Modeling Horizontal Flow Constructed Wetlands treating ... · SO4 S H2S X S X I CWM1 component f...
1
David David Sanchez Sanchez- Ramos Ramos a , , Guenter Guenter Langergraber Langergraber b a School of Civil Engineering, University of Castilla-La Mancha, Ciudad Real, Spain; e-mail: [email protected] (corresponding author) b Institute of Sanitary Engineering and Water Pollution Control, University of Natural Resources and Life Sciences, Vienna (BOKU), Vienna, Austria; e-mail: [email protected] INTRODUCTION & OBJECTIVES INTRODUCTION & OBJECTIVES INTRODUCTION & OBJECTIVES INTRODUCTION & OBJECTIVES INTRODUCTION & OBJECTIVES INTRODUCTION & OBJECTIVES INTRODUCTION & OBJECTIVES INTRODUCTION & OBJECTIVES MATERIALS & METHODS MATERIALS & METHODS MATERIALS & METHODS MATERIALS & METHODS MATERIALS & METHODS MATERIALS & METHODS MATERIALS & METHODS MATERIALS & METHODS RESULTS & CONCLUSIONS RESULTS & CONCLUSIONS RESULTS & CONCLUSIONS RESULTS & CONCLUSIONS RESULTS & CONCLUSIONS RESULTS & CONCLUSIONS RESULTS & CONCLUSIONS RESULTS & CONCLUSIONS Study area in the Upper Guadiana Bassin and location of the WWTP considered (green points) Modeling Horizontal Flow Constructed Wetlands treating Modeling Horizontal Flow Constructed Wetlands treating effluents of Wastewater Treatment Plants effluents of Wastewater Treatment Plants Tablas de Daimiel The TDNP is a floodplain wetland which covers 1,928 ha, located in the Upper Guadiana Basin over the West Mancha aquifer (5,500 km 2 ). This aquifer has been subject to intensive pumping since the late 1970s, resulting in the decrease of the phreatic level. Therefore, wetlands have been disconnected from the regional aquifer, producing serious environmental damage on TDNP. Introduction Introduction The HYDRUS Wetland Module (Langergraber & Šimůnek, 2006, 2011) has been used to estimate the efficiency of a horizontal subsurface flow (HSSF) constructed wetland (CW) system, designed for the treatment of effluents from wastewater treatment plants (WWTP) near the Tablas de Daimiel National Park (TDNP), Central Spain. The purpose is to restore this natural wetland with the effluents of several plants in the surroundings, improving the water quality throughout CWs, so the quality requirements are higher than standards. The Treated Sewage Effluents (TSE) volume available from the WWTP (near 10 hm 3 /year) could help to maintain a certain flooded area in the TDNP. The TSE quality has been analyzed, showing some risks for the direct use of the TSE: their quality is quite irregular and some parameter concentrations are sometimes higher than Spanish standards. Therefore, a CW system has been designed to improve the TSE quality before its use on the TDNP. Objectives Objectives The main objectives for the modeling of CWs in this study are: - To verify the treatment system design of HSSF CWs; - To estimate the CWs efficiency for the reduction of several contaminants, mainly COD and NH 4 ; - To better understand the processes concerning the development of bacteria and the reduction of contaminants. Constructed Wetlands sizing Constructed Wetlands sizing A Performance-Based Sizing for HSSF wetlands was used for the treatment design (Kadlec & Wallace, 2009). The system would be composed by several treatment cells. The HF beds have a length of 23 m (including a 0.4 m inlet area), a width of 22 m and a depth of 0.6 m for the treatment of 96 m 3 d -1 each one, resulting in an hydraulic loading rate of 19 cm d -1 . The water level in the bed is about 0.5 m (variable along the bed length, as represented in the next profile). TSE characterization TSE characterization The TSE characterization indicated that the system should be able to treat water with high concentrations of organic matter and ammonium. The influent concentrations have been considered 125 mg O 2 L -1 and 15 mg N L -1 for COD (mainly as slowly biodegradable particulate COD) and total nitrogen (as ammonium only, respectively. These values are the Spanish standards for WWTP, and they correspond approximately to the 90th percentile for both COD and ammonium concentrations records. Some kinetic and stoichiometric parameters have been then modified with respect to their original values in CWM1 to fix the X H development: maximum aerobic growth rate (μ H ), rate constant for lysis (b H ), saturation/inhibition coefficient (K OH ), fractions generated in biomass lysis (f BM,XI ) and hydrolysis rate constant (K h ). After several simulations combining different values for these parameters, a steady state was found. X X REFERENCES REFERENCES REFERENCES REFERENCES REFERENCES REFERENCES REFERENCES REFERENCES - Kadlec, R. & S. Wallace. 