Journal of Environmental Management 112 (2012) 384e391

Contents lists available at SciVerse ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Implementation of a knowledge-based methodology in a decision support system

for the design of suitable wastewater treatment process flow diagrams
Manel Garrido-Baserba a, *, Rubén Reif a, c, Francesc Hernández b, Manel Poch a, c
a
ICRA e Catalan Institute for Water Research, Scientific and Technological Park, H2O Building, Emili Grahit 101, 17003 Girona, Spain
b
Department of Applied Economics II, Faculty of Economics, Universitat de València, Campus dels Tarongers, Av. Tarongers s/n, 46022 Valencia, Spain
c
Laboratori d’Enginyeria Química i Ambiental (LEQUIA), Universitat de Girona, Facultat Ciències, Campus Montilivi, 17071 Girona, Spain

a r t i c l e i n f o a b s t r a c t

Article history: In light of rapid global change, the demand for wastewater treatment is increasing rapidly and will
Received 25 February 2011 continue to do so in the near future. Wastewater management is a complex puzzle for which the proper
Received in revised form pieces must be combined to achieve the desired solution, requiring the simultaneous consideration of
3 August 2012
technical, economic, social and environmental issues. In this context, a knowledge-based methodology
Accepted 10 August 2012
Available online
(KBM) for the conceptual design of wastewater treatment plant (WWTP) process flow diagrams (PFDs)
and its application for two scenarios is presented in this paper. The core of the KBM is composed of two
knowledge bases (KBs). The first, a specification knowledge base (S-KB), summarizes the main features of
Keywords:
Environmental decision support system
the different treatment technologies: pollutants removal efficiency, operational costs and technical
(EDSS) reliability. The second, a compatibility knowledge base (C-KB), contains information about the different
Wastewater treatment plants (WWTP) interactions amongst the treatment technologies and determines their degree of compatibility. The
Knowledge-based methodology (KBM) proposed methodology is based on a decision hierarchy that uses the information contained in both KBs
Conceptual design to generate all possible WWTP configurations, screening and selecting appropriate configurations based
Process flow diagrams (PFD) on user-specified requirements and scenario characteristics. The design of the most adequate treatment
train for small and medium sized wastewater treatment plants (2000 and 50,000 p.e. respectively)
according to different restrictions (space constraints, operation simplicity and cost optimization) was the
example in order to show the usefulness of the KBM.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction wastewater from populations higher than 2000 people equivalent

(p.e.) must set up biological treatment. Although there are many
In recent years, new and promising wastewater treatment process-specific technologies capable of adequately treating
technologies have been developed to address current and future wastewater, no single technology or group of technologies has been
scenarios such as the accomplishment of the Urban Wastewater developed to provide a global solution for the almost infinite
Treatment Directive (91/271/EEC), which states that all generated number of wastewater scenarios. Different units are combined and
adapted in a treatment train to meet specific requirements, and this
train is represented by a process flow diagram (PFD). However, the
Abbreviations: KBM, Knowledge-Based Methodology; IFAS, Integrated Fixed-
number of wastewater-related processes in a treatment train has
Film Activated Sludge; EDSS, Environmental Decision Support System; SBR, been growing steadily as well as the number of possible combi-
Sequencing Batch Reactor; PFD, Process Flow Diagram; MBR, Membrane Bioreactor; nations, which increases the difficulty of selecting the most suitable
S-KB, Specification Knowledge Base; MBBR, Moving Bed Bioreactor; C-KB, treatment train configuration (Hamouda et al., 2009; Joksimovic
Compatibility Knowledge Base; CAS, Conventional Activated Sludge; CBA, Cost-
et al., 2006). Traditional rules used by engineers to design PFDs
Benefit Analysis; CW, Constructed Wetlands; EBA, Environmental-Benefit Analysis;
RBC, Rotating Biological Contactors; MCDA, Multi-criteria Decision Analysis; ISF, are often inadequate for modern configurations. Thus, the selection
Intermediate Sand Filters; MU, Meta-units; LCA, Life Cycle Analysis; sMU, Submeta- of the most appropriate PFD design is a difficult task for even the
units; p.e., People Equivalent; O&M, Operation and Maintenance; NPV, Net Profit most experienced designers (Rivas et al., 2008) and the consider-
Value. ation of integrated approaches should be given preference in the
* Corresponding author. Tel.: þ34 649125197; fax: þ34 4972183248.
E-mail addresses: mgbaserba@gmail.com, mgarrido@icra.cat (M. Garrido-
decision-making process (Hidalgo et al., 2007). Expert knowledge,
Baserba), rreif@icra.cat (R. Reif), Francesc.Hernandez@uv.es (F. Hernández), mathematical models, statistical tools, life cycle analysis (Gutiérrez
manuel.poch@udg.edu (M. Poch). et al., 2010), environmental benefit analysis, and cost-benefit

