Athens 14-16 September 2016 Interlinkages between operational conditions and direct and indirect greenhouse gas emissions in a moving bed membrane biofilm reactor G. Mannina, M. Capodici, A. Cosenza, D. Di Trapani Università di Palermo Dipartimento di Ingegneria Civile, Ambientale, Aerospaziale, dei Materiali (DICAM)
Introduction Wastewater treatment entails: • direct emissions of greenhouse gases (GHGs), such as nitrous oxide (N 2 O) • indirect emissions resulting from power requirements N 2 O Unwanted even at small levels due to the high global warming potential 310 higher than CO 2
Introduction N 2 O Production Pathways Nitrification Denitrification reduction of NO 2 - as terminal intermediate of the incomplete electron acceptor to N 2 O (AOB heterotrophic denitrification denitrification) incomplete oxidation of hydroxylamine (NH 2 OH) to NO 2
Introduction Process operations aimed at the reduction of N 2 O could conflict with the effluent quality and increase the operational costs Operational costs GHG Effluent emission quality Challenge To identify GHG mitigation strategies as trade-off between operational costs and effluent quality index is a very ambitious challenge
Aim Simple model for interlinkage among operational conditions/influent features/effluent quality and emitted N 2 O. Performing a multivariate analysis + University Cape Town (UCT) moving bed (MB) membrane bioreactor (MBR) pilot plant.
Methods
Pilot plant Q R2 = 100 L h -1 Q IN = 20 L h -1 Q OUT = 20 L h -1 Q R1 = 20 L h -1 Q RAS = 80 L h -1 Gas Q out Funnel Suspended Gas Gas Gas Carriers Funnel Funnel Funnel Q RAS Clean In Place Tank Q in MBR Tank ODR Q R2 Anaerobic Tank Three experimental phases: Anoxic Tank Aerobic Tank Q R1 150 days of experimentation Phase I: SRT = ∞ Phase II: SRT = 30 days Mixture of real and synthetic Phase III: SRT = 15 days wastewater!
Pilot plant PURON 3 bundle ultrafiltration module (pore size 0.03 μm, surface 1.4 m 2 ) AMITECH carriers in anoxic and aerobic reactors with a 15 and 40% filling fraction respectively Measured data TSS, VSS, COD TOT , COD SOL , N-NH 4 ,N-NO 3 , N-NO 2 , TN, TP, P-PO 4 , DO, pH, T, N-N 2 O as gas and dissolved Two time per week in each tank
Indirect emissions The Operational Costs (OCs) were evaluated using conversion factors (Mannina and Cosenza, 2015 ): Pw [kWh m ‐ 3 ] energy required for the aeration Peff [kWh m ‐ 3 ] energy required for permeate extraction power,GHG e conversion factors, 0.7 gCO 2eq and 0.806 € kWh ‐ 1 EF [€ m ‐ 3 ] cost of the effluent fine including N 2 O
Indirect emissions The effluent fine (EF) was evaluated using: EFF Q OUT j Q C j t 2 1 1 n EFF C L , j OUT EF Heaviside C j dt EFF C L , j j j t 2 t 1 0, j C j Q IN j 1 t 1 Q IN and Q OUT are the influent and effluent flow, respectively; j is the slope of the curve EF versus C j EFF when C j EFF < C L,j (in this case, the function Heaviside =0); j represents the slope of the curve EF versus C j EFF when C j EFF > C L,j (in this case, the function Heaviside =1); 0,j are the increment of the fines for the latter case.
Indirect emissions The effluent quality index (EQI) was evaluated using: COD COD TOT TN TN PO PO t 1 1 EQI OUT Q dt N 2 Ogas N 2 O gas N 2 O , L T 1000 N 2 O L to COD , TN , PO , N2Ogas and N2O,L are the weighting factors of the effluent COD TOT , TN, PO, liquid N 2 O in the permeate and gaseous N 2 O.
