Академический Документы
Профессиональный Документы
Культура Документы
Abstract
This case study is one of the distinctive researches which highlight the factor ratings and process optimisation through Taguchi
design. This research has been conducted with a special reference to chemical plants in India. The objective of the present study was
to establish optimal setting of four factors, viz., (i) temperature (°C), (ii) percentage of chemical (Borex), (iii) pH and (iv) time (Hrs).
The present study was conducted to ensure optimisation of two response variables which were (i) Loss-on-Ignition (LoI) (lower the
better) and (ii) chemical Borate content (higher the better). The unique 10 step methodology was adopted for arriving at optimal set-
ting of four factors. Validation of optimal settings has been done by comparing actual values (means and signal to noise [S/N] ratios)
for optimal setting with predicted values (means and S/N ratios). The results show that the Taguchi parameter design improves the
response variables (i) LoI (lower is better) and (ii) chemical Borate content (higher is better), as the Taguchi L9 experiment resulted
in lower LoI (less than 15 per cent) and high Borate percentage of chemical (greater than 45 per cent). The optimisation of response
variables has ensured breakthrough improvement in the manufacturing process of Zinc Borate. The result underlines effectiveness of
Taguchi optimisation technique.
Keywords
Robust Taguchi design, main effects plots for means, main effects plots for signal to noise ratios
Taguchi method is a systematic way to determine cause and pharmaceuticals and paper. Zinc Borate is used as a flame
effect relationship between design factors and performance. retardant additive in poly vinyl chloride (PVC), epoxy
Many organisations have deployed Taguchi technique to resins and smoke suppressant for above mentioned indus-
optimise product and process design. The technique cate- tries (textile, rubber, etc.). The Zinc Borate finds its appli-
gorises independent and dependent variables with the help cation in paint (heat resistant coatings), pigment and
of identified factors and factor levels which lead to adhesive industries. It is used in cable wires and flooring as
enhanced quality performance. To understand the effective- a PVC additive.
ness of Taguchi design, the author(s) of the present case The annual turnover of the product (Zinc Borate) is
study approached an enterprise which manufactures Zinc about 20 per cent of total turnover of the firm, which
Borate and is located in Gujarat, India. The information is more important in terms of productivity and sales. The
about the organisation and the brief information about the ABC enterprise faced problem of higher customer com-
manufacturing process have been mentioned in the article. plaints because of high Loss-on-Ignition (LoI) and lower
The enterprise ABC is a chemical product manufac- percentage of chemical Borate content in Zinc Borate at
turer located in Gujarat, India. Out of the variety of prod- initial product development stage. The lower productivity
ucts manufactured by the organisation, the chemical of the Zinc Borate grabbed the attention of management
product Y (assumed name) was considered for the present and made an emphasis on maximisation of the percent-
study. This product Y has applications in various indust- age of chemical Borate content and the issue with higher
ries such as plastics, textiles, rubber, polymer, adhesive, LoI.
Vishal Singh Patyal, Research Scholar, Operation Management, National Institute of Industrial Engineering (NITIE), Mumbai, India.
E-mail: vishalpatyal2001@gmail.com
Sachin Modgil, Research Scholar, Operation Management, National Institute of Industrial Engineering (NITIE), Mumbai, India.
E-mail: sach.modgil@gmail.com
K. Maddulety, Associate Professor, National Institute of Industrial Engineering (NITIE), Mumbai, India. E-mail: koila@rediffmail.com
232 Vishal Singh Patyal, Sachin Modgil and K. Maddulety
The present study has four sections. The first section Moreover inferences resulted from reduced number of
introduces the rationale behind the study. The second experiments are applicable all over the experimental
section discusses the previous scholarly contributions, region spanned by control factors and levels (Phadke,
which are related to Taguchi technique and described 1989, pp. 19–26). The Taguchi technique helps in reducing
under the title ‘literature review’. The third section dis- time and cost. Taguchi technique involves three phases.
cusses the methodological details of the present study. The The first phase consists in selecting the performance
fourth section reports on the analysis of data followed by parameters (to be optimised), control factors and noise
interpretation, implications and conclusion of the present factors. The second phase includes collecting information
study. regarding product variability performance design and
execution of experiments using OAs.
