Abstract

The issues of freshwater scarcity in arid and semi-arid areas could be reduced via treated municipal wastewater effluent (TMWE). Artificial intelligence methods, especially the fuzzy inference system, have proven their ability in TMWE quality evaluation in complex and uncertain systems. The primary aim of this study was to use a Mamdani fuzzy inference system to present an index for agricultural application based on the Iranian water quality index (IWQI). Since the uncertainties were disregarded in the conventional IWQI, the present study improved this procedure by using fuzzy logic and then the fuzzy effluent quality index (FEQI) was proposed as a hybrid fuzzy-based index. TMWE samples of the Gheitarie wastewater treatment plant in Tehran city recorded from 2011 to 2017 were taken into consideration for testing the ability of the proposed index. The results of the FEQI showed samples categorized as ‘Excellent’ (21), ‘Good’ (10), ‘Fair’ (4), and ‘Marginal’ (1) for the warm seasons, and for the cool seasons, the samples categorized as ‘Excellent’, ‘Good’ and ‘Fair’ were 17, 18 and 1, respectively. Generally, a comparison between the IWQI and proposed model results revealed the FEQI's superiority in TMWE quality assessment.

INTRODUCTION

Municipal effluent discharged from a wastewater treatment plant is a reliable source of water supply, especially in arid and semi-arid areas where urban population growth and the sharp increase in water intake has raised concerns about water shortage as a critical point of water supply. Wastewater treatment and its use for different purposes have gained remarkable attention during recent years from governors and other authorities around the world (Verlicchi et al. 2011; Becerra-Castro et al. 2015). Using treated municipal wastewater effluent (TMWE) in Iran as a semi-arid country can be a suitable aid in dealing with water crisis (Nadiri et al. 2018). The largest volume of treated wastewater has been consumed in the irrigation field so far (Alfarra et al. 2011). Indeed, the TMWE has to regard all quality standards regulated by official entities for treated wastewater reutilization (Verlicchi et al. 2011). Different parameters should be considered in order to assign a single value as the water quality index (WQI). A method, therefore, is needed to represent a comprehensive quality index which includes all the determinants that becomes an applicable tool in the decision-making process. TMWE quality assessment with the employment of the WQI has been conducted in some studies (e.g., Silva et al. 2014; Nezhad et al. 2016). Providing a comprehensive view of water quality, Verlicchi et al. (2011) introduced a novel index as the wastewater polishing index (WWPI). Babaei Semiromi et al. (2011) proposed an Iranian water quality index (IWQI) and compared it with other well-accepted WQIs for the case study of Karoon River, Iran. The main flaw of the proposed indices was that they did not consider the uncertainty and subjectivity of quality assessment. To cover such a flaw, some novel methods were introduced for developing WQI based on artificial intelligence (AI) computational methods (Chau 2006; Nezhad et al. 2016). Among the AI methods, fuzzy logic was introduced in developing WQI for its ability in reflecting human thoughts and expertise to deal with non-linear and uncertain information (Falah Nezhad et al. 2015; Kizhisseri & Mohamed 2016; Li et al. 2016; Bórquez-Lopez et al. 2018; Nadiri et al. 2018; Rahimi et al. 2019). According to these studies, fuzzy set theory can give more flexibility in decision making for complicated circumstances, where the uncertainties are increased.

Reviewing the studies shows that WQI research has been mostly conducted for the groundwater and surface water fields. And, there is still no comprehensive view of TMWE based on fuzzy logic for agricultural irrigation (Ebrahimi et al. 2017; Tiri et al. 2018). More specifically, to the best of the authors' knowledge, no study has been carried out to evaluate municipal effluent using fuzzy logic for agricultural irrigation purposes.

Therefore, the novelty of the present study is to develop a hybrid fuzzy-based index of municipal effluent on the basis of the IWQI for agricultural irrigation specifically that is representative of overall effluent quality. The proposed index application was tested by the real data of the Gheitarie wastewater treatment plant in Tehran, Iran.

