Taguchi Optimization to Determine Optimal Denier of the Polyester Fibre

Shuwen Wang1, a and Te-Li Su2,b 1 Department of Health and Leisure Management, St. Mary's Junior College of Medicine, Nursing

and Management, Yilan 266, Taiwan

2 Department of Cosmetic Application and Management, St. Mary's Junior College of Medicine,

Nursing and Management, Yilan 266, Taiwan

apolarbearwen@yahoo.com.tw,

bsuteli66@gmail.com

Keywords: Polyester, Taguchi method, Optimization.

Abstract. Polyester fibre is the synthetic fiber made with man-made fibers. Its surface is smooth with the texture of fine linen. It has good flexibility and durability and is the best non-iron fibres. In order to promote the accuracy and reduce the quality deviation of melt spinning, the Taguchi method was adopted in the melt spinning process. The aim of the research is to enhance the optimization of the melt spinning process using the material polyester. First, the Taguchi method is applied to establish the design of experiment. An analysis of variance will be done to arrive at the significant factors that influence the melt spinning quality the most. Lastly, confirmation experiments verified the reproducibility of this experiment. Confirmed by the experiments, results showed that the obtained SN ratios were greater than the signal-to-noise ratio of nine orthogonal array experiments and this means the experiment is reliable.

Introduction

Polyester [1-5] is a thermoplastic material refined from crude oil through dissolution, continuous spinning process, drawing and thermosetting process to be able to have features, such as high energy absorption capacity, high elastic modulus, resistance to high temperature, excellent dielectric insulating properties, high resistance to chemicals and solvents.

In chemical, polyester is the general name of those polymers formed by ester binding with -CO.O- bond. These polymers are generated by condensation of dicarboxylic acid and divalent alcohol or polyesterification of oxy-acid. They are formed by polycondensation between carboxylic acid group and a hydroxyl group, removing water to form the ester bond. Because of the chemical characteristics of the polyester, clothing produced with polyester fibers is not affected by acid with acid resistance and alkali resistance, but will be decomposed at the high temperature.

Why is the polyester fiber used in clothing? It belongs to man-made fibers and has smooth surface without pores and cracks. In addition to its features, such as high strength, not easy to break, strong stretching, dry fast and so on, for the cleaning on storage, it also has easy to fold, not easy to shrink, not easy to have mold, and less likely to cause human allergies, etc. If the polyester fibers are spun closer during spinning process, it can even achieve effects of windproof and waterproof.

Melt spinning is the most commonly used method of synthetic fiber manufacturing. The most important physical characteristic of fiber is the denier affected by processing parameters, including spinneret temperature, screw speed, gear speed, and winding speed, etc. First, choose a suitable orthogonal array for the experimental programs to conduct experiments and understand the effect of processing parameters on the quality by using the signal-to-noise (SN) ratio and the main effect analysis. Also, confirm the reproducibility of experiments, and then obtain the optimal combinations of processing parameters.

Because of too many processing parameters during manufacturing and the expectation of getting the optimal processing parameters, there is difficulty in setting to achieve the optimal value of quality of fibre. However, the Taguchi method [6-9] can reduce the number of experiments and evaluate the effect of processing parameters on quality characteristic. In recent years, Taguchi quality engineering

Applied Mechanics and Materials Vol. 472 (2014) pp 688-692Online available since 2014/Jan/08 at www.scientific.net© (2014) Trans Tech Publications, Switzerlanddoi:10.4028/www.scientific.net/AMM.472.688

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of TTP,www.ttp.net. (ID: 130.207.50.37, Georgia Tech Library, Atlanta, USA-17/11/14,00:08:55)

combines with experimental design in the industry, quality control techniques, and statistical methods to significantly improve the quality of projects, and even receives affirmation by domestic and foreign industry and academia.

Consequently, the Taguchi quality method was applied to track the progress of the experiment. During the experimental process, the quality characteristic of the product was decided upon, and the processing parameters and theirs levels were initially set. Then, based on the adequate orthogonal array, the experiments were conducted and the data collected. Afterwards, the SN ratios and ANOVA were calculated to complete the optimization of the PET fibres.

