Statistical analysis to identify the main parameters to

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 03 Issue: 01 | Jan-2014, Available @ http://www.ijret.org 186
STATISTICAL ANALYSIS TO IDENTIFY THE MAIN PARAMETERS TO
EFFECTING WWQI OF SEWAGE TREATMENT PLANT AND
PREDICTING BOD
A.K.Khambete1
, R.A.Christian1
1, 2
Associate Professor, Faculty of civil Eng. S.V. National Institute of Technology, Surat
Abstract
The present study was conducted to determine the wastewater quality index and to study statistical interrelationships amongst
different parameters. The equation was developed to predict BOD and WWQI. A number of water quality physicochemical parameters
were estimated quantitatively in wastewater samples following methods and procedures as per governing authority guidelines.
Wastewater Quality Index (WWQI) is regarded as one of the most effective way to communicate wastewater quality in a collective way
regarding wastewater quality parameters. The WWQI of wastewater samples was calculated with fuzzy MCDM methodology. The
wastewater quality index for treated wastewater was evaluated considering eight parameters subscribed by Gujarat Pollution control
Board (GPCB), a governing authority for environmental monitoring in Gujarat State, India. Considerable uncertainties are involved
in the process of defining the treated wastewater quality for specific usage, like irrigation, reuse, etc.
The paper presents modeling of cognitive uncertainty in the field data, while dealing with these systems recourse to fuzzy logic. Also a
statistical study is done to identify the main affecting variables to the WWQI. The Statistical Regression Analysis has been found to be
highly useful tool for correlating different parameters. Correlation Analysis of the data suggests that TDS, SS, BOD, COD, O&G and
Cl are significantly correlated with WWQI and DO of wastewater. The estimated BOD from independent variance DO for maximum,
minimum and average is 25.35 mg/L, 2.65 mg/L and 13.56 mg/L respectively. While estimated WWQI from independent variance DO
for maximum, minimum and average is 0.6212, 0.3074 and 0.4581 respectively. Out of eight parameters, TDS-BOD, TDS-COD, TDS-
Cl, SS-BOD, SS-COD, and BOD-COD are significantly correlated. Present study shows that WWQI is influenced by BOD, COD, SS
and TDS.
Keywords: Wastewater quality, Uncertainty, fuzzy set theory, correlation, Regression.
------------------------------------------------------------------***----------------------------------------------------------------
1. INTRODUCTION
As urban and industrial development increases, the quality of
waste generated also increases (Mohmmed, 2006). Discharge
of untreated waste in to surface waters, such as rivers, lacks
and seas, determine the contamination of waters for human
purposes (Gabrieli et al., 1997). The treatment objectives were
concerned with the removal of suspended floatable material,
the treatment of biodegradable organics and elimination of
pathogenic organisms ( Jamrah, 1999). So the wastewater or
sewage treatment is one such alternative, where in many
processes are designed and operated in order to copy the
natural treatment process to reduce contaminant load to a level
that environment can hold ( Ravi kumar, 2010). The WQI is
presenting a cumulatively derived numerical expression
defining a certain level of water quality (Bordelo, 2006).
Water quality indices aim at giving a single value to the water
quality of a source reducing great amount of parameters into a
simpler expression and enabling easy interpretation of
monitoring data (Bharti, 2011). In other words, WQI
summarizes large amounts of water quality data into simple
terms (e.g., excellent, good, bad, etc.) for reporting to
management and the public in a consistent manner. Various
researchers have considered similar approaches which brought
changes to the methodology depending on the usage and
parameters under consideration. Prati et al. (1971) considered
13 different parameters of equal weight in their system. And
water quality index that is able to rapidly assess whether an
effluent—properly polished—is adequate for agricultural or
recreational purposes could be of great help to managers and
decision makers in water resources planning as well as for
comparing different wastewater treatment sequences
(Verlicchi et al., 2012).
Also there is a need to define one common parameter which
could determine the overall efficiency of plant in terms of
physical, biochemical and microbiological removal
efficiencies (Jamwal 2009). A new fuzzy approach initiated by
Zadeh 1965, has been adopted. i.e. Wastewater Quality Index
(WWQI), which represent the wastewater quality in terms of

