TREATMENT OF AGRO-FOOD INDUSTRIAL WASTEWATERS USING UAF AND HYBRID UASB-UAF REACTORS

Abstract

Objective: To evaluate the performance of upflow anaerobic filters (UAFs) and to develop a hybrid upflow anaerobic sludge blanket-upflow anaerobic filter (UASBUAF) reactor for the treatment of wastewaters from agro-food industries. Scope of the work: (i) The performance of UAFs filled with low density polyethylene supports for the treatment of soluble to complex synthetic wastewaters representing effluents from agro-food industries was evaluated. The variables studied were (a) size of filter media (Bioflow® 9, 30 and 40) and (b) nature and strength of wastewaters. The packing media viz. Bioflow® 9, 30 and 40 are grooved and baffled with specific surface area of800, 320, 305 m2/m3 respectively. (ii) The performance of hybrid UASB-UAF reactors for the treatment of high strength wine distillery wastewater was assessed. In this, proportion of reactor volume occupied by UAF and UASB was varied. Justification: Clogging is one of the major problems in continuous operation of packed upflow anaerobic filters (Jung and Choi, 1995; Escudie et al, 2005). Such a problem can be overcome by using UAF filled with low-density floating media. UASB reactors have sludge bed at the bottom (reaction zone) and settler at the top. Between bed and settler there is sludge blanket to attenuate sludge loading to the settler. The filter media in the blanket and settling zone may improve the sludge retention in the reactor. Therefore, the efficiency of UAF has been evaluated. UAF has then been combined with UASB to form a hybrid UASB-UAF reactor. The performance of hybrid UASB-UAF reactors wasevaluated. Methodology and Results: With this orientation, experiments were performed in three phases. In Phase- 1, the effect of the size of filter media namely Bioflow® 9, 30 and 40 on the performance of UAF was evaluated in three similar bench-scale reactors Rn, R12 and R)3 (each of 10 L capacity) respectively using winery wastewater of COD 20.2±0.9 111 g/L. These reactors equipped with continuous internal recirculation system were operated for a total period of 180 days at 33+l°C. Reactors were initially fed at an organic loading rate (OLR) of around 0.4 g COD/L.d and subsequently flow rate was increased to increase the OLR. The maximum OLR values in Rn, R12 and R13 corresponding to 80% soluble COD (CODs) removal were 42, 27 and 22 g COD/L.d respectively. The total COD (CODt) removal at these loadings was around 75%. The fraction of dead space assessed from the tracer study at the end of the experiments was 18, 12 and 11% for Rn, R12, and R13 respectively. The performance of the reactor Rn with Bioflow® 9 was best among the three reactors. However, due to practical difficulties such as escape of supports from the reactor, clogging of tubes and damage of pumps etc. associated with the small size of Bioflow® 9 (9 mm diameter x 7 mm height), Bioflow® 30 was subsequently used in the Phase- 2 and 3. Experiments in Phase-2 were designed to evaluate the feasibility of treatment of different agro-food wastewaters in UAF packed with Bioflow® 30. Three benchscale reactors R21, R22 and R23 (each of 10 L capacity) were fed with synthetic wastewaters of COD 1.0± 0.1 and 1.9± 0.2 (R2i), 9.0-11.6 (R22), 23-40 (R23) g/L respectively. The wastewater of COD 9.0-11.6 and 23-40 g/L representing highstrength fruit canning and complexcheese dairy wastewaters respectively were treated in R22 and R23. The other wastewaters referred as low and medium strength wastewater were treated in R2i. The UAF reactors R2i, R22 and R23 were operated for a total period of 215, 145 and 197 days respectively at 33±1°C. Similar to Phase-1, all the reactors were initially fed at an OLR of around 0.32-1.0 g COD/L.d and subsequently flow rate was increased to increase the OLR. UAFs with low-density polyethylene supports were efficient in the treatment of low to high strength agro-food industrial wastewaters. At the end, OLRs of 12 for R21, 19 for R22 and 17 g COD/L.d for R23 corresponding to 91, 81 and 80% COD removals were attained. The temperature of the reactor R21 fed with wastewater of COD 1.9 g/L was varied from 33±1 to 14.8±0.1°C (days 174 to 215). The COD removal efficiency of 84% was achieved at OLR of 6.6 g COD/L.d, HRT of 6.8 h and temperature of 14.8°C. IV The dead space assessed from the tracer studies at the end of the experiments was 68, 35 and 14% for R2i, R22, and R23 respectively. Reactor R2i from day 47 onwards was operated without internal recirculation which probably led to the development of dead space as high as 68%. Reactors R22 and R23 were operated with recirculation throughout. The observations suggest that UAF filled with Bioflow® 30 can be used for the treatment of low to high strength soluble and complex wastewaters. In Phase-3, a combination of UASB and UAF referred as hybrid reactor was studied. Three bench-scale hybridUASB-UAF reactors R31, R32 and R33 (each of-9.8 L capacity) were operated simultaneously for the treatment of wine distillery vinasse of COD ranging from 3.1 to 21.7 g/L. The UAF compartment of the UASB-UAF reactors were filled with 90, 68 and 40 numbers of Bioflow® 30 media to occupy 50, 35 and 20%of dry volume of the reactors R31, R32 and R33, respectively. The reactors equipped with continuous internal recirculation system were operated for a total period of 232 days at 33±1°C. The loading to these reactors (i.e. OLR) was increased to 19.5-19.9 g COD/L.d by increasing the feed COD at HRT of about 1.05 d. The COD removal efficiencies were 86-88%. Hydrodynamics of the reactor was assessed at day 0, 74, 123, 138, 165 and 223. The efficiencies of liquid mixing were good throughout the study with dead volume space of around 5% or less. The reactors were very close to the theoretical continuous stirred tank reactor (CSTR). The anaerobic digestion model No.l (ADM110) simulated well the dynamic evolutions ofthe main variables in the liquid as well as in the gas phases of a hybrid UASB-UAF reactor. There were no significant differences observed between the performances of the three reactors. The reactor R33 (20% UAF and 80% UASB) can be recommended in order to minimize the cost of packing media used in the UASB-UAF reactor. In addition to COD removal and tracer studies, following parameters have also been recorded during the reactor operations: (i) Performance of the reactors in terms of effluent VFA-COD and non VFACOD, alkalinity, VFA/alkalinity ratio, suspended solids washout, methane production etc; (ii) Quantification of biomass; (iii) Diversity of Archaea and Bacteria using polymerase chain reaction (PCR)- single strand conformation polymorphism (SSCP) molecular techniques (For Phase1 and 2); (iv) Kinetic model application for predicting the performance of the UAFs (For Phasel and 2); (v) Microscopic and granule size analysis (For Phase-3); and (vi) Application ofADM110 for simulating the experimental results (For Phase-3). Conclusions: This study reveals that the UAF and hybrid UASB-UAF reactors with lowdensity polyethylene supports are efficient in the treatment of low to high strength agro-food industrial wastewaters. The packingmediumhad a dual role in the retention of the biomass inside the reactor: i.e. entrapment of biomass within the support and filtration of the flocculent biomass, preventing it from going out of the reactor. Problems such as clogging and channelling were not significantly noticed. Thanikal et al (2007) described the unclogging procedures with fluidization by gas re-circulation or using liquid, which can be applied whenever clogging occurs. The effluent COD values were low with respect to the influent concentrations but not enough to approach effluent standards. It is therefore necessary to provide a post-treatment before effluent discharge or investigate probable reuse of the treated effluent. Thesis has been organized into five Chapters: Chapter-1 deals with the identification of the problem. The current state of research on topics of relevance to the work of the thesis has been compiled in Chapter-2. Chapter-3 deals with materials, methods and experimental protocols adopted. Results organized according to the sequence of experiments are discussed in Chapter-4. The conclusions and limitations of the work have been summarized in Chapter-5. References citied have been compiled in a separate section at the end. vi