1-s2.0-S0048969724000378-main

Science of the Total Environment 915 (2024) 169903
Contents lists available at ScienceDirect
Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv
Review
Indicator based multi-criteria decision support systems for wastewater

treatment plants
D. Renfrew a, V. Vasilaki a, E. Katsou b, *
a
Department of Civil & Environmental Engineering, Institute of Environment, Health and Societies, Brunel University London, Uxbridge Campus, Middlesex, UB8 3PH
Uxbridge, UK
b
Department of Civil & Environmental Engineering, Imperial College London, London SW7 2AZ, UK
H I G H L I G H T S G R A P H I C A L A B S T R A C T
• DSS indicators are not selected in line

with water sector sustainability targets.
• Process optimisation DSSs state much
clearer aims than technology selection
DSSs.
• Inconsistent categorisation of environ
mental, social, and technical KPIs
• Fuzzy-AHP and -TOPSIS are commonly
employed to reduce human error.
• Few examples of real-world WWTP
process control optimisation due to
reliance on BSM.
A R T I C L E I N F O A B S T R A C T
Editor: Paola Verlicchi Wastewater treatment plant decision makers face stricter regulations regarding human health protection, envi
ronmental preservation, and emissions reduction, meaning they must improve process sustainability and
Keywords: circularity, whilst maintaining economic performance. This creates complex multi-objective problems when
Decision support systems operating and selecting technologies to meet these demands, resulting in the development of many decision
Wastewater treatment plants
support systems for the water sector. European Commission publications highlight their ambition for greater
Multi-criteria decision-making
levels of sustainability, circularity, and environmental and human health protection, which decision support
Multi-objective optimization
Technology selection system implementation should align with to be successful in this region. Following the review of 57 wastewater
Key performance indicators treatment plant decision support systems, the main function of multi-criteria decision-making tools are tech
nology selection and the optimisation of process operation. A large contrast regarding their aims is found, as
process optimisation tools clearly define their goals and indicators used, whilst technology selection procedures
often use vague language making it difficult for decision makers to connect selected indicators and resultant
outcomes. Several recommendations are made to improve decision support system usage, such as more rigorous
indicator selection protocols including participatory selection approaches and expansion of indicators sets, as
well as more structured investigation of results including the use of sensitivity or uncertainty analysis, and error
quantification.
* Corresponding author.
E-mail address: e.katsou@imperial.ac.uk (E. Katsou).
https://doi.org/10.1016/j.scitotenv.2024.169903
Received 13 June 2023; Received in revised form 17 December 2023; Accepted 2 January 2024
Available online 8 January 2024
0048-9697/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
D. Renfrew et al. Science of the Total Environment 915 (2024) 169903
1. Introduction meaning process control is often intuitive with operators unable to un
derstand the real time impacts of their decisions (Ntalaperas et al.,
The wastewater sector is faced with many challenges that result from 2022). Another key area for DSS use in the water sector is online process
ageing and inefficient processes, including substantial carbon emissions, optimisation, however, it has not been widely applied in WWTPs as
high energy consumption, regulatory compliance failures, and loss of improvements to sensors, mathematical models (soft sensors), and data
public trust (Borzooei et al., 2019). Unfortunately they are only being visualisation are needed for precise operational monitoring and control.
worsened by the impacts of climate change, urbanisation, and popula However, a combination of data-driven models and artificial intelli
tion growth (Haldar et al., 2022). Although a plethora of technologies gence enables performance prediction that can be used to reduce energy
have been developed in recent years to combat these issues at a waste demand, decrease costs, improve effluent quality, and lower emissions
water treatment plant (WWTP) level by academia and industry (Kehrein (Matheri et al., 2022).
et al., 2020), water utilities are unable to make the required investment Considering the transformation that WWTPs face to improve per
decisions to shift towards sustainable wastewater treatment. Decision formance by reducing emissions, energy consumption, and operating
support systems (DSSs) have been used to support complex decision costs whilst meeting stricter regulatory targets, water utilities are ex
making in the water sector, including WWTPs with the aim of optimising pected to become ever more reliant on DSSs to solve multi-objective
technology selection procedures or process control to improve opera problems for optimal selection and operation of sustainable technol
tional performance (Wardropper and Brookfield, 2022). ogy. This study focuses on the use of multi-criteria DSSs to support these
Wastewater decision makers have additional considerations two functions for WWTP decision makers. Rather than focussing solely
compared with other industries, as on top of conventional technical, on their typology, analysis of the correct selection and utilisation of
economic, and environmental issues, the social and regulatory impli relevant KPIs during DSS application is prioritised, to ensure that out
cations of their actions must be considered (Ullah et al., 2020). comes fulfil decision maker requirements. This is a pertinent aspect of
Commonly public perception and social acceptance problems arise when complex multi-objective decision-making and one which is often over
utilising and recycling wastewater streams to generate resources (Keh looked or undervalued by the methodologies in the literature.
rein et al., 2020). Water provision and sanitation services are also highly
regulated and must be protected due to their importance for society, 2. Methodology
industry, and the environment (Preisner et al., 2022). Additionally, it is
proving difficult to create markets for new products recovered from 2.1. Wastewater sector goals
wastewater, such as tackling the end-of-waste status for their use in the
European Union (Palmeros Parada et al., 2022). Therefore, water utility Currently, there is a mismatch in terms of the decision maker goals
and WWTP decision makers are facing stricter regulations to improve and the KPIs selected when utilising DSSs at a WWTP level. Therefore,
operation regarding human health protection, environmental preser this section maps the wastewater ambitions of the European Commis
vation, and emissions reduction (Mannina et al., 2019), whilst simul sion which can be used to direct the utilisation of DSSs to meet water
taneously pursuing greater circularity and revenue generation through sector targets.
resource recovery strategies to improve business performance. This The European Commission has directives which act as the framework
creates complex multi-objective problems when operating and selecting for adequate wastewater treatment in the EU and are critical sources for
technologies for improving WWTPs, which are traditionally labour understanding high-level water sector goals. However, in many cases
intensive, trial-and-error experiments that rely on the judgement of they are decades old and do not reflect the regions recent sustainability
operators (Ntalaperas et al., 2022; Sucu et al., 2021). To ensure that all ambitions (European Commission, 2022). The Urban Wastewater
relevant information, performance trade-offs, and cause and effect re Treatment Directive (UWWTD) (91/271/EEC) published in 1991, acted
lationships are taken into consideration when dealing with complex as the basis for transforming European water systems by limiting
problems, DSSs must be utilised by the water sector for more robust pollutant levels in WWTP discharge. The Sewage Sludge Directive (86/
decision making (Ullah et al., 2020). 278/EEC) was introduced for the correct use of sewage sludge in agri
A DSS is a computational system that assists the user in choosing an culture. It details the requirements in terms of heavy metal concentra
optimal or consistent solution to a particular problem in a reduced tion, quantities of sludge applied per hectare, and the crops prohibited
timeframe, particularly when the solution is unclear, by aggregating from application (Council of the European Union, 1986). Although the
often conflicting values or preferences to examine the trade-offs between UWWTD and Sewage Sludge Directive have been successful in
solution objectives (Giupponi and Sgobbi, 2013; Mannina et al., 2019; improving environmental and human health, as 92 % of wastewater is
Wardropper and Brookfield, 2022). A review by Mannina et al. (2019) now treated satisfactorily (European Commission, 2022), the next gen
classified wastewater DSS intentions as; design, energy consumption, eration of wastewater treatment must go beyond this to achieve the EU's
operational optimisation, improvement of effluent quality, or environ sustainability goals, whilst ensuring this fundamental objective is still
mental sustainability. Of course, decision makers may want to investi maintained.
gate a combination of or all of these goals at once, which can often be To instigate further change to WWTPs, a proposal to update the
contradictory (Eseoglu et al., 2022). For example, WWTP direct emis UWWTD was published in October 2022 with the aim of introducing
sions and electricity consumption typically increase when improving new rules up to the year 2040 (European Commission, 2022). This up
effluent quality, however, this action negatively impacts any net zero date will be key for delivering the European Green Deal's zero pollution
targets. Therefore, when using DSSs to solve multi-objective problems target and highlights many water sector goals that decision makers will
the goal of the study must be defined with clear constraints for opti need to adopt in Europe. It expands regulatory compliance to smaller
misation, and an adequate number of relevant key performance in plants and introduces binding energy neutrality targets for the sector,
dicators (KPIs) chosen, to ensure the resulting decision is a true polluter pays for the treatment of toxic micropollutants, and minimum
reflection of the defined goals. recovery rates for phosphorus. Additionally, improved data monitoring
Multi-criteria decision-making (MCDM) tools for selecting the and usage are required for measuring and mitigating greenhouse gas
optimal technology for a specific scenario have been developed in (GHG) emissions and micropollutants, and making KPIs public to
literature (Eseoglu et al., 2022; Južnič-Zonta et al., 2022; Sucu et al., improve benchmarking and transparency (European Commission,
2021). Depending on MCDM application, the goals of the assessment 2022).
will impact the KPIs used to constrain the decision-making process and The European Commission is pursuing a CE to facilitate many of its
final outcome. Conventional WWTP operation is monitored using sustainability targets, therefore, it published the CE Action Plan in 2020
effluent quality and consequently controlled with a few key parameters, (European Commission, 2020). As part of this, the European
2
Commission aims to intensify nutrient recovery from wastewater by DSSs were found using the search term (“wastewater treatment” OR
establishing Integrated Nutrient Management Plans (Radini et al., WWT OR WWTP OR sludge) AND (DSS OR “decision support system” OR
2023). Another key element is the development of Water Reuse Regu MCA OR “multi criteria” OR MCDM OR “multi-criteria”) AND (selection
lation (2020/741) to facilitate the circular use of wastewater effluents. OR identification OR KPI). Whilst the multi-objective optimisation DSS
The document provides a classification system regarding the technology search used (WWTP OR “wastewater treatment plant” OR “wastewater
required to achieve the contaminant levels for application to specific treatment process”) AND (control OR operation OR monitoring OR
crop grades (European Parliament, 2020), relying on the use of Water optimisation OR optimization) AND multi AND (criteria OR objective)
Reuse Risk Management plans to ensure public and environmental terms.
health (Radini et al., 2023). WWTPs should improve effluent quality for
the circular use of water, also reducing the quantity of raw water 2.2.3. Selection of studies
abstracted. Therefore, it is clear that for a sustainable and circular Fig. 1 shows the steps taken to screen initial search results and collect
transition, WWTPs must focus on emissions reduction, resource recov articles used for review (Page et al., 2021). Results were exported to
ery, and water reuse, and acknowledge the importance of proper data Mendeley reference management software for processing, and after
usage, to align with water sector goals at a European level. removing duplicates 127 articles and 144 articles related to technology
Analysis of regional government wastewater strategies is vital for selection and process optimisation DSSs were identified respectively.
creating useful DSSs. However, their it remains challenging to imple They were then analysed to ensure the inclusion of only high-quality,
ment tangible decisions at WWTP level, as individual utilities have their peer-reviewed, original articles, thereby removing non-English, confer
own priorities based on local facets. Considering legislative constraints, ence proceedings, book chapter, and review paper sources. Next, sources
sector-wide ambitions, and local factors can make the identification of were primarily screened based on their title, and subsequently using the
