001.docx
CHEMICAL ENGINEERING TRANSACTIONS
VOL. 83, 2021
A publication of
The Italian Association
of Chemical Engineering
Online at www.aidic.it/cet
Guest Editors: Jeng Shiun Lim, Nor Alafiza Yunus, Jiří Jaromír Klemeš
Copyright © 2021, AIDIC Servizi S.r.l.
ISBN 978-88-95608-81-5; ISSN 2283-9216
Synthesis of a Sustainable Wastewater Treatment Plant for
Sago Industry using Fuzzy Optimisation
Jo Yee Hoa, Kelvin Teo Kai Wena, Yoke Kin Wana,*, Viknesh Andiappanb
aSchool of Engineering, Faculty of Innovation and Technology, Taylor’s University, Lakeside Campus, No. 1 Jalan Taylor’s,
47500 Subang Jaya, Selangor, Malaysia.
bSchool of Engineering and Physical Sciences, Heriot-Watt University Malaysia, 62200, Putrajaya, Wilayah Persekutuan
Putrajaya, Malaysia
yokekin.wan@taylors.edu.my
Wastewater treatment plant (WWTP) is an essential process in the manufacturing industry. However,
wastewater treatment process is not a profitable process as it requires a significant amount of investment. It is
important to design a WWTP that meets wastewater discharge legalisation with low investment costs. To do
this, area footprint (land area) occupied by technologies in a WWTP must be factored into the design decision.
Unfortunately, the area footprint is yet to be studied. As wastewater treatment processes involve multiple
treatment units, the different combinations of these treatment units will give a range of capital costs and costs
associated with the area occupied. In addition, the carbon footprint of technology resulting from power
consumption must be considered. Each technology possesses unique power consumption requirements and
these requirements may influence the total carbon footprint for a given WWTP design. Investment costs, area
footprint and carbon footprint must be considered simultaneously but are conflicting in nature. This work aims
to present a multi-objective decision-making tool to screen wastewater treatment technologies and to synthesise
a WWTP design with low investment cost, low area footprint, and low carbon footprint. Specifically, fuzzy multi-
objective optimisation (FMOO) is used to determine a desirable trade-off between investment costs, area
footprint, and carbon footprint. To demonstrate the developed approach, a sago-based WWTP case study is
solved. Based on the results, a trade-off between these optimisation objectives had reduced 5.35 m2 of area
footprint, 986 USD/d of total investment cost, and 108 kg CO2/d of carbon footprint of the synthesised WWTP.
1. Introduction
Sago industry is one of the major industries in Sarawak, Malaysia that generates a significant amount of organic
wastewater (Yunus et al., 2014). According to Adeni et al. (2010), every 1 t of sago starch produced,
approximately 10 t to 22 t of organic wastewater will be generated. This significant amount of wastewater
produced requires proper wastewater treatment. A proper wastewater treatment process can be synthesised
before actual installation and operation to avoid extra investment costs needed to rectify the process in the
future. In this respect, there are several published research works had incorporated cost optimisation (Ho et al.,
2019) and carbon footprint optimisation (Padrón-páez et al., 2020) of a WWTP. However, limited research works
had considered area footprint (land area) as an important design criterion in synthesising a WWTP. Shortage of
land area is an issue in some countries as the rapid development of cities and heavy industries had led to global
deforestation issues. Malaysia for instance, had the world’s highest rate of deforestation between year 2000
and 2012 (Butler, 2013). Consequently, authorised organisations and departments had implemented more
stringent regulations to control the rate of deforestation, resulting in land prices in Malaysia to rise significantly
for the past 10 y (WWF, 2020). The area footprint required to synthesise a WWTP is essential to be optimised
as the land area cost of a WWTP contributes to a high capital cost. To optimise the area footprint of a WWTP,
the size of equipment must be minimised. Nevertheless, smaller wastewater treatment technologies are usually
more advanced technologies and require higher cost compared to conventional wastewater treatment
technologies (Ang et al., 2019). In addition, smaller advanced technologies may consume higher power
consumption, resulting in a higher carbon footprint (Fernandez-Dacosta et al., 2016). Due to the extensive
DOI: 10.3303/CET2183063
Paper Received: 21/06/2020; Revised: 04/08/2020; Accepted: 10/08/2020
Please cite this article as: Ho J.Y., Wen K.T.K., Wan Y.K., Andiappan V., 2021, Synthesis of a Sustainable Wastewater Treatment Plant for
Sago Industry using Fuzzy Optimisation, Chemical Engineering Transactions, 83, 373-378 DOI:10.3303/CET2183063
373
selection of technologies, this becomes a challenge for industrial decision-makers to select the optimum
wastewater treatment pathway which fits all the desired objectives while ensuring the discharge effluent
complies with local discharge legislation. The combination of different technologies with various specifications
will affect the performance, investment cost, and area footprint (land area) of the entire WWTP. In this respect,
this work had employed fuzzy multi-objective optimisation (FMOO) in the developed mathematical model from
this study. FMOO provides a trade-off between multiple optimisation objectives by integrating these objectives
into a single parameter or degree of satisfaction, λ (Zimmermann, 1978). This method is useful to address the
vagueness and ambiguity present in quantifying the target range of each optimisation objective while having
more than one selection between the WWTP pathway alternatives. In this situation, FMOO is a better tool to
solve multi-objective optimisation problems with uncertainty in identifying the importance or contribution of each
optimisation objective (Pan et al., 2014). To illustrate the developed approach, this work aims to present a
FMOO tool to synthesise an optimum sago biorefinery WWTP with minimum cost, minimum area footprint and
minimum carbon footprint which meets the effluent quality discharge regulation (COD, BOD and TSS).