IJRET: International Journal of Research in Engineering and Technology
__________________________________________________________________________________________
Volume: 03 Issue: 01 | Jan-2014, Available @
numeric i.e. 0 to 1. Developing the WWQI on the line of WQI
is expected to be more practical in implementation and can
become effective decision making tool for authority.
The Multi Criteria Decision Making (MCDM) method is
applied to determine the WWQI at Surat WWTP, and
maximum, minimum and average values for WWQI were
found out, which are 0.6284, 0.2835 and 0.4433 respectively.
Once, WWQI was calculated the correlation between WWQI
and all the parameters was calculated. Factor Analysis (FA)
was adopted to assist the interpretation of elemental data
(Akobundu 2007). So, an attempt has been made to find out
the solution of this problem by developing
computing correlation and regression equations to assess
the wastewater quality of an treated wastewater on the basis of
few parameters.
For this SPSS tool was used. From this analysis, parameters
significantly affecting WWQI were found o
and DO etc.). At last, considering significant parameters, a
regression equation was developed using SPSS.
This equation was used to calculate WWQI once again.
Finally, the results obtained from both MCDM method and
further the regression equations were compared and the
regression equation was developed to estimate WWQI and
BOD from measured value of DO.
2. WATER QUALITY INDEX BY MCDM
METHOD
Due to fluctuations in organic and hydraulic loading in a
wastewater treatment plant, the application of fuzzy logic,
using linguistic variables gave a better description of
performance parameters. The number of criteria in multi
criteria decision making is virtually assumed to be finite
(Evangelos,2000 and Klir,1998). Balteiro and Romero (2004)
utilized a sustainability aggregating model in search an index
to natural systems sustainability. Chen et al. (2011) utilized
fuzzy MCDM approach and fuzzy AHP for selecting the best
environment-watershed plan in Taiwan.
2.1 Multi Criteria Decision Making (MCDM)
approach for Determination of the WWQI
The approach for the evaluating WWQI of treated wastewater
is based on somewhat analogous to the procedure suggested
by Singh and Tong (2005). The chart portrays an overview of
the fuzzy decision framework for evaluating the MPAI, which
is self-explanatory Figure 1.The first step is identification of
environmental experts.
IJRET: International Journal of Research in Engineering and Technology eISSN: 2319
__________________________________________________________________________________________
, Available @ http://www.ijret.org
numeric i.e. 0 to 1. Developing the WWQI on the line of WQI
is expected to be more practical in implementation and can
l for authority.
The Multi Criteria Decision Making (MCDM) method is
applied to determine the WWQI at Surat WWTP, and
maximum, minimum and average values for WWQI were
found out, which are 0.6284, 0.2835 and 0.4433 respectively.
the correlation between WWQI
and all the parameters was calculated. Factor Analysis (FA)
was adopted to assist the interpretation of elemental data
(Akobundu 2007). So, an attempt has been made to find out
the solution of this problem by developing WWQI and
computing correlation and regression equations to assess
the wastewater quality of an treated wastewater on the basis of
For this SPSS tool was used. From this analysis, parameters
significantly affecting WWQI were found out (TDS, COD,
and DO etc.). At last, considering significant parameters, a
regression equation was developed using SPSS.
This equation was used to calculate WWQI once again.
Finally, the results obtained from both MCDM method and
compared and the
regression equation was developed to estimate WWQI and
2. WATER QUALITY INDEX BY MCDM
Due to fluctuations in organic and hydraulic loading in a
on of fuzzy logic,
using linguistic variables gave a better description of
performance parameters. The number of criteria in multi-
criteria decision making is virtually assumed to be finite
(Evangelos,2000 and Klir,1998). Balteiro and Romero (2004)
d a sustainability aggregating model in search an index
to natural systems sustainability. Chen et al. (2011) utilized
fuzzy MCDM approach and fuzzy AHP for selecting the best
Multi Criteria Decision Making (MCDM)
Determination of the WWQI
The approach for the evaluating WWQI of treated wastewater
is based on somewhat analogous to the procedure suggested
(2005). The chart portrays an overview of
the fuzzy decision framework for evaluating the MPAI, which
The first step is identification of
Fig 1:Step Involved in a MCDM Model
weightage to the sub criteria of wastewater.
Weight for each of the sub criterion is developed by
normalized weight method by consulting the environmental
experts. In the present study, five fuzzy numbers are selected
to describe the level of performance on de
evaluation of MPAI. Five linguistic variables are used because
it is convenient for an expert to distinguish subjectively
between them.
The approach for developing MPAI was based on strength of
different parameters is somewhat analogous
suggested by Singh and Tong (2005). Linguistic variables
assigned for the study were five; Highly Significant (HS),
Significant (S), Average Significant (AV), Low Significant
(LS), and Not Significant (NS).Each were defined with four
fuzzy numbers average of the numbers assigned by experts
presented in Table below.
Table 1: Fuzzy numbers associated with the corresponding
Linguistic Variables.
The graphical presentation of fuzzy numbers for the linguistic
variables used by seven environmental experts to
weightage for each criterion is shown in Figure
Linguistic variables
Highly Significant (HS)
Significant (S)
Average Significant(AS)
Low Significant (LS)
Not Significant (NS)
eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
187
Step Involved in a MCDM ModelAssignimportance
tage to the sub criteria of wastewater.
Weight for each of the sub criterion is developed by
normalized weight method by consulting the environmental
experts. In the present study, five fuzzy numbers are selected
to describe the level of performance on decision criteria in
evaluation of MPAI. Five linguistic variables are used because
it is convenient for an expert to distinguish subjectively
The approach for developing MPAI was based on strength of
different parameters is somewhat analogous to the procedure
suggested by Singh and Tong (2005). Linguistic variables
assigned for the study were five; Highly Significant (HS),
Significant (S), Average Significant (AV), Low Significant
(LS), and Not Significant (NS).Each were defined with four
y numbers average of the numbers assigned by experts
Fuzzy numbers associated with the corresponding
The graphical presentation of fuzzy numbers for the linguistic
variables used by seven environmental experts to develop
weightage for each criterion is shown in Figure2.
Linguistic variables Fuzzy Number
Highly Significant (HS) (0.7,0.8,0.9,1.0)
(0.5,0.6,0.7,0.8)
Average Significant(AS) (0.3,0.40.5,0.6)
Low Significant (LS) (0.1,0.2,0.3,0.4)
Not Significant (NS) (0.0,0.0,0.1,0.2)