priorities at a WWTP-scale challenging for decision makers. Therefore, abstract and content in full, to establish the final list of articles. Tech
rigorous indicator selection and usage is needed to ensure DSS KPIs align nology selection DSSs were excluded if used for geographic location
with stakeholder goals at every level of decision making, or else WWTPs planning, source selection, resource allocation, performance assess
are at risk of undesirable future impacts and events. ment, or operation monitoring, and did not utilise multiple indicators for
decision making. Process optimisation DSSs were excluded if only used
2.2. Article collection method for performance monitoring, fault-detection, visualisation tasks, load
prediction, or sensor utilisation, and did not use multiple indicators to
2.2.1. Research question optimise control parameters. An additional six relevant articles were
There have been recent publications which discuss multi-criteria collected from a review paper by Mannina et al. (2019) investigating
analysis (MCA) DSSs for the wastewater sector (Ddiba et al., 2023; DSSs for WWTPs, to incorporate appropriate literature from outside the
Mannina et al., 2019), however they allude to issues that exist for the search time series.
assessment and selection of technologies. Mannina et al. (2019) states
that ‘sustainable aspects are incorporated in accordance to DSS developers, 3. Technology selection DSSs
as there is no standard that can be applied while developing the systems’,
whilst Ddiba et al. (2023) concludes that some sustainability implica The decision to invest in new technology at a WWTP is a complex and
tions are not adequately covered by decision support tools. This shows multi-faceted decision to fulfil business, sustainability, and regulatory
that a lack of standardisation has resulted in the development of targets. MCDM tools have been developed for this purpose, however,
indicator-based methodologies that do not fully consider the sector's there is often little emphasis on linking the goals of the assessment with
sustainability goals. However, the wastewater sector must meet the re indicator selection, weighting, and scoring methods. This potentially
quirements set out in Section 2.1 in the coming years, therefore, this leads to outcomes that do not truly satisfy all stakeholder and decision
review systematically analyses the specific indicators selected, and how maker goals at regional, national, utility, community, or WWTP scales.
they are used by DSSs, to understand the impact on WWTP outcomes. Literature collected in Section 2 is reviewed to understand conventional
This results in the research question of how are indicators selected and methods and highlight good practices regarding alignment of KPIs with
utilised in decision support tools for technology selection and process opti DSS goals. Table 1 summarises the MCDM WWTP technology selection
misation at WWTPs, and to what extent are sustainability and circularity DSSs collected from literature, resulting in a total of thirty-one articles.
pillars harmonised to meet decision maker goals? Additionally, the need to Table 1 summarises DSS properties namely the technologies selected,
construct standardised DSS procedures to facilitate sustainability out aim, case study of application, and categories used to group assessment
comes is highlighted, thus following literature review recommendations indicators. The four main technology groups selected using MCDM DSSs
are provided to act as the starting point for this. The types of MCA used are; wastewater treatment (WWT), sewage sludge treatment (SST),
to facilitate complex decision making have already been the subject of water reuse (WR) and resource recovery (RR), or a combination thereof.
systematic reviews (Kozłowska, 2022), meaning the methods available Since 2018, the development of DSSs for the selection of RR technologies
in literature have already established. Therefore, they do not require has emerged as a priority for decision makers. The aim of each DSS has
further generalised study and is why the focus of this review is on DSSs been directly quoted from the source, as this is key to understanding
implemented for wastewater technology assessment to understand cur specific goals of the DSS when selecting appropriate indicators to
rent practices and provide recommendations for improvement. facilitate desired outcomes. Lastly, the categories defined when select
ing indicators are provided, which is important for relating DSS goals to
2.2.2. Search strategy selected KPIs for technology assessment. The assessment category col
The evaluation of WWTP DSSs was completed using systematic re umn in Table 1 highlights the popularity of using economic, environ
view, following the guidelines of the Preferred Reporting Items for mental, social, and technical sustainability dimensions to group
Systematic Review and Meta-Analyses (PRISMA) method (Page et al., indicators. Steps of the reviewed MCDM DSSs are summarised in Fig. 2,
2021). Articles describing multi-criteria, indicator-based DSSs for including examples at each stage from the reviewed literature.
WWTP technology selection and process optimisation were collected The thirty-one papers developing MCDM technology selection DSSs
from Scopus (www.scopus.com) and Web of Science (www.webofscienc were categorised in Table 1 based on the type of technology being
e.com) databases. The configuration of this review required two inde assessed. Selection of WWT technologies is the most common with
pendent searches to collect data using a combination of Boolean con fifteen DSSs, as WWT decision making is complex so selection of treat
nectors, and a previous review in the area by Mannina et al. (2019) ment methods is challenging. RR is the second most common focus,
established the time series of 2018–2022. MCDM technology selection which can be attributed to the emphasis placed on selecting sludge
3
Fig. 1. Flowchart of the steps taken during the article selection procedure (Page et al., 2021).
treatment technologies for recovery purposes (usually energy). This most are not used across multiple case studies. Next, nine DSSs aim for
highlights the desire of decision makers to make use of a resource that the identification/selection/prioritisation/recommendation of technologies
was previously considered a waste during WWT, reflecting the modern for a specific function, including non-potable water reuse or resource
objective to enhance circularity of the sector. Four DSSs for WR tech recovery strategies. Although this instructs the user with regards to the
nology selection have been developed, acknowledging that due to global expected function of selected technologies, it does not provide any
warming, water stress is being exacerbated for many people, requiring justification as to the reasoning for their selection. Third, the aim of
more efficient use of water by decision makers. Lastly, two DSSs focused seven DSSs is to select sustainable or assess the sustainability of alter
solely on the selection of technologies for SST, whilst only one was natives. This is not useful unless a vision of sustainable wastewater
developed that combined technology selection for WWT and SST. It is treatment is defined by the DSS developers, as users cannot fully un
critical to list the type of technologies being selected by a DSS so that the derstand how to assess and compare the sustainability of alternatives
correct assessment criteria or indicators can be integrated. This is re (targets summarised in Section 2.1 are useful here). Evaluation/analysis
flected by the low number of WWT/SST DSSs, as it is difficult to select of technologies utilising specified criteria, such as environmental or
criteria that are suitable for the assessment of both treatment technol economic aspects is another common DSS aim, with three identified
ogies since their goals and expected outcomes differ. from the collated list. These highlight the assessment criteria used to
select technologies but does not provide the user with adequate
3.1. DSS goals reasoning of why they should implement the technologies. Finally, two
DSSs aim to optimise or find the optimum solution, which is difficult to
To ensure that selected technologies will result in the benefits ex comprehend unless the objectives being optimised are explicitly defined.
pected by stakeholders and decision makers, the aim of DSS application Without a clear definition of DSS aims, there is a disconnect in user
must be clearly defined. As shown in Table 1, the most common aims knowledge, as the aim is key for understanding why a DSS is imple
used vague and generic language for the selection of the most suitable/ mented and selecting the correct indicators to facilitate desired out
appropriate/viable/best-fitting technology in ten DSSs. Although this re comes or water sector targets. Therefore, vague language must be
veals the intention of the DSS, rarely are these terms explained in a way mitigated, and complete definition of aims is encouraged from DSS de
that enables the user to understand what these ‘suitable’ technologies velopers to help users implement technology selection tools correctly.
may look like considering the scenario of application. This lack of di
rection limits wider utilisation of developed DSSs and could explain why
4
Table 1
Summary of wastewater treatment MCDM technology selection DSSs.
Author Year Group Aim Case study Assessment categories Weighting KPIs selected
method
Molinos- 2014 WWT Assess the sustainability of WWT 1500 PE WWTP Economic, AHP CAPEX, OPEX, removal efficiency,
Senante technologies Environmental, Social energy consumption, land use,
et al. sludge production, potential for RR
and WR, reliability, odours, noise,
visual impact, public acceptance,
complexity
Garrido- 2015 SST Identification and assessment of 1,000,000 PE Economic, Fixed Annual cash flow, annual income,
Baserba the most appropriate sludge WWTP Environmental total annual equivalent costs, GWP
et al. treatment technologies
Castillo et al. 2016 WWT Analysis of the alternatives Retrofit vs Economic, User Defined Nitrogen removal, CAPEX, OPEX,
through a multi-criteria construction of Environmental, CBA, LCA, noise, visual impact,
approach, considering WWTP in Italy Operational need for specialised staff, flexibility
operational, economic, and
environmental criteria
Chhipi- 2017 WR Evaluating the potentiality of fit- Comparing non- Economic, User Defined Microbial concentration,
Shrestha for-purpose wastewater potable water uses Environmental quantitative microbial risk
et al. treatment and specific reuse for for 10,000 PE assessment, development of
a community community alternative treatment trains,
estimation of reclaimed water
quantity and its distribution, LCC,
energy use, carbon emissions
An et al. 2018 SST Helping the decision- makers/ Three sludge Economic, AHP CAPEX, OPEX, land use,
stakeholders to select the most management Environmental, environmental risk, resource
sustainable technology among strategies Social, Technical utilisation efficiency, operability,
multiple scenarios site selection, applicability, and
management level requirement
Arroyo and 2018 WWT Choice of the most sustainable Seven small-scale Economic, CBA CAPEX, OPEX, removal efficiency,
Molinos- WWT alternative WWT technologies Environmental, Social energy consumption, land use,
Senante sludge production, potential for RR
and WR, reliability, odours, noise,
visual impact, public acceptance,
complexity
Sadr et al. 2018 WR Selection of WWT technologies Large WWTPs in Economic, AHP CAPEX, OPEX, energy
for non-potable water reuse Brazil and Greece Environmental, consumption, environmental
applications in different contexts Social, Technical impact, community acceptance,
adaptability, ease of construction
and deployment, land requirement,
complexity, water quality
Oertlé et al. 2019 WR Promote water reuse in regions Thirteen treatment Economic, Technical, User Defined CAPEX, OPEX, distribution costs,
where it is still an emerging trains in different Requirements and energy demand, chemical demand,
concept locations Impacts odour generation, sludge
production, land required,
groundwater impact, reliability,
ease of upgrade, adaptability, ease
of operation/ construction/
demonstration
Đurđević et al. 2020 SST/ Technology selection for sewage WWTP planned for Socio-economic, AHP Material stabilisation, energy reuse,
RR sludge energy recovery Rijeka, Croatia Environmental, nutrient recovery, commercial
Technical acceptance, product transport/
storage, GHG reduction, pre-
treatment requirements, hazardous
by-products, heavy metal content,
public acceptance, OPEX, CAPEX,
labour requirements, energy
savings, societal contribution
Ali et al. 2020 WWT Evaluate and prioritise different Five WWT Undefined VIKOR Cost, land requirement, processing
wastewater treatment alternatives in time, manpower requirement,
technologies used in Pakistan Pakistan efficiency, environmental impact,
energy consumption, sludge
production, safety, chemical
requirement
Gherghel et al. 2020 WWT/ Identify the most suitable Large WWTP of Economic, AHP GHG emissions, running costs,
SST treatment scheme for the 720,000 PE in Italy Environmental, service landfill surface, electricity
management of wastewater and Energy, Land Use consumption, planimetric size,
sludge biorefinery capabilities, landfill
requirements
Chrispim et al. 2020 RR Support decision-making on WWTP in Sao Paulo Economic, Social, N/A Recovery potential, maturity,
resource recovery strategies; to serving 1.4 million Environmental and resource utilisation, skilled labour
recommend operational and PE Technical, Political requirements, product quality,
technological strategies positive environmental impact,
CAPEX, OPEX, revenue, logistics,
acceptance, accordance with policy
and legislation
