NAXOS 2018 Assessment of wastewater N2O generation using multivariate techniques Vasilaki V. 1 , ConcaV. 3 , Frison N. 3 , Mousavi A. 2 , Fatone F. 4 , Katsou E. 1 15/5/2018
Overview 21 June 2018 Introduction SMART‐Plant Aim and objectives Methodology Results DBSCAN Hierarchical clustering SAX Conclusions 2 Brunel University London
Scale-up of low-carbon footprint MAterial Recovery Techniques for upgrading existing wastewater treatment Plants (SMART-Plant) Sensor Smart Water Networks Management Agriculture Chemical Industry Struvite P ‐ rich compost Water ‐ energy nexus Chemical Intermediates (VFA, N,P Biogas WWTP derivatives) Construction Biocomposites Industry Biofuel from Cellulose & PHA from cellulosic sludge Water reuse Multipurpose water reuse Data Modelling and Management optimization Brunel University London Brunel University London
Scale-up of low-carbon footprint MAterial Recovery Techniques for Research Activities upgrading existing wastewater treatment Plants (SMART-Plant) C-FOOT-CTRL I NTCATCH P-Com post Struvite Urban W astew ater W ater Online Energy and Biopolym ers Energy GHG monitoring Cellulose Reduce Research Environmental Projects benefits Energy Economic benefits GHGs Ad-Bio Sm artPlant Brunel University London Brunel University London
Scale-up of low-carbon footprint MAterial Recovery Techniques for Research Activities upgrading existing wastewater treatment Plants (SMART-Plant) C-FOOT-CTRL I NTCATCH Data analysis Data Acquisition Outliers detection Equalization tank for Solid/Liquid separation and the carbon source storage tank Short ‐ cut Sequencing Batch Reactor (scSBR) Dynamic thickening Fermentation unit (SBFR) BACS storage anaerobic supernatant (screw press) Pattern recognition Conductivity MLSS MLSS in MLSSout Flow-rate in Flow-rate out polyelectrolyte Flow-rate in sludge Flow-rate in polyelectrolyte Date ‐ Time Level (m) Level (m) ORP (mV) DO (mg/L) pH N2O ( μ M) Level (m) pH Level (m) Level (m) (µS/cm) (mg/L) (mg/L) (mg/L) (m3/h) (m3/h) dosage (m3/h) (m3/h) Dependencies identification Enhanced Primary Sequencing Batch Fermentation Separation of municipal Ultrafiltration unit Crystallizer Nitritation SBR PHA ‐ accumulating biomass selection SBR Reactor wastewater unit Flow ‐ rate Flow ‐ rate Pressure Conductivity Conductivity Date ‐ Time Level (m) Level (m) pH Temperature (°C) Level (m) recycle Permeate pH Level (m) DO (mg/L) pH N2O ( μ M) Level (m) DO (mg/L) pH N2O ( μ M) (bar) (µS/cm) (µS/cm) (m3/h) (m3/h) PHA ‐ accumulating biomass selection SBR+U:V CF-EF assessment tool Research Cl N2O NH4- NO3- Influent NH4-N NO3-N DO1 DO2 DO3 NO2-N Temp C N PF N PF C C 1 0.01 10.38 1.50 4593 1.26 5.18 0.47 0.71 1.57 15.6 2 0.01 7.03 3.26 3559 0.69 3.86 0.01 0.30 0.99 1 3 0.1 15.00 1.50 3849 1.63 9.16 1.06 0.76 1.73 Projects 4 0.06 16.45 0.27 8665 9.77 4.79 1.36 0.68 1.55 2 5 0.73 15.05 1.92 3978 1.57 8.22 0.96 1.47 2.21 11.3 6 0.24 9.18 4.09 2994 0.71 5.71 0.02 0.59 1.34 7 0.19 11.21 0.90 11506 4.03 5.05 1.86 2.22 2.10 3 8 2.53 15.55 4.33 3111 1.23 11.33 0.65 2.42 2.39 11.6 9 1.18 10.22 3.34 3235 0.71 6.07 0.02 0.49 1.42 10 2.39 17.24 1.01 3539 1.78 6.73 0.98 1.25 2.25 11 1.72 21.99 0.11 9669 9.66 4.13 2.02 0.07 2.08 4 12 5.25 17.11 0.24 3422 1.31 3.87 0.85 2.37 2.27 11.6 13 3.39 11.42 0.65 3040 0.77 2.45 0.00 1.38 1.42 14 3.96 23.25 0.11 7489 9.56 3.10 1.96 1.74 1.96 15 0.96 8.62 0.14 9824 0.97 1.48 0.92 2.07 1.92 Identification of set-points that 5 16 16.2 7.10 19.37 0.85 3322 2.01 6.52 1.66 1.57 1.93 3.98 17 3.39 11.16 2.48 2665 0.90 3.81 0.06 0.71 1.63 0.77 18 2.12 8.29 6.08 2335 0.85 11.44 1.08 2.24 2.26 1.28 optimize carbon footprint Identification of patterns and dependencies Ad-Bio Sm artPlant Brunel University London Brunel University London
Aim and objectives 6/21/2018 Investigation N2O accumulation profiles and dependencies with operating parameters in a full ‐ scale reactor Implementation of appropriate statistical methods to identify operating conditions for mitigating N2O generation Motif and pattern analysis of critical parameters monitored online and i nvestigation of their effect on N2O generation in the bioreactor 6 Brunel University London
Methodology 21 June 2018 Flow ‐ chart Air 4571 m 3 Secondary clarifiers Effluent Influent Diffusers Rotating diffuser support PE : 40000 Return activated sludge Schreiber bioreactor simultaneous nitrification and denitrification process by a time ‐ based intermittent aeration. The excess phosphorus chemically removed poly ‐ alluminium ‐ chloride (PAC) Brunel University London 7 Presentation Title
