JournalChinnadurai of Chemical T., Arungalai Technology Vendan and S.,Metallurgy, Mohan Raj 52, N., 1, 2017,Prakash 41 N.- 51

STUDIES ON ULTRASONIC OF AND ACRYLONITRILE BUTADIENE STYRENE BLENDS

Chinnadurai T.1, Arungalai Vendan S.1, Mohan Raj N.2, Prakash N.3

1 School of Electrical Engineering (SELECT) Received 07 December 2015 VIT University, Vellore, India Accepted 07 October 2016 E-mail: [email protected] 2 Mechanical Engineering Shri Krishna College of Technology, Coimbatore 3 Department of Chemistry St. Martin’ s Engineering College, Dullapally, Secunderabad

ABSTRACT

Polycarbonate and arylonitrile butadiene styrene materials have several beneficial characteristics which facilitate their application in various industrial sectors. While developing products based on these polymers it is essential to enable the scope of its weld ability feature. Considering this critical factor, researchers are involved in investigating the weld ability behavior of the above mentioned polymers. This study reports an effective methodology developed to determine the optimum welding condition that maximizes the strength of joints produced by through the response surface method. A response surface methodology is employed to develop an effective model to predict the weld strength by incorporating process parameters such as pressure, weld time and amplitude. Experimental trials are conducted based on experiments design. An effective second order response surface model is developed based on the values measured. Subsequently, an analysis of the variances and inferences drawn from scanning electron microscope images are discussed. The research emphasizes that the pressure and welding time have a significant impact on the weld strength. Keywords: PC-ABS blend, ultrasonic welding, a response surface method, a weld strength.

INTRODUCTION has to be light as an excessive amount of energy is re- quired for the high frequency vibration. This in fact is a The polymers have found enormous applications noticeable lacuna of the method and limits its applica- in domestic use, packaging, food and beverages, auto- bility. Ultrasonic plastic welding plays a prominent role motive industry and various other industrial purposes. in the manufacturing of accessories used in automotive, Advanced, hybrid and polymer blends are play- aerospace, defense and electrical applications. There are ing an important role in producing vehicles interior and several research studies pertaining to the mechanism of exterior, body and structural components. Ultrasonic joint formation, the temperature distribution analysis welding (USW) is a solid state welding process in which at the weld interface, the joint strength, etc. Specific similar or dissimilar metallic components are joined by joint designs are the major concerns of USW process the application of high frequency vibrations under high employment as high ultrasonic energy is focused at the pressure. The high frequency relative motion between welding interfaces. Furthermore, the welding parameters the parts facilitates the solid progressive shearing and are highly coupled and show high sensitivity to toler- plastic deformation providing localized joining in a few ances and clamping of the parts to be welded. Ultrasonic seconds in absence of any significant amount of heat at welding typically provides a low-energy bonding tech- the work pieces. One of the parts in ultrasonic welding nique as compared with the conventional process. Two

41 Journal of Chemical Technology and Metallurgy, 52, 1, 2017 thin polymer sheets/plates having compatibility may be ing surface as illustrated in ref. [4]. In fact the energy welded by ultrasonic vibrations. Generally, amorphous directors are linear ridges with triangular, semicircular polymers are preferred over crystalline one aiming to or rectangular cross sections. Their number, shape, size achieve a good joint quality. It is worth adding that the and orientation have a significant effect on the welding application of high power ultrasonic energy creates oscil- process and its output parameters, as they influence heat lating shear forces that break the polymer chains, while generation and resin flow at the welding interface [5 - 6]. the static force present produces high strength solid-state Other important parameters of the USW process refer to bonds. Besides, the ultrasonic welding technology is also the holding pressure, the weld time, and the amplitudes the most appropriate method for producing hybrid joints of the vibration tip. for multi functional components. Fig.1 shows a scheme The study of joining a polymer blend containing of an ultrasonic plastic welding machine [1]. polycarbonate (PC) and acrylonitrile–butadiene–styrene Ultrasonic heat is created during the vibration phase (ABS) is imperative as it is of great practical impor- of the process upon the application of pressure and a tance. PC is a ductile when compared to vibration frequency ranging from 10 kHz - 70 kHz to other polymeric materials such as , , etc. 10 - 250 mm perpendicular to the parts to be welded [2]. ABS is classified as rubber toughened The process is realized through surface- and viscoelastic with scope for a wide industrial application. Due to the friction [3]. During the solidification phase, the weld is complementary properties of the components, PC-ABS allowed to cool down under pressure below the glass alloy has drawn considerable interest in engineering transition temperature (Tg) of the thermoplastic resin. applications requiring a high degree of toughness, as in Energy directors are embedded on the specimens to the automotive industry. Notomi et al. [7] investigate concentrate the heat generated at the interface between the tensile properties of PC-ABS alloy. Wang et al. [8] the parts to be joined, which in turn results in uniform examine the effect of tri-axial stress constraint on the de- energy distribution through the material solidification. formation and fracture of PC, ABS and PC-ABS alloys. But they can also bring about protrusions on the weld- Machmud et al. [9] report the strain, strain rate and