(continued on next page)
5
Table 1 (continued )
method
Liu et al. 2020 WWT Optimise the sewage treatment Town in Liao River Economic, AHP Construction cost, land cost, OPEX,
technologies and their Basin, China Environmental, Social removal rate, life expectancy,
combination technologies stability, resource recovery,
simplicity, ecological values, risk
assessment
Ullah et al. 2020 WWT Assist decision-makers to select Two sources of Undefined N/A Odour, removal efficiency, land
suitable WWTTs from a set of wastewater in use, manpower, financial resources,
alternatives Islamabad, Pakistan time availability, chemical
availability, oxygen requirement,
sludge management and disposal
Palma- 2020 SST/ Selection of the best-fitting WWTP in Spain Regional Level, Plant Fixed Viability, material circularity, self-
Heredia RR sewage sludge valorisation Level, Process Level sufficiency, risk assessment, NPV,
et al. strategies removal efficiency, sludge
production, biogas production,
efficiency, CAPEX, OPEX
Ling et al. 2021 WWT Assess and compare the Seven WWT options Economic, AHP Energy requirement, land use,
sustainability of different in UK Environmental, pollutant removal, sludge
wastewater treatment options Social, Resilience production, RR potential, GHG
emissions, public acceptance,
odour, noise, visual impact,
reliability, complexity, CAPEX,
OPEX
Fetanat et al. 2021 WWT/ Prioritise energy recovery from WW management in Water Security, LAM Water security (access, safety, and
RR wastewater treatment Behbahan City, Iran Energy Security, Food affordability), energy security
technologies Security (availability, accessibility,
affordability, acceptability,
applicability, and adaptability),
food security (availability, access,
utilisation, and stability)
Büyüközkan 2021 WWT Evaluate the most suitable WWT WWT selection for a Economic, AHP Water/energy/discharge/chemical
and Tüfekçi decision system company in Environmental, costs, monitoring, waste
Istanbul, Turkey Technical, production, environmental
Administrative, benefits, facility management, NPV,
Management volumetric capacity, water quality,
applicability and performance,
reliability and sustainability
Lizot et al. 2021 WWT Evaluation of WWT systems Twenty WWT Economic, AHP CAPEX, OPEX, NPV, Land,
considering relevant economic, options for a Environmental, manpower, reliability, replicability,
social, technical, and sanitation company Social, Technical complexity, removal efficiency,
environmental criteria in Brazil sludge production, GWP,
acceptance
Sucu et al. 2021 RR Find the optimum treatment Large and small Economic, User Defined Annual cost, potential income,
train consisting of compatible WWTP recovering Environmental, acceptability, affordability, land
unit processes which can recover irrigation water Social, Technical area, odour, noise, flexibility
water, energy and/or nutrients
de Almeida 2021 WWT Develop and apply a Benevente River Operational, Multi Attribute Removal efficiency, energy
et al. methodology for sewage watershed in Brazil Technical, Utility Theory demand, land use, CAPEX, OPEX,
treatment systems selection Environmental, Social sludge treated, sludge disposed,
reliability, simplicity, resistance,
odour, noise, aerosol generation,
insect attraction
Eseoglu et al. 2022 WWT Technology selection problem for Four WWTPs Economic, AHP Energy consumption, sludge
wastewater treatment that >100,000 m3/ Environmental, production, reuse of treated water,
integrates all aspects of d Istanbul, Turkey Social, Technical capital cost, land required, OM cost,
sustainability with the energy saving, sludge disposal cost,
behavioural characteristics of removal eff, maturity, simplicity,
decision makers applicability, replicability,
flexibility, reliability, odour,
manpower needed, social
acceptance, social benefit, aesthetic
Leoneti et al. 2022 WWT Choosing a WWTP for a Six 40,000 PE Economic, Social, Game Theory Cost, effluent quality, land area,
municipality WWTP alternatives Environmental (rank order sludge production
in Brazil centroid)
Liu and Ren 2022 SST/ Promote the sustainable Four sludge-to- Economic, BWM Climate change, acidification,
RR decision-making process of energy options Environmental, eutrophication, net cost, social
sludge management Social, Technical acceptance, governmental support,
educational significance, odour,
complexity, maturity, accessibility
Attri et al. 2022 WWT Sustainability assessment of Six alternatives for Economic, Fuzzy Stepwise Removal efficiency, effluent DO
wastewater treatment secondary WWT Environmental, Weighted and coliform, NP removal
technologies Social, Functional Assignment capabilities, area, power
Ratio Analysis requirement, OPEX, CPAEX, odour,
noise, visual impact, flexibility,
reliability, ease of operation, FOG
tolerance, waste sludge production
(continued on next page)
6
Table 1 (continued )
method
Renfrew et al. 2022 RR Identification of strategies for Priority resource Recovery, Market, User Defined RR potential, market, treatment,
resource recovery from identification for Cost, Carbon, cost, carbon, 6 capitals
wastewater UK water sector Treatment Impacts, 6
Capitals
Nkuna et al. 2022 SST/ Selection of the most viable Three technologies Economic, Technical AHP Dependability, maturity, sludge
RR thermochemical technology to converting WW use, energy production, energy
handle municipal WWS for sludge to energy consumption, CAPEX, OPEX
energy recovery
Južnič-Zonta 2022 RR Given a set of resource recovery Medium size WWTP Economic, User Defined Effluent quality, costs, maturity,
et al. and wastewater treatment in Manresa, Spain Environmental, GHG emissions, area
process units, quickly determine Technical
the best plant configuration
Silva Junior 2022 WWT Select the most appropriate WWT in urban and Economic, Socio- User Defined Area demand, quality performance,
et al. technologies for wastewater rural municipalities Environmental, mechanisation rates, electric power
treatment in Brazil Technical consumption, CAPEX, OPEX,
operational complexity, BOD
removal
Srivastava and 2022 WR Selection of an appropriate WWT technologies Economic, Full Consistency CAPEX, OPEX, land use, energy
Singh wastewater treatment chain that for water reuse in Environmental, Method consumption
produces effluent suitable for the Kanpur, India Technical
defined reuse
Salamirad 2023 WWT Select the most appropriate Seven WWTP Economic, Social, BWM Investment cost, reliability,
et al. municipal WWT technology alternatives in Iran Environmental efficiency, volume dependency,
requirement for additional
treatment, energy consumption,
sludge production, odour,
workforce requirement, law and
regulation compliance, salinity
removal, bacteria removal
Fig. 2. Generic steps of MCDM technology selection DSSs, including examples and techniques available for use at each stage.
7
3.2. Indicator selection selection to achieve desired outcomes by utilising external expert or
stakeholder opinions, for example to screen assessment criteria from a
As discussed, selection of assessment indicators or criteria when longlist identified during literature review (Ali et al., 2020; Salamirad
using any WWTP DSS is crucial to ensure the chosen technology fulfils et al., 2023). Ling et al. (2021a, 2021b) developed a method starting
decision maker and stakeholder goals. Therefore, methodologies with a round of literature review to collate indicators previously used to
implemented for indicator selection by DSSs are scrutinised and sum evaluate WWT performance. The list is then refined based on key ter
marised in Fig. 3. minology mentioned during interviews (thematic analysis using Nviva
Fig. 3 shows that it is common for DSS developers to self-select in software) with water company employees utilising the DSS. Eseoglu
dicators or provide a list from which users can choose indicators, with et al. (2022) employs the use of a questionnaire study by experts from
little methodological explanation given (An et al., 2018; Castillo et al., across many roles in WWTPs from design to operation, and combines
2016; Chhipi-Shrestha et al., 2017; Chrispim et al., 2020; Fetanat et al., this with other information including effluent discharge regulation,
2021; Garrido-Baserba et al., 2015; Gherghel et al., 2020; Južnič-Zonta environmental impacts, and design parameters. These DSSs acknowl
et al., 2022; Renfrew et al., 2022; Srivastava and Singh, 2022; Sucu edge that indicator selection is an important part of strategic MCDM,
et al., 2021). This results in a significant gap in DSS user knowledge, as and the combination of stakeholder views with technical appraisal of
they are unable to reason whether the selected indicators are relevant to local factors enables the user to select indicators which adequately
their scenario of application. Data availability can be used to guide in reflect their goals. Fig. 3 highlights that these more robust indicator
dicator selection, relying on primary data where possible or secondary selection methods are less popular, helping to answer the research
data acquired through reasonable effort, such as modelling, whilst question by reporting a lack of robust methods for indicator selection in
meeting data quality requirements. To improve the robustness of indi most of the DSSs developed for WWTPs.
cator selection, some authors define criteria or provide additional jus The specific indictors selected showed that only two DSSs did not
tifications to choose appropriate indicators from literature (Arroyo and utilise economic indicators (Chrispim et al., 2020; Fetanat et al., 2021),
Molinos-Senante, 2018; Attri et al., 2022; Liu and Ren, 2022; Molinos- with most the common being capital and operating expenditure, whilst
Senante et al., 2014; Nkuna et al., 2022; Palma-Heredia et al., 2020; others chose life cycle costing (LCC) (Chhipi-Shrestha et al., 2017) and
Sadr et al., 2018). For example, Molinos-Senante et al. (2014) reasons net present value analysis (Lizot et al., 2021). Removal efficiencies of
indicator selection using transparent, representative, relevant and quanti regulated wastewater constituents, including total suspended solids
fiable evaluation criteria, however, definitions of these terms are not (TSS), chemical oxygen demand (COD), biological oxygen demand
provided potentially resulting in ambiguity for the user. (BOD), nitrogen, and phosphorus, were commonly selected to determine
More complete approaches conducted structured literature reviews treatment performance (Arroyo and Molinos-Senante, 2018; Eseoglu
for indicator selection (da Silva Junior et al., 2022; Leoneti et al., 2022; et al., 2022; J Ling et al., 2021b; Liu et al., 2020; Molinos-Senante et al.,
Lizot et al., 2021). Lizot et al. (2021) describes the terms entered into 2015; Silva Junior et al., 2022). Indicators of environmental perfor
literature search engines to collect assessment criteria utilised by other mance covered GHG emission (Gherghel et al., 2020; Južnič-Zonta et al.,
WWT MCDM tools, and then lists specific information and data avail 2022; Jiean Ling et al., 2021a), carbon footprint (Chhipi-Shrestha et al.,
ability requirements applied to create indicator shortlists. However, 2017; Renfrew et al., 2022), and life cycle assessment (LCA) (usually
only a short description of shortlisting steps is given which focuses on eutrophication, climate change, and acidification) impacts (Castillo
technical aspects (such as plant load, location, or size), rather than et al., 2016; Lizot et al., 2021). Effort was made to consider the social
sustainability goals. Alternatively authors used knowledge of local fac impacts of technologies, commonly their odour and noise aspects
tors to select appropriate DSS indicators from literature (de Almeida (Eseoglu et al., 2022; Oertlé et al., 2019; Sucu et al., 2021), whilst some
et al., 2022; Oertlé et al., 2019). Đurđević et al. (2020) utilised their own quantified microbial (Chhipi-Shrestha et al., 2017) and ecological risks
judgement to select DSS indicators considering the state of wastewater (Liu et al., 2020). In most cases, circularity indicators were combined
and sewage sludge management, socio-economic standards, and avail with environmental KPI sets, including water reuse (Eseoglu et al., 2022;
able data (from national databases) in the local area. Liu et al. (2020) Lizot et al., 2021), resource or product recovery potential (Chrispim
provides an explanation of the local context for each indicator provided, et al., 2020; Renfrew et al., 2022), and material circularity (Palma-
such as using economic costs as the project may need some financial Heredia et al., 2020). Lastly, technology energy consumption was one of
support from the community or process simplicity due to the lack of pro the most commonly selected indicators, however, only a few DSSs
fessionals for operation. This strategy encourages the DSS user to consider renewable energy (Lizot et al., 2021), energy reduction
consider local factors during decision making, however, a more robust (Durdević et al., 2020), or self-sufficiency (Palma-Heredia et al., 2020)
approach is to use local stakeholder perspectives as well. dimensions.
Some DSS developers recognise the importance of rigorous indicator From this it is clear that DSS developers select indicators from across
the triple bottom line to support sustainable performance, but there is a
gap in terms of facilitating sustainability targets and circularity assess
ments. Few KPIs are explicitly selected to quantify progress towards the
high-level water sector targets of Section 2.1 by failing to link indicator
selection with targets such as GHG reduction, phosphorus recovery, or
energy neutrality. This even includes those DSSs with the aim of
selecting technologies for sustainable and circular actions, such as water
reuse or energy recovery.
3.3. Indicator categorisation
Often criteria or indicators are categorised to show user assessment