__________________________________________________________________________________________
Fig 2: Graphical representation of trapezoidal membership
function
The weightage factors are computed for eight sub criteria
(parameters) of wastewater for raw and treated wastewater. On
the basis of the experts’ opinion (linguistic variable) or
perception, a fuzzy decision matrix for the sub criteria of
wastewater is computed.
Weights are evaluated using the equation given below. The
average fuzzy numbers for all the environmental experts’
opinion can be expressed as
A ij 1
p a a a ⋯ … . . a for j
1, 2, … … , p 1
Where akij be the fuzzy number (weight) assigned to an
alternative Ai by DMj (Decision Maker i) for the decision
criteria Ck and p is the number of environment experts
involved in the evaluation process. The linguistic variables as
assigned by the experts are converted to fuzzy numbers used
in the above expressions through Figure. Now, the defuzzified
values for the sub-criteria are obtained by the equation below;
(Kaufmann; 1991).
e =X1 +X2 +X3+X4/4 (2)
The normalized weight for each sub criterion of raw and
treated wastewater is obtained by dividing the scores of each
sub criterion ( ijC
) by the total of all sub-criterion (∑ ijC
).
The next step is to convert the parametric values of raw and
treated wastewater to the fuzzy numbers (membership
functions) based on the specified statutory norms. Using
simple additive weighing method (Hwang and Yoon, 1981),
the overall score (OS) in terms of MPAI for raw and treated
wastewater criteria can be calculated using following
equation.
( )( ) ..,2,1kforCWX kkOS n=⊗= ∑ (3)
Where,
( )kCW =weight (importance value) of the sub criterion
k and
kX
= crisp score (defuzzified value) of treated data against
the sub criterion k.
2.2Stastical Analysis
To evaluate wastewater quality from a large number of ample
each having number of parameters is difficult. To analyze the
water quality, different approaches like statistical analysis of
individual parameter, multi- stressor water quality indices ect.
have been considered (Venkatesha, 2010). Statistical analysis
are generally used to identify the calculated WWQI is
appropriate or not. For this purpose Correlation Analysis and
Multiple Linear Regression analysis were performed.
Correlation Analysis: Correlation is a measure of degree of
linear relationship between two variables. It expresses the
extent to which two variables vary together in same directions
or opposite directions. Correlation coefficients reveal the
magnitude and direction of the relationships. A correlation has
value ranging from -1 to 1.Values that are closer to the
absolute value 1 indicate that there is a strong positive linear
relationship between the variables being correlated, whereas
values closer to 0 indicate that there is no liner relationship
between two variables under study. Below table 5 shows the
correlation between different parameters, which are, applied to
evaluate WWQI for treated wastewater. Correlation between
different parameters was found out by Florence P. et al(2010)
for Water Quality Index and Correlation Study at
YercaudTaluk, Salem, Tamil Nadu, India. Also, correlation
analysis for lower Maniar Reservoir was carried out by
Thirupathaiah M. et. Al.(2012)
Multiple Linear Regression Analysis: It is one of the very
important statistical tools which are used in almost all fields of
sciences. According to Rastogi, 2011, the Statistical
Regression Analysis has been found to be highly useful tool
for correlating different parameters. This has twofold
advantages. First with the help of correlation analysis it is easy
to find out interrelationship between different parameters and
also the main parameters affecting the WWQI can be easily
figured out. Later with the help of the regression equation
WWQI was calculated again and the degree of match is
checked. With help of this one can easily eliminate tedious
laboratory work of determining all parameters and can carry
out its work only with the parameters that are significantly
correlated in the regression equation. In regression analysis
there are two types of variables. The variable whose value is
influenced is called dependent variable and the variables
which influence the dependent variable are called explanatory
(regressors) variables. And the model equation is:
Y= β0 + β1X1 + β2 X2 + β3X3 + β4 X4 + β5X5 + β6 X6 +
β7X7+ β8X8 + e (4)
This model should be free from (as assumptions of applying
this criteria), Multicolinearity (regressors are correlated),