0301-4797/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jenvman.2012.08.013
 M. Garrido-Baserba et al. / Journal of Environmental Management 112 (2012) 384e391 385

analysis (Molinos-Senante et al., 2010) are some of the pieces that classify the number of choices available for different scenarios.
should be fitted for an accurate decision-making process during Therefore, this paper describes this methodology and discusses its
PFDs design. implementation addressing the design of two WWTP serving
Focusing on the number of processes, calculations by Joksimovic different-sized communities (2000 and 50,000 p.e. respectively)
et al. (2008) found that the number of possible combinations of considering both operational and economic criteria. Indeed, the
processes is in the millions. Therefore, similar number of decisions topic is of relevance in the context of the implementation of the
should also be considered during the engineering phase of Urban Wastewater Treatment Directive since the design of treat-
a wastewater project. Although differences are inevitable, three ment facilities used in medium and large cities usually do not give
common levels may be established for this design process. The first satisfactory results when they are implanted directly into small
level includes the generic definition of the scenario and leads to agglomerations (unaffordable operation and maintenance costs).
a description of a broad range of wastewater options and treatment The design of small facilities should rely on consolidated or more
objectives. During this phase, detailed studies of stakeholder innovative technologies that allow flexible operation, reliability and
preference should be conducted. Treatment facilities are only single low O&M costs, achieving sufficient effluent quality (Eusebi et al.,
components of this stage, and detailed treatment designs must be 2008). In this context, this example will show how the use of the
developed in subsequent stages (Daigger, 2011). In the second level, KBM in the NOVEDAR system systematize the selection of different
suitable technologies are selected to accomplish with the objectives designs that allow obtaining a good quality effluent in a sustainable
defined in the previous step. Engineering information is crucial in way depending on plant size and other indicators. The adoption of
this level, and knowledge of current technologies and their the aforementioned methodology constitutes a highlight for deci-
combinations must be considered. In Poch et al. (2004), a good sion makers, since it embraces a variety of different criteria, offering
example of integration of expert knowledge with technical and several technological alternatives (extensive or non-extensive
engineering information in the selection of wastewater treatment technologies) adapted for each specific situation.
technologies can be found. Similarly, Larsen et al. (2010) developed
a methodology to evaluate technologies devoted to nitrogen 2. Methods
removal and considered traditional and innovative aspects, such
as ethical, political, legal, and ecological issues. Gabaldón et al. 2.1. Hierarchical decision approach
(1998) and Flores-Alsina et al. (2008) attempted to satisfy user
requirements by using multiple-criteria and multiple-objective The proposed knowledge-based design methodology decom-
approaches, and sensitivity and uncertainty analysis. In the third poses the conceptual design process into a series of issues that are
and final level, specific designs and sizes of the chosen technolog- easier to analyze and evaluate individually. The hierarchical
ical units are provided in detail. At this stage, the implementation of decomposition is repeated for different levels of abstraction (López-
mathematical models, analytical tools and uncertainty analysis is Arévalo et al., 2007; Garrido-Baserba et al., 2012):
essential. This approach is not always structured as shown above
because development procedures are generally distinct for 1 Meta-Units (MU): Primary, secondary, tertiary, sludge, and
different projects (Gachet and Sprague, 2005). odour and head return treatments (based on Metcalf & Eddy,
A decision-making approach for identifying the optimal 2004).
combination of processes for specific scenarios has not been fully 2 Submeta-Units (sMU): For example, secondary treatment MU
developed yet (Larsen et al., 2010). Previous attempts had limita- is subdivided into: conventional activated sludge (CAS),
tions in the design stage and only evaluated a low number of nitrogen removal, phosphorus removal, attached-growth
alternatives because of the complexity associated with the large technologies, low-loaded technologies, hybrid processes,
number of possible combinations. Other approaches could not granular systems, anaerobic digestion, membrane bioreactor,
thoroughly evaluate all suitable treatment alternatives (the and secondary clarification. In this case, the total number of
response surface). Therefore, with the aim to address the problems elements is 46.
of WWTP design, a novel knowledge-based methodology (KBM) 3 Units (U): At this lowest abstraction level, 274 treatment
capable of generating all viable PFDs and supporting the selection alternatives are included. Continuing with the above example,
of the most suitable option for a specific scenario is proposed. the CAS meta-unit might be subdivided into the following
Thanks to recent research efforts in an environmental context, units: oxidation ditch, AB system, CAS (low SRT), conventional
powerful analytical tools able to provide information relevant to plug-flow, complete mix, step feed, sequential batch reactors
the decision-making process in the context of wastewater design and extended aeration.
have been selected, in order to satisfy with different criteria:
A graphic representation of the proposed design approach is
i) Economic criteria: cost-benefit analysis (CBA) and environ- shown in Fig. 1.
mental benefit analysis (EBA) (Molinos-Senante et al., 2011).
ii) Operational and Technical criteria: multi-criteria decision 2.2. Data bases: knowledge and compatibility
analysis (MCDA).
iii) Environmental and social criteria based on the life cycle The KBM is composed of two knowledge bases (KBs). Expert
analysis (LCA) tool. interviews, specialized literature and engineering experience are
incorporated into both KBs, which are repeated at the different
Their integrated implementation in a single tool, an Environ- levels of abstraction. The user can edit the information predefined
mental Decision Support System (named1 NOVEDAR), constitutes in both C-KB and S-KB, enabling customization according to his
the main highlight of the KBM and contributes to screen and own specifications and experience.
The compatibility knowledge-base (C-KB) contains exhaustive
information about the degree of compatibility amongst different
1
The development of the Novedar_DSS to systematize the design of wastewater
treatment technologies. As an example, a conceptual representa-
treatment plants was performed under the project “Conception of the WWTP of the tion of the assigned compatibilities is shown in Table 1, which
XXI century” http://www.novedar.com. represents the physical and logical relationships amongst specific
386 M. Garrido-Baserba et al. / Journal of Environmental Management 112 (2012) 384e391

