Thermophilic semi-continuous anaerobic digestion of palm-oil mill effluent

Thermophilic semi-continuous anaerobic digestion of palm-oil mill effluent

Agricultural Wastes 13 (1985) 295-304 Thermophilic Semi-continuous Anaerobic Digestion of Palm-oil Mill Effluent R. G. Cail & J. P. B a r f o r d De...

446KB Sizes 10 Downloads 198 Views

Agricultural Wastes 13 (1985) 295-304

Thermophilic Semi-continuous Anaerobic Digestion of Palm-oil Mill Effluent

R. G. Cail & J. P. B a r f o r d Department of Chemical Engineering,The University of Sydney, NSW 2006, Australia

ABSTRACT Palm-oil mill effluent was degraded in a thermophilic, semi-continuous digester. The operation of semi-continuous digesters allows the accumulation and retention of high biomass concentrations, such that treatment rates approaching those of sludge blanket reactors can be obtained. At a temperature of 57°C space loadings up to 52kg COD m -3 day -1 were achieved (hydraulic residence time of 1.3 days). These rates are significantly faster than the rates achieved in previously reported thermophilic studies. At the maximum loading, total COD removal was > 70 % and soluble COD removal was greater than 97%. The relationship between treatment rate, sludge production and gas production is discussed.

INTRODUCTION The effluent resulting from palm oil production can cause serious pollution if left untreated. This is a problem of considerable magnitude in a number of South East Asian countries, notably Malaysia, where increasingly stringent discharge regulations have been applied since 1978. The extraction of palm oil from the fruit of Elaeis guineensis involves a number of processing stages in which the fruit is first sterilized, digested to a mash at near boiling temperatures, followed by screw pressing and clarification of the resultant oil. The effluent discharged from the plant is a complex waste containing residual oils and greases, polysaccharides and proteins, as major components, with a high Chemical Oxygen Demand (COD) ranging from 45 000 to 75 000 mg litre-1 (Peyton et al., 1979). 295 Agricultural Wastes 0141-4607/85/$03.30 © Elsevier Applied Science Publishers Ltd, England, 1985. Printed in Great Britain

296

R. G. Cail, J. P. Barford

Davis & Reilly (1980) reviewed a variety of waste treatment practices. It was evident that no one system was completely satisfactory. The use of anaerobic lagoons or conventional digesters for the treatment of palm-oil mill effluent (POME) has been described by Sinnappa (1978), Southworth (1979) and Peyton et al. (1979). While such systems can be reasonably efficient, they are characterised by long residence times (over 20 days) and thus require large areas of land or large digesters. Barford et al.'(1985)land Cail & Barford (1985) have applied recent developments in anaerobic digestion technology to develop a semi-continuous digester process capable of achieving very high treatment rates. Such a system is relatively simple to operate and could potentially be adapted to conventional stirred tank reactors, thereby markedly improving their performance and efficiency. The digesters are operated on a fill and draw basis, with minimum mixing of the reactor contents. This method of operation enables considerable accumulation of biomass in the digester with consequently improved treatment rates. Operated in this way, the digester approaches the operation of a sludge blanket (UASB) system with respect to accumulating and retaining biomass, thus providing a rather similar performance to that attainable by the UASB. In studies of the mesophilic, semi-continuous digestion of POME, Cail & Barford (1985) achieved space loadings in excess of 12kg COD m - 3 day -1 (hydraulic residence time of 5.6 days), with > 97 ~ removal of the soluble COD. These values were two to three times faster than previously reported mesophilic studies. Although these results represented a significant improvement over conventional systems, it was considered that further reductions in treatment times could be achieved using thermophilic digestion. It is generally recognised that thermophilic operation has the potential for faster bacterial growth--and consequently higher treatment rates. However, not all systems are suited to such an approach, due to the need to heat the digester to maintain the temperature at approximately 55°C. Since POME is discharged at 60-70 °C, only minimal additional heating is required. In this context, Peyton et al. (1979) have estimated that, for a digester situated in Malaysia (ambient temperature, 35°C) and constructed of concrete insulated by dry earth, approximately 11 ~o of the digester gas produced would be required for digester heating. Such a requirement may be more than counterbalanced if significant improvements in treatment rates are realised. Peyton et al. also described the application of a 2700-1itre (600gallon) contact stabilization digester with solids recycle, operated at

