Winery wastewaters vary in quantity and quality during the annual cycle of winery activity. The disposal of winery effluent into rivers, streams and onto soil in South Africa is regulated by the National Water Act No. 36 of 1998. To meet the standards set, most winery wastewaters must undergo some form of pre-treatment before disposal. One method involves directing the effluent through a constructed wetland in which organic wastes are trapped by plants and decomposed by microbes. These microbial breakdown processes require copious amounts of oxygen. This oxygen may be taken either from the soil, in the case of land application, or water, where effluent is discharged directly into water bodies. Under oxygen deficient (anaerobic) conditions, decomposition is slow and noxious odours are produced. Anaerobic conditions are likely to arise where the rate of inflow is high and the organic contents of the wastewater are too concentrated. Constructed wetlands cease to be effective where the chemical oxygen demand (COD) of the wastewater exceeds 15 000 mg/ℓ (Mulidzi, 2005). COD is a measure of the amount of oxygen that is required to biologically oxidise a given quantity of residue in the wastewater. The higher the COD, the greater the risk to fish and other organisms in water, and the greater the likelihood that anaerobic, root inhibiting conditions will develop in soils. Clearly, high-COD effluents must undergo some form of pre-treatment before they enter a constructed wetland. A potentially viable method of pre-treatment involves passing the wastewater through containers filled with inert materials having very large surface areas. These materials form substrates on which the organisms that oxidise organic material become established, forming permanent coatings (biofilms). Such a device is known as a bioreactor.
This article reports on the effects on COD and turbidity of passing winery wastewaters of varying composition through aerated and non-aerated bioreactors.
Materials and methods
The bioreactor consisted of four, covered, 1 000 ℓ plastic tanks linked by a plastic pipe running from the base of one tank to the top of the next downstream tank (Figure 1A). Each tank was slightly lower than its predecessor to facilitate gravity flow, and was filled to 75% of its capacity with inert plastic spheroids moulded to create large internal surface areas (Bioballs) (Figure 1B & 1C). Commencing in May 2006, the tanks were filled through the opening in the top of Tank 1, from which effluent flowed into the successively lower tanks. The time period over which effluent remained in the reactor was regulated by adjusting the amount of effluent added to the top of Tank 1. The newly added wastewater was assumed to displace the effluent that already occupied it into the following tank, finally exiting from the outlet at the bottom of Tank 4. Since some mixing was inevitable, the time period for a hypothetical single drop of wastewater to pass through the reactor was regarded as the theoretical residence time (TRT). The wastewater used in this trial was obtained from a commercial winery in several bulk consignments (batches). These varied in composition. Additions of effluent were carried out at intervals and volumes sufficient to result in TRTs of four, eight and 12 days (820 ℓ/day for four days, 410 ℓ/day for eight days, and 270 ℓ/day for 12 days, respectively). Several batches of effluent were used for each TRT treatment. The effluents were tested to determine their COD and turbidity (a measure of clarity, or optical translucency) before entering Tank 1, and again at the outflow of Tank 4, one day after the final effluent addition of the treatment. Two bioreactors operated in parallel. One of these received no aeration. Ambient air was pumped into the tanks of the second reactor for periods of six, 15, 60 and 120 minutes per day.
FIGURE 1. Bioreactor, consisting of four 1 000 ℓ tanks. Arrows indicate the movement of wastewater. B. An individual Bioball. C. Bioballs on the surface of the wastewater in one of the tanks.
TABLE 1. Effect of theoretical retention time (TRT) on average reduction in COD and turbidity of winery wastewater flowing through a 4-tank reactor containing an inert plastic substrate having a large surface area.TABLE 2. Effect of influent wastewater quality (COD) on average reduction in COD and turbidity of winery wastewater flowing through a 4-tank reactor containing an inert plastic substrate having a large surface area. Residence time: 12 days.FIGURE 2. Reduction in COD and turbidity in 17 consecutive batches of winery wastewater where the rate of wastewater inflow into the bioreactor was adjusted to give a theoretical retention time of 12 days.FIGURE 3. Removal (reduction) of COD from wastewater in aerated and non-aerated bioreactors following aeration for six, 15, 60 and 120 minutes per day. Values are averages of multiple batches of wastewater and different retention times.