Experimental

In the Taguchi method [10-13], parameter design takes the interaction between control factors and the nuisance factors, and the advantages of non-linear effect to describe the relationship between the processing parameters and quality characteristic, and help achieve the purpose of parameter optimization. The method is to transfer quality characteristic into SN ratio, showing the effect of control factors and error factors on quality characteristic, to reduce variability and look for the optimal mean of quality characteristic. This is also the biggest difference from the traditional experimental design. Can be obtained by SN ratio definition: SN = log (mean / variance) dB

SN ratio can measure the stability of quality, but because of the needs for the analysis of results, uses different signal to noise ratio according to the type of quality characteristic. Thus, Smaller the Better is listed.

)yn

1log(10SN

2

i∑−= (1)

where, yi is the measured value and n is the total number of measurements. After conducting a Taguchi planned experiment, the derived experimental data are subjected to an

analysis of variance (ANOVA) in order to acquire the complete experimental results. The main goal of ANOVA is to evaluate the experimental error and to test its significance. Through using ANOVA in the experimental and applying a non-subjective assessment method, the effect of each control factor on the experimental error can be isolated. This allows the significance of each factor to be quantified numerically, ensuring that no significant factors are left out, thus increasing the accuracy of predictions.

Results and Discussion

The processing parameters used to control in this paper are spinneret temperature (A), screw speed (B), gear speed (C), and winding speed (D). If each parameter is divided into three a categorical value, each experiment requires at least 81 (34) times. The scale of experiments is so larger that cost and time are not allowed. Therefore, in this paper, using L9 orthogonal array as experimental planning is the basis of conducting experiments. As shown in Table 1, the letters of the first column represent each processing parameters, including temperature units, oC, and speed in rpm.

Using the formula of characteristics of quality calculates the SN ratio to show the denier of the PET fibres in Smaller the Better, as shown in Eq. 1. The SN ratios of nine groups were calculated respectively with the orthogonal array. The results are shown in Table 2.

Using L9 orthogonal array for planning, the SN ratio of each set of experimental data was substituted into Eq. 1. We calculated the average response of categorical values of each parameter, and draw the response table and response graph, as shown in Table 3 and Fig. 1. According to the response graph and response table, the optimal combination of factors is the A3, B2, C3, and D1, meaning spinneret temperature of 230 oC, screw speed of 25 rpm, gear speed of 45 rpm, the winding speed of 1,000 rpm. We can also understand the impact of processing parameters at different levels on the denier. Among these parameters, the most significant factor is screw speed, flowed by spinneret temperature.

Applied Mechanics and Materials Vol. 472 689

Table 1. Experimental control factors and their levels.

Factor Level 1 Level 2 Level 3

A. Spinneret temperature 190 °C 210 °C 230 °C

B. Screw speed 15 rpm 25 rpm 35 rpm

C. Gear speed 25 rpm 35 rpm 45 rpm

D. Winding speed 1,000 rpm 1,500 rpm 2,000 rpm

Table 2. Experimental results.

No. A B C D SN ratio (dB)

1 1 1 1 1 -18.7031

2 1 2 2 2 -19.6931

3 1 3 3 3 -21.2431

4 2 1 2 3 -21.1028

5 2 2 3 1 -17.6028

6 2 3 1 2 -20.4728

7 3 1 3 2 -16.6231

8 3 2 1 3 -16.5334

9 3 3 2 1 -19.9929

Table 3. Response table.

Factor A B C D

Level 1 -19.8797 -18.8097 -18.5698 -18.7662

Level 2 -19.7261 -17.9431 -20.2629 -18.9296

Level 3 -17.7164 -20.5696 -18.4896 -19.6264

Effect 2.1633 2.6265 1.7733 0.8602

Rank 2 1 3 4

-23

-21

-19

-17

-15

A1 A2 A3 B1 B2 B3 C1 C2 C3 D1 D2 D3

SN

rat

ios

Figure 1. Response graph.

690 Mechanical Science and Engineering IV

The ANOVA table used to calculate the SN ratios in Table 2 is shown in Table 4. From the ANOVA table it is possible to see that the control factors D had a smaller influence on the quality characteristic of the PET fibre. Thus, it could be listed as pooled error. On the other hand, control factors A, B and C had more significant effects. After five confirmation experiments the resultant average SN ratio was -15.9371dB. The SN ratios of PET fibres fell between -21.2431 dB and -16.5334 dB. The average SN ratio of PET fibres from confirmation experiments was -15.9371 dB, which was greater than the SN ratio of nine orthogonal array experiments. This indicates that the factor effects have very good reproducibility and that the results of the experiment are very reliable.