2. Problem statement
The problem definition in this work is as follows: Wastewater feed f 𝜖𝜖 F is treated by a series of treatment stages
beginning from preliminary treatment p 𝜖𝜖 P, chemical treatment r 𝜖𝜖 R, biological treatment s 𝜖𝜖 S, and tertiary
treatment t 𝜖𝜖 T to produce treated wastewater as shown in Figure 1. In chemical treatment and biological
treatment, a certain amount of sludge will be generated. These generated sludges will flow into the sludge
treatment process u 𝜖𝜖 U. Based on the problem shown in Figure 1, a mathematical model is formulated.
Figure 1: Generic superstructure of WWTP
3. Fuzzy multi-objective optimisation (FMOO) model
The mathematical model developed in this work consists of flowrate balance, component balance, area footprint,
cost computation, carbon footprint, and fuzzy equations as shown in each subsection. Based on Figure 1, these
equations will be repetitive at each treatment stage. A more generic formulation is presented in this paper where
index a represents the previous treatment stage, index b represents the present treatment stage and index c
would represent the subsequent treatment stage. For instance, to formulate equations for biological treatment
s, the present index b represents chemical treatment s (b = s). Index a will represent the preceding chemical
treatment r (a = r) and index c will represent the succeeding tertiary treatment t or sludge treatment u (c = t,u).
3.1 Flowrate and component balance
This work aims to synthesise an organic WWTP considering COD balance, BOD balance, and TSS balance. In
this respect, the formulation for volumetric flowrate balance and component balance for an organic WWTP can
be found in Ho et al. (2019). The concentration of components (COD, BOD and TSS) present in the treated
wastewater from the synthesised WWTP will always comply with the local discharge regulation as shown in the
constraint in Eq(1).
𝐶𝐶𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐,𝑐𝑐𝑡𝑡𝑐𝑐𝑡𝑡𝑐𝑐𝑐𝑐𝑡𝑡 𝑊𝑊𝑊𝑊𝑊𝑊 ≤ Ccomponent,std (1)
374
3.2 Area footprint
The formulation of area footprint is based on the area factor, P (m2) obtained from literature review or industrial
vendors as shown in Eq(2). 𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎𝑏𝑏 represents the area footprint for present treatment stage b while 𝑃𝑃𝑏𝑏
Area
represents the area factor for each technology in the present treatment stage b.
𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎𝑏𝑏 = 𝑃𝑃𝑏𝑏
Area ∀𝑏𝑏 (2)
3.3 Cost computation
The total investment cost for technology b, 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Total (USD/m3) includes the capital expenditure (CAPEX),
operating cost, and the land area cost formulated as shown in Eq(3) to Eq(6). The CAPEX for present technology
b, 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
CAPEX (USD) is formulated based on the CAPEX factor, 𝑃𝑃𝑏𝑏
CAPEX (USD) retrieved from literature reviews or
industrial vendors. The operating cost for present technology b, 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Opt (USD/m3) is contributed from the
material cost for technology b, 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Mat (USD/m3) and power cost for technology b, 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Power (USD/m3). In
addition, 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Area (USD) represents the total land area cost for present technology b and 𝑃𝑃𝑏𝑏
LandCost (USD/m2)
represents the area cost factor for technology b. To annualised the CAPEX and land area cost, an annualised
factor, 𝑃𝑃Annual is obtained based on the expected operating years and inflation rate.
𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
CAPEX = 𝑃𝑃𝑏𝑏
CAPEX ∀𝑏𝑏 (3)
𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Opt = Cost𝑏𝑏
Mat + Cost𝑏𝑏
Power ∀𝑏𝑏 (4)
𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Area = 𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎𝑏𝑏𝑃𝑃𝑏𝑏
LandCost ∀𝑏𝑏 (5)
𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Total = (𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
CAPEX + 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Area)𝑃𝑃Annual + 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑏𝑏
Opt ∀𝑏𝑏 (6)
3.4 Carbon footprint
Carbon footprint formulation is presented in Eq(7) where, 𝐶𝐶𝑎𝑎𝐴𝐴𝑏𝑏𝐶𝐶𝑛𝑛𝑏𝑏 (kg CO2/d) is the carbon footprint for present
technology b, 𝐸𝐸𝑏𝑏
power (kWh/m3) is the power consumption for technology b and 𝑃𝑃𝑏𝑏
carbon (kg CO2/kWh) is the
carbon footprint factor representing how much carbon dioxide will be released for every power consumption.
𝐶𝐶𝑎𝑎𝐴𝐴𝑏𝑏𝐶𝐶𝑛𝑛𝑏𝑏 = 𝐹𝐹𝑏𝑏
in𝐸𝐸𝑏𝑏
power𝑃𝑃𝑏𝑏
carbon ∀𝑏𝑏 (7)
3.5 Fuzzy optimisation
To incorporate multiple objectives that are contradictory in nature, fuzzy optimisation is used as shown in Eq(8)
to Eq(10). Fuzzy optimisation integrates multiple objectives into a single parameter called the degree of
satisfaction, λ. λ ranges from 0 to 1, whereby 0 indicates the total area footprint, cost, and carbon footprint are
approaching their upper limits (undesirable) while 1 indicates these optimisation objectives are approaching
their lower limits (desirable). Higher λ represents higher satisfaction for each objective. The total area footprint,
cost, and carbon footprint are represented by TotArea (m2), TotCost (USD/m3) and TotCarbon (kg CO2).
Superscript UL represents the predetermined upper limit each objective while superscript LL represents the
predetermined lower limit for each objective. These predetermined limits can be obtained based on decision-
makers’ interests or determined by optimising the model one objective at a time (Zadeh, 1965).
TotAreaUL−𝑊𝑊𝑐𝑐𝑐𝑐𝑇𝑇𝑡𝑡𝑐𝑐𝑡𝑡
TotAreaUL−TotAreaLL
≥ 𝜆𝜆 (8)
TotCostUL−𝑊𝑊𝑐𝑐𝑐𝑐𝑇𝑇𝑐𝑐𝑇𝑇𝑐𝑐
TotCostUL−TotCostLL
≥ 𝜆𝜆 (9)
TotCarbonUL−𝑊𝑊𝑐𝑐𝑐𝑐𝑇𝑇𝑡𝑡𝑡𝑡𝑏𝑏𝑐𝑐𝑐𝑐
TotCarbonUL−TotCarbonLL
≥ 𝜆𝜆 (10)
To solve the fuzzy model, the fuzzy degree of satisfaction, λ will be maximised as shown in Eq(11).
Maximise 𝜆𝜆 (11)
375
4. Case study
In this case study, a WWTP is synthesised to treat sago wastewater obtained from a sago-based biorefinery in
Sarawak, Malaysia (Wan and Ng, 2015). Sago wastewater enters at a flowrate of 276 m3/d with a COD level of
7,763 ppm, BOD level of 3,362 ppm, and TSS level of 4,942 ppm. Based on the local discharge legalisation
(Standard B), the synthesised WWTP in this work will reduce the level of COD, BOD and TSS present in the
sago wastewater to 200 ppm, 50 ppm, and 100 ppm (Department of Environment Malaysia, 2010). A series of
wastewater treatment technologies suitable for this case study were included as shown in Figure 2.
Figure 2: Case study superstructure of sago biorefinery WWTP in Sarawak, Malaysia.
Table 1 and Table 2 summarised the wastewater treatment technologies specifications such as removal
efficiency, dryness, CAPEX, material cost (MAT), power cost (POW), area footprint (AF), and power
consumption (E) obtained from industrial partners. To calculate the power cost, local electricity rates were
obtained. According to Sarawak Energy Berhad (2020), the electrical rate is USD 4.58/kWh for heavy industry
class I1. In addition, the carbon emission factor for Malaysia is 0.693 kg CO2/kWh (International Energy Agency
(IEA), 2016). The land area cost of the WWTP is calculated based on local industrial land rates in Sarawak
which is USD 212.25/m2 (Malaysian Investment Development Authority (MIDA), 2020).