Multiregression analysis Performed to point out general relationships for the N-N 2 O and the plant operation conditions or the available measured data Two type of analysis Simple linear regression Complex regressions
Simple linear regression N 2 O-N flux ANAER (N 2 O-N flux emitted from the anaerobic tank) N 2 O-N flux ANOX (N 2 O-N flux emitted from the anoxic tank) Dependent N 2 O-N flux AER (N 2 O-N flux emitted from the aerobic tank) variables N 2 O-N flux MBR (N 2 O-N flux emitted from the MBR tank) N 2 O-N dissolved OUT (N 2 O-N permeate dissolved concentration) Y = dependent variable; X 1 = independent variable; c 1, c 2 regression coefficients
Complex regressions Multiple linear (LINm) Multiple exponential (EXP) Sum of exponential (SumEXP) Dependent ∑N 2 O-N flux (sum of the N 2 O-N flux emitted from each tank) variables N 2 O-N dissolved OUT (N 2 O-N permeate dissolved concentration) Y = dependent variable; X 1 ,…,X m = independent variable; c 1 ,…,c n regression coefficients
Independent variables Influent concentration COD TOT, IN , N-NH 4,IN , P TOT,IN , P-PO 4,IN , C/N Effluent concentration COD TOT,OUT , BOD 5,OUT , N-NH 4,OUT , N-NO 3,OUT , NO 2 -N ,OUT , P-PO 4,OUT Intermediate concentration N-NO 2_AER , N-NO 2_ANOX, DO AER, DO ANOX, pH AER, pH ANOX, DO MBR Performance indicators COD,BIO, COD,TOT, NITR , DENIT , N TOT, P Operational conditions TSS*, SRT, Biofilm*
Numerical settings 10,000 Monte Carlo simulations varying coefficients Evaluation of Nash and Sutcliffe efficiency for each simulation Y meas,i = measured value of the ith dependent state variable; Y sim,i = simulated value of the ith dependent state variable; Y aver,meas,i = average of the measured values of the ith dependent state variable
Results
Simple linear regression analysis Maximum efficiency Varying the SRT different variables can be adopted to predict the N 2 O Dependent variables N 2 O-N N 2 O-N N 2 O-N NO 2 accumulation influence the N 2 O flux ANAER flux ANOX flux AER N 2 O-N flux MBR N 2 O-N dissolved OUT production Phase Independent TSS NO 2 -N ANOX NO 2 -N ANOX NH 4 -N IN NO 3 -N OUT I variable Efficiency 0.11 0.52 0.52 0.2 0.1 N 2 O dissolved in the permeate Independent NITR NO 2 -N ANOX NO 2 -N ANOX DO AER COD OUT II variable depend on COD OUT Efficiency 0.35 0.6 0.5 0.26 0.72 Independent pH AER Biofilm Biofilm PO 4 -P OUT NO 2 -N AER III variable Efficiency 0.12 0.36 0.67 0.52 0.94
Simple linear regression analysis Scatter plots Phase III 0.4 (a) (b) 0.3 Efficiency [-] 0.2 0.1 N 2 O-N flux ANOX N 2 O-N flux ANOX Combining 0.0 9 9.5 10 10.5 11 -6.5 -6 -5.5 -5 -4.5 Parameter- c 2 Parameter- c 1 effect 0.7 (c) (d) 0.6 Efficiency [-] 0.5 N 2 O-N flux AER N 2 O-N flux AER 0.4 34 34.5 35 35.5 36 -19 -18.5 -18 -17.5 -17 Parameter- c 2 Parameter- c 1 1.0 (e) (f) 0.8 Efficiency [-] 0.6 High dependence Poor dependence 0.4 0.2 N 2 O-N dissolved OUT N 2 O-N dissolved OUT 0.0 0 0.02 0.04 0.06 0 0.002 0.004 0.006 0.008 Parameter- c 1 Parameter- c 2
Complex multiregression analysis INm - Maximum efficiency ∑N 2 O-N flux N 2 O-N dissolved OUT Efficiency Efficiency dependent 0.015 0.244 variable C/N N-NH 4,IN TSS LINm poorly reproduces the measured Biofilm data for ∑N 2 O-N flux (efficiency 0.015). SRT Efficiency obtained for the N 2 O-N DO AER dissolved OUT is slightly higher than for N-NO 2_AER ∑N 2 O-N flux (equal to 0.244) pH AER DO ANOX
Complex multiregression analysis and SumEXP - Maximum efficiency EXP SumEXP ∑N 2 O-N flux N 2 O-N dissolved OUT ∑N 2 O-N flux N 2 O-N dissolved OUT Efficiency Efficiency Efficiency Efficiency Independent 0.125 0.164 0.198 0.178 variable C/N C/N N-NH 4,IN N-NH 4,IN TSS TSS Biofilm Poor efficiency values obtained Biofilm SRT for both the investigated SRT dependent variables DO AER DO AER N-NO 2_AER N-NO 2_AER pH AER pH AER DO ANOX
Conclusions asonable agreements for simple regression equations pendency of N 2 O flux with SRT and plant sections T of Phase III makes the conditions of N 2 O production re sharped ne of the investigated equations for complex multivariate alysis is able to provide satisfactory efficiencies
Message to take home! e interactions among the key factors affecting the make difficult to establish an unique equation valid different operational conditions for predicting N 2 O
Athens 14-16 September 2016 Thank you for your attention Giorgio Mannina giorgio.mannina@unipa.it Acknowledgements
Athens 14-16 September 2016 1 st International Conference www.ficwtmod2017.it FICWTMOD2017 - Frontiers International Conference on wastewater treatment and modelling 21 – 24 May 2017, Palermo, Italy ed by
Recommend
More recommend
