You are on page 1of 6

OPTIMIZATION OF MIXED PLASTIC IN VG-30 PAVING BITUMEN BY

TOPSIS BASED TAGUCHI METHOD

SANDIP KARMAKARa, TAPAS KUMAR ROYb and PRITAM SAHAb


a
Research Scholar, Department of Civil Engineering, IIEST, Howrah- 711103, India,
E-mail:karmakar38@gmail.com
b
Assistant professor, Department of Civil Engineering, IIEST, Howrah- 711103, India,
E-mail: tapash2000@hotmail.com, saha.pritam@gmail.com

ABSTRACT- Incorporation of 1wt% waste Mixed plastic (MP) had been


experienced to the virgin bitumen as eco-friendly and cost-effective method for
improving its performance in extreme weathering and traffic overloading condition
in pavement. But, optimized ingredients in that MP is in doubt till date considering
multi-criteria based performance analysis. Therefore, application of ‘Technique for
Order Preference by Similarity to an Ideal Solution (TOPSIS)’ based Taguchi
method for the purpose to dose optimisation of those ingredients in cited MP is the
state of art in this study. Experimental results indicate that 2wt% High-Density
Polyethylene (HDPE), 0.25wt% Low-Density Polyethylene (LDPE) and 1wt%
Polypropylene (PP) combinedly enhance the performance of the MP-modified
bitumen (MPMB) with less phase separation value. Further, homogeneity of the
MPMB enriched with the cited optimized proportion was checked by means of its
optical micrograph.
Keywords: MPMB, TOPSIS based Taguchi method, Phase separation, Optical
micrograph

INTRODUCTION
Over the past few decades, efforts probed to optimize quantity of the waste plastics
as bitumen modifiers (WPBMs) to depreciate polymer modified bitumen (PMB)
commercially. In that contrast, a study proved that 1wt% MP with randomly selected
ingredient in it shown significant result to do so resourced from municipal solid
waste (MSW) [1]. Thus, there is a need to determine the optimal quantity of
combinations of such modifiers with different performance measures of PMB is the
state of art in this study. Therefore, multiple performance criteria based TOPSIS
based Taguchi method have been considered in this study for optimizing the
ingredient in the MP for making low-cost MPMB as described in the literature [2]
and considering it as first phase of the investigation. Whereas second phase validates
that optimized ratio for MP with respect to the ratio of MP as found directly in the
municipal dust bin (see Figure 1) mainly by phase separation and optical microscopic
test as guided by different standard guide lines [3,4].
METHODOLOGY
Collection of Materials- Polymeric MSW was collected from the municipal dustbins
at different points in the city of Kolkata having ratio of [4(PB): 1(PC): 5.2(PMP)].
Body and hand protective measures were taken at the time of collecting them in order
to ensure personal safety (see Figure 1). For the purpose of blending them with
conventional bitumen, they were shredded into ‘1 Cm X 1 Cm’ in size to form MP.
Further, MP in different proportions (see Table 1) to constitute were varied and
blended with pre-heated (at a temperature of 1800C) 300 gm of VG-30 as base
bitumen by laboratory stirrer (a rotation of 45 minutes with an rpm of 3500) for
experimentation. This study first tries to optimize the ingredient of MP in base
bitumen and then compare it with the ratio collected from the municipal dust bins.

Fig. 1: Collection of Mixed Plastic from a Municipal Solid Waste dumping


yard in Kolkata.

Design of Experiments- To meet the objectives, the study applied ‘TOPSIS based
Taguchi method’ which is a unique multiple decision-making optimization
technique [1] and considered several bitumen’s attribute while forming the decision
matrix; they are penetration (R1), softening point (R2), specific gravity (R3),
penetration index or PI (R4) and storage stability i.e. ΔT (R5) test values respectively.
Since the attributes are of different units, all the values are converted into ‘signal to
noise (S/N) ratio’ by using Eq.1 and Eq. 2 for ease in analysis. Where, Eq. 1 refers
to smaller values that give the better performance and Eq. 2, similarly, refers to larger
values
 1 N
1 
i j  10 log 1 0   yi2j k

Eq.1

N k 1 

1 N
 Eq.2
 ij  10 log 10 
N
y
k 1
2
ijk 

Where, ηij is the S/N ratio of the cited attributes for i th (1to 16) experimental run of
jth response (R1 to R5), ‘y’ is the raw value of the attribute and ‘k’ is the number of
repetition of each tested attribute for different experimental run. Each experimental
run has been designed as per Taguchi's L16 (24) orthogonal array as described in the
Table 1. However, HDPE made PB, PP made PC and LDPE made PMP are selected
as the controlling factors with four-different levels of dosages (1 to 4) ranging from
0.5 wt % to 3 wt % of base bitumen in accordance with previous experiments [1].

Table 1. Taguchi's L16 (24) orthogonal array

Experiment Experiment
PB PC PMP PB PC PMP
No. (m) No. (m)

1 1 1 1 9 3 1 3

2 1 2 2 10 3 2 4

3 1 3 3 11 3 3 1

4 1 4 4 12 3 4 2

5 2 1 2 13 4 1 4

6 2 2 1 14 4 2 3

7 2 3 4 15 4 3 2

8 2 4 3 16 4 4 1

Thereafter, each S/N ratio is inserted in the decision matrix or ‘D’ (Eq. 3) after
getting every attribute enriched with different mixing ratio based on the designed
chart of orthogonal array (Table 1). Eq. 4 highlights such (m x n) decision matrix by
weighted normalized decision matrix (V) for the purpose of user’s priority
preference.

