Sand based system for physiochemical and biological treatment of winery wastewater

Abstract

In South Africa, many wineries use cellar effluent for irrigation after primary settling and pH adjustment. This poses an environmental risk as winery effluent can contain high concentrations of organic and inorganic pollutants. Most vine growing countries, including South Africa, are water-stressed, so it is important that winery effluent is remediated to the extent that it is safe for reuse as irrigation water. In order to align the South African wine industry with global views to achieve sustainable development, there is a need for low maintenance, low cost and effective solutions for treating winery wastewater which can be easily integrated into the existing infrastructure. This body of work presents the results of a number of fundamental and applied studies used to determine the validity of biosand reactors as a potential solution to fulfil this winery wastewater treatment function in small to medium sized cellars. In the first study, chapter 3, a pilot biosand reactor with a nodal design was installed, operated and monitored for 2 years at a medium sized winery (crushing 1600 tons of grapes per annum) in the Stellenbosch area, Western Cape. Two reactors were operated in alternating continuous and pulse modes, with COD removal efficiencies of 70% achieved in each mode. Higher hydraulic and organic loading rates were achieved in continuous mode (113 L.m-3sand.day-1 and 279 gCOD.m-3sand.day-1, respectively) than in pulse mode (90 L.m-3sand.day-1 and 192 gCOD.m- 3sand.day-1). In comparison to other passive systems treating winery effluent, the biosand reactor system achieved significantly higher loading rates with a smaller spatial footprint. This applied study was followed by three fundamental studies to elucidate the biodegradation, neutralization and hydraulic mechanisms involved in remediating winery wastewater in biosand reactors. These studies were conducted using a series of columns containing raw sand and various sized fractions of sand from the Philippi quarry site in Cape Town. Biosand reactors have been shown to effectively neutralize winery wastewater, and it was assumed that the major neutralization was via dissolution of calcite sand particles. The first flowthrough column experiment was conducted to determine whether there may be a biotic contribution to pH adjustment. Sand cores containing functional microbial communities were extracted from an operational biosand reactor system. One half of the cores were sterilized by irradiation and used to determine the abiotic contribution, while the biotic contribution was calculated by deducting the results obtained from the irradiated columns from those obtained from the non-irradiated columns. The cores were dosed with synthetic winery wastewater and hydrochloric acid at the same pH. Using the hydroxide ion concentration in the effluent as a proxy, it was found that the major neutralization mechanism was abiotic (average 95.5±0.16%) due to the dissolution of calcite, with a small biotic contribution (average 4.5±0.13%). The second series of column experiments was conducted in order investigate calcite dissolution kinetics and the effect of calcite dissolution on the hydraulic conductivity of the Philippi sand. The results were used in conjunction with those obtained from an operational pilot biosand reactor system to determine the temporal abiotic neutralization capacity of biosand reactors. Flow-through experiments using hydrochloric acid were conducted using columns containing sand with a variety of particles sizes. The larger particles (>0.425 mm) contained lower amounts of calcite (using Ca as a proxy), but exhibited higher hydraulic conductivities before and after calcite dissolution (3.0 ± 0.05 %Ca and 2.57 to 2.75 mm·s−1, respectively) compared to the fractions containing smaller particles and/or raw sand (4.8 ± 0.04 to 6.8 ± 0.03 %Ca and 0.19 to 1.25 mm·s−1, respectively). By measuring the amount of Ca removed from the sand as a proxy for calcite dissolution, a temporal abiotic neutralization capacity of 37 years was calculated for biological sand reactor systems containing Philippi sand with 5.4 % wt.wt Ca and an influent with a pH ranging from 2 to 3 and a hydraulic loading rate of 150 L·m−3 of sand.d−1. Based on the promising results obtained with the larger sand particles in the second series of columns experiments, a third set of column experiments was used to compare the raw (unfractionated) sand to fractionated sand (<0.425 mm removed) in terms of operational performance. The flow in biosand reactors slows down after start-up due to the growth of functional biomass within the sand pores that decreases the hydraulic conductivity. In this study, initial hydraulic conductivities of 0.285 mm.s-1 and 2.504 mm.s-1, were measured in the columns containing raw and fractionated sand, respectively. After operating the columns by feeding with winery effluent for 3 months, the hydraulic conductivities reduced to 0.129 mm.s-1 and 1.116 mm mm.s-1, respectively. Similar organic removal efficiencies were obtained with raw (94%) and fractionated (95%) sand, and both sands were able to effectively neutralize the acidic (pH 4.9) winery wastewater. It was calculated that one 5.6 m3 biosand reactor containing the fractionated sand can potentially treat 8102 L.d-1 of winery wastewater, which surpasses the design flow rate of raw sand of 1000 L.d-1. During the applied pilot study, it was noted that copious amounts of primary wastewater sludge is generated at the study site. Currently, the winery contacts a commercial company to remove this waste to landfill which is an economic and environmental burden. Biochemical methane potential tests were used to determine the potential for valorisation of this organic-rich wastestream via anaerobic digestion. The highest specific methane yield of sludge harvested during the crush season (206 ± 2.7 mLCH4/gVSadded) was obtained under mesophilic (37°C) conditions. The composition of the digestate compared favorably with commercial organic agricultural fertilizers. Finally, building on the results obtained, a zero-waste model is presented for treatment of winery effluent based on the use of biosand filters. This includes: (i) using fractionated sand in the filters and the sifted smaller sand particles in the concrete industry, (ii) remediating the winery wastewater in biosand filters, and using the treated effluent for safe irrigation, (iii) valorizing the primary wastewater sludge via anaerobic digestion and using the digestate as an agricultural fertilizer.