Fig. 1. Graphic representation of the design approach.

units. The full C-KB comprises a substantially wider number of the products and operation categories. The level of detail in the char-
different wastewater treatment unit processes, classified according acterization of the higher levels (MU and sMU) differs significantly
to the six meta-units. from that of the U level since these higher levels only consider
The specification knowledge-base (S-KB) summarizes the main information about functional issues, with the aim to decrease the
features of the technologies. At the highest level of detail (i.e., the number of feasible alternatives (preliminary screening step).
lowest level of abstraction), more than 270 unit processes are
thoroughly characterized. A wide range of factors (54) covering six 3. Application of the methodology: selection of alternatives
main topics are considered to provide the knowledge required for
the conceptual design of the PFD. For each technological process, A schematic representation of the methodology is shown in
information about expected water quality (influent characteristics Fig. 2. To generate a PFD adapted to a specific scenario, two main
and effluent quality), technical characteristics and operation data of steps are executed: (1) the complete response surface of suitable
the units, impacts, (odour, noise, visual and environmental impacts, PFD alternatives is created, and (2) the PFD that satisfies the user’s
etc.), costs and subproducts is provided. In this approach, it is crucial requirements is selected. Data must be entered to define the initial
to consider such high number of factors since most methodologies scenario previously to these main steps. In the scenario definition
considers only the major technical and economic factors associated step, the required information is collected and analysed to define
with the selection of a treatment process, such as contaminant the context in which the WWTP will be designed.
removal efficiency and capital cost (Hamouda et al., 2009). This This analysis includes exploration and study of the proposed
higher number of factors facilitates an integrated assessment and facility site, the composition of the wastewater to be treated, the
enables the use of key indicators and analytical tools, including the applicable legislation, and the restrictions that could affect the
aforementioned CBA, LCA, and MCDA. A conceptual representation design process, for example, budget constraints or land require-
of the knowledge assembled in the S-KB is shown in Table 2. In this ments (among many other options). It is important to clarify that
example, four different technologies are partially described after the scenario definition step (input data, specific user
according to the influent, effluent, impacts, economic costs, sub- requirements, desired effluent quality, etc.), the system does not

Table 1
Conceptual representation of the compatibilities. (4 e High Compatibility, 3 e Normal Compatibility, 2 e Low Compatibility, 1 e Potential Incompatibility, and 0 e
Incompatibility).