MATERIALS AND METHODS

Site characteristics

Gheitarie wastewater plant is the first of its kind in Iran to work with a deep extended aeration method. Gheitarie wastewater treatment plant currently responds to the needs of 40,000 people. To determine the quality of the TMWE, physicochemical and microbiological parameters were assessed in 72 TMWE samples collected monthly from April to September (warm months) and October to March (cool months).

Model structure

Theoretical concept

The deterministic indices for water quality assessment introduced by the wide scope of the studies and the implemented WQI approaches have not considered uncertainties. Fuzzy logic, introduced by Zadeh (1965), is a promising approach in addressing issues of vagueness and uncertainty due to its ability in applying the expert knowledge to address issues of uncertainty. Fuzzy inference systems (FISs) transfer expert knowledge (linguistic terms) through the reasoning processes of if-then rules. This rule-based system is useful in dealing with the intrinsic uncertainty of the results, especially when dealing with water quality problems. Since the input–output measurements of effluent quality are difficult to obtain and have inherent uncertainties that are difficult to address, the present study uses a fuzzy inference system (FIS) to develop a hybrid fuzzy effluent quality index (FEQI) based on expert knowledge.

The Mamdani is a popular model among the FISs that has been used in the field of water quality assessment due to its flexibility. Since the Mamdani FIS is an applicable approach in water quality evaluation, lots of researchers have used it, such as Dahiya et al. (2007) and Jinturkar et al. (2010). The ‘min’ and ‘max’ are then applied in the implication and for the aggregation of the consequence through the if–then rules, respectively. The proposed model used the Mamdani inference system for its transparency regarding the rule-based fuzzy system and membership functions. While the selection and weighting parameters in proposed WQIs have been conducted via mathematical equations, the present study selected the ranges of the input parameters based on Iranian Department of Environment (IDOE) and Food and Agriculture Organization (FAO) prescribed standard limits. The ranges of the output parameters were indicated based on the IWQI proposed by Babaei Semiromi et al. (2011). The proposed model is illustrated in Figure 1.

Figure 1

Model structure.

Figure 1

Model structure.

FIS usually consists of three phases, (1 – fuzzification, 2 – fuzzy rule inference and 3 – defuzzification), which can be utilized in dealing with various kinds of uncertainty. More information of fuzzy logic and FIS can be further studied in Ross (2012).

FEQI development

Fuzzification of input parameters

The parameter selection for any effluent quality assessment depends upon the application and the ability of an organization in carrying out accurate parameter evaluation. Local environment and background quality should be considered in selection of the parameters. Physicochemical and microbiological parameters should be indicated by considering the scale of the treatment procedure and the planned utilization. The selected parameters should cover an extended range of environmental and toxicological threats and probable technical problems (Salgot et al. 2006).

In this research, biochemical oxygen demand (BOD), chemical oxygen demand (COD), total suspended solids (TSS), total dissolved solids (TDS), pH, nitrate, NO2−3, dissolved oxygen (DO), fecal coliform (FC), nematodes, and PO4−3 were the selected initial parameters of the FEQI model which were collected from the Gheitarie wastewater treatment plant from 2011 to 2017. They were considered for the development of an FIS model to assess effluent quality. Although more parameters improve the suitability of a model, they also increase the uncertainty.

In the present study, a panel of well-known experts in wastewater effluent quality management was recruited from Universities, Municipalities, the Department of Environment, and water and wastewater section of the Ministries of Energy, Hygiene and Agriculture. The panelists were asked to select and rate the parameters that they would consider in a WQI through the questionnaires including the selected initial parameters. The questionnaires were completed by the specialists, following the Delphi procedure. Average weights were applied to the parameters afterwards. The indicated parameters were eight parameters including fecal coliform, nematodes, pH, TDS, TSS, COD, BOD and nitrate based on the questionnaires' results and as the input parameters of the proposed model. The statistics of selected parameters for both warm and cool seasons are presented in Table 1.