__________________________________________________________________________________________
Heteroscedasticity (error variance is nonconstant), and Auto-
correlation (error terms are correlated).
In this study WWQI is considered as dependent variable and
parameters are considered as explanatory variables. Chenini et
al(2009) have successfully applied multiple linear regression
for Evaluation of ground water quality. Rene et. al.(2008) has
also applied regression equation for prediction of water quality
indices.
3. MATERIALS & METHODOLOGY
3.1 Site Selection
As the study is concerned with the development of WWQI for
the treated wastewaters. All the samples were collected from
Anjana Treatment Plant located in Surat city itself. All the
samples were collected at fix time i.e. at 1 PM to 1.30 PM for
a time period of 1 year.
3.2 Sampling Procedure
For sampling, the procedure described in Standard Methods
and IS 3025 was adopted. DO fixation was performed at the
location itself by adding Manganese Sulfate and Alkaline KI
solution. For other physico – chemical parameters, samples
were collected in 1 liter plastic bottles. Samples were collected
in container rinsed with detergents, nitric acid and finally with
distilled water. These plastic bottles were placed in ice box to
maintain the temperature during travelling periods of
sampling.
3.3 Sample Analysis
The collected water samples are analyzed for eight different
parameters including Temperature, Total dissolved solids
(TDS), Suspended solids (SS), Biochemical demand (BOD),
Chemical oxygen demand (COD), Oil and grease (O&G),pH
and Chlorides (Cl). All the analyses are carried out as per the
guidelines provided in Standard Methods and IS: 3025
(1964).Temperature, DO and pH were measured at the timeof
sample collation.
4 RESULTS AND DISCUSSION
Constant monitoring of all the parameters described by CPCB
was carried out weekly throughout the year 2011.maximum,
minimum and average values of the analyzed data is
summarized in Table 2. As the treated wastewater was
chlorinated before discharge the FC and FS are omitted as the
wastewater was chlorinated before final discharge in to river
Tapi. Out of measured nine parameters total eight parameters,
except DO were considered to develop WWQI (Table3).
Table 2Minimum, maximum and average Data at Anjana STP
Note: Except Temperature and pH all parameters are in mg/L
Table 3 Minimum, maximum and average values for WWQI by MCDM method and regression equation
Data WWQI by MCDM
WWQI by regression equation (With 6
paramters)
WWQI by regression equation (With 5
parameters)
Max 0.6284 0.5711 0.6172
Min 0.2835 0.2992 0.3059
Average 0.4433 0.4241 0.4465
Then with the help of Pearson Correlation (Table 4)
statistically significantly correlated variables were identified.
Parameters Temp. TDS SS BOD COD O & G Cl pH DO
Treated Wastewater
Discharge Limits 40 2100 100 30 100 10 600 6.5-8.5 -
Max 33.80 792.00 40.00 23.00 128.00 0.80 360.00 8.05 0.74
Min 24.50 634.00 6.00 5.00 32.00 0.10 280.00 7.49 5.34
Average 28.90 732.07 22.14 12.28 67.81 0.41 317.93 7.70 3.13
Weights 0.108 0.117 0.141 0.166 0.154 0.08 0.141 0.092 -
Membership Grade Xk
Max 0.8450 0.3771 0.4000 0.7667 1.0000 0.0800 0.6000 0.6999
Min 0.6125 0.3019 0.0600 0.1667 0.3200 0.0100 0.4667 0.3266
Average 0.7225 0.3486 0.2214 0.4093 0.6781 0.0410 0.5299 0.4666

