E.
Applicationof anaerobicbiotechnology
to olive mill effluents,Proc.2nd Intl.
 Start-up of UASB Reactors with Activated and Granular Sludges 129

Symposium on Waste Management 13)APHA, AWWA and WEF.: Standard

Problems in Agro-industries, Istanbul, methods for the examination of water
Turkey, 39-49 (1992). and wastewater,19thed.(1995).
9) Bhatti Z.I., Furukawa K. and Fujita M. 14) Bhatti.Z. L: Studies on the biological
Feasibilityof methanolic waste treatment treatment of methanolicwaste in upflow
in UASB reactors, Water Research,30, anaerobicsludge blanket reactors,Ph.D.
2559-2568 (1996). thesis,Osaka University,1995.
10) Yang G. and Anderson G. K.: Effects of 15) BhattiZ. L, Furukawa K and FujitaM.:
wastewater composition on stabilityof Sludge loadingand productionresponse
UASB. ASCE J: Env. Eng. Div., 119, of an UASB reactor treating high
958-977 (1993). strength industrial wastewater, Jap.
11) Callander I.J., Clark T. A. and McFarlane Jour. Wat. Treat. Biol.,29, 25-29
P.N.: Anaerobic digestion of wood (1993).
ethanol stillageusing upflow anaerobic 16) LettingaG., Hobma S. W., HulshoffPol
sludge blanket reactor, Biotechnol. L.W., de Zeeuw W., de Jong P.,Grin P.
Bioeng., 30, 896-908 (1987). and Roersma R.:Design operationand
12) Yoda M., Imabayashi S. and Suzuki N.: economy of anaerobic treatment, Wat.
Pilot and full-scale experience in Sci. Technol.,15, 177-195 (1983).
anaerobic treatment of brewer's yeast (Submitted 1998.10.21)
processing wastewater, Wat. Sci. (Accepted 1999.2.10)
Technol., 23,1167-1177 (1991).