priorities and indicate potential benefits or impacts of selected tech
nologies. Table 1 defines the DSSs categories employed to separate in
dicators and shows that twenty of the thirty-one DSSs utilise discrete
sustainability pillars. The most popular combination uses four envi
Fig. 3. Methods of indicator selection used by MCDM technology selection ronmental, economic, social, and technical (assumed interchangeable
DSSs, with the number DSSs using each method in bold. with functional, operational, or resilience) categories. Nine DSSs used a
8
combination of other categories defined by the developers, and some did

use sustainability pillars, however, they were combined to create hybrid
socio-economic or -environmental categories. Many DSSs also utilise
circularity KPIs, however, as mentioned in Section 3.2 they are cat
egorised as environmental indicators. This is worrying as enhancing the
circularity of wastewater resources does not directly correspond to
improved environmental performance. This leads to a significant gap in
decision maker knowledge as circularity indicators are being used to as a
substitute for sustainability impacts. Therefore, DSSs with circularity
objectives, such as resource recovery, need standardised assessments
that use CE indicators to evidence enhanced resource circularity, sup
ported by sustainability analysis to quantify wider benefits. This will
facilitate technology selection that simultaneously meet the water sector
sustainability and circularity targets detailed in the European Green
Deal and CEAP.
Some DSS developers created hybrid categories including socio-
Fig. 4. Methods used to weight indicators for MCDM technology selec
economic (Đurđević et al., 2020) and socio-environmental (Silva Ju
tion DSSs.
nior et al., 2022), or combined environmental and technical categories
together (Chrispim et al., 2020). This suggests authors may be unsure as
to which categories some indicators belong. This is further perpetuated and then a weighted summation is simply calculated to create a com
by authors placing the same indicators in different sustainability pillar posite indicator used to analyse technologies. Garrido-Baserba et al.
categories. For example, WWT technology removal efficiencies have (2015) and Palma-Heredia et al. (2020) develop DSS weights for in
been placed in environmental sustainability (Liu et al., 2020; Lizot et al., dicators that are predetermined and fixed, and given equal weighting
2021; Molinos-Senante et al., 2014) and technical categories (Eseoglu respectively. Although this simplifies DSS usage, this weighting system
et al., 2022; Silva Junior et al., 2022). This may explain the increase in is not recommended as it does not provide users with the ability to tailor
popularity of using the four pillars of sustainability for categorisation in KPI impacts to reflect their goals, which can be viewed as undermining
recent years, as it enables delineation of operational and environmental the principles of MCDM. Another method commonly employed by DSS
KPIs, highlighting the desire of decision makers to understand the developers is to allow users to define weights themselves, (Castillo et al.,
environmental impacts of potential technologies more clearly. However, 2016; Chhipi-Shrestha et al., 2017; Južnič-Zonta et al., 2022; Oertlé
this seems to result in some confusion regarding the objectives of certain et al., 2019; Renfrew et al., 2022; Silva Junior et al., 2022; Sucu et al.,
indicators, such as GHG/carbon footprint, as Đurđević et al. (2020) 2021), however, DSS users can be faced with >20 indicators so assigning
defines this as a technical indicator, whereas Lizot et al. (2021) utilises it weights without a structured methodology of comparing indicators can
as an indicator of environmental performance. Similarly, odour and lead to inconsistencies during analysis. This can result in indicator
noise indicators are placed in both environmental (Sucu et al., 2021) and weightings that do not align adequately with aims and lead to bias in the
more commonly social categories (Eseoglu et al., 2022; Lizot et al., assessment. Therefore, techniques are employed by DSS developers
2021; Molinos-Senante et al., 2014). These differences evidence the enabling structured analysis of indicators using opinions of experts and
need to enhance sustainability/circularity assessment knowledge and stakeholders.
develop standardised methods for KPI selection and categorisation. The analytical hierarchy process (AHP) is the most common
The popularity of indicator categorisation using sustainability pillars weighting method used by reviewed DSSs with eleven. AHP was pro
has led to some DSS developers using this method even when their posed by Saaty (1987) for decision making influenced by multiple in
defined aims do not refer to sustainable technology selection. To combat dependent factors (Liu et al., 2020). It investigates the relationship
this, some authors generated their own indicator categorisation strate between criteria to create a hierarchy from which they can be prioritised
gies. Fetanat et al. (2021) developed indicator categories using the (Eseoglu et al., 2022), often utilising external experts and stakeholders
water-energy-food nexus framework to view wastewater as a renewable to create pair-wise comparisons. Therefore, many DSSs reviewed use
energy source, which aligns with the DSS goal to prioritise energy re standard AHP for weighting indicators (Đurđević et al., 2020; Gherghel
covery from WWT. Palma-Heredia et al. (2020) created an indicator et al., 2020; Jiean Ling et al., 2021a; Lizot et al., 2021; Molinos-Senante
hierarchy depending on the scale of the application, therefore allowing et al., 2014; Nkuna et al., 2022). However, Ling et al. (2021a, 2021b)
decision makers at regional, WWTP, and operational levels to prioritise reported rarely seeing extreme scores on the judgement scale, and when
certain indicators. Although these categorisation methods are not as the full scale was used the threshold consistency ratio (compares the
established in literature as sustainability pillars, developing indicator weighting matrix against a random matrix, acceptable value of ≤0.1) is
categories which consider DSS goals may be a more effective way for often not achieved. To overcome the uncertainty due to imprecise
users to understand the indicators required to achieve their aims, human judgements or ambiguity, fuzzy logic is implemented (Eseoglu
especially whenever sustainable technology selection is not the objec et al., 2022). Many DSSs employ fuzzy-AHP weighting (An et al., 2018;
tive. However, it can be concluded there is confusion when categorising Büyüközkan and Tüfekçi, 2021; Eseoglu et al., 2022; Liu et al., 2020;
selected indicators and how this activity aligns DSS outcomes with de Sadr et al., 2018), providing a structured method of indicator weighting
cision maker goals. whilst mitigating inconsistencies of human thinking.
Apart from AHP, DSS developers integrated a variety of weighting
3.4. Indicator weighting methods (Ali et al., 2020; Attri et al., 2022; de Almeida et al., 2022;
Fetanat et al., 2021; Leoneti et al., 2022). Arroyo and Molinos-Senante
Weighting of indicators is a critical stage for DSS users, as it enables (2018) implement Choosing-By-Advantages (CBA), citing several im
them to prioritise or mitigate criteria depending on their objectives. provements over AHP including that it does not assume linear trade-offs
Therefore, the frequency of each technique used by DSS developers for between criteria. CBA encourages DSS users to understand the differ
indicator weighting is provided in Fig. 4. The MCA discussed are those ences between criteria and assesses the importance of these differences,
currently employed by water sector DSSs and does not reflect best as supposed to AHP which can create conflicting questions. The Best-
practices for multi-attribute decision making. Worst Method (BWM) used by Liu and Ren (2022) and Salamirad
It is common for indicators to be weighted according to the DSS user, et al. (2023) provides a simpler weighting step for decision makers as the
9
number of comparisons is reduced, improving the consistency ratio of qualitative indicators to investigate social aspects, local stakeholder
results and removing much of the uncertainty during pairwise com views should be used to score technologies, due to their greater under
parisons. Srivastava and Singh (2022) simplify weighting even further standing of potential impacts in a given region.
by employing the Full Consistency Method, minimising the number of
comparisons to achieve consistent results. 3.6. Ranking
Lastly, some DSS developers recommend the use of ‘experts’ without
actually defining whom this might include (Attri et al., 2022; Liu et al., The final step is to rank technologies for selecting the technology
2020), collecting opinions from stakeholders with little knowledge of which supposedly best meets user requirements. Palma-Heredia et al.
the investigated system or local area, leading to inconsistent results. (2020) presents KPI results and recommends the DSS user to complete
Whereas Eseoglu et al. (2022) utilises expert opinions from every stage pairwise comparisons for technology selection. Although this is a simple
of WWT including design, construction and operation engineers, and method of completing the final ranking, extensive indicator lists create
Gherghel et al. (2020) acknowledges the viewpoints of stakeholders complexity and inconsistencies in user judgement. Therefore, the most
from six different specialities, such as political, environmental, and common method of technology ranking employed by DSS developers is
plant operator stakeholders to ensure the holistic collection of view to create a composite indicator using the weighted sum method (Castillo
points. Therefore, stakeholders with a range of expertise that understand et al., 2016; de Almeida et al., 2022; Garrido-Baserba et al., 2015;
local factors for indicator weighting should be used to reduce bias and Gherghel et al., 2020; Liu and Ren, 2022; Molinos-Senante et al., 2014).
inconsistency. This synthesises indicator scores and their corresponding weights into a
Generally, the majority of DSSs in this study rely on AHP to weight single performance index used to rank and select technologies (Jiean
criteria, which is corroborated by other reviews in the area (Kozłowska, Ling et al., 2021a).
2022; Zolghadr-Asli et al., 2021), potentially incorporating high levels In the cases where multiple experts or stakeholders are used to
of uncertainty. Therefore, to ensure better indicator utilisation fuzzifi weight or score assessment indicators, systematic analysis of results is
cation and consultation of relevant experts can be used to reduce un needed to rank and select technologies. The technique for order of
certainty. Additionally, methods recommended in literature, not utilised preference by similarity to ideal solution (TOPSIS) is commonly coupled
by the water sector DSSs reviewed, to mitigate weighting procedure with AHP. TOPSIS selects the best alternative based on the shortest
errors are the entropy method for objective weight assignment or distance to the ideal solution and the farthest distance from the negative-
analytical network process (ANP) to account for correlations between ideal solution in geometric terms (Južnič-Zonta et al., 2022), to intensify
criteria (Zolghadr-Asli et al., 2021). the correctness and validate selection of the most appropriate technol
ogy (Nkuna et al., 2022). When fuzzification of data has occurred during
3.5. Indicator scoring indicator weighting, to improve the robustness of outcomes, this can be
continued to complete fuzzy-TOPSIS (Attri et al., 2022; Büyüközkan and
A range of methods to score assessment indicators have been utilised Tüfekçi, 2021; Eseoglu et al., 2022; Liu et al., 2020; Sadr et al., 2018).
due to the variety of scales and units of indicator results, and often the Another method employed to overcome the uncertainty of comparative
mix of quantitative and qualitative indicators selected. Linguistic (such analysis, is fuzzy-VIKOR as used by Ali et al. (2020), which utilises
as very bad to very good) or numerical (can be from 0 up to 10) series are positive and negative characteristics to define compromises when con
commonly integrated to normalise results enabling their combination. flicting views cause issues with decision making. This is achieved by
Several DSSs rely on the experts used for indicator weighting to assign calculating three variables to establish the summation and maximum
numerical ratings directly based on their opinion (Đurđević et al., 2020; distance from the best value, which are then combined to calculate an
Jiean Ling et al., 2021a; Renfrew et al., 2022), usually when there is lack overall score.
of empirical data (Jiean Ling et al., 2021a). Alternatively, Fetanat et al. Alternatively, Leoneti et al. (2022) implements game theory to
(2021) relied on linguistic terms to rate technology alternatives as the determine the preferred option from the list of acceptable outcomes,
indicators selected were immeasurable (such as energy security selecting the technology that maximises the Nash equilibria social
availability). welfare function. Lastly, Fetanat et al. (2021) utilised the linear
Literature searches were used to establish numeric ranges of indi assignment method (LAM) to rank technologies for energy recovery
cator results for each technology assessed (Attri et al., 2022). Silva Ju from WWTs. This method is chosen as it ranks alternatives according to
nior et al. (2022) collected data from technical-scientific literature conflicting criteria, by analysing the trade-offs between the ranking of
relevant to case study location and assigned the final result by calcu each indicator for each technology. Therefore, LAM may be beneficial as
lating the mean of the data range found. Before combination of indicator wastewater and sewage sludge treatment shifts to prioritise other
results, they were normalised to a value between 0 and 1 using the functions, such as resource recovery or water reuse. Studies in this area
lowest and highest value observed for each parameter. Rather than agree that TOPSIS is the most common ranking procedure (Štilić and
quantitatively normalising values collected from literature, Liu and Ren Puška, 2023), however, best practice depends on the scenario of appli
(2022) utilised a linguistic scale of five from very good to very poor, whilst cation. Methods such as PROMETHEE and ELECTRE are suited to handle
Lizot et al. (2021) created ranges for each indicator to assign a numeric conflicting stakeholder priorities (Štilić and Puška, 2023), whereas fuzzy
value to normalise quantitative indicator scores. logic and use of experts from as many specialities as possible should be
Lastly, a common method for assigning scores to indicators is to used to tackle subjective ranking issues (Garcia-Garcia, 2022).
directly quantify results (except for the indicators which are inherently
qualitative). It was observed that most environmental and economic 3.7. Uncertainty
indicators were quantifiable, whilst technical and social indicators were
qualitatively scored (Castillo et al., 2016; Leoneti et al., 2022; Liu and There are various types of uncertainty that exist in MCDM that can
Ren, 2022; Molinos-Senante et al., 2014). Quantitative calculation of arise at each step of DSS utilisation resulting from: variation, ambiguity,
each indicator investigating technology performance is recommended, and incomplete preferences of human inputs; lack of system, parameter,
as it incorporates specific details and local factors of the case study. data, external factor, or model knowledge; and prediction of outcomes
Relying on the judgement of DSS users or external experts enables un or future events (climatic or socio-economic changes) (Walling and
certainty through the ambiguity or bias of human decision making to Vaneeckhaute, 2020). There are many methods to deal with MCDM
incorrectly score technologies. Furthermore, the use of values extracted uncertainty, one being fuzzification of scoring, weighting, and ranking
from literature can mitigate the influence of local factors which can be procedures reliant on human judgement, as previously discussed in
pertinent for economic and technical indicators. Of course, when using Sections 3.4, 3.5, and 3.6. Alternatively, sensitivity analysis is able to
10
provide decision makers with insights into the uncertainty resulting Table 2
from erroneous modelling of the assessed system or potential/future Summary of issues, recommendations, and beneficial outcomes related to the
scenarios. reviewed MCDM technology selection DSSs.
Scenario investigation is a widely applied method of sensitivity Issue Recommendation Outcome
analysis, in which MCDM indicator weighting is altered to reflect Few DSSs provide a clear Defining the goal and Help decision makers
different viewpoints or future situations. For example, Molinos-Senante definition of aims or scope of the assessment understand the desired
et al. (2014) and Salamirad et al. (2023) conducted scenario analysis by goals should become common outcome of DSS
favourably weighting environmental, economic, and social KPIs in turn, practice, as the first step of utilisation