2009. Treatment Wetlands. 2nd edition. CRC Press, Boca Raton, Florida (U.S.A.). - Langergraber, G., D. Rousseau, J. García and J. Mena. 2009. CWM1: A general model to describe biokinetic processes in subsurface flow constructed wetlands. Water Sci. Technol. 59:1687-1697. - Langergraber, G. & J. Šimůnek. 2006. The multi-component reactive transport module CW2D for constructed wetlands for the HYDRUS software package. Hydrus Software Series 2. Department of Environmental Sciences, University of California, Riverside, CA, USA. - Langergraber, G. & J. Šimůnek. 2011. HYDRUS wetland module, version 2. Hydrus Software Series 4. Department of Environmental Sciences, University of California, Riverside, CA, USA. 5 105 0 14 0 15 5 5 5 4 Influent conc. (mg l --1 ) X I X S S H2S S SO4 S NO S NH S I S A S F S O CWM1 component -80 0,02 0,1 f BM,XI (g COD XI /g COD BM ) -33,33 2 3 K h (d -1 ) 100 0,4 0,2 K OH (mg O 2 /l) 0 0,4 0,4 b H (d -1 ) -58,33 2,5 6 μ H (d -1 ) Variation (%) Simulation value CWM1 original value Parameter X 0.00 20.000 40.000 80.000 160.000 320.000 640.000 1280.000 2560.000 5120.000 10240.000 20743.082 X X H X A X X FB FB FB FB Model calibration Model calibration The first simulation results showed an excessive heterotrophic bacteria (X H ) growth, especially near the inlet and above the water table, with unrealistic values over 50000 mg kg -1 after 6 days simulation. Implementation in HYDRUS Implementation in HYDRUS In the HYDRUS implementation, the vertical domain of 23 m (length) and 0.6 m (depth) was discretized into 25 columns and 25 rows, resulting in a 2D finite element mesh consisting of 625 nodes and 1152 triangular finite elements. Nodes are regularly distributed in depth, but the distribution is irregular in length to densify the mesh near the inlet and outlet of the bed. An atmospheric boundary condition was applied at the top of the mixing zone, whereas the effluent boundary was represented using a constant head boundary condition at the bottom of the right end of the bed. CWM1 CWM1 The simulation used the CWM1 biokinetic model formulation simulating reactive transport in CWs (Langergraber et al., 2009) since it is considered more appropriate than CW2D for HF CWs (because CWM1 also models anaerobic processes). The influence of plants in wetlands has been considered in this implementation, with the typical values reported for common reed for transpiration rate and specific O 2 release: 7.4 mm d -1 and 5 g m -2 d -1 . As root biomass is normally located in the upper half of the bed, the effects of vegetation were simulated only in this area (in red), with an transition area (in green) in which is considered only half the effect. In CWM1 16 components are considered: 8 soluble and 8 particulate (2 COD components and 6 bacteria species). At the beginning of the simulation, it is assumed an homogeneous distribution of the bacteria in the filter (1 mg/kg for all types of bacteria). The initial conditions for the rest of components are the same than the influent concentrations (left table). Guadiana Bassin SPAIN X I O 0.000 0.250 0.500 1.000 2.000 4.000 8.000 16.000 32.000 64.000 128.000 342.491 X H concentration (mg/l) The 30 days simulation shows better treatment efficiency than expected in the design: 66% of COD removal and 39% of NH 4 removal. This may be due to the high bacteria development (especially X H ). The high hydraulic loading rate (19 cm d -1 ) produces an important O 2 and COD supply in the inlet, enabling a fast growth of heterotrophic bacteria. * Stretching factor in Z-Direction = 5 (for every bed profile) Simulation results Simulation results After a 30 days simulation based on this implementation, X H concentration is still high near the inlet (higher available O 2 ), but with logical values except a few nodes. Autotrophic bacteria (X A ) are concentrated also near the inlet and the surface, where there is enough O 2 for their development. In the deeper half of the bed (without plant root growth) the X H and X A concentration is almost inexistent. Fermenting bacteria (X FB ) don’t need O 2 for their development, and they consume fermentable COD (S F ), which is produced by hydrolisis of slowly biodegradable COD (X S ). Thus, X FB are concentrated downstream of X H (responsible of hydrolisis). 0 20 40 60 80 100 120 140 0 5 10 15 20 25 30 Time (days) Concentration (mg/l) SF SA SI XS XI COD total 0 2 4 6 8 10 12 14 16 0 5 10 15 20 25 30 Time (days) Concentration (mg/l) SNH SNO
Transcript of Modeling Horizontal Flow Constructed Wetlands treating ... · SO4 S H2S X S X I CWM1 component f...