Table 1

The descriptive statistics of the agricultural water parameters

WWTP nameCriteriaSeasonFC (CFU/100 ml)Nematodes (N/L)pH (−)TDS (mg/L)TSS (mg/L)COD (mg/L)BOD (mg/L)Nitrate (mg/L)
Gheitarie minimum Warm 4/3 0/0 6/1 329/0 6/0 10/6 0/0 114/2 
maximum 929/1 1/0 7/4 615/0 208/7 402/0 235/0 163/2 
mean 315/8 0/5 6/8 415/4 56/5 85/0 43/2 131/3 
minimum Cool 3/0 0/0 6/2 332/0 8/0 10/6 3/0 112/7 
maximum 836/8 1/0 7/4 500/0 216/3 253/0 168/0 160/3 
mean 63/0 0/5 6/9 433/3 48/4 57/0 33/0 136/8 
Standard – – 400 (IDOE) 1 (IDOE) 6.5–8.5 (IDOE) 450 (FAO) 100 (IDOE) 200 (IDOE) 100 (IDOE) 30 (FAO) 
WWTP nameCriteriaSeasonFC (CFU/100 ml)Nematodes (N/L)pH (−)TDS (mg/L)TSS (mg/L)COD (mg/L)BOD (mg/L)Nitrate (mg/L)
Gheitarie minimum Warm 4/3 0/0 6/1 329/0 6/0 10/6 0/0 114/2 
maximum 929/1 1/0 7/4 615/0 208/7 402/0 235/0 163/2 
mean 315/8 0/5 6/8 415/4 56/5 85/0 43/2 131/3 
minimum Cool 3/0 0/0 6/2 332/0 8/0 10/6 3/0 112/7 
maximum 836/8 1/0 7/4 500/0 216/3 253/0 168/0 160/3 
mean 63/0 0/5 6/9 433/3 48/4 57/0 33/0 136/8 
Standard – – 400 (IDOE) 1 (IDOE) 6.5–8.5 (IDOE) 450 (FAO) 100 (IDOE) 200 (IDOE) 100 (IDOE) 30 (FAO) 

The efficiency of the FIS model depends on the relevant inputs, number of membership functions for each input and the corresponding numerical data. The subject of the study can affect the shape of the selected membership function (Klir & Yuan 1995). Membership functions exist in various shapes, however, the simple ‘trapezoidal’ shape is widely chosen due to its simplicity in practical issues (Barua et al. 2014). Figure 2 depicts the input and output parameters' membership functions. In the present study, trapezoidal and Gaussian membership functions were used based on the physicochemical data and past research to represent fuzzy sets using crisp input by considering the standard limits prescribed by the IDOE and FAO. The parameters of the fuzzy rules besides the shape of the membership functions have been adjusted in order to obtain an optimal fuzzy system. Table 1 shows the descriptive statistics of the TMWE parameters. The if–then rules and the output parameters determine a fuzzy domain of the selected inputs with the corresponding outputs, respectively. The FIS functionality, thus, depends on the pertaining selected input parameters, the number of the membership functions, assigned to each one, and the relevant measured data.

Figure 2

The input and output membership functions of FEQI.

Figure 2

The input and output membership functions of FEQI.

Fuzzification of output parameters

In this study, the IWQI, which is the Iranian water quality index, was selected to develop the new fuzzy-logic-based index: FEQI. Output membership functions were defined based on IWQI.

The overall IWQI is calculated as below: 
formula
(1)
where , and are the weights and sub-index for the th parameter, respectively.

The output membership functions of the FEQI model were categorized based on the IWQI classes. The IWQI as a dimensionless number from 0 to 100 can be represented in five classes: ‘Poor’ (95–100); ‘Marginal’ (75–94); ‘Fair’ (50–74); ‘Good’ (25–49); and ‘Excellent’ (0–25) (Babaei Semiromi et al. 2011).