Fig.1. Schematic presentation of ultrasonic plastic welding equipment [1].

42 Chinnadurai T., Arungalai Vendan S., Mohan Raj N., Prakash N.

temperature dependent deformation behavior of PC- interaction effects. The general model equation of RSM ABS alloy on the ground of quasi-static uniaxial tension is as follows: tests carried out. Yin, Z. N. et al [10] study further the b b b b 2 viscoelastic deformation of PC-ABS blend with the y = ε 0 + ∑ε i xi + ∑ ∑ε ij xi x j + ∑ε ii xi (1) application of dynamic mechanical analysis (DMA) i=1 i=1 j=i+1 i=1 techniques. Yin, Z. N., and T. J. Wang [11] evaluate where y is the response or the weld strength in this , experimentally the deformation of PC-ABS blend in xi stands for the pressure, the weld time and the ampli- -1 case of application of strain rates of 8.0 × 102 s , 2.7 tude, while ε 0, ε i, ε ij and ε ii represent the constant, × 103 s-1 and 1.0 × 104 s-1. The study reveals that ABS linear, quadratic and interaction terms respectively. fractional content and strain rate play a vital role in the In engineering and allied industries it becomes joining process. O’Connell et al. [12] perform tests to imperative to find regions of response improvement study the creep tension properties and the slow crack instead of defining those providing an optimum re- growth. The joining of two dissimilar polymers namely sponse [15]. Gaitonde et al. [16] develop second order ABS and PC becomes an essential process in the auto- mathematical models for minimization of burr height mobile industry. The front and tail lights of several car and burr thickness using RSM. In this investigation, types are made of components containing PC and ABS five-level, half replicate, second order rotatable central [13]. The stress cracking is an important mode of fail- composite designs are adopted to study the interactions ure in joining PC-ABS. Internal tensile stresses induce effect. The RSM model developed is evaluated on the temperature in the range of 230°C–420°C, which leads ground of the variance analysis (ANOVA) and is found to cracks. Hence, the attainment of high weld strength adequate. The study carried out by N. Bradley [17] for these types of polymers is significant, especially for focuses on the design, modeling and analysis of RSM the automobile industry. and elucidates in depth the first-order, the second-order, There are discussions in the literature on polymer and three-level fractional factorial. Kumar et al. [18] hybrids and the need of joining them. Experimental propose a methodology to improve the mechanical investigations on the strength aspects and failure modes properties of AA 5456 aluminum alloy welds obtained are also reported. The techniques applied, especially the by the magnetic arc oscillation welding process. The response surface methodology (RSM) used for paramet- authors use the Taguchi’s method to optimize the pro- ric analysis, are further discussed. RSM is a combination cess parameters. The percentage of error between the of mathematical and statistical tools for modeling and experimental and predicted values is found negligible. analysis of problems where the response expected is The welds microstructures are investigated and cor- influenced by several variables and the objective is to related with the mechanical properties displayed. Fang optimize this response [14]. The process modeling by et al. [19] investigate the large deformation of PC-ABS RSM uses a statistical design of the experiments. It is plastics using digital image correlation (DIG) methods. based on the central composite face centered design. They find a relationship between the necking ratio and This method not only reduces the cost and time but the tensile strain, which depends on the geometry but also gives the required information about the main and not on the loading ratios.