__________________________________________________________________________________________
Table 4 Pearson Correlation Coefficients showing relation between all the parameters
Temp TDS SS BOD COD O&G Cl pH DO WWQI
Temp 1
TDS .989**
1
SS -.522**
-.513**
1
BOD -.945**
-.953**
.584**
1
COD .862**
.887**
-.245**
-.806**
1
OG -.985**
-.996**
.524**
.955**
-.887**
1
Cl .991**
.997**
-.509**
-.947**
.885**
-.991**
1
pH -.887**
-.897**
.517**
.970**
-.797**
.895**
-.893**
1
DO -.985**
-.996**
.522**
.958**
-.888**
1.000**
-.991**
.899**
1
WWQI .056 .287**
.720**
.920**
.893**
.089 .321**
.194 -.984**
1
Out of eight parameters, Out of eight parameters, TDS-SS,
TDS-BOD, TDS-COD, TDS-OG, TDS-pH, TDS-DO, SS-
BOD, SS-COD, SS-OG, SS-Cl, SS-pH, SS-DO, BOD-COD,
BOD-OG, BOD-Cl, BOD-pH, BOD-DO, COD-OG, COD-Cl,
COD-pH, COD-DO, OG-Cl, OG-pH, OG-DO, Cl-pH, Cl-DO,
pH-DO are found to be significantly correlated with each other
(Table 4). Also it was found that WWQI is significantly
correlated with TDS, SS, BOD, COD, Cl and DO which
means that out of these nine parameters, only Six parameters
are responsible for the deviation or change in the values of
WWQI. Now as our main motive is to check the WWQI is
correctly drawn or not, we go for regression. From Table 5,
unstandardized coefficients for the equation constant, TDS,
SS, BOD, COD, Chloride and DO as 0.456, 0.0000103,
0.000451, 0.002448, 0.000530, 0.000128, and -0.0412
respectively. These coefficients were replaced with the
coefficients of equation 4 and Y predicted is presented in
equation 2. Table 6 shows that the model fit with 98.5% of
accuracy.
Y Estimated = 0.456 + 0.0000103 X1+ 0.000451 X2 +
0.002448 X3 + 0.000530 X4 + 0.000128 X5 - 0.0412 X6
………………….(5)
Where Y= WWQI, X1= TDS, X2= SS, X3= BOD, X4 =
COD, X5 =Cl, and X6 = DO
Table 5: Coefficients for regression analysis (Model 1)
Coefficients Standard Error t Stat Significance
Constant 0.456091134 0.029309116 15.56140879 6.45896E-26
TDS 1.03817E-05 3.97293E-05 0.261309917 0.794525072
SS 0.000451482 0.000158837 2.842434083 0.005679048
BOD 0.002448483 0.000387202 6.323524967 1.36726E-08
COD 0.000530319 7.9732E-05 6.651263435 3.29148E-09
Cl 0.000128486 6.10444E-05 2.104788562 0.038448341
DO -0.041153394 0.003067748 -13.41485266 3.58124E-22
Table 6: Model Summary (Model 1)
Model R R Square Adjusted R Square Std. Error of the Estimate Durbin-Watson
1 .993 .986 .985 .00642 1.937

