Study and optimization of sequencing batch reactors for the biological nitrogen removal from anaerobic digestate of swine slurry.

Livestock sector is rapidly changing in response to pressures from globalization and growing demand for meat products, especially in developing countries. This implies a world-wide increase of livestock production and critical manure management. Livestock manure is a valuable source of nutrients for crops and can improve soil productivity, but when nutrients content exceeds the amount that can be used for land fertilization it is necessary to treat the produced manure and manage nutrient surplus load. Besides, ammonia emission in atmosphere from zootechnical farms is an emergent environmental problem that will be taken into account in next years. Livestock effluents represent one of the most significant contributions to nitrogen sources in Europe and, in 1991, the Commission of the European Communities published the Nitrates Directive to regulate the land application of livestock effluents and establish the Nitrate Vulnerable Zones –NVZ. Anaerobic digestion (AD) process is commonly used to treat swine slurries and it is considered a successful and convenient treatment, particularly from the energetic and economic point of view, but this process causes increase in ammonia nitrogen content in the produced digestate. In order to deal with these issues, research is focused on finding and improving advanced technologies to minimize the environmental impacts of digestates. Among the different available techniques, biological nitrogen removal is particularly reliable and diffuse in wastewater treatment; for this reason, in the proposed thesis we chose to put in practice a biological process using a Sequencing Batch Reactor (SBR), that is an alternative technology to conventional activated sludge systems. The SBRs have a higher flexibility and controllability, allowing for more rapid adjustment to changing wastewater characteristics. Lower investment and recurrent cost are necessary because secondary settling tanks and sludge return systems are not required. In this scenario, this thesis analyzes the comparison between a conventional nitrification-denitrification process and an innovative partial nitrification-denitritation treatment in a bench-scale SBR, for management of anaerobic supernatant of swine slurry. Special attention was given to biological nitrogen removal, in order to reduce an influent ammonia concentration of 1.2±0.3 gN/L and a high conductivity content (about 15 mS/cm), due to the elevate salts presence as dissolved solids (average concentrations: 3.5 ± 0.8 g/L, more than 50% of total solids content). The conventional process was conducted in a 10L-glass SBR with dissolved oxygen (DO) concentration of 3 mgO2/L, in two experimental runs with volumetric NLR (Nitrogen Loading Rate) increase, 0.19 and 0.28 KgN/m3d, respectively. The N-NOX reduction was provided by acetic acid addition, since the low biodegradability of influent COD from the digestate, keeping the influent COD/N ratio about 4.5. The vNLR increase determined a biomass growth in terms of suspended solids amount in the mixed liquor of about 40% of MLSS and about 50% of MLVSS, but the specific NLR remained constant for both periods. The process we obtained was a conventional nitrification/denitrification reaction, with total oxidation of ammonia nitrogen to nitrate, as demonstrated by cycle analyses measuring N forms and online signal trends. From these cycle analyses, the sludge kinetic constants were calculated: the sAUR (specific Ammonia Utilization Rate) was 6.5 mgN-NH4 gVSS-1 h-1, while the sNUR (specific Nitrogen Uptake Rate) measured 7.5 mgN-NO3 gVSS-1 h-1, demonstrating that the active biomass had got good proprieties of removing completely ammonia nitrogen in nitrate form and reducing it in the anoxic phase. The mass balances demonstrated the good results of the process in terms of removal efficiencies of nutrients: the averaged COD removal efficiency was 40% and for phosphorous was 25%; as regards nitrogen removal, the efficiency was very high up to values of 100 % for the denitrified N and >98% for nitrified N. Besides, this thesis describes the study and the development of a real-time control system using indirect parameters that are directly connected to biological reaction in the SBR: DO, oxidation reduction potential (ORP), pH and conductivity. By monitoring the control parameters, it is possible to determine the termination points of the nitrification/denitrification reaction. As regards the SBR studied, the ORP profile was more suitable for the anoxic phase control, with the nitrate knee indicator that is the point where the nitrates end in the denitrification reaction. This indicator point was found in 75% of analyzed cycles. In the oxic phase, the ammonia termination was detected by the “ammonia valley” in pH trends and “DO elbow” in DO profiles. The appearance frequency of these indicators was 75%. An interesting result was given by the conductivity signal: in 90% of analyzed cycles we assisted to a variation in conductivity trend and we calculated the incremental ratio ΔCond/ΔN-NH4+: it was about 17 mScm-1/(gN-NH4+L-1)in the first part of oxic phase, while in the second part it was about 9.5 mScm-1/(gN-NH4+L-1) and approximately zero at the end of nitrification. In the second part of the thesis, the achievement of an advanced process for nitrogen removal is described. In a microaerophilic environment (DO=0.3 mgO2/L) partial nitrification (PN) occurred in a 25L-plexiglass SBR, due to AOB (Ammonia Oxidizer Bacteria) higher affinity for oxygen than NOB (Nitrite Oxidizer Bacteria), that are inhibited at low oxygen concentration. In this way, the ammonia oxidation process stopped to N-NO2 with air supply saving and external carbon reduction for denitritation. During the start-up of the process we assisted to sludge granulation together with floccular fraction: the hybrid process maximize the hold-up of AOB operating at low DO concentrations and inhibiting nitrite oxidation. After the SBR start-up, the reactor was conducted increasing the vNLR from 0.01 to 0.08 kgN/m3d, in six experimental runs to evaluate the process performance. The process was conducted with dosage of acetic acid as external carbon source. The PN was performed for more than 300 days performing cycle analyses to measure the biomass kinetics and evaluate the online signals trends. At lower DO concentrations, the measured sAUR values were weak, with an average value of 1.5 mgN-NH4 gVSS-1 h-1, while the denitritation process was guaranteed by an average COD/N ratio with acetic acid of 3.5 and gave high sNUR values, up to 28±3.7 mgN-NO2 gVSS-1 h-1. With specific reference to nutrients removal performances, the averaged COD removal efficiency was 40% and for phosphorous was 50%; as regards nitrogen removal, the efficiency was very high up to values of 97% for nitrified and denitrified N. The nitrite concentration represented up to 98% of the N-NOX formed within the process, while in one run we assisted to the reactivation of NOB activity due to long aeration time and a decrease of influent ammonia concentration. This confirmed that NOB were only inactivated by operational conditions and not washed-out. Also in PN process, the online signals were analyzed to find process control indicators: the nitrite knee in the ORP profile had 70% of positives during anoxic phase; in microaerobic phase DO increase and ammonia valley from pH signals had appearance frequencies of 70% and 60%, respectively. The incremental ratio ΔCond/ΔN-NH4+ measured was 3.6±1.6 mScm-1/(gN-NH4+L-1). Finally, the molecular biology analyses performed (PCR-DGGE and FISH) demonstrated that, in piggery digestate, anaerobic digestion bacteria and sterols oxidizer microorganisms were present; both in conventional and advanced process the denitrifying biomass was predominant and, in PN bacterial community, AOB bacteria were present and active.