MBR Fine Screens Oxidation Imhoff tank Trickling Nanofiltration Rectangular

Drum Screens ditch filter tank
MBR 2 0 2 0 0 4
Fine screens Drum screens 0 0 0 0 0 0
Oxidation ditch 0 2 2 0 0 4
Imhoff tank 0 4 0 0 0 0
Trickling filter 0 2 0 4 0 4
Nanofiltration 4 0 1 0 1 0
Rectangular tank 0 4 0 0 0 0

Scores are assigned considering that treatment units from the row precede the ones from the column in the PFD. Gray shade indicates same technology in row and column.
 M. Garrido-Baserba et al. / Journal of Environmental Management 112 (2012) 384e391 387

Table 2
Conceptual representation of a table summarizing knowledge compiled in the S-KB.

Technology 1 Technology 2 Technology 3 Technology 4

Influent
People equivalent (p.e.) 150e1500 25e1000 1200e20,000 <300
Hydraulic loading (m3/m2 d) 0.01e0.08 0.015e0.06 0.01e0.3 0.02e0.005
.
Effluent
DBO Removal (%) 50e85 80e90 55e95 90e99
Nt Removal (%) 10e20 30e70 55e85 65e98
.
Costs
Investment costsa y ¼ 4617,x0,43 y ¼ 3292,x0,32 y ¼ 1642,x0,22 y ¼ 8966,x0,45
O&M costsa y ¼ 136,1x0,38 y ¼ 211,5x0,40 y ¼ 258,6x0,41 y ¼ 15,543x1,32
x is peq; y is total cost
expressed as V/peq
Undesirable outputs
Subproducts Sludge and vegetation that Vegetation to be collected No Crops harvested
can develop in the lakes and digested sludge
Nuisances From odours sporadically and Odours can occur, but rarely No Very rarely can occur odours
mainly in anaerobic lagoons
...
Impacts
Visual Low Low High Fair
Noise Very low Very low Very high Average
...
Operation
Staff specialization level Low. Does not require Low. Does not require Low. Does not require LoweMedium. Does not
skilled labor skilled labor skilled labor require skilled labor. Knowledge
of agriculture are needed
a
Ortega de Ferrer et al. (2011), Tchobanoglous et al. (2003) and Comas et al. (2004).

perform a direct selection of a single secondary treatment unit. screening, propagation and evaluation of the entire set of PFD
Instead, it ranks different feasible options among which the user alternatives described by the network. The flow paths (edges)
has to select one (normally, the one that achieved the highest score) between units, which are obtained from the C-KBs, can be used as
in order to proceed with Step 1. functional connections to send and save information between
nodes. The network then becomes a functional system capable of
3.1. Step 1: generation of the complete response surface of PFDs conducting an integrated assessment of treatment trains.
A pre-screening stage is used to simplify the evaluation of
The information summarized in the KBs (S-KBs and C-KBs) is multiple alternatives (Loetscher and Keller, 2002). This stage is only
converted into a network structure composed of edges (C-KBs) and used for the MU and sMU levels. Using information on local
nodes (S-KBs). This network structure (Fig.1) is repeated for the circumstances and water quality collected during the scenario
three levels of abstraction and represents the complete response definition, this screening stage identifies and discards inappro-
surface of suitable alternatives. priate PFD alternatives that do not satisfy user requirements. The
Edges represent the degree of compatibility between technol- propagation step transfers information through the nodes. During
ogies. As the degree of compatibility increases, the number of propagation, data from the data entry step are transferred through
feasible PFDs in the response surface decreases and vice versa. the combinations of nodes that represent all feasible PFDs. This
Nodes contain the knowledge summarized in the S-KBs. Every procedure is called recursive evaluation (Fig. 3). The information
node, which is linked to its specification information in the generated by each node exposed to the scenario-specific data is
knowledge base, has direct access to mathematical equations, saved. This process is repeated for all nodes until an end node
expressions and models, indicators calculated during the evalua- terminates the propagation. Finally, a complete evaluation of the
tion process (e.g., LCA, CBA, and EBA), and qualitative data. In case different combinations of nodes (PFDs) clustered in the response
further diagram evaluations were required, edge properties might surface is produced, and the output from analysis of the entire set of
be analyzed to obtain several node or eigenvalue centralities, embedded treatment trains is finally generated as a reduced
a measure based on the largest positive eigenvalue of the network network structure.
adjacency matrix (Bañares-Alcántara, 2010). Node centralities can The proposed PFDs are then analyzed with the aim to compare
then be used to measure relative importance within the network quantitatively the alternatives, following a weighed sum model.
(Saaty and Shih, 2009) and to obtain a more exhaustive evaluation. However, it is feasible to use improved MCDA methodologies
Although node centralities are not explored further in this paper, similar to those shown in previous studies (Flores-Alsina et al.,
network characteristics and analytical techniques can be used to 2008; Ashley et al., 2008). More concretely, it is planned to
provide a more thorough and objective assessment of the final implement the analytical hierarchy process (AHP), which was
selected alternatives. designed to subjectively evaluate a set of alternatives based on
multiple factors arranged in a hierarchical structure (Saaty et al.,
3.2. Step 2: selection of PFDs for a specific scenario 2003).

