Bioresource Technology 93 (2004) 209–212

Feasibility studies on the treatment of dairy wastewaters withupflow anaerobic sludge blanket reactors

E.V. Ramasamy, S. Gajalakshmi, R. Sanjeevi, M.N. Jithesh, S.A. Abbasi *

Centre for Pollution Control and Energy Technology, Pondicherry University, Kalapet, Pondicherry 605 014, India

Abstract

The feasibility of using upflow anaerobic sludge blanket (UASB) reactors for the treatment of dairy wastewaters was explored.

Two types of UASBs were used––one operating on anaerobic sludge granules developed by us from digested cowdung slurry

(DCDS) and the other on the granules obtained from the reactors of M/s EID Parry treating sugar industry wastewaters. The

reactors were operated at HRT of 3 and 12 h and on COD loading rates ranging from 2.4 kg perm3 of digester volume, per day to

13.5 kgm�3 d�1. At the 3 h HRT, the maximum COD reduction in the DCDS-seeded and the industrial sludge-seeded reactors was

95.6% and 96.3%, respectively, better than at 12 h HRT (90% and 92%, respectively). In both the reactors, the maximum, the second

best, and the third best COD reduction occurred at the loading rates of 10.8, 8.6 and 7.2 kgm3 d�1, respectively. At loading rates

higher than 10.8 kg, the reactor performance dropped precipitously.

Whereas in the first few months the reactors operating on sludge from EID Parry achieved better biodegradation of the waste,

compared to the reactors operated on DCDS, the performance of the latter gradually improved and matched with the performance

of the former.

� 2003 Elsevier Ltd. All rights reserved.

Keywords: Sludge granules; Biogas; Methane; Energy recovery; Organic loading rate; UASB

1. Introduction

Dairy industry has grown in most countries of the

world because the demand for milk and milk productshas steadily risen. Simultaneously, the production of

milk per head of cattle has also grown as a result of

advancements in veterinary science (Poompavai, 2002).

Among nations, India is one of the largest, and is

projected to become the largest, producer of milk and

dairy products in the world (Gupta, 1997). With annual

milk production crossing 85 million tonnes in the year

2002, and growing at the rate of 2.8% per annum(DAH&D, 2003). India is also by far the largest pro-

ducer of dairy-based wastewaters.

The dairy industry wastewaters are generated pri-

marily from the cleaning and washing operations in the

milk processing plants and are estimated to be 2.5 times

the volume of the milk processed. Thus, some 200 mil-

lion tonnes of wastewaters are generated annually from

*Corresponding author. Tel.: +91-413-2655263/2655262.

E-mail address: prof_abbasi@vsnl.com (S.A. Abbasi).

0960-8524/$ - see front matter � 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/j.biortech.2003.11.001

the Indian dairy industry. Similarly large volumes of

dairy industry wastewaters are generated in other parts

of the world as well.

These wastewaters are rich in biodegradable organicsand nutrients (Poompavai, 2002; Ramasamy and Ab-

basi, 2000). If not treated, they cause gross pollution of

land and water with their high BOD and COD. But they

also have the potential to supply carbon in a form that

anaerobic microorganisms can convert into methane

(Franklin, 2001). This opens up the possibility of gener-

ating clean fuel (methane) with concomitant pollution

control.We have been exploring the possibilities of develop-

ing dairy wastewater treatment systems with simul-

taneously recovery of energy (Ramasamy and Abbasi,

2000; Ramasamy, 1997). In one such attempt the effi-

ciency of the conventional anaerobic CST (continuously

stirred tank) reactors was enhanced by incorporat-

ing inexpensive yet effective biofilm support systems

(Ramasamy and Abbasi, 2000). But, inspite of beingused world-wide to the present day, anaerobic CSTRs

are increasingly giving way to faster and more effi-

cient higher-rate anaerobic digesters, notably anaerobic

210 E.V. Ramasamy et al. / Bioresource Technology 93 (2004) 209–212

filters, expanded/fluidized bed reactors, or upflow an-

aerobic sludge blanket (UASB) reactors (Poompavai,

2002; Abbasi and Ramasamy, 1999; Datar et al., 2001).