DSS development or
validating the selected technology (constructed wetlands and integrated
application
fixed-film activated sludge respectively) still ranked highest under Rigorous indicator Utilise participatory Technology selection
alternative weighting schemes. Alternatively, Renfrew et al. (2022) selection is often methods to incorporate using KPIs that
improved the robustness of technology selection by weighting KPIs overlooked by DSS local stakeholder, business adequately reflects
based on potential future scenarios, including legislative changes for developers and do not (water utility), and desired results and
consider high level regional/governmental facilitate sector
emissions compliance and carbon footprint reduction, and selecting water sector goals objectives transformation
technologies based on their average performance across the scenarios. Indicator categorisation Use categorises that reflect Help to select and group
Furthermore, global sensitivity analysis (GSA) was utilised to verify that is often unclear the intentions of the DSS, relevant indicators, such
technology ranking is robust to fluctuating inputs over a ±10 % range resulting in helping to create more as using sustainability
inconsistencies across robust weighting strategies pillars when selecting
and investigate which parameter's uncertainty have the largest impact
DSSs, mitigating and consider CE targets of sustainable
on MCDM outcomes, educating future assessments (Renfrew et al., circularity objectives the water sector technologies, and
2022). Lastly, Južnič-Zonta et al. (2022) aimed to use Monte-Carlo (MC) mitigate the alignment
simulations to overcome probabilistic uncertainty of bio-chemical of CE metrics with
modelling processes to configure design parameters, before technology sustainability impacts
Expert or user defined Stakeholders with an Ensures that DSSs select
ranking is calculated and verified over each iteration (however this was
weighting schemes can understanding of the local technology that will
not included in case study). Therefore, if potential errors are likely to be lead to a lack of local area from a range of job meet the local demands
introduced by MCDM structure or case study that impact outcomes, then factor consideration roles should be used for in each scenario of
sensitivity analysis (scenario or global) should be used to validate the indicator weighting application and reduce
uncertainty of results
robustness of DSS results.
Unstructured or Consider the specific issues Remove the
subjective weighting of each DSS application to inconsistency and
3.8. Recommendations and ranking methods decide which method reduce uncertainty that
can lead to uncertain should be used to reduce can arise when human
As discussed, there are already many reviews of DSS typologies in the outcomes uncertainty, such as inputs are used to weight
entropy methods to and rank indicators
literature, therefore, the review focuses on how indicator usage can be enhance the objectivity of
improved based on the methods currently implemented for WWTP weighting, and either fuzzy
technology selection. Therefore, following the review of thirty-one logic to reduce human
MCDM DSSs final recommendations and comments are provided in error or PROMETHEE/
ELECTRE to overcome
Table 2. Unfortunately, Sections 3.2 and 3.3 highlight the significant gap
conflicting priorities
related to the utilisation of circularity and sustainability indicators, during ranking
mainly that circularity aspects are used to investigate environmental There is little critical Techniques such as Ensure that the method
performance and the lack of alignment with water sector goals reported analysis of final sensitivity analysis should is consistent across
as part of European Green Deal and CEAP. Additionally, WWTP DSSs technology selection in be applied to investigate alternative scenarios,
relation to decision DSS outcomes enhancing robustness of
still rely on user defined weighting, scoring, and ranking procedures, or maker goals final technology
structured methods, such as AHP and TOPSIS, which have issues with selection
introducing uncertainty to the assessment. Generally, decision making
in the water sector is still some distance from standardisation and har
monisation of sustainability and circularity assessments. control. It is necessary to conduct distinct analysis of these DSS types as
It is worth noting that analysis of DSS case studies showed that they are utilised differently by decision makers. Therefore, alternative
economic indicators were commonly prioritised during the weighting methods and indicators are required, as it was seen that technology
stages (Eseoglu et al., 2022; Liu et al., 2020; Lizot et al., 2021; Sadr et al., selection DSSs focus on sustainability KPIs whereas operational opti
2018). The CBA method employed by Arroyo and Molinos-Senante misation DSSs target cost and regulatory (effluent quality) aspects.
(2018) excluded economic indicators during the initial assessment, Table 3 summarises the multi-objective process optimisation WWTP
prioritising environmental and social factors, as monetary resources DSSs collected from literature, resulting in the review of twenty-six
available are usually the constraint for any project. Environmental and articles.
social indicator results are then plotted against cost to facilitate the se Table 3 shows an increase in the number of publications in this area,
lection of the best technology option. Authors highlight the impacts of growing from four in 2018 to seven in 2022 which coincides with the
this by comparing AHP with CBA and showed that by considering eco availability of Benchmark Simulation Model (BSM) 1 and BSM2 (IWA,
nomic factors alongside environmental and social indicators, unfav 2018) for testing WWTP control strategies. Some authors have recently
ourable impacts were offset by low capital and operating costs. published multiple papers in this area, testing different algorithms to
Therefore, as governments demand improved environmental and social find the optimal control strategy on the same simulation platform. DSSs
performance of WWT in the coming years, to achieve targets such as net were categorised depending on their ability to optimise the control of
zero, the exclusion of economic indicators from initial assessment may process operation dynamically (respond to changes in real-time) or
be favoured. statically (user defined inputs followed by KPI calculation). Most DSSs
are dynamic, which corresponds with use of BSMs as time series data
4. Multi-objective optimisation control across three weather conditions is available for simulation testing (IWA,
2018). Generally, DSS aims were stated in clearer terms than those for
Following the selection of technologies, another type of DSS is technology selection, often stating which performance parameters or
needed for multi-objective optimisation of WWTP process operation and KPIs are targeted for optimisation. Most DSSs were not applied to real
11
Table 3
Summary of multi-objective DSSs for optimisation of WWTP operation.
Author Year Control Aim Application Objective function
Qiao et al. 2018 Dynamic Achieving the effluent quality (EQ) BSM1 EC and EQI
requirements and minimizing the energy
consumption (EC)
Díaz- 2018 Static Develop more sustainable water systems 2500 PE WWTP in Total connections costs, total freshwater use, and total regenerated
Madroñero Alicante, Spain freshwater use
et al.
Han et al. 2018 Dynamic Optimal control operation with EC reduction BSM1 EC and EQI
while retaining standard EQ
Qiao and Zhou 2018 Dynamic Acquire the balance between EC and EQ with BSM1 EC and EQI
the usage of the best set points
Qiao et al. 2019 Dynamic Suitable set-points to balance the treatment BSM1 EQI and EC
performance and the operational costs
Zhou and Qiao 2019 Dynamic Optimal control strategy is designed to reduce BSM1 EQI and OCI
EC without violating effluent standards
Pisa et al. 2019 Dynamic Reduction of the number of violations as well BSM2 EQI and OCI
as the improvement of WWTP's EQI and OCI
metrics
Dai et al. 2019 Dynamic Optimal modification of an ASM-2D EQ, operating cost, and total volume
anaerobic–anoxic/nitrifying/ induced
crystallization (A2N-IC) process
Borzooei et al. 2019 Static Evaluate and improve existing process 2 million PE EQI and ECI
performance in addition to optimise the Castiglione Torinese
production of renewable energy WWTP, Italy
Mannina et al. 2020 Static Optimization … in terms of operational costs Pilot plant MBR Effluent Fine, EQI (liquid and gas), oxygen-to-total-
and direct greenhouse gases emissions. Kjeldahl‑nitrogen ratio, ratio nitrate-ammonia, CO2 and N2O
emissions, and direct and indirect GHG emissions.
Revollar et al. 2021 Static Improving the eco-efficiency of WWTPs BSM2 EQI, OCI, Net energy, Excess heating energy, Electricity
consumption, Energy/Pollution removed, Energy net/Pollution
removed, Violations of the permit limits of effluent N, NH4 and
COD
Heo et al. 2021 Dynamic Operate at cost-efficient and sustainable BSM2 EQI, OCI, CH4 reutilised as energy source
WWTP
Ortiz- 2021 Dynamic Optimise an economic cost term and an BSM1 EQI and economic cost
Martínez effluent quality index
et al.
Han et al. 2021 Dynamic Achieve excellent treatment performance for a BSM1 and 10,000 m3/ EC and EQI
WWTP d WWTP Beijing, China
Tejaswini 2021 Dynamic Enhance the performance of the WWTP by BSM1 EQI and OCI
et al. optimising the parameters of the default
control strategy
Chen et al. 2021 Static Obtain sustainable control strategies 10,000 PE WWTP LCC and three LCA impact indicators (energy consumption,
Jiangsu Province, eutrophication, GHGs)
China
Campana et al. 2021 Static Reduce WWTP operating costs, improving at 86,400 PE WWTP, Italy Self-sufficiency ratio and net present cost
the same time treated effluent quality
Li et al. 2021 Dynamic Meet the requirements of effluent quality and BSM1 EC and EQI
maintain sustainable operation with the
lowest energy cost
Fox et al. 2022 Dynamic Best setup that can enable optimal Residential NH4 removal, prediction error, treatment time reduction
operational, environmental and energy development SBR
performance
Xie et al. 2022 Dynamic Achieve tracking control of the main operating BSM1 EC and EQI
variables of the WWTP
Niu et al. 2022 Dynamic Optimise EQ and EC in wastewater treatment BSM1 EC and EQI
process
Han et al. 2022 Dynamic Optimal control strategy is proposed to BSM1 EQI, pumping energy, aeration energy
improve the performance of WWTP
Caligan et al. 2022 Static Minimise the system's overall economic costs Wastewater sludge to Cost and GHG emissions
and environmental greenhouse gas emissions bioenergy park
F. Li et al. 2022 Dynamic Optimise the control of WWTPs BSM1 EC and EQI
Han et al. 2022 Dynamic Guarantee satisfactory EQ and EC with the BSM1 EC and EQI
excellent control accuracy of WWTP
Du and Peng 2023 Dynamic Optimal control of wastewater treatment BSM1 EC and EQI
process
case studies and instead utilised BSMs due to the complexity and non- manipulating the oxygen transfer coefficient, and nitrate level setpoints,
linearity of WWTP modelling. This reliance results in little variation of by changing the internal recycle rate in the fifth and second compart
KPIs selected (type or number), as BSMs have predefined indicators ments respectively (IWA, 2018). The performance assessment of the
related to effluent quality and energy consumption/cost. plant is based on two main KPIs; the effluent quality index (EQI) and
As shown in Fig. 5, the BSM1 plant is a 5-compartment activated overall cost index (OCI). The EQI is the weighted sum (weightings from
sludge reactor modelled using ASM1, with configuration facilitating literature) of effluent contaminant TSS, COD, BOD, Kjeldahl nitrogen
nitrification-denitrification for biological nitrogen removal. The model (TKN), and nitrate (NO). The OCI combines cost factors of sludge pro
utilises PI controllers to control the dissolved oxygen (DO) level by duction, aeration energy, pumping energy, mixing energy, and external
12
Fig. 5. BSM1 process flow diagram and control systems based on a figure from the IWA (2018) where Q is the flowrate, S is the set point, PI is the controller, and KLa
is the transfer coefficient.
carbon consumption (Alex et al., 2018). BSM2 utilises the same waste multi-objective optimisation calculation for EQI, OCI, and biogas gen
water treatment process, with the addition of sludge anaerobic digestion eration performance indicators, and finally the NSGA-II algorithm
to the balance heat energy required and energy generated from methane searches for the optimal setpoint of each controller. Therefore, the
production. The number of effluent limit violations and duration must WWTP can maintain optimal performance and respond to fluctuations in
be reported, meaning operation is constrained by discharge limits of influent composition. The use of DSSs for dynamic optimisation means
NH4 ≤4 mg/l; TN ≤18 mg/l; TSS ≤30 mg/l; BOD ≤10 mg/l; COD ≤100 that indicator selection must focus on KPIs that are calculated from data
mg/l (Alex et al., 2018). that is easily and reliably monitored over a given period.
4.1. DSS goals 4.3. Modelling platform
In Section 3 it was revealed that MCDM technology selection DSS Of the nineteen DSSs that dynamically control WWTP operation,
aims are too generic, meaning it is difficult to relate indicator selection fourteen are implemented in BSM1 without being tested on real pro
to desired outcomes. Many of the multi-objective process optimisation cesses. In most cases these DSSs are made up of two algorithms, one
DSSs reviewed in Table 3 take the opposite approach, as twelve stated responsible for the multi-objective optimisation of KPIs (commonly a
the KPIs targeted for optimisation in their aims. This definition enables neural network) and another for determining the set point of controllers
users to clearly understand the outcomes that can be expected when (such as a NSGA-II or AMODE algorithm (Heo et al., 2021; Ortiz-Mar
implementing this optimisation technique, however, many of these DSSs tínez et al., 2021; Qiao et al., 2019, 2018; Tejaswini et al., 2021)). The
relied on BSMs meaning there is little flexibility in the indicators uti repeated investigation of different algorithm combinations is necessary
lised. Another helpful method of defining DSS aims for the user is to to which results in the best EQI and OCI outcomes, and lowest controller
identify its specific function. For example, Borzooei et al. (2019) and error (Du and Peng, 2023). Two DSSs are used to control the operation of
Revollar et al. (2021) aimed to optimise the production of renewable energy BSM2, enabling users to optimise operation considering biogas pro
and improve the eco-efficiency of WWTPs respectively, making it clear duction as part of the OCI. Two DSSs are utilised for the dynamic control
to users the reasons for implementing this DSS and selecting indicators of actual processes including the work of Han et al. (2021) which runs
for optimisation. Still, a significant number of multi-objective optimi initial tests on BSM1 then uses data extracted from the SCADA system of
sation DSS developers use vague language when stating their aims. Eight a 10,000 m3/d plant in Beijing, China to run experimental tests. Lastly,
DSSs aim to either optimise or improve performance of WWTPs, whilst Dai et al. (2019) developed their own optimisation models, using ASM-
three DSSs aim for sustainable operation or control of WWTPs, without 2D to optimise a WWTP for inducing crystallisation. Therefore, few DSSs
explicitly stating which areas are targeted. Therefore, DSSs clearly have been tested on real systems so may not perform as expected when
define their aims, but few explicitly relate this to sustainability or applied at different scales or locations, especially under unexpected
circularity objectives, aiming to generally ‘optimise WWTP perfor influent loadings. It is recommended that users test DSSs in real systems
mance’ or improve conventional operation KPIs. or on models that represent the specific process it will be applied to,
ensuring optimisation reflects the operational expectations of decision
makers.
4.2. Static vs dynamic control
In the cases of static control, the DSSs developed usually rely on
simulation software or the development of process models. Four DSSs
Of the twenty-six DSSs reviewed in Table 3, six provided users with
used their own models which facilitated the selection and utilisation of
static control strategies for improving the operation of WWTPs, meaning
less conventional KPIs, including regenerated water usage (Díaz-
the results are used by operators to make decisions rather than the DSS
Madroñero et al., 2018), energetic self-sufficiency (Campana et al.,
dynamically altering operation. Borzooei et al. (2019) created a simu
2021), and environmental impact (kg CO2eq) (Caligan et al., 2022). Two
lation of a large-scale WWTP and altered the SRT between 10 and 40
DSSs simulated WWTPs in Hydromantis's GPS-X software, with config
days, then plots the EQI and EC results to establish the optimal SRT for
urations based on real-world processes (Chen et al., 2021) and fed with
process operation. Mannina et al. (2020) goes a step further by using
historic data taken from plant SCADA systems (Borzooei et al., 2019).