Digestion of palm-oil mill effluent

297

55 °C. Stable digestion was reported over a period of 2 weeks at 8 days hydraulic retention time with approximately 73 ~ removal of the total COD (estimated from their data). Operation at hydraulic residence times as low as 5 days was reported but no other performance data was available. In addition to thermophilic operation, it was considered that the application of sludge blanket technology to the thermophilic digestion of P O M E might enable further rate increases to be achieved. Although little attention has been paid to the development of bacterial granules at high temperatures, Bochem et al. (1982) reported the development of methanogenic bacterial granules of up to 3 mm diameter in a thermophilic culture grown anaerobically on a defined medium consisting of mineral salts, vitamins and acetic acid. This investigation was undertaken to assess the potential of the thermophilic digestion of P O M E using a high rate semi-continuous digester.

METHODS

Equipment The digestion was carried out at 57 °C in a 2-1itre semi-continuous reactor, shown in Fig. I. By feeding the digester intermittently in accordance with the 3-h cycle shown in Fig. 2, it was possible to achieve much greater biomass accumulation than would be possible in a continuously stirred reactor. The stirrer speed was limited to 40-50 rpm and it was used only for a short period of time after feeding, so as to get good substrate to organisms contact. Digester pH proved to be stable at 7.2-7.6 (averaging 7.5) and did not require adjustment of the feed. Although provision was made for flocculant (polyelectrolyte) dosage, it was found that this was not necessary and that good settling was achieved in the 2.5 h or so that the digester stirrer was off. At the end of the settling period feed was pumped to the digester and clarified effluent was simultaneously drawn off from the surface.

Analytical Volatile Fatty Acid (VFA) concentrations were measured daily by gas chromatography (Holdeman & Moore, 1975). Gas production was

298

R. G. Cail, J. P. Barford

ST 20/2 Stirring 'Gland Sealed With Water

MF18 Socket Cone Adapter With "T'~. % Connection Feed |



Gas 31aS

Surface Rake

Quickfit 2L Round B o t t o m Flask

Peddle BladeStirrer

Fig. 1. Two-litre semi-continuous digester.

18Drain

J Om,n "i

1Omin

I'- 2Drain

/-~ Feed M i x e r F e e d / E f f l u e n t Pump Digester Mixer Flocculant Pump

i"- 1rain

I

JJ~ on off

Fig. 2.

Semi-continuous operation timing sequence.

Digestion of palm-oil mill effluent

299

monitored by wet gas meter and the gas composition was determined using gas chromatography on a Poropak N column. The COD, total Kjeldahl nitrogen and Volatile Suspended Solids (VSS) were estimated by Standard Methods (1975). The palm-oil mill effluent (POME) was digested in aqua regia, and its elemental composition analysed by ICP emission spectrophotometry. Soluble COD removal efficiencies were determined on the supernatant of samples centrifuged at 10 000 rpm for 10min in a Sorval RC2-B. In order to estimate the improvement in digester effluent quality that might be achieved if an additional external clarifier were used, a settled COD value was determined on the supernatant of a sample of the total digester effluent which was let stand for I h in a 100-ml measuring cylinder.

Palm-oil mill effluent (POME) The effluent was obtained from a commercial mill. It was thoroughly mixed and dispensed into 20-1itre drums and frozen to prevent deterioration. Subsamples were thawed and stored at 4°C prior to use each day. The composition of the effluent (per litre) was: 70 g COD, 32-8 g SS, 28.8g VSS, 920mg Kjeldahl N, 171 mg P, 482mg Fe, 296mg S, 506 mg Ca, 17 mg Na, 1260 mg K, 446 mg Mg, 8 mg Cu, 18 mg Zn, 0-8 mg Mo, 0-05 mg Co, 3 mg Mn, 3 mg Ni, 229 mg A1, 5 mg B, 1 mg Ba and 131 mg Si. As such, the strength of the effluent is near the top of the range of literature values.

Seed sludge The sludge was obtained from a mesophilic anaerobic lagoon treating POME.