David David SanchezSanchez--RamosRamos aa, , GuenterGuenter LangergraberLangergraber bb
aa School of Civil Engineering, University of Castilla-La Mancha, Ciudad Real, Spain; e-mail:
[email protected] (corresponding author)
bb Institute of Sanitary Engineering and Water Pollution Control, University of Natural Resources and Life
Sciences, Vienna (BOKU), Vienna, Austria; e-mail: [email protected]
INTRODUCTION & OBJECTIVESINTRODUCTION & OBJECTIVESINTRODUCTION & OBJECTIVESINTRODUCTION & OBJECTIVESINTRODUCTION & OBJECTIVESINTRODUCTION & OBJECTIVESINTRODUCTION & OBJECTIVESINTRODUCTION & OBJECTIVES
MATERIALS & METHODSMATERIALS & METHODSMATERIALS & METHODSMATERIALS & METHODSMATERIALS & METHODSMATERIALS & METHODSMATERIALS & METHODSMATERIALS & METHODS
RESULTS & CONCLUSIONSRESULTS & CONCLUSIONSRESULTS & CONCLUSIONSRESULTS & CONCLUSIONSRESULTS & CONCLUSIONSRESULTS & CONCLUSIONSRESULTS & CONCLUSIONSRESULTS & CONCLUSIONS
Study area in the Upper Guadiana Bassin and location
of the WWTP considered (green points)
Modeling Horizontal Flow Constructed Wetlands treating Modeling Horizontal Flow Constructed Wetlands treating
effluents of Wastewater Treatment Plantseffluents of Wastewater Treatment Plants
Tablas de
Daimiel
The TDNP is a floodplain wetland which covers 1,928 ha, located in the Upper Guadiana Basin over the West Mancha aquifer (5,500 km2). This aquifer has been subject to intensive pumping since the late 1970s, resulting in the decrease of the phreatic level. Therefore, wetlands have been disconnected from the regional aquifer, producing serious environmental damage on TDNP.
IntroductionIntroduction
The HYDRUS Wetland Module (Langergraber & Šimůnek, 2006, 2011) has been used to estimate the efficiency of a horizontal subsurface flow (HSSF) constructed wetland (CW) system, designed for the treatment of effluents from wastewater treatment plants (WWTP) near the Tablas de Daimiel National Park (TDNP), Central Spain. The purpose is to restore this natural wetland with the effluents of several plants in the surroundings, improving the water quality throughout CWs, so the quality requirements are higher than standards.
The Treated Sewage Effluents (TSE) volume available from the WWTP (near 10 hm3/year) could help to maintain a certain flooded area in the TDNP. The TSE quality has been analyzed, showing some risks for the direct use of the TSE: their quality is quite irregular and some parameter concentrations are sometimes higher than Spanish standards. Therefore, a CW system has been designed to improve the TSE quality before its use on the TDNP.
ObjectivesObjectives
The main objectives for the modeling of CWs in this study are:
- To verify the treatment system design of HSSF CWs;
- To estimate the CWs efficiency for the reduction of several contaminants, mainly COD and NH4;
- To better understand the processes concerning the development of bacteria and the reduction of contaminants.
Constructed Wetlands sizingConstructed Wetlands sizing
A Performance-Based Sizing for HSSF wetlands was used for the treatment design (Kadlec & Wallace, 2009). The system would be composed by several treatment cells. The HF beds have a length of 23 m (including a 0.4 m inlet area), a width of 22 m and a depth of 0.6 m for the treatment of 96 m3 d-1 each one, resulting in an hydraulic loading rate of 19 cm d-1. The water level in the bed is about 0.5 m (variable along the bed length, as represented in the next profile).
TSE characterization TSE characterization
The TSE characterization indicated that the system should be able to treat water with high concentrations of organic matter and ammonium. The influent concentrations have been considered 125 mg O2 L
-1 and 15 mg N L-1 for COD(mainly as slowly biodegradable particulate COD) and total nitrogen (as ammonium only, respectively. These values are the Spanish standards for WWTP, and they correspond approximately to the 90th percentile for both COD and ammonium concentrations records.
Some kinetic and stoichiometric parameters have been then modified with respect to their original values in CWM1 to fix the XH development: maximum aerobic growth rate (µH), rate constant for lysis (bH), saturation/inhibition
coefficient (KOH), fractions generated in biomass lysis(fBM,XI) and hydrolysis rate constant (Kh). After several simulations combining different values for these parameters, a steady state was found.