Development of fuzzy rules

Rules were developed based on expert knowledge for the construction of the model. Input parameters and membership functions are the basis of indicating the number of rules. For large numbers of rules, the complexity of the model increases. Therefore, removal of less important rules results in a compact fuzzy model with better generalizing ability and an overall simplification of the system architecture (Yen & Wang 1999).

The model developed here consisted of eight input parameters with a different number of membership functions for each parameter, including two for nematodes, five for pH, and three for the other parameters respectively. To avoid model complexity, a total of 178 most important rules, of which four examples are presented below, were selected for constructing the fuzzy model on the basis of the available datasets and the experience of experts.

  • (1)

    If fecal coliform, nematodes, pH, TDS, TSS, COD, BOD and nitrate are low; then: FEQI is ‘Excellent’.

  • (2)

    If fecal coliform, nematodes, pH, TSS, COD and BOD are low, whereas the TDS and nitrate are moderate; then: FEQI is ‘Good’.

  • (3)

    If TSS, COD and BOD are low, while the pH is moderate, but the fecal coliform, nematodes, nitrate and TDS are high; then: FEQI is ‘Fair’.

  • (4)

    If fecal coliform, nematodes, TDS, TSS, COD, BOD, pH and nitrate are high; then: FEQI is ‘Poor’.

Performance criteria

Common performance criteria methods including the variance accounted for (VAF), root mean square error (RMSE), and coefficient of determination, were applied to evaluate the proposed model performance (Shams et al. 2015).

The RMSE reveals the bias between the simulated and measured data, whereas the VAF indicates the degree of difference between the measured and predicted data variances. The R2 is a good indicator of the model performance, which is between measured and predicted values. A high level of R2 indicates conformity between the output of the model and the real measured values. RMSE and VAF are calculated as below (Equations (2) and (3)): 
formula
(2)
 
formula
(3)
where x, and N are the measured and simulated data and number of values, respectively.

RESULTS AND DISCUSSION

FEQI results

The 72 TMWE samples of the Gheitarie wastewater treatment plant were assessed by using the FEQI model besides assessment by the crisp index. The results show that 21 samples were categorized as ‘Excellent’, ten as ‘Good’, four as ‘Fair’, and one as ‘Marginal’ for the warm seasons. The results also show 17 samples categorized as ‘Excellent’, 18 as ‘Good’ and one as ‘Fair’ for the cool seasons, respectively (Table 2).

Table 2

Results of TMWE quality evaluation using IWQI and FEQI models

WWTP nameRankWarm season
Cool season
IWQI
FEQI
IWQI
FEQI
SamplesPercentSamplesPercentSamplesPercentSamplesPercent
Gheitarie Excellent 22 0.61 21 0.58 14 0.39 17 0.47 
Good 0.25 10 0.28 21 0.58 18 0.50 
Fair 0.11 0.11 0.03 0.03 
Marginal 0.03 0.03 0.00 0.00 
Poor 0.00 0.00 0.00 0.00 
WWTP nameRankWarm season
Cool season
IWQI
FEQI
IWQI
FEQI
SamplesPercentSamplesPercentSamplesPercentSamplesPercent
Gheitarie Excellent 22 0.61 21 0.58 14 0.39 17 0.47 
Good 0.25 10 0.28 21 0.58 18 0.50 
Fair 0.11 0.11 0.03 0.03 
Marginal 0.03 0.03 0.00 0.00 
Poor 0.00 0.00 0.00 0.00 

Crisp and hybrid fuzzy index comparison

The main aim of this research was to present a new hybrid fuzzy index (FEQI) for TMWE quality assessment based on the accepted Iranian index (IWQI).

Figure 3 shows numerical values of determinist index and hybrid fuzzy index (IWQI and FEQI) involved in decision-making for both warm and cool seasons.

Figure 3

The comparison between the calculated crisp index and predicted hybrid fuzzy index.

Figure 3

The comparison between the calculated crisp index and predicted hybrid fuzzy index.