Table 4. Analysis of variance table.

Source of Variable DOF SS Variance F ratio

A 2 8.7422 4.3711 6.9818

B 2 10.7467 5.3733 8.5826

C 2 6.0178 3.0089 4.8060

D 2* - - -

Pooled error 2 1.2521

Total 8 26.7588

* The pool-up terms

Conclusions

The Taguchi method is used in this study to plan experiments in order to identify the optimal combination of the processing parameters of PET fibres melting spinning with minimum experiments. The optimal combination of processing parameters includes spinneret temperature of 230 oC, screw speed of 25 rpm, gear speed of 45 rpm, the winding speed of 1,000 rpm. The SN ratios from confirmation experiments were greater than the SN ratio of nine orthogonal array experiments, meaning that the selection of significant factors is appropriate and the experimental results are reproducible. Moreover, the results show the experiment is trustworthy. In addition, from the main effect analysis of Table 3, spinning temperature and screw speed are significant factor affecting denier. Therefore, this research method is also suitable for other process parameter optimization to benefit the improvement of the processing efficiency.

Acknowledgement

This research was supported by the National Science Council of the Republic of China under contract no. NSC 102-2221-E-562-002.

References

[1] S. Şardağ and Ö. Özdemir: Fibres Text. East. Eur. Vol. 21 (2013), p. 49.

[2] J.Czekalski, D.Cyniak, T.Jackowski and K. Sieradzki:Fibres Text. East. Eur. Vol. 15(2007), p. 38.

[3] P. Gnanavel and T. Ananthakrishnan: Fibres Text. East. Eur. Vol. 21 (2013), p. 137.

[4] X. Liu, J. Liu1and X. Su: Fibres Text. East. Eur. Vol. 20 (2012), p. 102.

[5] E. C. Sabir and Y. Yeşil: Fibres Text. East. Eur. Vol. 19 (2011), p. 125.

Applied Mechanics and Materials Vol. 472 691

[6] T.L. Su, Y.L. Kuo, T.J. Wu and F.C. Kung: J. Chem. Technol. Biot. Vol. 87 (2012), p. 102.

[7] G.B. Hong and T.L. Su: Fibres Text. East. Eur. Vol. 20 (2012), p. 75.

[8] T.L. Su, C.C. Lai and P.C. Tsai: J. Surfactants Deterg. Vol. 14 (2011), p. 363.

[9] G.B. Hong and T.L. Su: Iran. Polym. J. Vol. 21, p. 877.

[10] T.L. Su, H.W. Chen and C.F. Lu: Adv. Polym. Tech. Vol. 29 (2010), p. 54.

[11] S. Mavruz and R.T. Ogulata: Fibres Text. East. Eur. Vol. 18 (2010), p. 78.

[12]T.L. Su, C.S. Chiou and H.W. Chen: Int. J. Photoenergy Vol. 2012 (2012), Article ID909678,p. 1.

[13] T.L. Su, C.F.J. Kuo and H.J. Wei: Fiber.Polym. Vol. 12 (2011), p. 657.

692 Mechanical Science and Engineering IV

Mechanical Science and Engineering IV 10.4028/www.scientific.net/AMM.472 Taguchi Optimization to Determine Optimal Denier of the Polyester Fibre 10.4028/www.scientific.net/AMM.472.688

DOI References

[8] T.L. Su, C.C. Lai and P.C. Tsai: J. Surfactants Deterg. Vol. 14 (2011), p.363.

http://dx.doi.org/10.1007/s11743-011-1266-3 [10] T.L. Su, H.W. Chen and C.F. Lu: Adv. Polym. Tech. Vol. 29 (2010), p.54.

http://dx.doi.org/10.1002/adv.20181 [13] T.L. Su, C.F.J. Kuo and H.J. Wei: Fiber. Polym. Vol. 12 (2011), p.657.

http://dx.doi.org/10.1007/s12221-011-0657-4