Of these, UASB reactors hold particular attraction be-

cause such reactors can handle higher suspended solid

loads and shock loads, besides wastewaters of a greater

range of strengths, than other type of reactors (Lettinga

and Pol, 1991, 1992; El-Mamouni et al., 1998; Brinkmanand Hack, 1996; Uemura et al., 2002; Xu et al., 2003).

The presence of carbohydrates, as in dairy wastewaters,

promotes the production of extracellular polysaccha-

rides, which enhance bacterial agglomeration and hence

are believed to be essential to the formation of granules

necessary for the success of UASB reactors (Quarmby

and Forster, 1995; Batstone and Keller, 2001). Some

additives have been shown to enhance sludge granula-tion, especially natural and cationic polymers (El-

Mamouni et al., 1995) which may also be present in

dairy wastewaters.

In this paper we have explored the feasibility of using

UASB reactors for dairy wastewater treatment and en-

ergy recovery. The paper describes studies on the setting

up of the reactors, performance of two different types of

sludge granules, and effect of different COD (chemicaloxygen demand) loading rates on the reactor perfor-

mance.

2. Methods

The reactors were fabricated in borosilicate glass and

were housed in thermostated (30± 2 �C) chambers. The

feed loading rates were controlled with peristalticpumps(Ravel Hi-tech, Rh-P120 MC), pH was measured

with systronic grip-D pH meters (accuracy ±0.01 pH

units), chemical oxygen demand (COD) was assessed by

the open reflux method (APHA, 1998) and volatile fatty

acids (VFAs) were analyzed by the method of Dilallo

and Albertson (1961). Biogas quantity was measured

with wet gas flow meters (Gemini, GSI 032) and its

quality was determined by AIMIL-make gas liquidchromatograph using thermal conductivity detectors.

All chemicals were of analytical reagent grade unless

otherwise stated. Deionized water, double distilled in all-

glass units of NPL design, was used for all purposes.

2.1. Feed

In order to have a feed of consistent composition,

synthetic dairy waste (SDW) was generated in the lab-

oratory by dissolving 1g milk powder (Amulya brand,

manufactured by Mehsana District Cooperative Milk

producer’s Union Ltd., Gujarat) per litre of water. Itprovided a COD of 1440 mg/l. From this stock solution,

SDWs of different strengths were made by appropriate

dilution. During the period when the reactors were

operated to develop active, granular, sludge, the feed

was fortified with appropriate nutrients using the recipe

of Zehnder and Wuhrmann (1977).

2.2. Reactor operation

Four sets of reactors, each consisting of duplicates

were started simultaneously. Two of the sets were seeded

with digested cowdung slurry (DCDS) obtained from

the exit sump of a low-rate, plug-flow, 40 d HRT di-

gester of KVIC (Khadi and Village Industries Com-

mission) design operated with cowdung-water mixture

as feed. The other two sets of digesters were seeded withgranular sludge taken from the UASBs functioning at

the sugar processing unit of EID Parry, Pondicherry.

A set each of DCDS and EID Parry seeded digesters

were fed with synthetic dairy waste of strength 300

mg l�1 COD and the other two similar sets were fed with

the waste of 1200 mg l�1 COD. This enabled us to operate

the reactors at different HRTs while at the same time

maintaining identical space loading rates per day (Tables1 and 2). The performance of duplicates agreed to within

±5%; the data in Tables 1 and 2 present the average.

3. Results and discussion

All the four reactors were started with a low organic

loading rate (OLR) of 2.4 kgCODm�3 d�1 (Tables 1and 2) so as to give the microorganisms in the sludge an

opportunity to get acclimatized with the SDW feed. It

also enabled development of sludge granules in the

reactors.

During the first 40 d of operation the COD reduction

was high in the reactors with HRT 12 h––73.5% in

DCDS seeded reactors and 85% in EID Parry seeded

reactors at the outset––whereas in the 3 h HRT reactorsthe COD reduction gradually built up to a little above

60% in both types of reactors (Tables 1 and 2).