Samples 38 and 40 were categorized as ‘Fair’ as a result of both IWQI and FEQI models in which most parameters were in the ‘High’ category. For sample 9 (cool season), most of the parameters (FC, nematodes, TSS, COD, BOD) were in the ‘Low’ classification and the sample was categorized as ‘Excellent’ based on both IWQI and FEQI. The quality of sample 20 for the cool season was ‘Excellent’ by the FEQI. FC, nematodes, TSS, COD, BOD were ‘Low’, while pH was ‘Medium’ and TDS and nitrate were in the ‘High’ category. Sample 29 was classified as ‘Excellent’ and ‘Good’ by IWQI and FEQI, respectively. Sample 41 (warm seasons), having most parameters (FC, nematodes, nitrate, TDS, TSS, BOD) in the ‘High’ category, was classified as ‘Marginal’ according to both IWQI and FEQI. Samples 55, 56 and 57 were categorized as ‘Excellent’ for the cool seasons while with the deterministic method, FC, nematodes, TSS, COD, BOD were in the ‘Low’, pH in the ‘Medium’ and TDS, nitrate in the ‘High’ category. The results of samples 43 and 44 (cool seasons) were classified in the ‘Good’ category according to both indices.

With cool season samples 43 and 44, differences in input parameter between the two models are evident: the TSS input of sample 43 was classified as ‘High’ and for sample 44 it was classified as ‘Low’ but values of IWQI and FEQI were good. Therefore, results of the comparison between IWQI and FEQI showed that despite the differences that occurred in sample parameters, the output category was the same according to both indices (IEQI and FEQI). This indicated that the ranges of output categories can be well defined by the fuzzy quality index as the crisp model can give a deterministic numerical value of the effluent quality while the fuzzy index is capable of dealing with inherent uncertainties by considering all the border values or near-identical qualities. Therefore, where input parameter values fall on the margin of the standard, both expert judgment and the prescribed limits are useful to provide a better effluent quality assessment via fuzzy membership functions and if–then rules instead of a single value of the standard limits. The FEQI overcomes the mentioned deficiency by placing the sample in the appropriate category by using the present data. Based on expert judgment and the sensitivity of the model to border values, the FEQI model clearly exhibited a better performance than the IWQI, which supports the superiority of the fuzzy model compared with the crisp model in effluent quality assessment.

Model performance

VAF and RMSE, which are the reliable performance indices, were calculated to evaluate the applicability of the developed hybrid fuzzy model. The VAF value for the FEQI is 88 percent in the Gheitarie plant. The RMSE value for the FEQI is 6.9. The value of R2 between IWQI and FEQI is 0.88 (Figure 3). The results showed that the hybrid fuzzy index, FEQI, presents a reasonable accordance with the crisp index, which subsequently reveals the reliability of the hybrid fuzzy index.

Based on these results, it can be inferred that the proposed Mamdani fuzzy model is a good alternative tool for minimizing the inherent uncertainties in effluent quality assessment in comparison with the crisp effluent quality index.

CONCLUSIONS

The main aim of the current paper was to present a new fuzzy index for municipal effluent quality assessment for agricultural irrigation. FIS was created according to the standard limits as inputs of the model and IWQI ranges as the model outputs. A total of 72 monthly effluent samples from 2011 to 2017 were used, which were measured and recorded by Gheitarie Wastewater Treatment Plant, Iran. The FIS model was developed based on the selected parameters as the model inputs and IWQI ranges as the outputs of the model. The resulting fuzzy values were compared with the crisp numerical values of IWQI. Most of the samples were categorized as ‘Excellent’ for warm and cool seasons according to both IWQI and FEQI models. The results of the FEQI model in Gheitarie wastewater treatment plant indicated 21, 10, 4 and 1 samples categorized as ‘Excellent,’ ‘Good’, ‘Fair’ and ‘Marginal’ for the warm seasons, respectively. Also, the results showed 17, 18 and 1 samples categorized as ‘Excellent’, ‘Good’ and ‘Fair’. The same results of the two indices showed that the FEQI can decrease uncertainty by considering the border values and placing the sample in the appropriate category. It can be concluded that the FEQI model could be a good alternative tool in minimizing the inherent uncertainties in effluent quality assessment in comparison with the crisp IWQI. Therefore, where input parameter values fall in the margin of the standard, FIS is useful to provide a better effluent quality assessment via fuzzy membership functions and if–then rules instead of a single value of the standard limits. The proposed model could be a great aid for decision makers, particularly in those kinds of plants in which comparison of the effluent quality is crucial and fast decision-making in evaluation of the different scenarios is vital. The VAF, RMSE and R2 performance indices for the FEQI showed close accordance between crisp and fuzzy indices using fuzzy inference models vs the other developed models. The proposed FIS model of the study can be used in further research to investigate the role of the related indices, the different standards and permissible limits prescribed by different organizations for industrial, recreational, groundwater recharge, and aquaculture applications. Further studies can be also conducted by using the proposed FIS in assessing industrial effluent for different applications.