RESULTS

Space loading The digester was initially fed once daily to allow the mesophilic seed sludge to acclimatise to the higher temperature. By day 13, the Volatile Fatty Acid (VFA) concentrations had fallen to zero and the space loading

300

R. G. Cail, J. P. Barford

was thus increased rapidly up to a loading of 11-7 kg C O D m - 3 d a y - 1 by day 72, with the VFA levels remaining low, at about 100 mg litre-1 Automatic semi-continuous feeding every 3 h was commenced at this stage. Loading rates were increased in steps of 1-2 kg C O D m - 3 d a y - 1, depending on the VFA concentrations. VFA levels were not allowed to rise above 500-600 mg litre- 1 and were controlled by either keeping the loading rate steady or reducing it slightly until the concentrations had fallen to below 100 mg litre-1. It can be seen from Fig. 3 that a space loading of 52.4 kg C O D m - 3 d a y - 1 (hydraulic residence time of 1.3 days) was attained within 180 days of start up. Even this rate, which is four times greater than previously reported thermophilic rates, appeared to be non-maximal, as VFA concentrations were still quite low at < 100 mg litre-1. The experiment was ceased when supplies of P O M E ran out. 55 50



45 40

35 E

8 30 U

"~ 25 d0

g

20

03

15 10

5

oo

I 50

I 100 Time

Fig. 3.

I 150

i

200

(day)

The performance of a semi-continuous digester treating palm-oil mill effluent at 57°C.

Digestion of palm-oil mill effluent

301

COD conversion efficiency The C O D removal efficiencies are shown in Fig. 4. Soluble C O D removals were relatively constant at 94-97 ~ at loadings up to 47 kg C O D m - 3 day - 1. However, total C O D removal, based on mass balance studies, decreased from approximately 76 ~o at a loading of 23 kg C O D m -3 d a y - 1 to around 62 ~o at 41 kg C O D m - 3 d a y - 1. As a result of these lower conversions, higher quantities of undigested material were discharged from the digester. In a full-scale operation this material could be readily removed in an external clarifier, as indicated by the high settled C O D removal etficiencies of 90-93 ~ , even at space loadings up to 47 kg C O D m - 3 d a y - 1. 100

~

Soluble Settled

w

Total >

o

E 50 c~ 0 (3

0 0

I 10 Space

Fig. 4.

I 20 Load

I 30

I 40

I 50

(kgCOD/m3. day)

The effect of treatment rate on effluent COD removal efficiencies during the thermophilic digestion of POME.

Gas production The methane concentration of the evolved biogas decreased from 68-69 ~o at space loadings below 10 kg C O D m - 3 d a y - 1 to 60 ~ at space loadings of 40-50kg C O D m - 3 day-1. C O D conversion to methane ranged from 346 ml CH 4 per gram of C O D applied, at a space loading of 9.7kg C O D m -3 day -1, down to 214ml CH 4 at a space loading of 51.9 kg C O D m - 3 d a y - 1. This reduction in C O D conversion efficiency is the result of the slow microbiological degradation of the more recalcitrant

302

R. G. Cail, J. P. Barjord

compounds, such as cellulose and fibre, present in POME compared with the more rapid degradation (particularly in modern high-rate digesters) of the soluble fraction. Hence, any digester system which significantly reduces the retention time (e.g. UASB, Upflow Floc, semi-continuous, fluidized bed or filter) achieves a more rapid soluble waste treatment rate but also has an increased undigested insoluble waste component. Thus, while treatment rates increase overall, treatment efficiencies decrease. Microscopic examination of the digester effluent verified this, with considerable quantities of partially degraded plant cell material being present. This material settles very rapidly and could be easily removed in an external clarifier before final effluent discharge.

Sludge discharge The digester VSS concentration varied between 39 and 57 g VSS per litre, averaging about 48 g litre- 1. The discharge of VSS increased from 0.15 g VSS per g COD applied, at a space loading of 33.1, up to 0.20 g VSS per g COD applied at a loading of 41-1 kg COD m -3 day -1. As discussed previously, at the higher loadings the shorter solids retention times result in incomplete assimilation of the plant tissues and other difficult to degrade substances. As a result, the reduction in solids is less and higher quantities of sludge are discharged.

Sludge loading One measure of the bacterial activity of the sludge is the sludge loading rate (kg COD applied per kg VSS in digester-day). At steady state, the higher the value, the more active is the biomass. In this investigation, at a space loading of 33.1 kg COD m - 3 day- 1 (day 122) the sludge loading was 0.73 kg COD per kg VSS per day. By day 150, at a space loading of 41.1 kg COD m-3 day-1, the sludge loading had increased to 0.94 kg COD per kg VSS per day. Since the digester VSS was the same and VFA concentrations were still relatively low at the higher space loading, the biomass would appear to be still adapting to the conditions. Pol et al. (1982) have reported sludge loadings as high as 2.2 kg COD per kg VSS per day on acetate/propionate feeds and 1.0-1.4 kg COD per kg VSS per day for sugar beet and potato-processing wastes. Consequently, further improvements in the rates reported here may be possible.