__________________________________________________________________________________________
Table 7: ANOVA (Model 1)
Model Sum of Squares df Mean Square F Sig.
1
Regression .240 6 .040 972.248 .000
Residual .003 80 .000
Total .244 86
This model is fitted significantly as F= 972.248 (Sig F: 0.000)
(Table 7), also root mean square error (RMSE) is 0.000. This
model is not having problem of auto correlation as Durbin-
Watson test statistic is 1.937 which is nearer to 2 (Table 6).
As TDS was found to be correlated significantly with most of
the other parameters, TDS was neglected and again the
regression analysis was executed in order to develop a new
equation. So TDS is dropped and re calculated the WWQI and
again fitted the regression model, and got the regression
equation as follows with 98.2% of accuracy (Table 9).
Y Estimated = 0.4595 + 0.000457 X1+ 0.002464 X2 + 0.000527
X3 + 0.000141X4 -0.0412X5 (Table 8) (6)
Where Y= WWQI after dropping TDS, X1=SS, X2=BOD,
X3=COD, X4 = Cl, and X5 =DO
Table 8: Coefficients for Regression Analysis (b)
Coefficients Standard Error t Stat P-value
Constant 0.459523038 0.026050707 17.63956079 1.79428E-29
SS 0.000456715 0.00015666 2.915317416 0.004593801
BOD 0.002464727 0.000379976 6.486540962 6.48043E-09
COD 0.000527356 7.84664E-05 6.720784388 2.32032E-09
Cl 0.000140821 3.84823E-05 3.659358498 0.000448766
DO -0.041114604 0.003046481 -13.49576953 1.99668E-22
Table 9: Model Summary (b)
Model R R Square Adjusted R Square Std. Error of the Estimate Durbin-Watson
1 .993 .986 .986 .00638 1.954
Table 10: ANOVA (b)
2
Regression .240 5 .048 1180.260 .000b
Residual .003 81 .000
Total .244 86
This model is fitted significantly as F= 1180.260 (p-value:
0.000) (Table 10), also root mean square error (RMSE) is
0.000. Each parameter has significant effect on WWQI (Table
8). This model is not having problem of auto correlation as
Durbin-Watson test statistic is 1.954 which is nearer to 2
(Table 9).
Now DO depends on the strength of wastewater i.e. WWQI.
As the value of WWQI is close to 1, poor is the quality of
treated wastewater and measured DO will be less in quantity
and for the WWQI nearer to zero , quality of wastewater is
very good for reuse and the DO will be more. In order to
predict BOD, regression equation can be fitted with BOD as
explanatory variable and DO as dependent variable. Following
equation was developed to calculate BOD.
Y Estimated = 29.182 + 0.951 X1+e (Table 11)
……..……. (7)
Where Y=BOD, X1= DO and e is error term.This model is
fitted significantly as F= 356.781 (p-value: 0.000) (Table 13),
also root mean square error (RMSE) is 0.000 with 96.9% of
accuracy (Table 12).

__________________________________________________________________________________________
Table 11: Coefficients for Regression Analysis (c)
Table 12: Model Summary (c)
Table 13: ANOVA (c)
1 Regression 1237.090 1 1237.090 356.781 .000a
Residual 294.726 85 3.467
Total 1531.816 86
a. Predictors: (Constant), DO
b. Dependent Variable: BOD
Similarly the WWQI can be estimated from regression
equation fitted with WWQI as explanatory variable and DO as
dependent variable. Following equation was developed to
calculate WWQI.
Y Estimated = 0.672 - 0.068 X1+e (Table 14)
Where Y=WWQI, X1= DO and e is error term.
This model is fitted significantly as F= 2800.788 (p-value:
0.000) (Table 16), also root mean square error (RMSE) is
0.000 with 97.1 % of accuracy (Table 15).
Table 14: Coefficients for Regression Analysis (d)
Table 15: Model Summary (d)
Model R R Square Adjusted R Square Std. Error of the Estimate R Square Change Durbin-Watson
1 .985a
.971 .970 .0091875 .971 1.608
Model Unstandardized Coefficients Standardized
Coefficients
t Sig.
B Std. Error Beta
1 (Constant) 29.182 .951 30.698 .000
DO -4.966- .273 -.892- -18.203- .000
Model R R Square Adjusted R Square Std. Error of the Estimate
3 .985(a) .969 .969 4.18460
Model
Unstandardized Coefficients Standardized Coefficients
t Sig.
B Std. Error Beta
1
(Constant) .672 .004 150.041 .000
DO -.068 .001 -.985- -52.922- .000
a. Dependent Variable: WWQI

__________________________________________________________________________________________
Table 16: ANOVA (c)
1
Regression .236 1 .236 2800.788 .000a
Residual .007 85 .000
Total .244 86
a. Predictors: (Constant), DO
b. Dependent Variable: WWQI
The graphical comparison for WWQI developed by MCDM
Method and both regression equations is shown in Figure 3.
The Figure 3shows that the result slightly deviates with TDS
and it almost matches without TDS.
Fig 3 Comparison in WWQI calculated by MCDM method and two regression equations
Figure 4 shows the comparison between BOD and BOD
Estimated from regression equation and Figure 5 shows the
Comparisonbetween WWQIand WWQI Estimated.
Figure 4 shows the comparison between BOD and BOD Estimated
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85
WWQI
Number of samples
No. of Samples V/S WWQI WWQI by MCDM
WWQI by Regression Equation (With TDS)
0
5
10
15
20
25
3.21
3.27
1.89
3.76
4.85
3.55
4.47
4.29
3.06
2.34
2.27
3.83
2.83
3.82
3.64
4.59
4.21
1.53
3.23
2.65
3.62
3.04
BOD&EstimatedBOD
DO mg/L
DO v/s BOD
BOD BOD Estimated