Methodology – N 2 O monitoring 21 June 2018 Clark-type electrode (Unisense, Aarhus, Denmark). Dissolved N 2 O Online Offline DO COD influent NH4-N effluent ORP TN influent NO3-N effluent TSS TP influent NO2-N effluent N2O COD/Ntot TP influent Flow-rate COD effluent pH effluent Blowers flow-rate TN efffluent 8 Presentation Title Brunel University London
Methodology - analysis 21 June 2018 Anomalous event detection Uncommon diurnal patterns Sudden changes in the underlying system Sensor faults Density ‐ Based Spatial Clustering of Applications with Noise – DBSCAN (1) Widely used anomaly detection problems (2) Neighborhood distance epsilon (eps) : The radius of the neighborhoods around a data point p. minPts: The minimum number of data points required in a neighborhood to define a cluster. Effective method with medium sized data sets (3) . 9 Presentation Title Brunel University London
Methodology - analysis Hierarchical clustering (4) . Agglomerative: clusters incrementally, producing a dendrogram Identify groups of similar days online operating variables Ward's method (5) Motif discovery – Symbolic Aggregate approXimation (SAX) (6) Repeated motifs in a time series Convert 24h hourly series to piecewise aggregate approximation (PAA) representation Convert the PAA into a string of symbols *Data standardization adfefdaa Brunel University London
Results 21 June 2018 0,4 Mean: 9.4 18,0 0,35 18,0 16,0 16,0 0,3 14,0 14,0 N2O COD/TN 12,0 N2O (mg/L) 12,0 0,25 COD/TN 10,0 10,0 accumulation 8,0 8,0 0,2 6,0 6,0 0,15 4,0 4,0 profile 2,0 2,0 0,1 0,0 0,0 0 0,05 0,1 0,15 12/7/2017 27/7/2017 11/8/2017 26/8/2017 10/9/2017 25/9/2017 0,05 N2O (mg/L) Date 0 Lowest COD/N 13/7/2017 2/8/2017 22/8/2017 11/9/2017 1 Nitification efficiency (%) Monitoring period (h) 0,95 0,9 Variables Values 0,85 0,8 Offline variables Average Std 0,75 0,7 COD influent (mg COD/ L) 234.6 88.5 0,65 TN influent (mg/L) 24.6 3.3 0,6 0 0,05 0,1 TP influent (mg/ L) 4.8 0.9 N2O (mg/L) COD effluent (mg/ L) 26.8 3.7 Nitrification efficiency: ~80% and TN efffluent (mg/ L) 6.4 `1 Denitrification efficiency: ~75%. TP influent (mg/ L) 1.6 0.7 pH effluent 7.1 0.1 11 Presentation Title Brunel University London
Results 21 June 2018 N2O accumulation ORP 0,07 20 N2O Emissions N2O (mg/L)- 0,06 0 ORP and DO 0,05 -20 ORP (mV) 0,04 -40 Spearman’s correlation: 0,03 -60 0,02 -80 DO, N2O 0.62 0,01 -100 0 -120 25 35 45 55 65 75 40 0,07 Operating time (h) ORP slope ORP, N2O 0.72 30 0,06 N2O Emissions 20 ORP slope 0,05 N2O (mg/L) 10 0,04 0 ORP, log(DO) 0.80 25 35 45 55 65 75 -10 0,03 -20 0,02 -30 0,01 -40 -50 0 Strong connection is 0,3 0,07 Operating time (h) do indicated. However, is 0,06 0,25 N2O Emissions 0,05 N2O (mg/L DO (mg/L) this relationship 0,2 0,04 0,15 consistent in the system? 0,03 0,1 Under which conditions 0,02 0,05 0,01 exists? 0 0 25 35 45 55 65 75 Operating time (h) 12 Presentation Title Brunel University London
Results 21 June 2018 Outliers detection Influent flow-rate DO ORP Each point: d-dimensions, where d=24 is Red points: outliers the number of elements in the daily time series. Outliers Influent DO ORP TSS Days 7 2 5 1 Sensor errors 0 1 0 0 13 Brunel University London
Results 21 June 2018 Outliers detection Sensor fault Unusual behavior Outliers Influent DO ORP TSS Days 7 2 5 1 Sensor errors 0 1 0 0 14 Brunel University London
Results 21 June 2018 Outliers detection Linked with precipitation events N2O accumulation = 0 Unusual diurnal profile of ORP linked with high influent flow-rates 15 Brunel University London
Results 21 June 2018 Group Cluster Cluster Days ORP DO Mean (mg/L) Cluster 1 1 1 19 ‐ 32 2 1 2 14 ‐ 2 3 1 3 1 4 2 2 3 i) specific ranges of DO and ORP N2O accumulation and ii) examine the effect of other parameters (monitored online and 5 2 3 15 offline) that could impact the N2O measurements in periods Disturbances - - 10 with similar ranges of ORP and DO. i) variations impacted the N2O accumulation Cluster Mean (mg/L) ii) specific ranges of DO and ORP high N2O accumulation 0.19 iii) examine the effect of other parameters with 0.23 steady profiles of ORP and DO . 0.29 16 Presentation Title Brunel University London
Recommend
More recommend