Digestion of palm-oil mill eJfluent

303

DISCUSSION A semi-continuous digester was used for the thermophilic digestion of palm-oil mill effluent (POME). By operating the digester in a manner which encouraged biomass retention and accumulation, high treatment rates were obtained. Space loading rates over 52 kg COD m - 3 day - ~were reached. This rate is more than four times faster than that achieved by Cail & Barford (1985) using a similar digester operated at 35 °C, and it is at least four times faster than the thermophilic system reported by Peyton et al. (1979). Although the soluble COD removal was relatively constant at about 97 ~o over a range of loading rates from 20 to 55 kg COD m - 3 day- 1, total COD removal decreased from 74 ~o at 33 kg COD m - 3 d a y - ~to approximately 62 ~ at a loading of 41 kg COD m - 3 day- 1. Although no pelletization and little flocculation occurred in the short duration of the experiment, the sludge settled rapidly, as evidenced by the high digester VSS and the settled COD removal rate ( > 90 ~o). Further studies on flocculation development and, in particular, pellet formation, are required in order to ascertain the environmental conditions and time involved. The quantity of sludge discharged increased, as the loading rate was increased (i.e. at shorter HRT), due to the incomplete digestion of significant portions of plant cells and other debris. This was also reflected in a decrease in the amount of gas produced per kilogram of COD applied. Any economic evaluation of this system would need to take into account this decrease in gas yield and the need for supplementary digester heating and balance these against the very significant rate increases which would markedly reduce the required digester size and hence capital costs. A further consideration, not investigated here but observed by Peyton et al. (1979), was that, although more COD was removed in a thermophilic system (8 day HRT) than in a mesophilic digester (20 day HRT), more soluble BOD remained in the settled effluent. They assumed that this was due to the increased liberation of cell components, such as oils, which were not degraded at the shorter HRT. However, it would appear to be unlikely, in this instance, as soluble COD removal efficiencies (which would reflect soluble BOD concentrations) remained very high. Longer term running at high loading rates and larger scale is now required in order to assess the stability and the potential of the system described in this paper.

304

R. G. Cail, J. P. BarJord

ACKNOWLEDGEMENTS The assistance of Mr E. Floyd is gratefully recognised. The financial assistance of Bizorba-Sanamatic is acknowledged.

REFERENCES Barford, J. P., Cail, R. G., Callander, I. J. & Floyd, E. J. (1985). Anaerobic digestion of high strength cheese whey utilizing semi-continuous digesters and chemical flocculant addition. Biotech. Bioeng. (submitted). Bochem, H. P., Schoberth, S. M., Sprey, B. & Wengler, P. (1982). Thermophilic biomethanation of acetic acid: Morphology and ultrastructure of a granular consortium. Can. J. Microbiol. 28, 500-10. Cail, R. G. & Barford, J. P. (1985). Mesophilic semi-continuous anaerobic digestion of palm oil effluent. Biomass, 7 (1985), 287 95. Davis, J. B. & Reilly, P. J. A. (1980). Palm oil mill effluent--A summary of treatment methods. Oleagineaux, 35(6), 323-9. Holdeman, L. V. & Moore, W. E. C. (1975). Anaerobe Laboratory Manual. (3rd edn), Virginia Polytechnic Institute and State University, Blacksburg, Virginia, USA, I 13 pp. Peyton, T. O., Cooper, I. W. & Quah, S. K. (1979). Mesophilic and thermophilic anaerobic tank treatment of palm oil mill waste waters. Proc. 34th Indust. Waste ConJl Purdue Univ., Lajayette, Indiana, USA, 473-82. Pol, L. H., Dolfing, J., de Zeeuw, W. & Lettinga, G. (1982). Cultivation of well adapted pelletized methanogenic sludge. Bioteeh. Letters, 4(5), 329-32. Sinnappa, S. (1978). Treatment studies of palm oil mill waste effluent. Int. ConJl on Water Poll. Control in Devel. Countries, Bangkok, Thailand, February, 1978, 525-37. Southworth, A. (1979). Palm oil factory effluent treatment by anaerobic digestion in lagoons. Proc. 34th lndust. Waste Conjl, Purdue Univ., Lafayette, Indiana, USA, 421-34. Standard Methods Jor the Examination of Water and Wastewater (1975). (14th edn.), American Public Health Association.