After 40 d the OLR was stepped up to 3.6

kgCODm�3 d�1 and the reactors were operated at this

influent loading for another 50 d. This period witnessed

improvement in the reactor performance in all cases, the

COD reduction reaching 92% in the EID Parry seeded

reactors and 83–90% in DCDS seeded reactors.Encouraged by this we doubled the OLR to 7.2

kgCODm�3 d�1. But this led to some sludge wash-out

from the DCDS seeded reactors of 3 h HRT. In the EID

Parry seeded reactors, which had heavier sludge, there

was no wash-out but the entire sludge in the 3 h HRT

reactor beds rose to the top portion of the reactors. The

shock load also effected the performance of 12 h HRT

reactors as the COD reduction in them dropped byover 30%. We therefore stepped the OLR down to

5.4kgm�3 d�1 whereupon the sludge wash-out as well as

its polarization to the top of the reactor stopped. The

Table 1

Treatment of synthetic dairy waste in UASB reactors operating on anaerobic sludge granules derived from digested cowdung slurry

Number of days HRT 12 h HRT 3 h

COD loading rate (kgm�3 d�1) COD reduction (%) COD loading rate (kgm�3 d�1) COD reduction (%)

10 2.4 67 2.4 25

20 2.4 69 2.4 45

30 2.4 73.5 2.4 56

40 2.4 74 2.4 61

50 3.6 83 3.6 63

70 3.6 87 3.6 66

80 3.6 89 3.6 68

90 3.6 90 3.6 83

100 7.2 53 7.2 83

130 5.4 53 5.4 84

145 5.4 63 7.2 85

160 5.4 76 7.2 91

175 6.48 81.7 8.64 94.5

190 6.48 83 10.8 95.6

210 6.48 83 13.5 86

The COD reduction data is the average of duplicate reactors (with a maximum relative error ±5%) and pertains to the last day of each set of

observations.

Table 2

Treatment of synthetic dairy waste in UASB reactor operating on anaerobic sludge granules loaned from M/s EID Parry

Number of days HRT 12 h HRT 3 h

COD loading rate (kgm �3 d�1) COD reduction (%) COD loading rate (kgm �3 d�1) COD reduction (%)

10 2.4 85 2.4 37

20 2.4 87 2.4 62

30 2.4 89 2.4 68

40 2.4 92 2.4 66

50 3.6 86 3.6 74.6

70 3.6 87.3 3.6 91

80 3.6 91.5 3.6 92

90 3.6 92 3.6 92

100 7.2 51 7.2 82

130 5.4 62 5.4 93

145 5.4 69 7.2 90

160 5.4 80 7.2 93

175 6.48 62 8.64 94

190 6.48 62 10.8 96.3

210 6.48 61 13.5 83

The COD reduction data is the average of duplicate reactors (with a maximum relative error ±5%) and pertains to the last day of each set of

observations.

E.V. Ramasamy et al. / Bioresource Technology 93 (2004) 209–212 211

rate of COD reduction also improved. Thereafter we

gradually increased the OLR to 6.48 kgm�3 d�1 in 12 h

HRT reactors and to 13.5 kgm�3 d�1 in 3 h HRT

reactors. With the increase in OLR upto 10.8

kgCODm�3 d�1 in lower HRT reactors, the reactor

performance attained a peak of 95.6% COD removal.

Further increase in OLR to 13.5 kgCODm�3 d�1 caused

a significant drop in the performance. On the other handin reactors with 12 h HRT, increase in OLR beyond 3.6

kgCODm�3 d�1 did not cause improvement in the ex-

tent of COD reduction and the reactors did not even

attain the levels of COD reduction achieved with the

OLR of 3.6 kgCODm�3 d�1.

It may be seen that the UASB reactors seeded with

sludge taken from M/s EID Parry achieved significantly

better COD reduction during the first few months. But

gradually the granulation and settling characteristics of

DCDS seeded sludge improved; with it improved the

performance of these reactors. By the end of the 7-

month long experiments there was little to choose be-

tween the two types of reactors.The studies establish the feasibility of UASB reactors

in treating dairy wastewaters to the extent of >90%

reduction in COD. It is also seen that digested cow-

dung slurry ensuing from low-rate biogas digesters can

be used to develop active anaerobic sludge with easily

212 E.V. Ramasamy et al. / Bioresource Technology 93 (2004) 209–212

settleable granules appropriate for UASB reactors. The

quality of DCDS sludge developed in this manner was

as good as that of the sludge taken from mature reactors

of EID Parry. The granulation process might have

been helped by the presence of carbohydrates, and

polymer additives in the dairy wastewaters, as noted

earlier. The optimal performance was given by reactors

with 3-h HRT and OLR 13.5 kgCODm�3 d�1.