Table 1: Specifications of case study wastewater treatment technologies (Lakghomi et al., 2015)
Table 2: Specifications of case study sludge treatment technologies (Wang et al., 2019)
Technologies
Removal efficiency
(%) CAPEX
(USD)
MAT
(USD/m3)
POW
(USD/m3)
AF
(m2)
E
(kWh/m3)
TSS COD BOD
Bar Screen 65 0 0 4,128 0 0 15.00 0
Grit Removal 40 0 0 7,523 0 0 20.00 0
DAF 91 70 65 3,440 0.34 1.02 15.95 0.222
CAAS 85 85 88 22,936 0.60 12.17 - 2.652
MBR 99 90 90 68,807 0.46 16.35 - 3.565
MBBR 85 90 92 57,339 0.37 9.27 - 2.022
Chlorination 16 56 55 18,349 0.73 0.12 7.84 0.026
Ozone
disinfection
60 83 70 22,936 0.46 4.82 3.92 1.051
Activated carbon
system
58 65 60 11,009 1.15 0.50 6.09 1.110
Technologies
Dryness
(kg SS/m3)
CAPEX
(USD)
Operating cost (USD/m3) Area
(m2)
E (kWh/m3)
MAT POW
Filter press 25.0 18,349 0.11 8.26 2.80 1.800
Grit Removal 29.9 22,936 0.11 6.78 4.30 1.478
Activated carbon system 28.5 27,523 0.13 11.47 3.00 2.500
376
The FMOO model developed for this case study is a mixed integer non-linear programming (MINLP) model.
Using a commercial optimisation software, LINGO (version 18.0) the case study is solved with computer
specification of Intel ® Core ™ i7-6500U @ 8 GB RAM, x64-based processor. An optimum wastewater treatment
pathway with minimal area footprint, total investment cost, and carbon footprint will be synthesised. This will be
done by optimising each objective individually to determine the upper and lower limits for each optimisation
objective as summarised in Table 3.
Table 3: Summarised results for each optimised objective WWTP pathways.
The area footprint, total investment cost, and carbon footprint for each optimised pathway were tabulated in
Table 4. By optimising each objective individually, the best (lower limit) values and worst (upper limit) values
can be identified to set a range for trade-offs. A global optimised wastewater treatment pathway at a degree of
satisfaction, λ of 0.5813 is synthesised as shown in Figure 3. The optimum wastewater treatment process
consists of bar screen, dissolved air flotation (DAF) tank, moving bed biofilm reactor (MBBR), activated carbon
system, and filter press. Comparing to the worst scenarios, by applying a maximum trade-off percentage of
58.13 %, the area footprint is decreased from 60.10 m2 (Pathway C) to 54.75 m2. At the same time, the cost
needed for the WWTP is decreased from 20,273.43 USD/d (Pathway B) to 19,286.77 USD/d where lesser
advanced technology in preliminary treatment and sludge treatment is chosen. In addition, a significant amount
of carbon footprint is reduced from 657 kg CO2/d (Pathway A) to 549 kg CO2/d.
Table 4: Comparison of wastewater treatment pathways at different optimisation objectives.
*UL – Upper limit LL – Lower limit
Figure 3: Optimal configuration of case study WWTP with traded off cost, area footprint, and carbon footprint.
5. Conclusion
In conclusion, this work had developed a decision-making tool to synthesise WWTP based on different
optimisation objectives such as minimum cost, minimum area footprint, and minimum carbon footprint. The
contradicting problem between these optimisation objectives is solved via FMOO. The case study in this work
Pathways
Preliminary
treatment
Chemical
treatment
Biological
treatment
Tertiary
treatment
Sludge
treatment
Min. area footprint (Pathway A) Bar screen DAF MBBR
Ozone
disinfection
Centrifugal
Min. total investment cost
(Pathway B)
Bar screen DAF MBBR
Ozone
disinfection
Filter press
Min. carbon footprint (Pathway C)
Grit
removal
DAF MBBR
Activated
carbon system
Belt filter
press
Objective function Area footprint (m2)
Total investment
cost (USD/d)
Carbon footprint
(kg/d)
Min. area footprint (Pathway A) 52.33LL 19,915.93 657.38UL
Min. total investment cost (Pathway B) 53.13 18,576.11LL 631.75
Min. carbon footprint (Pathway C) 60.10UL 20,273.43UL 472.04LL
Final optimised results (λ=0.5813) 54.75 19,286.77 549.64
377
had demonstrated that area footprint is an important criterion to be considered during the synthesis of WWTP
as the cost associated with the land area is significant. Results from the case study indicate that a maximum
trade-off at approximately 58 % would reduce 5.35 m2 of area footprint, 986 USD/d of total investment cost, and
108 kg CO2/d of carbon footprint. For future work, the weightage of each optimisation objective can be included
in the model to allow a more extensive decision making during a WWTP synthesis.