Fig. 2. Prior to joining PC/ABS blend in presence and absence of an energy director. 43 Journal of Chemical Technology and Metallurgy, 52, 1, 2017

A second order RSM model is developed in this study for predicting the weld strength of PC-ABS joints. The model developed is employed to determine the optimum welding conditions leading to maximum weld strength.

EXPERIMENTAL

The trials were carried out using PC/ABS hybrid specimens to determine the range of ultrasonic welding parameters. The experiments were planned as per Central Com- posite Face Centered (CCF) design with the star points at the center of each face of the factorial space. This design fits accurately [20] the second order response surface. The appropriate parametric ranges were identified and listed in Table 1. Table 1. Variable levels used for the experiments conducted.

Factor Unit Levels Level 1 Level 2 Level3 Pressure (P) Bar 3.5 4 4.5 Fig. 3. Illustration of a digitally controlled ultrasonic plastic Weld time (T) Sec 4 4.5 5 welding machine (40 KHz). Amplitude (A) µsec 30 40 50 of inclination and positioning are illustrated in Fig. 2. The welding experiments were carried out using a Fig. 3 shows the equipment and positioning of speci- conventional ultrasonic welding machine (4800 W, 40 mens in a lap configuration. There were four phases of kHz) providing different ranges of weld parameters. The ultrasonic welding. During the welding itself, the horn design matrix presented in Table 1 was used. vibration was perpendicular to the joint surface and the The PC-ABS grains were procured in a 60 % - 40 point of the energy director. The latter melted and flew % ratio using Chennai.chemicals. They were subjected into the interface. The heat generation was high. The to preheating followed by mould preperation using second phase corresponded to the meeting of the part an equipment. Two sets of mould surfaces, while the melting rate increased. Steady-state specimens were prepared. They differred on account of melting occurred during phase 3 and a melt layer formed the energy directors presence. The specimens of the first at the interface. A maximum displacement was reached series contained triangular energy directors. Their angle during phase 4, and an additional melt was squeezed

Fig. 4. Welded PC/ABS blend specimen.

44 Chinnadurai T., Arungalai Vendan S., Mohan Raj N., Prakash N.

Table 2. ANOVA for PC-ABS weld meets with 95% confidence interval.

Analysis of Variance

Source DF Adj SS Adj MS F-Value P- Value Model 9 8.1198 0.9022 7.75 0.002 Linear 3 4.2731 1.4243 12.23 0.001 Pressure(bar) 1 3.6445 3.6445 31.30 0.000 Weld time (sec) 1 0.2099 0.2099 1.80 0.009 Amplitude (µm) 1 0.4186 0.4186 3.60 0.087 Square 3 0.1236 0.0412 0.35 0.787 Pressure (bar)*Pressure (bar) 1 0.0944 0.0944 0.81 0.389 Weld time (sec)*Weld time (sec) 1 0.0000 0.0000 0.00 0.999 Amplitude (µm)*Amplitude (µm 1 0.009 0.0096 0.08 0.780 2-Way Interaction 3 0.7230 1.2410 10.66 0.002 Pressure (bar)*Weld time (sec) 1 0.7527 0.7527 6.47 0.029 Pressure (bar)*Amplitude (µm) 1 0.2278 0.2278 1.96 0.192 Weld time (sec)*Amplitude (µm) 1 2.7424 2.7424 23.55 0.001 Error 10 1.1643 0.1164 Lack of Fit 5 0.8725 0.1745 2.99 0.127 Pure error 5 0.2917 0.0583 Total 19 9.2841 out of the joint interface in the form of a spruce. The conducted at different levels of parameters (Table 1) aim- intermolecular diffusion during the melt flow resulted in ing to obtain ultrasonic welded joints of PC-ABS blends. new polymer chain entanglements between the two parts The values of weld strength obtained from experiments being welded, serving to produce high weld strength. and those predicted from the response surface model are After a set weld time, and/or power level, and/or dis- tabulated in Table 4. tance was reached, the welder was turned off but the horn pressure was maintained as the weld cooled [13]. RESULTS AND DISCUSSIONS Fig. 4 shows PC-ABS weld metes after the welding. The corresponding welding input relations and output values The ANOVA technique [22] is employed to check are summarized in Table 2. Twenty experiments were the adequacy of the models developed at 95% confidence