Once the compatible PFDs have been created, feasible solutions 3.3. Case studies results
that meet the user overall degree of satisfaction are selected. The
previously generated network structure is used as a functional Table 3 shows the selection of adequate technologies for two
structure for the transfer of information. Step 2 includes the scenarios, 1) small (2000 p.e.) and 2) medium (50,000 p.e.) sized
388 M. Garrido-Baserba et al. / Journal of Environmental Management 112 (2012) 384e391

Fig. 2. Operation flow diagram of the methodology.

treatment plants, assessed from the point of view of 8 different can be assigned proportionally. This feature is crucial, particularly
criteria (AeH): when three or more different criteria are applied simultaneously.
For example, in F (30:70) the space constraints have a relative
(i) No criteria (A): the selection will be carried out only consid- influence of 30% over the final score, whereas the cost optimization
ering wastewater characteristics has a 70%. When no restrictions are applied (Cases A1 and A2), the
(ii) Technical (B and C): space requirements and operation EDSS randomly proposed different alternatives capable of treating
simplicity. a middle strength wastewater. Only in this specific case, the use of
(iii) Economic assessment (D): cost-benefit analysis (CBA and a KBM for selecting secondary treatment units might be considered
EBA), operation and maintenance costs and investment. of limited utility for reducing the number of available alternatives.
(iv) Combination of criteria from ii) and iii): (EeH) The difference in population size between scenarios 1 and 2 had
a considerable impact over the secondary treatment selection,
Influent characteristics, and therefore part of the input data, since the information compiled in S-KB also includes detailed
were those typically found on a medium strength wastewater descriptions of technologies particularly appropriate for the design
(COD: 430 mg L1; BOD: 190 mg L1; Total N: 40 mg L1; Total P: of small facilities. Thereby, in the list of technologies selected for
7 mg L1; TSS: 210 mg L1). scenario 1 we find extensive wastewater treatment systems. These
Therefore, a total number of 16 case studies, combining sce- systems require lower energy demand and can be easily operated.
narios þ criteria (for example, A2, H1, C1, etc.), were assessed, being When restrictions and/or priorities are applied, the options are
the final result the proposal of 4 specific treatment units ranked slightly or drastically changed depending on each case. For
according to a final score (although the Novedar_EDSS is able to example, comparing 1B and 1C, it can be observed that the EDSS
provide a higher number of feasible units). The priority order does not highlight extensive technologies when the space avail-
(based on the score) is calculated according to the weighed sum ability is restricted. Instead, alternatives such as IFAS, SBR or MBR
model, in which a different weight (or relevance) for each criteria are considered more advisable whereas constructed wetlands or
 M. Garrido-Baserba et al. / Journal of Environmental Management 112 (2012) 384e391 389

Fig. 3. Schematic representation of the recursive evaluation process: Data transfer and process functioning within the directed network structure.