Acknowledgements

The authors thank All India Council of Technical

Education, New Delhi, for infrastructural support under

their MODROBS grant. SG thanks Council for Scien-

tific and Industrial Research, New Delhi for a SeniorResearch Fellowship.

References

Abbasi, S.A., Ramasamy, E.V., 1999. Biotechnological methods of

pollution control. Orient Longman, Universities Press of India

Ltd., p. 168.

APHA, 1998. Standard Methods for the Examination of water. In:

Rand, M.C., Greenberg, A.R., Taras, M.J. (Eds.), Wastewaters,

20th ed., Washington.

Batstone, D.J., Keller, J., 2001. Variation of bulk properties of anaerobic

granules with wastewater type. Water Res. 35 (7), 1723–1729.

Brinkman, J., Hack, P.J.F.M., 1996. The paques anaerobic digestion

process: a feasible and flexible treatment for solid organic waste.

Encology 10, 31–36.

Datar, M.T., Dubey, D.R., Singh, A.N., 2001. Quick start-up and

quick steady-state techniques for UAFFB reactor. Ind. J. Environ.

Health 43 (3), 92–103.

DAH&D, 2003. Department of Animal Husbandry and Dairy,

Ministry of Agriculture, Government of India, information posted

on www.dairy.com.

Dilallo, R., Albertson, O.E., 1961. Estimation of volatile fatty acids by

direct titration. J. Water Pollut. Control Federat. 33, 556.

El-Mamouni, R., Guiot, S.R., Mercier, P., Safi, B., Samson, R., 1995.

Liming impact on granules activity of the multiphase anaerobic

reactor (MPAR) treatingwhey permeate. Bioprocess Eng. 12, 47–53.

El-Mamouni, R., Leduc, R., Guiot, S.R., 1998. Influence of synthetic

and natural polymers on anaerobic granulation. Water Sci.

Technol. 28 (8–9), 341–347.

Franklin, R.J., 2001. Full scale experience with anaerobic treatment

wastewater. Water Sci. Technol. 44 (8), 1–6.

Gupta, P.R., 1997. Dairy India, 5th ed., National Dairy Development

Board, Anand.

Lettinga, G., Pol, H.L.W., 1991. UASB-process design for various

types of wastewaters. Water Sci. Technol. 24, 87–107.

Lettinga, G., Pol, H.L.W., 1992. Economy of anaerobic wastewater

treatment. In: Anaerobic Reactor Technology. Wageningen Agri-

culture University, Wageningen, The Netherlands, pp. 56–63.

Poompavai, S., 2002. Treatment of different industry wastewaters.

MPhil thesis, Pondicherry University, Pondicherry, p. 103.

Quarmby, J., Forster, C.F., 1995. An examination of the structure of

UASB granules. Water Res. 29 (11), 2439–2452.

Ramasamy, E.V., 1997. Biowaste treatment with anaerobic reactors,

Ph.D. thesis, Pondicherry University, Pondicherry, p. 300.

Ramasamy, E.V., Abbasi, S.A., 2000. Energy recovery from dairy

wastewaters: impacts of biofilm support systems on anaerobic CST

reactors. Appl. Energy 65, 91–98.

Uemura, S., Takahashi, K., Takaishi, A., Machdar, I., Ohashi, A.,

Harada, H., 2002. Removal of indigenous coliphages and fecal

coliforms by a novel sewage treatment system consisting of UASB

and DHS units. Water Sci. Technol. 46 (11–12), 303–309.

Xu, H.L., Wang, J.Y., Zhang, H., Tay, J.H., 2003. A comparative

study of anaerobic digestion of food waste in a single pass, a

leachate recycle and coupled solid/liquid reactors. Water Sci.

Technol. 47 (1), 319–324.

Zehnder, A.J.N., Wuhrmann, K., 1977. Physiology of a Methanobac-

terium strain AZ. Arch. Microbiol. 111, 199–205.