X
X
REFERENCESREFERENCESREFERENCESREFERENCESREFERENCESREFERENCESREFERENCESREFERENCES- Kadlec, R. & S. Wallace. 2009. Treatment Wetlands. 2nd edition. CRC Press, Boca Raton, Florida (U.S.A.).- Langergraber, G., D. Rousseau, J. García and J. Mena. 2009. CWM1: A general model to describe biokinetic processes in subsurface flow constructed wetlands. Water Sci. Technol. 59:1687-1697.- Langergraber, G. & J. Šimůnek. 2006. The multi-component reactive transport module CW2D for constructed wetlands for the HYDRUS software package. Hydrus Software Series 2. Department of Environmental Sciences, University of California, Riverside, CA, USA.- Langergraber, G. & J. Šimůnek. 2011. HYDRUS wetland module, version 2. Hydrus Software Series 4. Department of Environmental Sciences, University of California, Riverside, CA, USA.
51050140155554Influent conc.
(mg l--1)
XIXSSH2SSSO4SNOSNHSISASFSO
CWM1
component
-800,020,1fBM,XI
(g CODXI
/g CODBM
)
-33,3323Kh
(d-1)
1000,40,2KOH
(mg O2/l)
00,40,4bH
(d-1)
-58,332,56µH
(d-1)
Variation (%)Simulation
value
CWM1 original
valueParameter
X
0.000
20.000
40.000
80.000
160.000
320.000
640.000
1280.000
2560.000
5120.000
10240.00020743.082
X
XXXXHHHH
XXXXAAAA
X
XXXXFBFBFBFB
Model calibrationModel calibration
The first simulation results showed an excessive heterotrophic bacteria (XH) growth, especially near the inlet and above the water table, with unrealistic values over 50000 mg kg-1
after 6 days simulation.
Implementation in HYDRUSImplementation in HYDRUS
In the HYDRUS implementation, the vertical domain of 23 m (length) and 0.6 m (depth) was discretized into 25 columns and 25 rows, resulting in a 2D finite element mesh consisting of 625 nodes and 1152 triangular finite elements. Nodes are regularly distributed in depth, but the distribution is irregular in length to densify the mesh near the inlet and outlet of the bed.
An atmospheric boundary condition was applied at the top of the mixing zone, whereas the effluent boundary was represented using a constant head boundary condition at the bottom of the right end of the bed.
CWM1CWM1
The simulation used the CWM1 biokinetic model formulation simulating reactive transport in CWs (Langergraber et al., 2009) since it is considered more appropriate than CW2D for HF CWs (because CWM1 also models anaerobic processes).
The influence of plants in wetlands has been considered in this implementation, with the typical values reported for common reed for transpiration rate and specific O2 release: 7.4 mm d-1 and 5 g m-2 d-1. As root biomass is normally located in the upper half of the bed, the effects of vegetation were simulated only in this area (in red), with an transition area (in green) in which is considered only half the effect.
In CWM1 16 components are considered: 8 soluble and 8 particulate (2 COD components and 6 bacteria species). At the beginning of the simulation, it is assumed an homogeneous distribution of the bacteria in the filter (1 mg/kg for all types of bacteria). The initial conditions for therest of components are the same than the influent concentrations (left table).
Guadiana Bassin
SPAIN
X
I
O
0.000
0.250
0.500
1.000
2.000
4.000
8.000
16.000
32.000
64.000
128.000342.491X
H c
on
ce
ntr
ati
on
(mg
/l)
The 30 days simulation shows better treatment efficiency than expected in the design: 66% of COD removal and 39% of NH4 removal. This may be due to the high bacteria development (especially XH). The high hydraulic loading rate (19 cm d-1) produces an important O2 and COD supply in the inlet, enabling a fast growth of heterotrophic bacteria.
* Stretching factor in Z-Direction = 5 (for every bed profile)
Simulation resultsSimulation results
After a 30 days simulation based on this implementation, XH concentration is still high near the inlet (higher available O2), but with logical values except a few nodes. Autotrophic bacteria (XA) are concentrated also near the inlet and the surface, where there is enough O2 for their development. In the deeper half of the bed (without plant root growth) the XH and XA concentration is almost inexistent. Fermenting bacteria (XFB) don’t need O2 for their development, and they consume fermentable COD (SF), which is produced by hydrolisis of slowly biodegradable COD (XS). Thus, XFB are concentrated downstream of XH (responsible of hydrolisis).
0
20
40
60
80
100
120
140
0 5 10 15 20 25 30
Time (days)
Concentr
ation (m
g/l)
SF
SA
SI
XS
XI
COD total
0
2
4
6
8
10
12
14
16
0 5 10 15 20 25 30
Time (days)
Concentr
ation (m
g/l)
SNH
SNO
![Bm D% o 9 ] F 0% Bm +f D% Bm D%](https://static.fdocuments.in/doc/165x107/62bed0ece1d6637c2a6a1a76/bm-d-o-9-f-0-bm-f-d-bm-d.jpg)