__________________________________________________________________________________________
Fig 5 shows the comparison between WWQI and WWQI Estimated.
CONCLUSIONS
First WWQI was found and later the correlation for the
interrelation between different parameters and their correlation
with WWQI was determined with help of SPSS. From this
analysis WWQI is significantly correlated with TDS, SS,
BOD, COD, Cl and DO which means that out of these nine
parameters, only Six parameters are responsible for the
deviation or change in the values of WWQI. Also it was found
that TDS, SS, BOD, COD, Cl and DO were significantly
correlated with WWQI at 0.01 Level.
When 6 parameters were considered and regression equation
was developed, it was found nearly 98.2 % accurate.
Y Estimated = 0.456 + 0.0000103 X1+ 0.000451 X2 +
0.002448 X3 + 0.000530 X4 + 0.000128 X5 - 0.0412 X6 .
It has the standard error of 0.00642. Now, as it has R Square >
.75, the equation is acceptable and can be applied to determine
WWQI instead of performing whole tedious procedure of
MCDM Method.
Further omitting TDS and considering five parameters and
developed regression equation it was found 98.6 % fitting with
the original WWQI. The equation was like:
Y Estimated = 0.4595 + 0.000457 X1+ 0.002464 X2 +
0.000527 X3 + 0.000141 X4 - 0.0412X5.
It has the standard error of 0.969. Now, as it has R Square >
.75, this equation is also acceptable and can be applied to
determine WWQI instead of performing whole tedious
procedure of MCDM Method. But, as accuracy increases only
about 0.04%, we can neglect determination of TDS also which
may result in considerable time saving and reduce
complications in determination of WWQI.
The estimated BOD from independent variance DO for
maximum, minimum and average is 25.35 mg/L, 2.65 mg/L
and 13.56 mg/L respectively. While estimated WWQI from
independent variance DO for maximum, minimum and
average is 0.6212, 0.3074 and 0.4581 respectively.
Hence, instead of analyzing all 9 parameters (neglecting FC
and FS) only 6 or 5 parameters can provide results with
sufficient accuracy around 98-99 %. This eliminates tedious
procedure of determining and testing each and every
parameter which ultimately results in considerable amount of
time, chemicals and money savings. Also higher accuracy
eliminates any chances of deviation form original WWQI
expect shock loadings.
REFERENCES
[1] Akobundu A. N., Quality assessment of Aba River
using heavy metal pollution index, Am. J. Environ.
Engin., 2012, 2(1), 45-49.
[2] American public health association, (APHA). Standard
methods for the examination of water and wastewaters,
2003. (21th
edn). Washington, DC, USA.
[3] Balteiro, L., and Romero, C., In search of a natural
systems sustainability index, 2004, Ecological
Economics, 49 (3), 401-405.
[4] Bharti N. and Katyal D., Water quality indices used for
surface water vulnerability assessment, International
Journal of Environmental Science, 2011, 2(1), 154-173.
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
3.21
3.27
1.89
3.76
4.85
3.55
4.47
4.29
3.06
2.34
2.27
3.83
2.83
3.82
3.64
4.59
4.21
1.53
3.23
2.65
3.62
3.04
WWQI
DO
DO v/s WWQI
WWQI
WWQI Estimated