pond systems are among the most recommended units when more proposed advanced technologies (VIP, UCT, PhoStrip.) for the
operation simplicity is required. design of medium-sized plants. Apparently, it might not be correct
Another good example can be shown comparing cases based on to consider such technologies as feasible from an economic point of
a cost optimization criteria (1D and 2D). In this case, the EDSS view, since their high degree of complexity (compartments with
varying redox conditions, internal recyclings, need for meticulous
Table 3
control, etc.) usually entails higher costs. However, four different
Selection of suitable secondary treatment technologies according to different
criteria. indicators are taken into account by the EDSS for the economic
assessment: Operation and maintenance, investment, CBA and EBA.
Criteria Scenario 1 Score Scenario 2 Score
In this specific study, we established that those four indicators
(small) (medium)
influenced the score in the same proportion. The EBA methodology
A No restrictions MBBR e Ludzack-Ettinger e
IFAS e SBR e
(Molinos-Senante et al., 2011) estimates an additional economic
MBR e Oxidation Ditch e benefit based on the amount of pollution removed. Therefore,
SBR e Steep Feed e solutions which enhance the overall quality of treatment removing
B Space constraints IFAS 10 SBR 10 both nitrogen and phosphorus are profitable from this point of
MBR 10 CAS 5
view. Since such technologies are not particularly recommended
SBR 10 Complete Mix 5
MBBR 7.5 Ludzack-Ettinger 2.5 for small WWTPs, the EDSS only proposed them in the case 2D.
C Operation simplicity CW 10 CAS 7.5 When the user included a higher number of preferences (case H2,
Pond 7.5 Complete Mix 7.5 operation simplicity and low space requirements), no technologies
System specific for nutrients removal were suggested by the EDSS with the
RBC 7.5 Ludzack-Ettinger 5
Ext. Aer. 7.5 Oxidation Ditch 5
exception of PhoStrip, although it achieved a substantially lower
D Cost optimization ISF 6.3 SBR 8.9 score compared with the other alternatives. Finally, in case F1,
Pond 6 VIP 5.3 space availability and cost optimization were simultaneously
System selected, and the MBR was the technological choice due to their
CW 5.3 PhoStrip 5.3
well-known advantages: production of a high quality effluent and
SBR 4.6 UCT 5.3
E B þ C (30:70) RBC 7.5 CAS 6.75 intense biological treatment within a reduced space. Although high
CW 7 Complete Mix 6.75 O&M costs have been pointed out as a major drawback for MBRs
Trick Filt 6.75 MBR 6.5 widespread implementation, in this case such costs are counter-
IFAS 6.5 Ludzack-Ettinger 4.25 acted again by the environmental damage avoided (EBA method-
F B þ D (30:70) MBR 6.2 SBR 9.3
MBBR 4.62 CAS 5
ology), due to the total absence of solids in MBR permeates and
RBC 4.6 Complete Mix 5 their high treatment quality.
ISF 4.4 VIP 4.4 After the selection of a specific secondary treatment technology,
G C þ D (30:70) CW 6.7 CAS 7.7 the EDSS will use both S-KB and C-KB to propose a complete
Pond 6.4 SBR 5.8
treatment train (Table 4), providing the user with a set of estimated
System
IFAS 6 Complete Mix 5.8 outputs (Table 5). Although the output data can be considered as
ISF 5.9 PhoStrip 5.2 estimative, it indeed constitutes a value-added differentiation of
H B þ C þ D (50:25:25) SBR 7.7 SBR 8.3 the Novedar software and therefore, of the proposed methodology.
MBR 6.6 CAS 5.8 On the one hand, it permits to the user to establish the most suit-
MBBR 6 Complete Mix 5.8
Trick Filt 5.3 PhoStrip 3.8
able operational strategies for a concrete selection. On the other
hand, this vast amount of information can be used for comparison
390 M. Garrido-Baserba et al. / Journal of Environmental Management 112 (2012) 384e391

Table 4 1. The ability to generate a response surface of WWTP alterna-

Example of treatment trains proposed for two of the cases considered. tives that explore all possible technological combinations and
Case Pretreatment Primary Secondary Sludge line do not miss potential solutions that could maximize benefits.
G1 1 Well with 4 Imhoff tank 5 Constructed e 2. The creation of customized wastewater treatment schemes
grinder wetlands based on a set of design requirements and initial conditions.
2 Manual fine 3. The incorporation of multi-criteria decision methods in the
screens
evaluation of treatment trains, which enables an integrated and
G2 1 Well with 4 Stacked 5 SBR a) Blending
grinder clarifiers comprehensive analysis of all parameters (e.g., environmental,
2 Fine screens b) Rotary drum social, economic and technical) and indicators (e.g., LCA, CBA,
(drums) thickening and environmental benefit analysis).
3 Aerated grit c) Mesophilic
chamber phase anaerobic
digestion
The proposed systematic approach, which explores WWTP
d) Filter press configurations and can be adapted to user-defined conditions,
accelerates the synthesis and evaluation of WWTP designs.

