The University of Queensland, Australia - … · 2015. 7. 22. · Model&based*selec,on*of*the*best*...
Transcript of The University of Queensland, Australia - … · 2015. 7. 22. · Model&based*selec,on*of*the*best*...
Model-‐based selec,on of the best viable opera,onal strategy for a
full scale MBR Ignasi Rodriguez-Roda, Sara Gabarrón, Montse Dalmau,
Hèctor Monclús, Julian Mamo, Jose Porro and Joaquim Comas
9th IWA Symposium on Systems Analysis and Integrated Assessment 11-‐14 June 2015, Gold Coast, Australia
MEMBRANES FULL SCALE Area (3 lines): 22.752 m2
Cycles: 10’/30’’
Flux: 27 LMH
Design flow: 90.000 m3/day
IFAS : 75.000 m3/day
MBR : 15.000 m3/day
Pre-‐treatment 1 mm screening
Anoxictank 1
Anoxictank 2
Aerobictank 1
Aerobictank 2
Membranes tank
Membranes tank
Membranes tank
Internal recirculation
Primary Settler
BIOLOGICAL TANK – IFAS Secondary Settler
THE PROCESS: WWTP Terrassa (Catalonia)
MOTIVATION
No specific process problems… No interest nor budget..
Interest from academia: LEQUiA-‐UdG and ICRA (Girona) ExperMse in MBR operaMon and in process opMmizaMon 1 on going PhD in MBR opMmizaMon Some R+D projects to cover iniMal expenses
END OF THE STORY
Opportunity (saving cost.. Pay back soluMon… no environmental risk)
THERE IS A POTENTIAL STORY
METHODOLOGY
1 PROCESS CHARACTERIZATION 2 MODEL CALIBRATION 3 EVALUATION OF ALTERNATIVES 4 NEGOTIATION (DAM and ACA) 5 IMPLEMENTATION AND VALIDATION
PROCESS CHARACTERIZATION
Membrane filtraMon process (TMP, flux, Permeability)
Pre-‐treatment 1 mm screening
Anoxictank 1
Anoxictank 2
Aerobictank 1
Aerobictank 2
Membranes tank
Membranes tank
Membranes tank
External recirculation
Internal recirculation
Primary Settler
Influent / effluent quality (N, C, P)
Sludge quality
Bioreactor condiMons (DO, Redox, Temperature, …)
Historical data (never enough)
3 days experimental campaign
METHODOLOGY
1 PROCESS CHARACTERIZATION 2 MODEL CALIBRATION 3 EVALUATION OF ALTERNATIVES 4 NEGOTIATION (DAM and ACA) 5 IMPLEMENTATION AND VALIDATION
MODEL CALIBRATION
Simulation platform: WEST (Mike by DHI)
Caracterización del agua de entrada: del 12/5/2013 12:00 al 15/5/13 12:00 (posterior a la decantación)
Model layout:
V= 872.84m3
DO = 0
V= 867.68m3DO = 0
V= 1013m3DO = 0.5
V= 1840m3DO = 1.2
V= 248.25m3DO = 5
Recirculación del tanque de membranas al tanque aeróbico = 4 Q
Recirculación del aeróbico al
anóxico: 4,5 Q
Waste flow: 160 m3/d
Biological model: ASM2d
Membrane behaviour: resistance –in series model
METHODOLOGY
1 PROCESS CHARACTERIZATION 2 MODEL CALIBRATION 3 EVALUATION OF ALTERNATIVES 4 NEGOTIATION (DAM and ACA) 5 IMPLEMENTATION AND VALIDATION
ROOM FOR OPTIMITZATION
Reduce energy costs
Improve N removal efficiency
EVALUATION OF ALTERNATIVES
The MBR was working properly, but…
1 ReducMon of aerobic tank DO 2 ModificaMon of external/internal recirculaMon flow rates 3 C Source addiMon 4 DO control of external recirculaMon Effluent quality (EQI)
Opera,onal costs (OC) AeraMon energy (Nopens et al., 2010; Maere et al., 2011)
Pumping energy (Maere et al., 2011) Bulking Risk (Comas et al., 2008)
BSM Criteria
NitrificaMon was achieved before the membrane tanks Punctual poor denitrificaMon