IJRET: International Journal of Research in Engineering and TechnologyeISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
[5] Bordalo A. A., Teixeira R. and Wiebe W. J., A water
quality index applied to an international shared river
basin : The case of the Douro River, Environ. Manag.,
38(1), 2006, 910–920.
[6] Chen,Y., C., Lien, H., P., Tzeng, G., H., and Yang, L.,
S., Fuzzy MCDM approach for selecting the best
environment-watershed plan, Applied Soft Computing
Journal, 2011,11(1), 265-275.
[7] Chenini I. and Khemiri S., Evaluation of ground water
quality using multiple linear regression and structural
equation modelling, Int.J. Environ. Sci. Tech., 2009, 6
(3), 509-519.
[8] Evangelos., P mastering visual basic 6, BPP
Publications, New Delhi, (2000).
[9] Florence P. L., Paulraj A.,Ramachandramoorthy T.,
Water Quality Index and correlation study for the
assessment of water quality and its parameters of
YercaudAluk, Salem District, Tamil Nadu, India,
Chem. Sci.Trans., 2010,1(1), 139-14.
[10] Gabrieli, R.; Divizia, M.; Donia, D.; Ruscio, V.;
Bonadonna, L.; Diotallevi, C.; Villa, L.; Manzone, G.;
Panà, A., Evaluation of the wastewater treatment plant
of Rome airport. Water Science and Technology, 1997,
vol. 35, issue 11-12. p. 193-196
[11] Hwang, C., and Yoon, K. Multiple attribute decision
making – methods and applications, A state of the art
survey, Springer-Verlag, Berlin. ,1981.
[12] Jamrah, A., I., Assessment of characteristics and
biological treatment technologies of Jordanian
wastewater, 1999, Bioprocess Engineering, 21: 331-
340.
[13] Jamwal, P., Mittal, A., K. and Mouchel,J., M.
Efficiency evaluation of sewage treatment plants with
different technologies in Delhi, (India). 2009,
Environmental Monitoring assessment,153, 293-305.
[14] Kaufman, A., and Gupta, M., M. , Fuzzy mathematical
modelling in engineering and management sciences,
Fuzzy Set and Systems, 1991, 43, 17–32.
[15] Klir, G., J., and Yuan, B. , Operation of fuzzy sets in:
Ruspini, E. H., Bonissone, P. P. and Pedryez W.,
Handbook of Fuzzy Computation, Institute of Physics
Publishing, Dirac House, temple Bath, Bristol , 1998.
[16] Mohammed,B , Design and Performance Evaluation of
a Wastewater Treatment Unit.” Australian Journal of
Technology. , 2006, 9(3): 193-198 .
[17] Prati L., Pavanello R. and PesarinF.,Assessment of
surface water quality by a single index of pollution, J.
Wat. Res.,1971, 5(1), 74.
[18] Rastogi. G.K. and Sinha D. K., A novel approach to
water quality management through correlation study,
Journal of Environmental Research and Development ,
2011, 5 ,(4), 1029-1035.
[19] Ravikumar, P., Liza, B., Pinto. andSomashekar, R., K.,
Assessment of the Efficiency of Sewage Treatment
plants: A Comparative Study between
NagasandraandMailasandra Sewage Treatment Plants.
2010, Kathmandu University, Journal of Science
Engineering and Technology, 6(2), 115-125.
[20] Rene E. and Saidutta M. B., Prediction of Water
Quality indices by regression analysis and artificial
neural networks, Int. Environ. Res., 2008,2(2), 183-
188.
[21] Singh, D. and Robert L. K.Tiong, A fuzzy decision
framework for contractor selection, 2005, Journal of
Construction Engineering and Management (ASCE),
131(1), 62-70.
[22] Thirupathaiah M., Samatha C. andSammaiah C.,
Analysis of water quality using physico-chemical
parameters in lower manair reservoir of
Karimnagardistrict,Andhra Pradesh, Int. J. Environ.
Sci., 2012,3(1), 172-180.
[23] Venkatesharaju K., Somashekar R. K and Prakash K.
L., Study of seasonal and spatial variation in surface
water quality of Cauvery River Stretch in Karnataka, J.
Ecol. Nat. Environ., 2010, 2(1), 1-9.
[24] Verlicchi P, Masotti L, Galletti A. Wastewater
polishing index: a tool for a rapid quality assessment of
reclaimed wastewater. Environtal Monitoring and
Assessment; 2011, 173(1-4):267-77. doi:
10.1007/s10661-010-1386-7. Epub 2010 Mar 9.
BIOGRAPHIES
Mrs. Anjali Khambete has received her
bachelor’s and master’s from S V National
Institute of Technology, Surat. Currently
working as Associate Professor in the department of Civil
engineering. At SVNIT, Surat. Her research interests are in the
field of water and wastewater engineering. Fuzzy logic,
Dr. Robin Christian has received his master’s
and PhD from S V National Institute of
Technology, Surat. Currently working as
Associate Professor in the department of Civil
engineering. At SVN IT, Surat. His research interests are noise
pollution, application of fuzzy logic in environmental field.
Published many papers in international journals related to
environmental engineering.

Statistical analysis to identify the main parameters to

More Related Content

What's hot (18)

Viewers also liked (16)

Similar to Statistical analysis to identify the main parameters to (20)

More from eSAT Publishing House (20)

Statistical analysis to identify the main parameters to