The third phase results in arriving at an optimal setting
Literature Review by using analysis of variance (ANOVA) (Wang, Liu, Chen
Design of experiments (DOE) is a technique, which facili- & Yang, 1999). Ghani, Choudhury and Hassan (2004) have
tates development to achieve information for optimised applied OA, signal to noise (S/N) ratio and Pareto ANOVA
levels through numerous variables (Taguchi, 1987). The to examine the effect of milling parameters. The results of
Taguchi facilitates development in process which further this study showed that Taguchi technique is appropriate for
offers systematic, simple and efficient method for optimi- problem-solving with the lowest number of trials in com-
sation (Taguchi, 1986). Taguchi orthogonal array (OA) is parison with full factorial designs. Further, Taguchi tech-
a factorial based approach which combines statistical and nique helps find optimal cutting parameters for surface
engineering techniques (Mitra, 1998). The conventional roughness.
experimental design approach which makes an emphasis S/N ratio and ANOVA are employed to study the per-
on average process performance characteristics, while in formance characteristics in turning operations (Nalbant,
contrast Taguchi approach focuses on the effect of varia- Gokkaya & Sur, 2007). Table 1 represents information
tion on the process characteristics (Ross, 1989) and makes about the deployment of Taguchi technique for problem-
the process/product performance robust. The concept solving in different industrial settings.
of Taguchi’s philosophy is the loss function defined as In this study author(s) have used a 10 step methodology
‘The quality of a product is the (minimum) loss imparted (Maddulety, Modgil & Patyal, 2012) for deploying robust
by the product to society from the time the product is Taguchi design in process optimisation of a moulding
shipped’ (Byrne & Taguchi, 1987). Taguchi technique operation by using Minitab software. These steps have
employs OAs to analyse numerous variables with rela- been used as a methodology for the present study and
tively lesser number of experiments (Pignatiello, 1988). diagrammatically presented in Figure 1.
Figure 1. Schematic Representation of the Steps Involved in Step 4: Identification of Zinc Borate Quality
the Taguchi Optimisation Procedure Characteristics Z that is to be Optimised
The brainstorming technique was used by involving con-
cerned employees and executives to minimise LoI percent-
age and to maximise the percentage of chemical Borate
content.
Step 7: Conducting Experimentation (1) and ‘smaller the better’ (2) for S/N ratio are calculated
(see Tables 5–8):
As per the design shown in Table 3, three experiments were
S
Ratio = - 10 * log n d | 2 n(1)
conducted for each of nine treatments (3 * 9 = 27) and Zinc 1 1
Borate was produced. The Zinc Borate is tested and the N y
values of the Loss on Ignition (LoI) and Borate content
Ratio = - 10 * log n _| y 2 i(2)
S 1
predicted by Minitab for optimal setting after analyses are
N
shown in Table 4.
Here in case of Zinc Borate percentage of chemical where yr is the average of the experimental data, Sy2 is the
borate is considered as ‘higher the better’ and LoI sample variance of y, n is the number of observations and y
as ‘smaller the better’. Taguchi ‘higher the better’ is the experimental data (Pai, Rao & Shetty, 2012).
LoI B2O3
Te Br pH T R1 R2 R3 R1 R2 R3
60 1 1.5 6 25.55 25.66 25.77 42.29 42.19 42.09
60 3 2.5 10 30.38 30.48 30.28 30.8 34.8 32.8
60 6 3 15 25.76 25.36 25.56 41.25 41.15 41.35
80 1 2.5 15 19.9 19.95 20 38.72 39.32 39.02
80 3 3 6 20.58 20.08 20.33 39.82 39.72 39.62
80 6 1.5 10 19.22 19.02 19.42 30.8 38.8 34.8
100 1 3 10 15.34 15.24 15.14 42.3 42.15 42
100 3 1.5 15 14.03 14.18 14.33 45.35 45.3 45.4
100 6 2.5 6 12.9 12.6 12.3 46.63 46.43 46.23
Table 5. Response Table for S/N Ratio for LoI vs. Te, Br, pH, T Table 6. Response Table for Means for LoI vs Te, Br, pH, T
‘Smaller the Better’ ‘Smaller the Better’
Level Te Br pH T Level Te Br pH T
1 28.66 25.95 25.63 25.45 1 27.20 20.28 19.69 19.53
2 25.95 26.28 25.89 26.33 2 19.83 21.63 20.98 21.61
3 22.90 25.28 25.99 25.73 3 14.01 19.13 20.38 19.90
Delta 5.76 1.00 0.36 0.88 Delta 13.19 2.50 1.29 2.08
Rank 1 2 4 3 Rank 1 2 4 3
Table 7. Response Table for S/N Ratio for Borate Content vs Table 8. Response Table for Means for Borate Content vs Te,
Te, Br, pH, T ‘Higher the Better’ Br, pH, T, ‘Higher the Better’
Level Te Br pH T Level Te Br pH T
1 38.75 41.12 40.78 42.78 1 47.92 45.83 51.67 38.33
2 37.85 39.29 39.42 36.58 2 47.92 47.92 48.75 64.17
3 44.64 40.83 41.04 41.87 3 54.17 56.25 49.58 47.50
Delta 6.80 1.83 1.62 6.20 Delta 6.25 10.42 2.92 25.83
Rank 38.75 41.12 40.78 42.78 Rank 3 2 4 1
Step 8: Analysis of Data Table 10. ANOVA General Linear Model for LoI
Analysis of data of ‘Zinc Borate’ for LoI minimisation and Source DF Seq. SS Adj. SS Adj. MS F P
Borate maximisation by ANOVA and Taguchi approach, Temp 2 786.85 786.85 393.42 122,226.54 0.000
interpretation of analyses and selection of the optimum Borex 2 28.25 28.25 14.13 439.03 0.000
levels:
pH 2 7.50 7.50 3.75 116.55 0.000
27 Hrs 2 22.26 22.26 11.13 345.97 0.000
SS Total = | (Xr - Xi ) 2 (3)
i=1
Error 18 0.58 0.58 0.03
Total 26 845.44
SS Temp = 3 # 3 * [(Xr - Xr Temp (60)) 2 Source: S = 0.179382, R-Sq = 99.93 per cent, R-Sq(adj) = 99.90 per cent.