 
Eq.3
V  vij mxn

vij  W  D 
j

Where, ‘V’ is the product of ‘D’ and priority vector, ‘Wj’. However, such vector has
been obtained from pair wise comparison matrix on the basis of ‘Saaty’s 1–9 scale’
[5].
11 12 ... 1n 
  22 ...  2 n 
D   21
 ... ... ... ... 
 
 m1  m2 ...  mn  mn
Eq.4
Further, positive ideal (A*) and negative-ideal (A−) solutions have been chosen from
the V- matrix abiding by Eq. 5 and Eq. 6 to calculate “separation measures”
corresponding to each solution (i.e. S*i and S- i). Where (J) and (J') be the best and
worst values in the V- matrix for each column.
A  {min v j  J , max v j  J ' }
i
ij  i
ij 
Eq.5


A  min vij j  J , max vij j  J 
i
 i
 Eq. 6
However, separation measures are calculated as per Eq. 7 and Eq. 8 to estimate
ranking score of each experimental run as per their importance in the matrix for
making low cost PMB.
Eq.7
S i*   vij  v *j 
n
2

j 1

Eq.8
 v 
n
2
S i  ij  v j
j 1

Thereafter, the ‘ranking score’ (C*i) was obtained from Eq. 9. which is based-on
from weighted normalized decision matrix and that has been adopted by similar type
of approaches [5].
Si
Ci*  Eq.9
Si  Si*

RESULTS AND DISCUSSIONS


First phase: Table 2 shows the descending trend of ranking scores of the MPMB on
the basis of TOPSIS -based Taguchi method. In that process of dose optimisation,
highest-ranking score has been observed against experiment number 3.
Phase 2: A look into the phase separation value (see table 3) reveals that such value
for the studied blend is 1.06 when TOPSIS -based Taguchi method proportionate
MP [2 (PB): 1(PC): 0.25 (PMP)] mixed in the virgin bitumen namely MPMB1.
Whereas, the value goes as high as 3.18 in the event of Municipal dustbin
proportionate MP [4 (PB): :1(PC): 5.2 (PMP)] incorporated in the cited bitumen as
MPMB2. Such elevated value depicts heterogeneity for the blend while the former
value portrays homogeneous blend. Also, optical micrograph (see figure 3) makes
it clear that TOPSIS -based Taguchi method proportionate MP would show the
homogeneous dispersion of MP droplets in the bitumen medium than that of the
Municipal dustbin proportionate MP. Different prime properties for the studied
blends have been comply with the standard guide lines in table. 3.
Table.2 Dose Optimization of the MP in producing MPMB

Dosing by weight of 3
Experiment Ranking Level of Dosing
gm MP in wt.%
Number score, (C*i)
PB PC PMP PB PC PMP
3 0.831 1 3 3 2 1 0.25
12 0.6812 3 4 2 1 0.5 0.375
2 0.6748 1 2 2 2 2 0.375
16 0.5725 4 4 1 0.5 0.5 0.5
13 0.5696 4 1 4 0.5 3 0.125
14 0.5535 4 2 3 0.5 2 0.25
10 0.5218 3 2 4 1 2 0.125
9 0.5179 3 1 3 1 3 0.25
15 0.5158 4 3 2 0.5 1 0.375
11 0.4459 3 3 1 1 1 0.5
8 0.4298 2 4 3 1.5 0.5 0.25
7 0.4217 2 3 4 1.5 1 0.125
5 0.4181 2 1 2 1.5 3 0.375
4 0.4056 1 4 4 2 0.5 0.125
6 0.4048 2 2 1 1.5 2 0.5
1 0.3802 1 1 1 2 3 0.5

Table. 3 Property Comparison of the studied modified bitumen


Type of Blend Properties of the modified bitumen
Penetration Softening Specific Phase Penetration
(0.1mm) point gravity separation index
(0C, min) (0C)
PMB 40 as per 30-50 60 1.05 3 -
IS: 15462 (2004)

Optimized 31.33 63.7 1.03 1.06 0.0075


proportionate MP
modified bitumen
Municipal dustbin 18.67 68.9 1.04 3.18 -0.1375
proportionate
MPMB
a b

Fig. 3: Optical micrograph of modified bitumen [a] MPMB1 [b] MPMP2

CONCLUSIONS
Results of the study has explored that experiment number 3 gives the highest-ranking
score in the process of dose optimization. That result indicates that 2wt% HDPE,
0.25wt% LDPE and 1wt%PP combinedly enhance the studied performance of the
bituminous binder having phase separation value of 1.06. Optical microscopic
images show that cited proportion gives the homogeneous blend as compare to the
other studied blend.
The study is limited to defining the optimized ratio of different ingredients in MP to
form MPMB. Investigation on fluctuation of adhesion and rheological change on the
that material is considered as a future scope of work.

ACKNOWLEDGEMENT
The authors are grateful to Prof. Ashoke K. Sarkar, BITS Pilani for his suggestions
to improve the paper.
REFERENCES
1. Karmakar S. and Roy T. K., Effect of Waste Plastic and Waste Tires Ash on
Mechanical Behaviour of Bitumen, J. of Mat.s in Civil Eng., 28(6), 04016006-1-10,
2016.
2. Simsek B., Tansel Y. Ic, Simsek E. H., A TOPSIS-based Taguchi optimization to
determine optimal mixture proportions of the high strength self-compacting
concrete, Chemo. and Intel. Lab. Sys., 125, 18–32, 2013.
3. Li Jun, Zhang Y., Zhang Y., The research of GMA-g-LDPE modified Qinhuangdao
bitumen, Const. and Build Mat., 22:1067–1073, 2008.
4. IS: 15462, “Polymer and rubber modified bitumen—Specification.” Bureau of
Indian Standards, New Delhi, India, 2004.
5. Saaty Thomas L., How to make a decision: The Analytic Hierarchy Process,Euro.
J. of Op. Res., 48,9-26, 1990. 5

You might also like