Scenario analysis (modelling)
DO: 1.2 DO: 1 DO: 0.8 DO: 0.5
Cos
ts (€·d
ay-1
)
40
50
60
70
80
90
100
EQ
(mg
pollu
tant·L
-1)
2500
2600
2700
2800
2900
3000
3100Costs EQ
0
1
2
3
4
5
6
7
8
9
Bulking risk (%
)
Best scenario
1 Reduc,on of 2nd aerobic tank DO set point
Scenarios chosen according to the blowers possibili,es
EVALUATION OF ALTERNATIVES (1)
Cos
ts (€·d
ay-1
)
40
50
60
70
80
90
100
Effl
uent
qua
lity
(mg
pollu
tant·L
-1)
2500
3000
3500
4000Internal Recirculation:
4.5 QInternal Recirculation:
6 QInternal Recirculation:
2 QInternal Recirculation:
1 Q
PE EQI
External recirculation (Q)
PE Real operational conditions
6Q 6Q 6Q 6Q4Q 4Q4Q 4Q2Q 2Q 2Q 2Q1Q 1Q 1Q 1Q
NO3-‐:
8,76 mg·∙L-‐1
NO3-‐:
18,07 mg·∙L-‐1
2 Modifica,on of the external and internal recycle flow rate
EVALUATION OF ALTERNATIVES (2)
0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00
NO
3- -N (m
g·L-1
)
0
5
10
15
20
Hours
Simulated DO:0.8 mg/L
A NO3-‐-‐N effluent
concentra,on reduc,on of 30%
0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00
NO
3- -N (m
g·L-1
)
0
5
10
15
20
DO: 0.8 mg·L-1 and Dynamic methanol addition
Hours
Simulated DO:0.8 mg/L
Cost analysis predicted an increment of 41 % of the total costs (aeraMon and pumping energy costs)
EVALUATION OF ALTERNATIVES (3)
3 C source addi,on
DO > 2 mg/L DO < 2 mg/L To the AEROBIC TANK To the ANOXIC TANK
Pre-‐treatment 1 mm screening
Anoxictank 1
Anoxictank 2
Aerobictank 1
Aerobictank 2
Membranes tank
Membranes tank
Membranes tank
External recirculation
Internal recirculation
Primary Settler
DO sensor
ANOXIC AEROBIC MEMBRANES
DO concentra,on ≈ 5 ppm (Tan et al., 2008)
EVALUATION OF ALTERNATIVES (4)
4 Control of the external Sludge recircula,on DO
METHODOLOGY
1 PROCESS CHARACTERIZATION 2 MODEL CALIBRATION 3 EVALUATION OF ALTERNATIVES 4 NEGOTIATION (DAM and ACA) 5 IMPLEMENTATION AND VALIDATION
IMPLEMENTATION AND VALIDATION
Characteriza,on
Membrane filtraMon process (TMP, flux, Permeability) Influent / effluent quality
Sludge quality
Bioreactor condiMons (DO, Redox, Temperature, …)
10:00 22:00 10:00 22:00 10:00 22:00 10:00 22:00
N-N
O3-
(mg·
L-1)
0
10
20
30
401st Experimental campaign DO: 1.2 mg/L 2nd Experimental campaign DO:0.8 mg/LAfter implementation of DO: 0,8 mg/L
Effluent
1 Reduc,on of DO set point to 0.8 mg/L in the second aerobic reactor 2 Maintain int/ext rec flow rates 3 DO control system for the external recircula,on in order to ensure the anoxic condi,ons in the anoxic tank 4 Monitor results!!
CONCLUSIONS
More interes,ng for academia than for the companies
→ Improved the Nitrogen removal efficiency (27 %)
→ Reduced the biological aera,on costs (7 %)
→ Shown the efficiency of the DO control system for the external recircula,on in order to ensure the anoxic condi,ons in the anoxic tank
The op,miza,on strategies implemented in the full-‐scale MBR have….
Model based approach MBR (N/D, but filtra,on is s,ll an issue… rela,onship with biology)
Op,mal vs. viable (implementa,on vs publica,on)
Indirect R+D funding.. Competence to consul,ng companies?
9th IWA Symposium on Systems Analysis and Integrated Assessment 11-‐14 June 2015, Gold Coast, Australia
THANK YOU VERY MUCH