Results and discussion
Effects of wastewater quality and retention time
The effect of TRT was inconsistent (Table 1). Outflow COD was appreciably lower after eight than after four days TRT (49.3% and 8.3% reduction, respectively). Turbidity followed a similar pattern (82.6% and 49.3% decrease, respectively). For most batches within a given TRT treatment outflow COD and turbidity were similar. These values did not improve appreciably between the first batch of the trial (4-day TRT, reduction in COD, 6.5%; turbidity reduction, 51.5%) and seventh batch (COD reduction 7.8%; turbidity reduction 39.2%), implying that efficiency did not improve over successive batches. After 12 days TRT the decreases in COD and turbidity were lower than after eight days TRT. This anomaly stemmed from the fact that the effectiveness of the bioreactor increased as the COD and turbidity of the inflowing wastewater increased, i.e. with worsening wastewater quality (Table 2). The appreciably lower CODs observed after four compared with eight days TRT may therefore have been at least partly due to the fact that the average inflow CODs of the wastewaters used in the 4-day TRT trial runs (1 327 mg/ℓ) were lower than in the 8-day trial runs (19 974 mg/ℓ). It is also probable that the biofilms were less well established, and less effective during the 4-day trial runs, which were the first trial runs after the bioreactor was commissioned, than during the subsequent 8-day runs. A second factor which contributed to the poor performance in the 12-day, compared with the 8-day TRT was that, after some batches, both COD and turbidity were worse after, than before, treatment in the bioreactor (denoted by negative values in Table 2 and shown for batches 3 and 11 in Figure 2). This occurred most frequently where the influent quality was reasonably high (low CODs, i.e. <6 000 mg/ℓ, as in the 4-day trial runs and in batches 3, 5, 11 and 16 in the 12-day TRT trial runs), and where the bioreactor was refilled many times during a single trial run, as in the 12-day TRT treatments). These anomalies arose, because aggregates of organic material from the wastewater, and fragments of biofilm, periodically broke free from the substrate, probably when fresh wastewater was added to Tank 1. These aggregates were then flushed through the bioreactor in the wastewater flow. Most of these aggregates could probably have been removed if appropriate sieves had been fitted on the outflow side of each tank. The sieves would have needed to be cleaned frequently.
Effects of aeration on COD
Averaged over a range of TRTs and wastewater compositions, COD reduction was not consistently affected where aeration times were less than two hours per day (Figure 3). However, where aeration was carried out for two hours per day, COD removal averaged 78%, compared with 66% in the non-aerated reactor.
Conclusions and recommendations
Over the range of CODs tested (4 606 to 10 437 mg/ℓ), passage through a 4-tank bioreactor brought about a c. 50% reduction in COD, and a c. 80% reduction in turbidity after a retention time of eight days. Aerating the reactor for periods of two hours per day promoted bioreactor efficiency. These results suggest that bioreactors may be a practical way to reduce COD in winery wastewaters before they are discharged into wetlands, particularly if prolonged aeration is provided. The primary disadvantage of bioreactors is that aggregates of organic material periodically break free from the substrate and may exit the bioreactor. If adequate filtration were provided to remove these aggregates from the bioreactor it is likely that greater efficiencies would be achieved at retention times in excess of eight days, possibly permitting a continuous flow system to be designed and effectively operated.
The ability of a bioreactor to reduce the chemical oxygen demand (COD) and turbidity of winery wastewater was tested at theoretical retention times of four, eight and 12 days. At these intervals COD reduction averaged 9%, 51% and 16%, respectively. Associated decreases in turbidity were 48%, 87% and 56%. Effectiveness decreased with decreasing inflow wastewater quality and with frequency of refilling. Filtration is needed to remove planktonic fragments of organic material.
This project was funded by Winetech and by the Agricultural Research Council.
Mulidzi, A.R., 2005. Winetech Progress Report for 2004/2005. Project WW19/01: Investigation into the chemical composition of effluent from wine cellars and distilleries for the evaluation of artificial wetlands as a secondary treatment for purification. ARC Infruitec-Nietvoorbij, Stellenbosch, South Africa.
– For more information, contact Reckson Mulidzi at firstname.lastname@example.org.