Microbial community analysis of a UASB reactor and application of an evolutionary algorithm to enhance wastewater treatment and biogas production

Abstract

Anaerobic digestion, a proven and highly efficient biological process for treating industrial wastewater and biogas generation is an underutilized technology in South Africa. Some of the industries that have on-site anaerobic reactors tend to face problems in operating these reactors due to poor understanding of the process and implementation of the technology. This has resulted in high pollutant loads in their final effluents and low energy recovery. In this study, an on-site full–scale upflow anaerobic sludge blanket (UASB) reactor treating brewery wastewater was extensively monitored in order to evaluate the efficiency in terms of effluent quality, biogas production and microbial structure. Furthermore, developed and adopted kinetic models were used to optimize the performance of the full–scale UASB reactor using a combined Pareto differential evolution (CPMDE) algorithm. A preliminary analysis of the raw wastewater has shown that the wastewater produced from the brewery industry was high in organic matter with a total chemical oxygen demand (COD) between 1096.41 to 8926.08 mg/L. The average removal efficiency of COD from the UASB reactor after treatment was 79% with a methane (CH4) production of 60-69% at temperature ranges of 28-32˚C and hydraulic retention time (HRT) of 12 h within the optimal pH range for anaerobic bacteria (6.6 and 7.3) under various organic loading rates. However, the results also showed an increase in total suspended solids (TSS), nitrogen (N2), ammonia (NH3) and orthophosphate concentrations when comparing the influent to the effluent, which indicated the necessity for further optimization of the reactor condition in order to reduce these effluent parameters to acceptable standards and to increase CH4 production. In order to optimize the process, a thorough understanding of microbial interaction was essential. A combination of different molecular techniques viz., fluorescence in–situ hybridization (FISH), polymerase chain reaction (PCR) and quantitative real-time PCR (QPCR) were employed to understand the microbial community structure of the granular sludge samples using species specific primers and probes. The results revealed that the dominance of diverse groups of eubacteria belonging to phyla Proteobacteria, Firmicutes and Chloroflexi and an uncultured candidate division WS6 with four different orders of methanogenic Archaea viz., Methanomicrobiales, Methanococcales, Methanobacteriales and Methanosarcinales belonging to hydrogenotrophic and aceticlastic methanogens were within the reactor samples. Quantification of the 16S rDNA copies of eubacteria and methanogenic Archaea using species-specific primers further confirmed the spatial distribution of these microorganisms within the different compartments of the reactor where, the upper compartments were dominated by eubacteria and the lower compartments by methanogenic Archaea. The concentration of Archaea per nanogram of DNA was much higher (96.28%) than eubacteria (3.78%) in lower compartments, while, the eubacteria concentration increased to 98.34% in upper compartments with a decrease in Archaea quantity (1.66%). A modified kinetic methane generation model (MMGM) was developed on the basis of mass balance principles with respect to substrate (COD) degradation and the endogenous decay rate to predict CH4 production efficiency of the reactor. Furthermore, a Stover–Kincannon kinetic model was adopted with the aim of predicting the final effluent quality in terms of COD concentration and model coefficients were determined using the data collected from the full–scale reactor. Thereafter, a model-based multi-objective optimization was carried out using the CPMDE algorithm with three–objective functions namely; maximization of volumetric CH4 production rate; minimization of effluent substrate concentration and minimization of biomass washout, in order to increase the overall efficiency of the UASB reactor. Important decision variables and constraints related to the process were set for the optimization. A set of non-dominated solutions with high CH4 production rates of between 2.78 and 5.06 L CH4/g COD/day at low biomass washout concentrations were obtained at almost constant solution for the effluent COD concentration. A high COD removal efficiency (85-87%) at ~30-31˚C and 8-9 h HRT was obtained for the multi-objective optimization problem formulated. This study could significantly contribute towards optimization of a full–scale UASB reactor treating brewery wastewater for better effluent quality and biogas production. Knowledge on the activity and performance of microbial community present in the granular sludge taken from the full–scale UASB reactor would contribute significantly to future optimization strategies of the reactor. In addition, optimization using an evolutionary algorithm under different operational conditions would help to save both time and resources wasted in operating anaerobic bioreactors.