Energy optimization of biological wastewater treatment using rotating belt filters upstream BNR

Abstract

Treatment of wastewater to acceptable permit standards requires energy, mostly as electricity. The typical energy demand for various wastewater treatment technologies vary from 0.30 to 1.50 kWh/m3 of treated water. For conventional activated sludge, the energy demand is 0.30 – 0.70 kWh/m3, in which 50 - 60 % is used for the aeration of the aerobic reactors. As a high fraction of wastewater, COD appear as total suspended solids (TSS), primary treatment has an impact on the performance and energy demand of the downstream processes. Consequently, efficient TSS removal during primary treatment will result in reduced organic load and a reduction in oxygen demand in the downstream biological treatment and resulting in significant energy cost savings. In addition, enhanced primary treatment generates more sludge suitable for anaerobic digestion and corresponding biogas and energy production. The goal of this research was to define the particle size cut-off for TSS and particulate COD removal prior to biological nitrogen removal. The main question would be how much TSS and associated COD removal is acceptable in order to maintain sufficient nitrogen removal and to maximize biogas production. Laboratory and pilot experiments performed at the laboratory of Aquateam in Oslo and at Nordre Follo WWTP (NFR) near Oslo, using wastewater and sludge from NFR and from Bekkelaget WWTP (BRA) in Oslo. Anoxic batch tests were with both activated sludge and biofilm processes in laboratory scale sequencing batch reactors (SBRs) in order to determine the impact of TSS on denitrification rates. Filtration of wastewater upstream SBR was by several fine mesh sieves, from 150 μm to 1.2 μm pores. The TSS and COD removal were inversely proportional to the filter pores. COD removal was from 43 % with 18 μm and 21 % with 150 μm sieves for NFR wastewater. For BRA wastewater the removal was from 42 % with 18 μm to 32 % with 90 μm sieves. By analyzing the slope of the curve for nitrate reduction in the batch tests, identification of the denitrification rates according to readily biodegradable COD (RBCOD), K1, slowly biodegradable COD (SBCOD), K2, and endogenous denitrification, K3 came about. The tests with wastewater from BRA had higher K1, between 0.18 and 0.26 gNOx-N/gVSS-d (Test 2) compared to wastewater from NFR with K1 between 0.05 and 0.09 gNOx-N/gVSS-d (Test 1). One reason for the difference could be that the activated sludge was collected from BRA and was adapted to that wastewater compared to the wastewater from NFR. However, the K2 and K3 rates were similar for the two wastewaters. In the tests with MBBR, K1 varied between 0.80 – 2.43 gNOx-N/m2-d for wastewater from NFR (Test 3) and between 1.22 – 2.69 gNOx-N/m2-d for wastewater from BRA (Test 4). The K2 rate was slightly higher for NFR wastewater compared to BRA, probably caused by the biofilm media from NFR, while the K3 rates were quite similar during Test 3 and Test 4. Regarding the effect of TSS removal on the specific denitrification rates, it appeared to be of minor importance, while the main effect was on the overall denitrification potential. Three laboratory scale SBRs at three liter each were operated during three periods, investigating the effect of TSS removal with different sieves on biological nitrogen removal. In period 1 (P1) the wastewater was filtered with 1.2 and 18 μm sieve, in period 2 (P2) filtered with 33 and 90 μm sieves and in period 3 (P3) filtered with 55 and 150 μm sieves. In addition, one SBR had raw wastewater in all periods as control. The comparison of the performances showed that the SBRs fed filtered wastewater removed between 65 and 75 % COD while the SBRs fed raw wastewater removed between 70 and 91 % COD. This indicates that reducing the COD load on the SBR will affect the performance of the process. However, when including the removal of COD in the primary treatment, similar or slightly higher TSS and COD removals were observed in the SBRs fed with filtered wastewater compared to the control reactors. The nitrogen removal was about 60 % for the SBRs fed raw and wastewater filtered at 33 μm and larger pore sizes. The SBRs fed wastewater filtered with smaller pore sizes had reduced nitrogen removal efficiency. SBRs fed with filtered wastewater produced more sludge compared to the control reactor, about 70 – 184 % more in P1, up to 139 % in P2 and 41 – 64 % more in P3. The calculations show that the SBRs fed filtered wastewater required less oxygen compared to the control SBR. The oxygen requirement decreased by 37 % in the SBR fed wastewater filtered at 18 μm and by 59 % in the reactor fed wastewater filtered at 1.2 μm. The difference in the total oxygen demand during the biological process was mainly due to the oxygen consumed for degradation of COD. [...]