Acknowledgement
The authors would like to gratefully acknowledge the financial support provided by the Taylor's Internal Research
Grant Scheme - Emerging Research Funding Scheme (TIRGS-ERFS) (TIRGS-ERFS/1/2019/SOE/002) and
LINDO Systems for providing academic licenses to complete this research work.
References
Ang M.S., Duyag J., Tee K.C., Sy C.L., 2019, A multi-period and multi-criterion optimization model integrating
multiple input configurations, reuse, and disposal options for a wastewater treatment facility, Journal of
Cleaner Production, 231, 1437–1449.
Adeni D.S., Abd-Aziz S., Bujang K., Hassan M.A., 2010, Bioconversion of sago residue into value added
products. African Journal of Biotechnology, 9(14), 2061–2021.
Butler R.A., 2013, Malaysia has the world’s highest deforestation rate, reveals Google forest map, Mongabay
News & Inspiration from Nature’s Frontline accessed 06.06.2020.
Cai W., Liu J., Zhang X., Jern W., Liu Y., 2016, Generation of dissolved organic matter and byproducts from
activated sludge during contact with sodium hypochlorite and its implications to on-line chemical cleaning in
MBR, Water Research, 104, 44–52.
Department of Environment Malaysia, 2010, Environmental Requirements: A guide for investors, Putrajaya,
Malaysia.
Drews A., 2010, Membrane fouling in membrane bioreactors — Characterisation , contradictions , cause and
cures, Journal of Membrane Science, 363, 1–28.
Fernandez-Dacosta C., Posada J.A., Ramirez A., 2016, Techno-economic and carbon footprint assessment of
methyl crotonate and methyl acrylate production from wastewater-based polyhydroxybutyrate (PHB),
Journal of Cleaner Production, 137, 942–952.
Ho J.Y., Wan Y.K., Andiappan V., Ng D.K.S., 2019, Material flow cost account-based approach for synthesis
and optimisation of wastewater treatment plant, Chemical Engineering Transactions, 76, 529–534.
International Energy Agency (IEA), 2016, Carbon dioxide emission factor accessed 06.06.2020.
Lakghomi B., Lawryshyn Y., Hofmann R., 2015, A model of particle removal in a dissolved air flotation tank:
Importance of stratified flow and bubble size, Water Research, 68, 262–272.
Malaysian Investment Development Authority, 2020, Cost of Industrial Land, Starting a Business
accessed 06.06.2020.
Padrón-páez J.I., Almaraz S.D., Román-martínez A., 2020, Sustainable wastewater treatment plants design
through multiobjective optimization, Computers and Chemical Engineering, 140, 1–16.
Pan W., Yu L., Wang S., Wang X., 2014, A fuzzy multi-objective model for provider selection in data
communication services with different QoS levels, International Journal of Production Economics, 147,
689–696.
Sarawak Energy Berhad., 2020, Tariff Category Units accessed 06.06.2020.
Wan Y.K., Ng D.K.S., 2015, Synthesis and Optimisation of Total Water Network via Material Flow Cost
Accounting (MFCA)-Based Approach, Chemical Engineering Transactions, 45, 1519–1524.
Wang H., Xu G., Qiu Z., Zhou Y., Liu Y., 2019, Chemosphere NOB suppression in pilot-scale mainstream
nitritation-denitritation system coupled with MBR for municipal wastewater treatment, Chemosphere, 216,
633–639.
World Wide Fund Nature (WWF), 2020, Forests
accessed 27.05.2020.
Yunus N., Jahim J., Anuar N., 2014, Batch fermentative hydrogen production utilising sago (Metroxylon sp.)
starch processing effluent by enriched sago sludge consortia, International Journal of Hydrogen Energy,
39(35), 19937–19946.
Zadeh L.A., 1965, Fuzzy Sets, Information and Control, 8, 338–353.
Zimmermann H.J., 1978, Fuzzy programming and linear programming with several objective functions, Fuzzy
Sets and System, 1(1), 45–55.
378