TOPSIS to optimise five operational parameters using ten KPIs and
Lastly, Revollar et al. (2021) specify four scenarios (fluctuating DO,
combining this with E-FAST sensitivity analysis to understand the in
NH4, and internal recycle setpoints) in BSM2, which enables the calcu
fluence of operating parameters on performance. These static DSSs allow
lation of eco-efficiency indicators for comparing control strategies.
users to observe and understand what an optimised system may look
Therefore, static control systems are able to optimise a greater variety of
like, enabling them to derive and implement the WWTP control strategy.
KPIs, like the EQI and OCI, and operational parameters, including solids
The remaining twenty DSSs are able to dynamically alter operation
retention time (Borzooei et al., 2019; Mannina et al., 2020) or process
parameters without user interference. For example, Heo et al. (2021)
flowrates (Caligan et al., 2022; Revollar et al., 2021).
uses the fuzzy c-mean algorithm to process and cluster influent data to
predict initial BSM2 setpoints, a deep neural network then completes the
13
4.4. Indicators selected indicator results and uses weights defined in literature, whilst Mannina
et al. (2020) weights all ten objective functions selected equally. The use
The reliance of DSS developers on BSM platforms results in little of decision maker weighting strategies is recommended, as it enables the
variability of selected indicators. In fact, Table 3 shows eighteen goals of local stakeholders to be integrated within optimisation out
reviewed DSSs used only the inbuilt indicators of BSMs, including EQI, comes. Lastly, some DSS developers did not provide a method for
OCI, or its sub indicators (pumping, aeration, and total energy con selecting the optimal strategy, leaving it to the interpretation of the user
sumption). Although indicators are fixed in the platform, little justifi to compare KPI results (Revollar et al., 2021), such as Borzooei et al.
cation or reasoning for selecting these indicators is given by DSS or BSM (2019) which relies on optimisation curves showing EQI vs OCI to select
sources, except that they cover both economic and environmental im the best operational SRT parameter. Although weighting strategies are
pacts (Li et al., 2021), reflect the operational state of the WWTP, and can useful, the dynamic optimisation of KPIs is now accepted as best practice
evaluate process performance (Han et al., 2022a). EQI and OCI in to enable automatic, supervisory control of plants, placing greater
dicators reflect the traditional goals of water related literature and emphasis on proper selection of KPIs to reflect decision maker needs
regulations, such as for human health protection and cost functions, that during WWTP operation.
will always be important to maintain WWTP performance. However,
modern water sector targets relate to areas such as GHG emissions and 4.6. Error and uncertainty
resource recovery, therefore, expansion to include KPIs that reflect these
goals is recommended for further development of BSMs. This is needed These DSSs aim to provide an optimised control strategy for the
as inclusion of sustainability and circularity dimensions would enable operation of WWTPs, however, alternative controllers, KPIs, or condi
users to optimise WWTP operation considering the wider impacts to tions may result in differing performance. Therefore, it is critical to test
stakeholders and achieve targets defined in Section 2.1, such as those the sensitivity of DSS performance on results. One of the main strategies
defined in the CEAP. employed was to compare the optimised KPI results with alternative
Subsequently, the eight remaining DSSs developed integrated other controller algorithms, to ensure the adopted method achieves the best
indicators to optimise process operation considering impacts other than performance. Controller performance metrics including the Integral of
cost and effluent quality. Three DSSs calculate process GHGs, including Absolute Error (IAE) (no error weighting), Integral of Squared Error
Mannina et al. (2020) that consider a combination of CO2 and N2O (ISE) (penalises larger errors), and Root Mean Square Error (RMSE) were
emissions, direct and indirect GHGs, and air-EQI to understand how utilised. In fact, six DSSs compare the controller algorithm deployed
MBR operational parameters impact emissions. Caligan et al. (2022) also using the IAE with other algorithms (Han et al., 2022a, 2021; Li et al.,
considered GHGs emissions and compared this with cost functions, 2022; Qiao et al., 2019, 2018; Xie et al., 2022), one utilised both ISE and
whilst Chen et al. (2021) conducted full LCC and LCA to investigate the IAE for comparison (Han et al., 2018), and another implemented RSME
impact of indicator prioritisation on a 10,000 PE WWTP. Other DSSs (Qiao and Zhou, 2018) to investigate whether the method used results in
selected indicators to investigate a specific function of a WWTP, namely the lowest error. Additionally, six DSSs compared controller algorithms
freshwater and regenerated water use (Díaz-Madroñero et al., 2018), using KPI results only (Han et al., 2022b; Li et al., 2021; Mannina et al.,
eco-efficiency (Revollar et al., 2021), treatment time reduction (Fox 2020; Niu et al., 2022; Pisa et al., 2019; Zhou and Qiao, 2019), which is a
et al., 2022), and energetic self-sufficiency (Campana et al., 2021). useful exercise to reassure the user their DSS will produce the best
These indicators align better with modern water sector sustainability outcomes. However, investigating errors is important as it indicates the
goals compared with EQI and OCI indicators. However, they all size and longevity of potential disruptions to system performance.
employed self-selection methods and generally circularity indicators Multi-objective optimisation DSSs utilise similar approaches to those
were mitigated from optimisation DSSs, showing this is yet to become a discussed in Section 3.7 for MCDM for uncertainty analysis. For
priority of WWTP operators. example, the DSS developed by Caligan et al. (2022) formulated sce
narios to investigate the impacts of events that WWTP operators may
4.5. Indicator prioritisation face, including how the fluctuation of biofuel prices, inlet wastewater
quality, and requirements for wastewater and sludge disposal, impact on
Again, there is a difference between static and dynamic DSS in cost and GHG emission KPIs. Ortiz-Martínez et al. (2021) created sce
dicators in how they are analysed to produce the optimal solution. The narios simulating lack of aeration due to process error and mitigation of
majority of dynamic calculations aim to minimise the performance in flow recirculation due to maintenance, to investigate the effect on pro
dicators selected, including BSMs trying to minimise both EQI and OCI cess optimisation. Alternatively, some authors investigated optimisation
(or energy consumption). This results in an optimisation problem (Heo strategies through prioritisation of certain indicators to see how the
et al., 2021) since decreasing one of these KPIs increases the other, for system responds. DSSs were tested by optimising either the environ
example greater removal efficiency requires additional energy con mental (i.e. EQI) or economic (i.e. OCI) KPI, and comparing this with
sumption from aeration and recirculation pumping. Therefore, DSS al when both are optimised (Chen et al., 2021; Tejaswini et al., 2021).
gorithms must cope with KPI trade-offs, known Pareto sets, which Mult-objective optimisation DSSs have other inherent uncertainties
derives a sub-optimal solution for the chosen KPIs but establishes that to deal with when modelling WWTP systems, such as climatic changes
both results are better than the rest of the potential outcomes in the and fluctuating wastewater concentration (Chen et al., 2018). DSS de
search space (Qiao et al., 2018). Fox et al. (2022) developed one of the velopers tackled this uncertainty by investigating the effects of the
only dynamic optimisation DSSs to employ a weighting method, with wastewater influent on performance using fluctuation of the TKN/COD
the hope of considering site-specific requirements. Local plant operators inlet ratio (Heo et al., 2021) and fuzzification of inlet composition (Díaz-
assigned weights, which were combined with KPI result rankings to Madroñero et al., 2018). However, further uncertainty analysis is rec
decide on the soft sensors that produces the best control strategy. In ommended to test how WWTP optimisation models respond to external
previous years it was common to weight KPIs to create a single objective factors. MC simulations are commonly used for modelling input uncer
optimisation (Niu et al., 2022), however, this necessitates real-time su tainty as different probability distributions (normal, parametric etc.) can
pervision by plant operators to achieve optimal control (Han et al., be selected depending on error attributes and case study characteristics
2014). (Haag et al., 2019). Testing the uncertainty of DSS performance is crit
Of the static DSSs, three utilise KPI weighting to achieve the optimal ical and for a complete study it is recommended to make comparisons in
solution. Díaz-Madroñero et al. (2018) used fuzzy goal programming to KPI performance and controller error with other systems, and investi
incorporate decision maker preferences and trade-offs between objec gate fluctuations to influent load and process operation to ensure the
tive functions. Alternatively, Chen et al. (2021) normalises LCA impact DSS will meet all user expectations when deployed at a real WWTP.
14
4.7. Recommendations means it is currently very difficult to assess wastewater systems for the
selection and optimisation of technology that facilitates decision maker
Following the review of twenty-six multi-objective DSSs for optimi goals. It has been shown in the literature, and this review, that CE in
sation of WWTP process operation, some final recommendations and dicators have been aligned with sustainability dimensions to validate
comments are provided in Table 4. However, it can again be concluded decision making (Harris et al., 2021), whilst the opposite can be said
that although these DSSs aim to optimise WWTP performance there is when using LCA impacts to evidence enhanced circularity (Corona et al.,
little attention given to how this results in the indicators selected for 2019). Therefore, there is a large knowledge gap regarding how circu
optimisation, nor an explanation of how subsequent operation aligns larity and sustainability indicators can be combined. Overcoming this
with sustainability aims, and mitigate circularity dimensions entirely. issue requires the development of methods that assess the circularity of
resource flows and is supported using wider sustainability analysis to
5. Future work quantify economic, environmental, and social benefits. Only then can
wastewater decision making facilitate governmental circularity and
5.1. Indicator selection sustainability targets, whilst maintaining WWTP performance, and
meeting the customer expectations.
Unfortunately, it has been shown that the indicators utilised by most
wastewater DSSs do not align with water sector sustainability and 5.2. Data
circularity targets. Currently, there is a lack of standardised assessment
methodologies that combine circularity and sustainability dimensions, Inefficient use of data is one of the main problems in plant man
and confusion due to the excess of indicators developed (Valls-Val et al., agement which many WWTPs struggle to solve. The WWT process is
2022). Most methods rely on the user cherry picking indicators from a complicated and decentralised, so data is scattered, and managers can
predetermined list, therefore, when an organisation adopts an innova struggle to supervise the whole plant leading to poor management. The
tive activity it can result in redundant indicator selection and efforts or water industry is still developing data collection, management, ana
resources being allocated incorrectly (Peral et al., 2017). There are lytics, and controls to more effectively use this data to inform decision
additional difficulties to consider in the water sector, particularly for making across all management and operational functions (Corominas
circularity assessments, as the main focus until now has been enhancing et al., 2018). As a result, much of this data is relatively untapped to
the performance of technical systems. For example, processes that safely support decisions that would enable higher levels of performance and
return biotic and water resources to the environment can be seen as control. Subsequently, online optimisation of WWTP control has not
effective waste management (Chojnacka et al., 2020), whereas this can been widely applied to real-world systems, due to the complex, non-
occur during landfilling or incineration for abiotic resources which are linear behaviour of biological WWT systems (increasing the computa
non-circular actions, meaning many indicators are not appropriate. This tional requirements), lack of visualisation techniques, and low-quality
sensor measurements (Matheri et al., 2022). Types of advanced con
trol known as model predictive controllers, use data-driven techniques
Table 4
for early correction of process operation to reduce process faults and
Summary of issues, recommendations, and beneficial outcomes related to the
therefore costly downtime, effluent violations, and resource consump
reviewed wastewater treatment multi-objective process optimisation DSSs.
tion (Ntalaperas et al., 2022). The combination of this with effectively
Issue Recommendation Outcome
constructed multi-objective optimisation DSSs results in powerful and
Few DSSs are applied to Test DSSs in realistic User achieves the desirable tools for the water sector to achieve its goals. Finally, the use of
real WWTP systems, process models or trial expected performance data-driven techniques can also be extended to improve the selection of
mitigating the impacts them in real-world systems when DSS is applied to
of local climate and their system
indicators, including the use of techniques combined with expert
influent composition knowledge, to find precise KPIs for monitoring specific strategic goals.
Although KPI selection is Develop process models DSS will optimise WWTP This would enable the differentiation between performance (lead) and
fixed for many of the that utilise KPIs in a way that generates result (lag) indicators, and create numeric thresholds and benchmarks
DSSs reviewed, considering local desired benefits for
(del Mar Roldán-García et al., 2021), providing more knowledge for
rigorous indicator stakeholder and business stakeholders
selection is often objectives for WWTP decision making purposes.
overlooked optimisation, rather than
depending on those 5.3. Uncertainty
integrated within BSMs
Focussing on EQI and Expansion of indicators to Align WWTP operation
OCI (or energy include environmental, with modern
There are many levels of uncertainty associated with WWTP DSSs,
consumption) KPIs social, circularity, and sustainability and namely model structure (misrepresented boundaries, inaccuracy of
provides a narrow technical aspects circularity aims of the construction, subjective judgement, and mitigation of important mech
view of ‘optimal’ or water sector anisms), data (quality, processing, measurement error, and reliability),
‘sustainable’ WWTP
and relationship with the natural environment (knowledge gaps, dy
performance
Dynamic control and Implement robust Responsive systems that namic system, and uncertain future) (Uusitalo et al., 2015; Walling and
optimisation of indicator selection to optimise performance in Vaneeckhaute, 2020). It is important to map the source and magnitude
WWTPs aligns better ensure optimal terms of selected KPIs, of uncertainty in DSSs, for which many techniques have been discussed
with the water sector's performance facilitates rather than relying on in Sections 3 and 4, however, uncertainty is still recognised and treated
digitisation goals, decision maker goals at a intuitive decision
mitigating plant plant level making of operators
differently by decision makers in the water sector. Even though the
operator decision water sector has additional complexities that result from its strong
making capabilities relationship and dependency with the natural environment, a stand
Many DSSs did not IAE and ISE are Better understanding of ardised methodology to identify, quantify, reduce, and report uncer
investigate the recommended for how the investigated
tainty to support decision making is still missing (Walling and
performance of understanding the WWTP will respond to
controller algorithms response of the selected external stressors Vaneeckhaute, 2020). A large number of techniques have been imple
using appropriate algorithm to process mented to reduce uncertainty, depending on the MCDM method or
metrics alterations, especially as model utilised, including the use of expert assessments, sensitivity
dynamic operation of analysis (scenario, GSA, or MC based), model emulators (Gaussian
WWTPs evolves
processes), deterministic models (temporal and spatial variability), and
15
heterogeneous data assimilation (meteorological and hydrological) Acknowledgements