REFERENCES

REFERENCES
Alfarra
A.
Kemp-Benedict
E.
Hötzl
H.
Sader
N.
Sonneveld
B.
2011
A framework for wastewater reuse in Jordan: utilizing a modified wastewater reuse index
.
Water Resources Management
25
(
4
),
1153
1167
.
https://doi.org/10.1007/s11269-010-9768-8
.
Babaei Semiromi
F.
Hassani
A. H.
Torabian
A.
Karbassi
A. R.
Lotfi
F. H.
2011
Evolution of a new surface water quality index for Karoon catchment in Iran
.
Water Science and Technology
64
(
12
),
2483
2491
.
https://doi. 10.2166/wst.2011.780
.
Barua
A.
Mudunuri
L. S.
Kosheleva
O.
2014
Why Trapezoidal and Triangular Membership Functions Work So Well: Towards a Theoretical Explanation
.
Department Technical Reports: Paper 783/Technical Report UTEP-CS-13-50, University of Texas at El Paso, El Paso, TX, USA
. .
Becerra-Castro
C.
Lopes
A. R.
Vaz-Moreira
I.
Silva
E. F.
Manaia
C. M.
Nunes
O. C.
2015
Wastewater reuse in irrigation: a microbiological perspective on implications in soil fertility and human and environmental health
.
Environment International
75
,
117
135
.
https://doi: 10.1016/j.envint.2014.11.001
.
Bórquez-Lopez
R. A.
Casillas-Hernandez
R.
Lopez-Elias
J. A.
Barraza-Guardado
R. H.
Martinez-Cordova
L. R.
2018
Improving feeding strategies for shrimp farming using fuzzy logic, based on water quality parameters
.
Aquacultural Engineering
81
,
38
45
.
https://doi.org/10.1016/j.aquaeng.2018.01.002
.
Chau
K. W.
2006
A review on integration of artificial intelligence into water quality modelling
.
Marine Pollution Bulletin
52
(
7
),
726
733
.
https://doi.org/10.1016/j.marpolbul.2006.04.003
.
Dahiya
S.
Singh
B.
Gaur
S.
Garg
V. K.
Kushwaha
H. S.
2007
Analysis of groundwater quality using fuzzy synthetic evaluation
.
Journal of Hazardous Materials
147
(
3
),
938
946
.
https://doi.org/10.1016/j.jhazmat.2007.01.119
.
Ebrahimi
M.
Gerber
E. L.
Rockaway
T. D.
2017
Temporal performance assessment of wastewater treatment plants by using multivariate statistical analysis
.
Journal of Environmental Management
193
,
234
246
.
https://doi.10.1016/j.jenvman.2017.02.027
.
Falah Nezhad
M.
Mehrdadi
N.
Torabian
A.
2015
Definition of a new domestic effluent quality index using TOPSIS decision making tool
.
Canadian Journal of Pure and Applied Sciences
9
(
2
),
3467
3472
.
Jinturkar
A. M.
Deshmukh
S. S.
Agarkar
S. V.
Chavhan
G. R.
2010
Determination of water quality index by fuzzy logic approach: a case of ground water in an Indian town
.
Water Science and Technology
61
(
8
),
1987
1994
.
https://doi.10.2166/wst.2010.095
.
Kizhisseri
M. I.
Mohamed
M. M. A.
2016
Fuzzy-based wastewater quality indices for pollution classification: a case study in the United Arab Emirates
.
Environment Systems and Decisions
36
(