Table 5
Acknowledgments
Example of the estimation of outputs from two selected treatment trains.

Output data (effluent) and design parameters (SBR) The authors would like to thank the financial support from the
Equivalent Population 50,000 Flow Rate (m3 d1) 11,250 Spanish Ministry of Science and Innovation (NOVEDAR CON-
DBO 18 MLSS range (g L1) 3e4 SOLIDER CDS2007-00055) and our computer engineering collabo-
COD 81.5 Reactants consumption 605,930 V/year
TSS 21 Investment 5.05 MV
rators Adrià Riu, Raül Clemente and Albert Benzal for their helpful
Total N 4 O&M Costs 0.13 MV/year support and suggestions. ICRA headquarters have been co-financed
Total P 0.35 Total cost 9.04 MV by the Ministry of Science and Innovation (MICINN) and the
(lifetime 30 years) European Regional Development Fund (ERDF) under the ERDF
HRT (h) 15e30 Sludge production 2720 kg d1
Operational Programme for Catalonia 2007e2013.
SRT (d) 25e35 NPV 26.43 MV
Output Data and Design Parameters (Constructed Wetlands)
Equivalent Population 2000 Flow Rate (m3 d1) 450
DBO 14.2 MLSS range e References
COD 37.6 Reactants costs 35,500 V/year
TSS 5.6 Investment 0.43 MV Ashley, R., Blackwood, D., Butler, D., Jowitt, P., Davies, J., Smith, H., Gilmour, D.,
Total N 12 O&M Costs 0.03 MV/year Oltean-Dumbrava, C., 2008. Making asset investment decisions for wastewater
Total P 4.9 Total cost 1.43 MV systems that include sustainability. J. Environ. Eng. 134, 200.
Bañares-Alcántara, R., 2010. Perspectives on the potential roles of engineers in the
(lifetime 30 years)
formulation, implementation and enforcement of policies. Comput. Chem. Eng.
HRT (h) 90e350 Sludge production e
34 (3), 267e276.
SRT (d) e NPV 27.3 MV Comas, J., Alemany, J., Poch, M., Torrens, A., Salgot, M., Bou, J., 2004. Development of
a knowledge-based decision support system for identifying adequate waste-
water treatment for small communities. Water Sci. Technol. 48, 393e400.
purposes, in case the user needs to consider different option for Daigger, G.T., 2011. A practitioner’s perspective on the uses and future developments
for wastewater treatment modelling. Water Sci. Technol. 63 (3), 516e526.
a preliminary proposal of a wastewater management project.
Eusebi, A.L., Carletti, G., Cola, E., Fatone, F., Battistoni, P., 2008. Switching small
WWTPs from extended to intermittent aeration: process behaviour and
4. Conclusions performances. Water Sci. Technol. 58 (4), 865e872.
Flores-Alsina, X., Rodríguez-Roda, I., Sin, G., Gernaey, K.V., 2008. Multi-criteria
evaluation of wastewater treatment plant control strategies under uncertainty.
The design of integrated wastewater treatment plants is Water Res. 42 (17), 4485e4497.
a complex exercise that must consider a wide range of objectives Gabaldón, C., Ferrer, J., Seco, A., Marzal, P., 1998. A software for the integrated design
of wastewater treatment plants. Environ. Modell. Softw. 13 (1), 31e44.
(e.g., economic, technical, social, and environmental issues) and
Gachet, A., Sprague, R., 2005. A context-based approach to the development of
select a combination of treatment processes to achieve a desired decision support systems. In: International Workshop on Context Modeling and
effluent quality. In recent years, the number of novel, optimized Decision Support, Paris, France.
technologies, analytical techniques, and innovative indicators has Garrido-Baserba, M., Reif, R., Rodríguez-Roda, I., Poch, M., 2012. A knowledge
management methodology for the integrated assessment of WWTP configu-
rapidly increased. The proposed methodology is a significant rations during conceptual design. Water Sci. Technol. 66 (1), 165e172.
contribution to the integration of various aspects of wastewater Gutiérrez, E., Lozano, S., Moreira, M.T., Feijoo, G., 2010. Assessing relationships
management, particularly to the decision-making process for the among life-cycle environmental impacts with dimension reduction techniques.
J. Environ. Manage. 91 (4), 1002e1011.
selection and design of customized PFDs. The methodology facili- Hamouda, M.A., Anderson, W.B., Huck, P.M., 2009. Decision support systems in
tates the simultaneous consideration of different levels of decision water and wastewater treatment process selection and design: a review. Water
making, quantitative analyses (e.g., MCDA, mathematical models, Sci. Technol. 60 (7), 1767e1770.
Hidalgo, D., Irusta, R., Martinez, L., Fatta, D., Papadopoulos, A., 2007. Development of
and control strategies), and innovative indicators (e.g., LCA, cost- a multi-function software decision support tool for the promotion of the safe
benefit analyses and EBA). Therefore, the proposed approach effi- reuse of treated urban wastewater. Desalination 215 (1e3), 90e103.
ciently explores design alternatives, which should contribute to the Joksimovic, D., Savic, D.A., Walters, G.A., 2006. An integrated approach to least-cost
planning of water reuse schemes. Water Sci. Technol. 6 (5), 93e100.
development of more efficient and environmentally friendly Joksimovic, D., Savic, D.A., Walters, G.A., Bixio, D., Katsoufidou, K., Yiantsios, S.G.,
WWTPs. The proposed KBM simultaneously creates a complete 2008. Development and validation of system design principles for water reuse
response surface and produces an integrated assessment of treat- systems. Desalination 218 (1e3), 142e153.
López-Arévalo, I., Bañares-Alcántara, R., Aldea, A., Rodríguez-Martínez, A., 2007.
ment trains. Diagrams are generated by combining a database that
A hierarchical approach for the redesign of chemical processes. Know. Info. Syst.
contains information about the different treatment technologies 12 (2), 169e201.
with a database that contains information about the compatibility Larsen, T.A., Maurer, M., Eggen, R.I.L., Pronk, W., Lienert, J., 2010. Decision support in
of the technologies. urban water management based on generic scenarios: the example of NoMix
technology. J. Environ. Manage. 91 (12), 2676e2687.
The main contributions of this methodology include the Loetscher, T., Keller, J., 2002. A decision support system for selecting sanitation
following: systems in developing countries. Socioecon. Plann. Sci. 36 (4), 267e290.
 M. Garrido-Baserba et al. / Journal of Environmental Management 112 (2012) 384e391 391