+ (Xr - Xr Temp (80)) 2 + (Xr - Xr Temp (100)) 2] (4)
Figure 3. Main Effect Plots of B2O3 for S/N Ratio Figure 4. Main Effects Plot of LoI for Means
Predicted values
S/N Ratio Mean
–21.7540 11.31
Predicted values
S/N Ratio Mean
33.6376 47.7933
LoI B2O3
S. No. Te Br pH T (Hr) R1 R2 R3 R1 R2 R3
1 100 6 1.5 6 12.00 11.58 11.38 45.80 45.68 46.08
2 100 6 1.5 6 11.36 12.38 11.56 46.00 46.58 46.94
3 100 6 1.5 6 11.32 11.58 12.06 47.72 47.77 47.10
4 100 6 1.5 6 13.26 12.78 12.58 46.48 46.90 47.56
5 100 6 1.5 6 11.46 11.58 11.66 46.92 46.50 46.58
6 100 6 1.5 6 11.38 11.86 12.08 46.50 46.80 46.94
7 100 6 1.5 6 11.42 12.22 11.78 47.12 46.80 46.72
8 100 6 1.5 6 12.20 12.32 12.76 47.02 47.14 47.58
9 100 6 1.5 6 11.50 11.68 11.74 47.66 47.42 47.54
10 100 6 1.5 6 12.58 12.70 12.88 47.10 47.46 47.58
11 100 6 1.5 6 11.78 12.18 12.12 47.18 47.28 46.80
12 100 6 1.5 6 11.34 11.68 11.84 47.06 46.88 46.86
13 100 6 1.5 6 11.58 11.78 12.26 45.40 45.80 45.78
14 100 6 1.5 6 12.18 12.34 12.56 47.44 47.80 47.68
15 100 6 1.5 6 11.88 11.50 12.02 47.48 47.20 47.22
16 100 6 1.5 6 12.26 12.56 12.44 47.56 47.26 47.38
17 100 6 1.5 6 11.58 11.96 11.72 47.74 47.56 47.48
18 100 6 1.5 6 13.10 12.68 12.50 46.56 46.50 46.80
19 100 6 1.5 6 11.54 11.58 11.78 46.58 47.10 47.20
20 100 6 1.5 6 11.58 11.82 11.70 47.22 46.80 47.06
Oliveira, D. & Oliveira, J.V. (2001). Enzymatic alcoholysis Xylitol production by Candida sp.: Parameter optimiza-
of palm kernel oil in n-hexane and SCCO2. The Journal of tion using Taguchi approach. Process Biochemistry, 39(8),
Supercritical Fluids, 19(2), 141–148. 951–956.
Pai, Dayanand, Rao, Shrikantha & Shetty, Ravira (2012). Sudharsan, N.M. & Ng, E.Y. (2000). Parametric optimization
Application of Taguchi and response surface methodolo- for tumour identification: bioheat equation using ANOVA
gies for metal removal rate and surface roughness in grind- and the Taguchi method. Proceedings of the Institution of
ing of drac’s. International Journal of Engineering and Mechanical Engineers. Part H, Journal of Engineering in
Management Sciences, 3(1), 1–8. Medicine, 214(5), 505.
Phadke, M.S. (1989). Quality engineering using robust design. Taguchi, G. (1986). Introduction to quality engineering. Tokyo:
Englewood Cliffs, New Jersey: Prentice-Hall. Asian Productivity Organization.
Pignatiello, J.J., Jr. (1988). An overview of the strategy and tac- ———. (1987). System of experimental design. New York:
tics of Taguchi. IIE Transactions, 20(3), 247–254. UNIPUB Kraus, International Publications.
Ross, Philip J. (1989). Taguchi techniques for quality engineer- Wang, H.T., Liu, Z.J., Chen, S.X. & Yang, J.P. (1999). Appli
ing. New York: Prentice Hall. cation of Taguchi method to robust design of BLDC motor
Sreenivas Rao, R., Prakasham, R.S., Krishna Prasad, K., performance. IEEE International Magnetics Conference,
Rajesham, S., Sarma, P.N. & Venkateswar Rao, L. (2004). 35(5), 3700–3702.