(Pelissari et al., 2021; Uusitalo et al., 2015). Consideration of future
climate impacts is particularly challenging for water decision makers, This research was supported by the Horizon 2020 research and
when making operational decisions in the face of uncertain conditions. innovation programme DEEP PURPLE. The H2020 DEEP PURPLE proj
In this case, stochastic modelling frameworks and decision tree ap ect has received funding from the Bio-based Industries Joint Undertak
proaches are recommended to assign future conditions a probability of ing (JU) under the European Union's Horizon 2020 research and
occurrence (Horne et al., 2016). Furthermore, to achieve more stand innovation programme under grant agreement No 837998. The JU re
ardised protocols these methods must be aligned with the type and ceives support from the European Union's Horizon 2020 research and
magnitude of uncertainties, ensuring that resultant decisions are with innovation programme and the Bio-based Industries Consortium. Au
acceptable ranges, thereby permitting dissemination. Identifying un thors would also like to acknowledge ICT4WATER and the AG Intelli
certainties that have the greatest impact on results will help water sector gent Smart Systems.
decision makers create knowledge bases with which to refine future
studies and models (Horne et al., 2016). References
6. Conclusions Alex, J., Benedetti, L., Copp, J., Gernaey, K.V., Jeppsson, U., Nopens, I., Pons, M.N.,
Steyer, J.P., Vanrolleghem, P., 2018. Benchmark Simulation Model no. 1 (BSM1).
Ali, Y., Pervez, H., Khan, J., 2020. Selection of the most feasible wastewater treatment
WWTP decision makers face stricter regulations regarding human technology in Pakistan using Multi-Criteria Decision-Making (MCDM). WATER
health, environmental protection, and emissions reduction, meaning Conserv. Sci. Eng. 5, 199–213. https://doi.org/10.1007/s41101-020-00094-6.
An, D., Xi, B.D., Ren, J.Z., Ren, X.S., Zhang, W.S., Wang, Y., Dong, L.C., 2018. Multi-
they must optimise performance and replace infrastructure, whilst criteria sustainability assessment of urban sludge treatment technologies: method
maintaining positive economic performance. This creates complex and case study. Resour. Conserv. Recycl. 128, 546–554. https://doi.org/10.1016/j.
multi-objective problems when operating and selecting technologies for resconrec.2016.08.018.
Arroyo, P., Molinos-Senante, M., 2018. Selecting appropriate wastewater treatment
improving WWTPs, meaning many DSSs have been developed for the technologies using a choosing-by-advantages approach. Sci. Total Environ. 625,
water sector. Currently, there is a mismatch in terms of the decision 819–827. https://doi.org/10.1016/j.scitotenv.2017.12.331.
maker goals and KPIs selected for DSSs, so water sector objectives at a Attri, S.D., Singh, S., Dhar, A., Powar, S., 2022. Multi-attribute sustainability assessment
of wastewater treatment technologies using combined fuzzy multi-criteria decision-
European level were summarised. The regulation and action plans from
making techniques. J. Clean. Prod. 357, 131849 https://doi.org/10.1016/j.
the European Commission highlight their recent ambition for greater jclepro.2022.131849.
levels or sustainability, circularity, and environmental and human Borzooei, S., Campo, G., Cerutti, A., Meucci, L., Panepinto, D., Ravina, M., Riggio, V.,
health protection. Following this, DSS literature was reviewed and Ruffino, B., Scibilia, G., Zanetti, M., 2019. Optimization of the wastewater treatment
plant: from energy saving to environmental impact mitigation. Sci. Total Environ.
showed the main function of MCDM tools was for WWT technology 691, 1182–1189. https://doi.org/10.1016/j.scitotenv.2019.07.241.
selection, whereas multi-objective optimisation DSSs focused on optimal Büyüközkan, G., Tüfekçi, G., 2021. A multi-stage fuzzy decision-making framework to
set-point control to improve effluent and cost indicators. A large contrast evaluate the appropriate wastewater treatment system: a case study. Environ. Sci.
Pollut. Res. 28, 53507–53519. https://doi.org/10.1007/s11356-021-14116-w.
was found regarding the aims of DSS typologies, as optimisation stra Caligan, C.J.A., Garcia, M.M.S., Mitra, J.L., San Juan, J.L.G., 2022. Multi-objective
tegies tend to clearly define their goals in terms of the KPIs used, how optimization for a wastewater treatment plant and sludge-to-energy network.
ever, MCDM tools often use vague language making it difficult for users J. Clean. Prod. 368, 133047 https://doi.org/10.1016/j.jclepro.2022.133047.
Campana, P.E., Mainardis, M., Moretti, A., Cottes, M., 2021. 100% renewable wastewater
to make a connection between indicators selected and resultant out treatment plants: techno-economic assessment using a modelling and optimization
comes. Considering these issues several recommendations were made to approach. Energ. Conver. Manage. 239 https://doi.org/10.1016/j.
improve DSS deployment, such as more rigorous indicator selection enconman.2021.114214.
Castillo, A., Porro, J., Garrido-Baserba, M., Rosso, D., Renzi, D., Fatone, F., Gómez, V.,
protocols including participatory approaches and expansion of in Comas, J., Poch, M., 2016. Validation of a decision support tool for wastewater
dicators sets (specifically for multi-objective optimisation), or greater treatment selection. J. Environ. Manage. 184, 409–418. https://doi.org/10.1016/j.
analysis of results whether it is the use of sensitivity/uncertainty anal jenvman.2016.09.087.
Chen, X., Xu, Z., Yao, L., Ma, N., 2018. Processing technology selection for municipal
ysis or ISE/IAE indicators. Lastly, to facilitate the success of DSSs
sewage treatment based on a multi-objective decision model under uncertainty. Int.
implementation, development should focus on standardised methods of J. Environ. Res. Public Health. https://doi.org/10.3390/ijerph15030448.
indicator selection that directly links outcomes with decision maker Chen, K., Wang, H., Valverde-Pérez, B., Zhai, S., Vezzaro, L., Wang, A., 2021. Optimal
goals, and the water sector's circularity and sustainability targets. control towards sustainable wastewater treatment plants based on multi-agent
reinforcement learning. Chemosphere 279, 130498. https://doi.org/10.1016/j.
chemosphere.2021.130498.
CRediT authorship contribution statement Chhipi-Shrestha, G., Hewage, K., Sadiq, R., 2017. Fit-for-purpose wastewater treatment:
conceptualization to development of decision support tool (I). Sci. Total Environ.
607–608, 600–612. https://doi.org/10.1016/j.scitotenv.2017.06.269.
D. Renfrew: Conceptualization, Data curation, Formal analysis, Chojnacka, K., Moustakas, K., Witek-Krowiak, A., 2020. Bio-based fertilizers: a practical
Investigation, Methodology, Visualization, Writing – original draft, approach towards circular economy. Bioresour. Technol. 295, 122223 https://doi.
Writing – review & editing. V. Vasilaki: Conceptualization, Methodol org/10.1016/j.biortech.2019.122223.
Chrispim, M.C., de M. de Souza, F., Scholz, M., Nolasco, M.A., 2020. A framework for
ogy, Validation, Writing – review & editing. E. Katsou: Conceptuali sustainable planning and decision-making on resource recovery from wastewater:
zation, Funding acquisition, Project administration, Resources, showcase for São Paulo megacity. Water 12. https://doi.org/10.3390/w12123466.
Supervision, Writing – review & editing. Corominas, L., Garrido-Baserba, M., Villez, K., Olsson, G., Cortés, U., Poch, M., 2018.
Transforming data into knowledge for improved wastewater treatment operation: a
critical review of techniques. Environ. Model. Software 106, 89–103. https://doi.
Declaration of competing interest org/10.1016/j.envsoft.2017.11.023.
Corona, B., Shen, L., Reike, D., Rosales Carreón, J., Worrell, E., 2019. Towards
sustainable development through the circular economy—a review and critical
The authors declare that they have no known competing financial
assessment on current circularity metrics. Resour. Conserv. Recycl. https://doi.org/
interests or personal relationships that could have appeared to influence 10.1016/j.resconrec.2019.104498.
the work reported in this paper. Council of the European Union, 1986. COUNCIL DIRECTIVE of 12 June 1986 on the
protection of the environment, and in particular of the soil, when sewage sludge is
used in agriculture (86/278/EEC).
Data availability da Silva Junior, L.C.S., de Sá Salomão, A.L., Pereira Santos, A.S., 2022. MATTI - a multi-
criteria decision analysis framework for assessing wastewater treatment
Data will be made available on request. technologies. Water Sci. Technol. 86, 2764–2776. https://doi.org/10.2166/
wst.2022.374.
Dai, H.L., Chen, W.L., Peng, L.H., Wang, X.G., Lu, X.W., 2019. Modeling and performance
improvement of an anaerobic-anoxic/nitrifying-induced crystallization process via
16
the multi-objective optimization method. Environ. Sci. Pollut. Res. 26, 5083–5093. Heo, S., Nam, K., Tariq, S., Lim, J.Y., Park, J., Yoo, C., 2021. A hybrid machine
https://doi.org/10.1007/s11356-018-3971-1. learning–based multi-objective supervisory control strategy of a full-scale
Ddiba, D., Andersson, K., Dickin, S., Ekener, E., Finnveden, G., 2023. A review of how wastewater treatment for cost-effective and sustainable operation under varying
decision support tools address resource recovery in sanitation systems. J. Environ. influent conditions. J. Clean. Prod. 291, 125853 https://doi.org/10.1016/j.
Manage. 342, 118365 https://doi.org/10.1016/j.jenvman.2023.118365. jclepro.2021.125853.
de Almeida, K.N., dos Reis, J.A.T., Mendonça, A.S.F., Silva, F.G.B., 2022. Methodology to Horne, A., Szemis, J.M., Kaur, S., Webb, J.A., Stewardson, M.J., Costa, A., Boland, N.,
support the planning of sewage treatment systems within river basins. Int. J. 2016. Optimization tools for environmental water decisions: a review of strengths,
Environ. Sci. Technol. 19, 3741–3756. https://doi.org/10.1007/s13762-021-03387- weaknesses, and opportunities to improve adoption. Environ. Model. Software 84,
y. 326–338. https://doi.org/10.1016/j.envsoft.2016.06.028.
del Mar Roldán-García, M., García-Nieto, J., Maté, A., Trujillo, J., Aldana-Montes, J.F., IWA, 2018. Benchmarking [WWW Document]. Model. Integr. Assess. URL
2021. Ontology-driven approach for KPI meta-modelling, selection and reasoning. http://iwa-mia.org/benchmarking/#BSM1 (accessed 3.2.22).
Int. J. Inf. Manag. 58, 102018 https://doi.org/10.1016/j.ijinfomgt.2019.10.003. Južnič-Zonta, Ž., Guisasola, A., Baeza, J.A., 2022. Smart-plant decision support system
Díaz-Madroñero, M., Pérez-Sánchez, M., Satorre-Aznar, J.R., Mula, J., López-Jiménez, P. (SP-DSS): defining a multi-criteria decision-making framework for the selection of
A., 2018. Analysis of a wastewater treatment plant using fuzzy goal programming as WWTP configurations with resource recovery. J. Clean. Prod. 367, 132873 https://
a management tool: a case study. J. Clean. Prod. 180, 20–33. https://doi.org/ doi.org/10.1016/j.jclepro.2022.132873.
10.1016/j.jclepro.2018.01.129. Kehrein, P., van Loosdrecht, M., Osseweijer, P., Garfí, M., Dewulf, J., Posada, J., 2020.
Du, X., Peng, Y., 2023. Multi-objective pity beetle algorithm based optimal control of A critical review of resource recovery from municipal wastewater treatment plants –
wastewater treatment process. Process. Saf. Environ. Prot. 170, 188–206. https:// market supply potentials{,} technologies and bottlenecks. Environ. Sci. Water Res.
doi.org/10.1016/j.psep.2022.12.005. Technol. 6, 877–910. https://doi.org/10.1039/C9EW00905A.
Durdević, D., Trstenjak, M., Hulenić, I., 2020. Sewage sludge thermal treatment Kozłowska, J., 2022. Methods of multi-criteria analysis in technology selection and
technology selection by utilizing the analytical hierarchy process. Water technology assessment: a systematic literature review. Eng. Manag. Prod. Serv. 14,
(Switzerland) 12. https://doi.org/10.3390/W12051255. 116–137. https://doi.org/10.2478/emj-2022-0021.
Đurđević, D., Trstenjak, M., Hulenić, I., 2020. Sewage sludge thermal treatment Leoneti, A.B., Banares-Alcantara, R., Pires, E.C., de Oliveira, S., 2022. A multi-criteria
technology selection by utilizing the analytical hierarchy process. Water 12. https:// and multi-agent framework for supporting complex decision-making processes. Gr.
doi.org/10.3390/w12051255. Decis. Negot. 31, 1025–1050. https://doi.org/10.1007/s10726-022-09785-y.
Eseoglu, G., Yapsakli, K., Tozan, H., Vayvay, O., 2022. A novel fuzzy framework for Li, F., Su, Z., Wang, G., 2021. An effective integrated control with intelligent
technology selection of sustainable wastewater treatment plants based on TODIM optimization for wastewater treatment process. J. Ind. Inf. Integr. 24, 100237
methodology in developing urban areas. Sci. Rep. 12, 8800. https://doi.org/ https://doi.org/10.1016/j.jii.2021.100237.
10.1038/s41598-022-12643-1. Li, F., Su, Z., Wang, G., 2022. An effective dynamic immune optimization control for the
European Commission, 2020. A New Circular Economy Action Plan for a Cleaner and wastewater treatment process. Environ. Sci. Pollut. Res. 29, 79718–79733. https://
more Competitive Europe. doi.org/10.1007/s11356-021-17505-3.
European Commission, 2022. Questions and Answers on the new EU rules on treating Ling, Jiean, Germain, E., Murphy, R., Saroj, D., 2021a. Designing a sustainability
urban wastewater [WWW Document]. URL https://ec.europa.eu/commission/pressc assessment framework for selecting sustainable wastewater treatment technologies
orner/detail/en/QANDA_22_6281 (accessed 1.16.23). in corporate asset decisions. Sustainability 13. https://doi.org/10.3390/
European Parliament, 2020. REGULATION (EU) 2020/741 OF THE EUROPEAN su13073831.
PARLIAMENT AND OF THE COUNCIL of 25 May 2020 on Minimum Requirements Ling, J., Germain, E., Murphy, R., Saroj, D., 2021b. Designing a sustainability assessment
for Water Reuse. tool for selecting sustainable wastewater treatment technologies in corporate asset
Fetanat, A., Tayebi, M., Mofid, H., 2021. Water-energy-food security nexus based decisions. Sustain 13. https://doi.org/10.3390/su13073831.
selection of energy recovery from wastewater treatment technologies: An extended Liu, Y., Ren, J., 2022. Developing a sustainability-oriented multi-criteria game
decision making framework under intuitionistic fuzzy environment. Sustain. Energy theoretical decision analysis framework: a case study of sludge management.