Metcalf & Eddy, 2004. Wastewater Engineering: Treatment and Reuse, fourth ed. Poch, M., Comas, J., Rodríguez-Roda, I., Sanchez-Marre, M., Cortés, U., 2004.
McGraw-Hill, New York. Designing and building real environmental decision support systems. Environ.
Molinos-Senante, M., Hernández-Sancho, F., Sala-Garrido, R., 2010. Economic Modell. Softw. 19 (9), 857e873.
feasibility study for wastewater treatment: a cost-benefit analysis. Sci. Total Rivas, A., Irizar, I., Ayesa, E., 2008. Model-based optimisation of wastewater treat-
Environ. 408 (20), 4396e4402. ment plants design. Environ. Modell. Softw. 23 (4), 435e450.
Molinos-Senante, M., Hernández-Sancho, F., Sala-Garrido, R., 2011. Cost-benefit Saaty, T.L., Shih, H., 2009. Structures in decision making: on the subjective geometry
analysis of water-reuse projects for environmental purposes: a case study of hierarchies and networks. Eur. J. Oper. Res. 199 (3), 867e872.
for Spanish wastewater treatment plants. J. Environ. Manage. 92 (12), Saaty, T.L., Vargas, L.G., Dellmann, K., 2003. The allocation of intangible resources:
3091e3097. the analytic hierarchy process and linear programming. Socioecon. Plann. Sci.
Ortega de Ferrer, E., Ferrer, Y., Salas, J.J., Aragón, C., Real, A., 2011. Manual para la 37 (3), 169e184.
implementación de sistemas de depuración en pequeñas poblaciones, first ed. Tchobanoglous, G., Burton, F.L., Stensel, H.D., Metcalf & Eddy, 2003. Wastewater
CENTA, Madrid. Engineering: Treatment and Reuse, fourth ed. McGraw-Hill, Boston.