1
),
62
71
.
https://doi.org/10.1007/s10669-015-9579-9
.
Klir
G. J.
Yuan
B.
1995
Fuzzy Sets and Fuzzy Logic
, 4th edn.
Prentice Hall
,
Upper Saddle River, NJ, USA
.
Li
R.
Zou
Z.
An
Y.
2016
Water quality assessment in Qu River based on fuzzy water pollution index method
.
Journal of Environmental Sciences
50
,
87
92
.
https://doi.org/10.1016/j.jes.2016.03.030
.
Nadiri
A. A.
Shokri
S.
Tsai
F. T.-C.
Moghaddam
A. A.
2018
Prediction of effluent quality parameters of a wastewater treatment plant using a supervised committee fuzzy logic model
.
Journal of Cleaner Production
180
,
539
549
.
https://doi.org/10.1016/j.jclepro.2018.01.139
.
Nezhad
M. F.
Mehrdadi
N.
Torabian
A.
Behboudian
S.
2016
Artificial neural network modeling of the effluent quality index for municipal wastewater treatment plants using quality variables: south of Tehran wastewater treatment plant
.
Journal of Water Supply: Research and Technology – Aqua
65
(
1
),
18
27
.
https://doi.org/10.2166/aqua.2015.030
.
Rahimi
M.
Ebrahimi
K.
Araghinejad
S.
2019
A novel approach in determination of the effluent applications
.
Journal of Water Supply: Research and Technology – Aqua
68
(
3
),
157
170
.
https://doi.org/10.2166/aqua.2019.079
.
Ross
T. J.
2012
Fuzzy Logic with Engineering Applications
.
John Wiley & Sons, Chichester, UK
.
Salgot
M.
Huertas
E.
Weber
S.
Dott
W.
Hollender
J.
2006
Wastewater reuse and risk: definition of key objectives
.
Desalination
187
(
1–3
),
29
40
.
https://doi.org/10.1016/j.desal.2005.04.065
.
Shams
S.
Monjezi
M.
Majd
V. J.
Armaghani
D. J.
2015
Application of fuzzy inference system for prediction of rock fragmentation induced by blasting
.
Arabian Journal of Geosciences
8
(
12
),
10819
10832
.
https://doi.org/10.1007/s12517-015-1952-y
.
Silva
C.
Quadros
S.
Ramalho
P.
Rosa
M. J.
2014
A tool for a comprehensive assessment of treated wastewater quality
.
Journal of Environmental Management
146
,
400
406
.
https://doi.org/10.1016/j.jenvman.2014.03.028
.
Tiri
A.
Belkhiri
L.
Mouni
L.
2018
Evaluation of surface water quality for drinking purposes using fuzzy inference system
.
Groundwater for Sustainable Development
6
,
235
244
.
https://doi.org/10.1016/j.gsd.2018.01.006
.
Verlicchi
P.
Masotti
L.
Galletti
A.
2011
Wastewater polishing index: a tool for a rapid quality assessment of reclaimed wastewater
.
Environmental Monitoring and Assessment
173
(
1–4
),
267
277
.
https://doi: 10.1007/s10661-010-1386-7
.
Yen
J.
Wang
L.
1999
Simplifying fuzzy rule-based models using orthogonal transformation methods
.
IEEE Transactions on Systems, Man, and Cybernetics, Part B
29
(
1
),
13
24
.
Zadeh
L. A.
1965
Fuzzy sets
.
Information Control
8
(
3
),
338
353
.
https://doi: 10.1016. S0019-9958
.