Technol. Assessments 43, 100937. https://doi.org/10.1016/j.seta.2020.100937. J. Clean. Prod. 354, 131807 https://doi.org/10.1016/j.jclepro.2022.131807.
Fox, S., McDermott, J., Doherty, E., Cooney, R., Clifford, E., 2022. Application of neural Liu, B., Tang, J., Li, Z., Yan, Y., Chen, J., 2020. Optimal selection of sewage treatment
networks and regression modelling to enable environmental regulatory compliance technologies in town areas: a coupled multi-criteria decision-making model.
and energy optimisation in a sequencing batch reactor. Sustainability 14. https:// Environ. Manag. 66, 709–721. https://doi.org/10.1007/s00267-020-01338-w.
doi.org/10.3390/su14074098. Lizot, M., Goffi, A.S., Thesari, S.S., Trojan, F., Afonso, P.S.L.P., Ferreira, P.F.V., 2021.
Garcia-Garcia, G., 2022. Using multi-criteria decision-making to optimise solid waste Multi-criteria methodology for selection of wastewater treatment systems with
management. Curr. Opin. Green Sustain. Chem. 37, 100650 https://doi.org/ economic, social, technical and environmental aspects. Environ. Dev. Sustain. 23,
10.1016/j.cogsc.2022.100650. 9827–9851. https://doi.org/10.1007/s10668-020-00906-8.
Garrido-Baserba, M., Molinos-Senante, M., Abelleira-Pereira, J.M., Fdez-Güelfo, L.A., Mannina, G., Rebouças, T.F., Cosenza, A., Sànchez-Marrè, M., Gibert, K., 2019. Decision
Poch, M., Hernández-Sancho, F., 2015. Selecting sewage sludge treatment support systems (DSS) for wastewater treatment plants – a review of the state of the
alternatives in modern wastewater treatment plants using environmental decision art. Bioresour. Technol. 290, 121814 https://doi.org/10.1016/j.
support systems. J. Clean. Prod. 107, 410–419. https://doi.org/10.1016/j. biortech.2019.121814.
jclepro.2014.11.021. Mannina, G., Ni, B.-J., Ferreira Rebouças, T., Cosenza, A., Olsson, G., 2020. Minimizing
Gherghel, A., Teodosiu, C., Notarnicola, M., De Gisi, S., 2020. Sustainable design of large membrane bioreactor environmental footprint by multiple objective optimization.
wastewater treatment plants considering multi-criteria decision analysis and Bioresour. Technol. 302, 122824 https://doi.org/10.1016/j.biortech.2020.122824.
stakeholders’ involvement. J. Environ. Manage. 261, 110158 https://doi.org/ Matheri, A.N., Mohamed, B., Ntuli, F., Nabadda, E., Ngila, J.C., 2022. Sustainable
10.1016/j.jenvman.2020.110158. circularity and intelligent data-driven operations and control of the wastewater
Giupponi, C., Sgobbi, A., 2013. Decision support systems for water resources treatment plant. Phys. Chem. Earth, Parts A/B/C 126, 103152. https://doi.org/
management in developing countries: learning from experiences in Africa. Water 5, 10.1016/j.pce.2022.103152.
798–818. https://doi.org/10.3390/w5020798. Molinos-Senante, M., Gómez, T., Garrido-Baserba, M., Caballero, R., Sala-Garrido, R.,
Haag, F., Reichert, P., Maurer, M., Lienert, J., 2019. Integrating uncertainty of 2014. Assessing the sustainability of small wastewater treatment systems: a
preferences and predictions in decision models: an application to regional composite indicator approach. Sci. Total Environ. 497–498, 607–617. https://doi.
wastewater planning. J. Environ. Manage. 252, 109652 https://doi.org/10.1016/j. org/10.1016/j.scitotenv.2014.08.026.
jenvman.2019.109652. Molinos-Senante, M., Gomez, T., Caballero, R., Hernandez-Sancho, F., Sala-Garrido, R.,
Haldar, K., Kujawa-Roeleveld, K., Acharjee, T.K., Datta, D.K., Rijnaarts, H., 2022. Urban 2015. Assessment of wastewater treatment alternatives for small communities: an
water as an alternative freshwater resource for matching irrigation demand in the analytic network process approach. Sci. Total Environ. 532, 676–687. https://doi.
Bengal delta. Sci. Total Environ. 835, 155475 https://doi.org/10.1016/j. org/10.1016/j.scitotenv.2015.06.059.
scitotenv.2022.155475. Niu, G., Li, X., Wan, X., He, X., Zhao, Y., Yi, X., Chen, C., Xujun, L., Ying, G., Huang, M.,
Han, H.-G., Qian, H.-H., Qiao, J.-F., 2014. Nonlinear multiobjective model-predictive 2022. Dynamic optimization of wastewater treatment process based on novel multi-
control scheme for wastewater treatment process. J. Process Control 24, 47–59. objective ant lion optimization and deep learning algorithm. J. Clean. Prod. 345,
https://doi.org/10.1016/j.jprocont.2013.12.010. 131140 https://doi.org/10.1016/j.jclepro.2022.131140.
Han, H.-G., Zhang, L., Liu, H.-X., Qiao, J.-F., 2018. Multiobjective design of fuzzy neural Nkuna, S.G., Olwal, T.O., Chowdhury, S.P.D., 2022. Assessment of thermochemical
network controller for wastewater treatment process. Appl. Soft Comput. 67, technologies for wastewater sludge-to-energy: an advance MCDM model. Clean. Eng.
467–478. https://doi.org/10.1016/j.asoc.2018.03.020. Technol. 9, 100519 https://doi.org/10.1016/j.clet.2022.100519.
Han, H.-G., Liu, Z., Lu, W., Hou, Y., Qiao, J.-F., 2021. Dynamic MOPSO-based optimal Ntalaperas, D., Christophoridis, C., Angelidis, I., Iossifidis, D., Touloupi, M.-F.,
control for wastewater treatment process. IEEE Trans. Cybern. 51, 2518–2528. Vergeti, D., Politi, E., 2022. Intelligent tools to monitor, control and predict
https://doi.org/10.1109/TCYB.2019.2925534. wastewater reclamation and reuse. Sensors 22. https://doi.org/10.3390/s22083068.
Han, H.-G., Chen, C., Sun, H.-Y., Qiao, J.-F., 2022a. Multi-objective integrated optimal Oertlé, E., Hugi, C., Wintgens, T., Karavitis, C.A., 2019. Poseidon—decision support tool
control for a wastewater treatment process. Control. Eng. Pract. 128, 105296 for water reuse. Water 11. https://doi.org/10.3390/w11010153.
https://doi.org/10.1016/j.conengprac.2022.105296. Ortiz-Martínez, V.M., Martínez-Frutos, J., Hontoria, E., Hernández-Fernández, F.J.,
Han, H.-G., Zhang, L., Qiao, J., 2022b. Dynamic optimal control for wastewater Egea, J.A., 2021. Multiplicity of solutions in model-based multiobjective
treatment process under multiple operating conditions. IEEE Trans. Autom. Sci. Eng. optimization of wastewater treatment plants. Optim. Eng. 22, 1–16. https://doi.org/
1–13 https://doi.org/10.1109/TASE.2022.3189048. 10.1007/s11081-020-09500-3.
Harris, S., Martin, M., Diener, D., 2021. Circularity for circularity’s sake? Scoping review Page, M.J., McKenzie, J.E., Bossuyt, P.M., Boutron, I., Hoffmann, T.C., Mulrow, C.D.,
of assessment methods for environmental performance in the circular economy. Shamseer, L., Tetzlaff, J.M., Akl, E.A., Brennan, S.E., Chou, R., Glanville, J.,
Sustain. Prod. Consum. 26, 172–186. https://doi.org/10.1016/j.spc.2020.09.018. Grimshaw, J.M., Hróbjartsson, A., Lalu, M.M., Li, T., Loder, E.W., Mayo-Wilson, E.,
17
McDonald, S., McGuinness, L.A., Stewart, L.A., Thomas, J., Tricco, A.C., Welch, V.A., support tool for the evaluation of advanced wastewater treatment trains: a novel
Whiting, P., Moher, D., 2021. The PRISMA 2020 statement: an updated guideline for approach to improve urban sustainability. Environ. Sci. Policy 90, 1–10. https://doi.
reporting systematic reviews. Syst. Rev. 10, 89. https://doi.org/10.1186/s13643- org/10.1016/j.envsci.2018.09.006.
021-01626-4. Salamirad, A., Kheybari, S., Ishizaka, A., Farazmand, H., 2023. Wastewater treatment
Palma-Heredia, D., Poch, M., Cugueró-Escofet, M.À., 2020. Implementation of a decision technology selection using a hybrid multicriteria decision-making method. Int.
support system for sewage sludge management. Sustainability 12. https://doi.org/ Trans. Oper. Res. 30, 1479–1504. https://doi.org/10.1111/itor.12979.
10.3390/su12219089. Silva Junior, L.C.S. da, Salomão, A.L. de S., Santos, A.S.P., 2022. MATTI – a multi-criteria
Palmeros Parada, M., Kehrein, P., Xevgenos, D., Asveld, L., Osseweijer, P., 2022. Societal decision analysis framework for assessing wastewater treatment technologies. Water
values, tensions and uncertainties in resource recovery from wastewaters. J. Environ. Sci. Technol. 86, 2764–2776. https://doi.org/10.2166/wst.2022.374.
Manage. 319, 115759 https://doi.org/10.1016/j.jenvman.2022.115759. Srivastava, R.R., Singh, P.K., 2022. Reuse-focused selection of appropriate technologies
Pelissari, R., Oliveira, M.C., Abackerli, A.J., Ben-Amor, S., Assumpção, M.R.P., 2021. for municipal wastewater treatment: a multi-criteria approach. Int. J. Environ. Sci.
Techniques to model uncertain input data of multi-criteria decision-making Technol. 19, 12505–12522. https://doi.org/10.1007/s13762-021-03803-3.
problems: a literature review. Int. Trans. Oper. Res. 28, 523–559. https://doi.org/ Štilić, A., Puška, A., 2023. Integrating multi-criteria decision-making methods with
10.1111/itor.12598. sustainable engineering: a comprehensive review of current practices. Eng. https://
Peral, J., Maté, A., Marco, M., 2017. Application of data mining techniques to identify doi.org/10.3390/eng4020088.
relevant key performance indicators. Comput. Stand. Interfaces 54, 76–85. https:// Sucu, S., van Schaik, M.O., Esmeli, R., Ouelhadj, D., Holloway, T., Williams, J.B.,
doi.org/10.1016/j.csi.2016.11.006. Cruddas, P., Martinson, D.B., Chen, W.-S., Cappon, H.J., 2021. A conceptual
Pisa, I., Santín, I., Morell, A., Vicario, J.L., Vilanova, R., 2019. LSTM-based wastewater framework for a multi-criteria decision support tool to select technologies for
treatment plants operation strategies for effluent quality improvement. IEEE Access resource recovery from urban wastewater. J. Environ. Manage. 300, 113608 https://
7, 159773–159786. https://doi.org/10.1109/ACCESS.2019.2950852. doi.org/10.1016/j.jenvman.2021.113608.
Preisner, M., Smol, M., Horttanainen, M., Deviatkin, I., Havukainen, J., Klavins, M., Tejaswini, E.S.S., Panjwani, S., Gara, U.B.B., Ambati, S.R., 2021. Multi-objective
Ozola-Davidane, R., Kruopienė, J., Szatkowska, B., Appels, L., Houtmeyers, S., optimization based controller design for improved wastewater treatment plant
Roosalu, K., 2022. Indicators for resource recovery monitoring within the circular operation. Environ. Technol. Innov. 23, 101591 https://doi.org/10.1016/j.
economy model implementation in the wastewater sector. J. Environ. Manage. 304, eti.2021.101591.
114261 https://doi.org/10.1016/j.jenvman.2021.114261. Ullah, A., Hussain, S., Wasim, A., Jahanzaib, M., 2020. Development of a decision
Qiao, J., Zhou, H., 2018. Modeling of energy consumption and effluent quality using support system for the selection of wastewater treatment technologies. Sci. Total
density peaks-based adaptive fuzzy neural network. IEEE/CAA J. Autom. Sin. 5, Environ. 731, 139158 https://doi.org/10.1016/j.scitotenv.2020.139158.
968–976. https://doi.org/10.1109/JAS.2018.7511168. Uusitalo, L., Lehikoinen, A., Helle, I., Myrberg, K., 2015. An overview of methods to
Qiao, J.-F., Hou, Y., Zhang, L., Han, H.-G., 2018. Adaptive fuzzy neural network control evaluate uncertainty of deterministic models in decision support. Environ. Model.
of wastewater treatment process with multiobjective operation. Neurocomputing Software 63, 24–31. https://doi.org/10.1016/j.envsoft.2014.09.017.
275, 383–393. https://doi.org/10.1016/j.neucom.2017.08.059. Valls-Val, K., Ibáñez-Forés, V., Bovea, M.D., 2022. How can organisations measure their
Qiao, J.-F., Hou, Y., Han, H.-G., 2019. Optimal control for wastewater treatment process level of circularity? A review of available tools. J. Clean. Prod. 354, 131679 https://
based on an adaptive multi-objective differential evolution algorithm. Neural doi.org/10.1016/j.jclepro.2022.131679.
Comput. Appl. 31, 2537–2550. https://doi.org/10.1007/s00521-017-3212-4. Walling, E., Vaneeckhaute, C., 2020. Developing successful environmental decision
Radini, S., González-Camejo, J., Andreola, C., Eusebi, A.L., Fatone, F., 2023. Risk support systems: challenges and best practices. J. Environ. Manage. 264 https://doi.
management and digitalisation to overcome barriers for safe reuse of urban org/10.1016/j.jenvman.2020.110513.
wastewater for irrigation – a review based on European practice. J. Water Process Wardropper, C., Brookfield, A., 2022. Decision-support systems for water management.
Eng. 53, 103690 https://doi.org/10.1016/j.jwpe.2023.103690. J. Hydrol. 610, 127928 https://doi.org/10.1016/j.jhydrol.2022.127928.
Renfrew, D., Vasilaki, V., McLeod, A., Lake, A., Danishvar, S., Katsou, E., 2022. Where is Xie, Y.B., Wang, D., Qiao, J.F., 2022. Dynamic multi-objective intelligent optimal control
the greatest potential for resource recovery in wastewater treatment plants? Water toward wastewater treatment processes. Sci. China Technol. Sci. 65, 569–580.
Res. 220, 118673 https://doi.org/10.1016/j.watres.2022.118673. https://doi.org/10.1007/s11431-021-1960-7.
Revollar, S., Meneses, M., Vilanova, R., Vega, P., Francisco, M., 2021. Eco-efficiency Zhou, H., Qiao, J., 2019. Multiobjective optimal control for wastewater treatment
assessment of control actions in wastewater treatment plants. Water 13. https://doi. process using adaptive MOEA/D. Appl. Intell. 49, 1098–1126. https://doi.org/
org/10.3390/w13050612. 10.1007/s10489-018-1319-7.
Saaty, R.W., 1987. The analytic hierarchy process—what it is and how it is used. Math. Zolghadr-Asli, B., Bozorg-Haddad, O., Enayati, M., Chu, X., 2021. A review of 20-year
Model. 9, 161–176. https://doi.org/10.1016/0270-0255(87)90473-8. applications of multi-attribute decision-making in environmental and water
Sadr, S.M.K., Saroj, D.P., Mierzwa, J.C., McGrane, S.J., Skouteris, G., Farmani, R., resources planning and management. Environ. Dev. Sustain. 23, 14379–14404.
Kazos, X., Aumeier, B., Kouchaki, S., Ouki, S.K., 2018. A multi expert decision https://doi.org/10.1007/s10668-021-01278-3.
18

1-s2.0-S0048969724000378-main

Uploaded by

Copyright:

Available Formats

1-s2.0-S0048969724000378-main

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1-s2.0-S0048969724000378-main

Uploaded by

Copyright:

Available Formats

Science of the Total Environment 915 (2024) 169903

Contents lists available at ScienceDirect

Science of the Total Environment

Indicator based multi-criteria decision support systems for wastewater

• DSS indicators are not selected in line

3.3. Indicator categorisation

Often criteria or indicators are categorised to show user assessment

combination of other categories defined by the developers, and some did

4.1. DSS goals 4.3. Modelling platform

heterogeneous data assimilation (meteorological and hydrological) Acknowledgements

You might also like