Table 3. Coded coefficients for PC-ABS weld meets. Coded Coefficients Term Effects Coef SE Coef T- Value P-Value VIF Constant 3.449 0.117 29.40 0.000 Pressure (bar) 1.207 0.604 0.108 5.59 0.000 1.00 Weld time (sec) 0.290 0.145 0.108 1.34 0.009 1.00 Amplitude (µm) 0.409 0.205 0.108 1.90 0.087 1.00 Pressure (bar)*Pressure (bar) 0.371 0.185 0.206 0.90 0.389 1.82 Weld time (sec)*Weld time (sec) 0.001 0.000 0.206 0.00 0.999 1.82 Amplitude (µm)*Amplitude (µm) -0.118 -0.059 0.206 -0.29 0.780 1.82 Pressure (bar)*Weld time (sec) -0.614 -0.307 0.121 -2.54 0.029 1.00 Pressure (bar)*Amplitude (µm) -0.338 -0.169 0.121 -1.40 0.192 1.00 Weld time (sec)*Amplitude (µm) -1.171 -0.585 0.121 -4.85 0.001 1.00

45 Journal of Chemical Technology and Metallurgy, 52, 1, 2017 interval. As per the ANOVA technique, the model is ade- in the model have a significant effect on the response. quate within the confidence level, provided the calculated Similarly, the main effect of the pressure (p), the weld time value of the F ratio of the lack of fit to the pure error does (s), the two-level interaction pressure*weld time, weld not exceed the standard tabulated value of F-ratio [16]. time*amplitude and pressure*amplitude are significant Here, the pressure, weld time and amplitude models are model terms. Other model terms have an insignificant adequate in a non-linear form as illustrated in Table 4. contribution. These insignificant model terms (not counting those required for the hierarchy support) may be eliminated Regression Equation to evolve an improved model. The lack-of-fit appears also Weld strength (N mm-2) = insignificant. This is a desirable model to achieve a proper = - 41.1 + 2.15 Pressure (bar) + 9.87 Weld time (Sec) fit. The pressure and the weld time are the most significant + 0.730 Amplitude (µm) + 0.741 Pressure (bar)*Pressure (bar) factors associated with PC-ABS welding. The results + 0.001 Weld time (Sec)*Weld time (Sec) presented in Table 3 show the coefficient of determination - 0.00059 Amplitude (µm)*Amplitude (µm) used to examine the weld strength with respect to the pres- - 1.227 Pressure (bar)*Weld time (Sec) sure and weld time. That provides a measure of variability - 0.0337 Pressure (bar)*Amplitude (µm) in the observed response values and may be explained by - 0.1171 Weld time (Sec)*Amplitude (µm) the controllable factors and their interactions. The final It is essential to perform tests for significance of the mathematical model in coded forms used to predict better regression model, a test for significance of individual welding conditions for PC-ABS blends is presented as: model coefficients and a test for lack-of-fit. ANOVA table Y = -41.1 + 2.15 X1 + 9.87 x2 + 0.730 x3 + 0.741 X1 is used to compare the tests performed. Table 2 shows the * X1 + 0.001 x2* x2 - 0.00059 x3* x3 - 1.227 X1 * x2

ANOVA table for the surface roughness. The value of F - 0.0337 X1* x3 - 1171 x2* x3 in Table 2 for the model is less than 0.05, which indicates where Y is the weld strength, X1 is the pressure (bar), x2 that the model is significant. It emphasizes that the terms is the weld time (sec), while x3 is the amplitude (µsec). Table 4. Experimental weld strength values.

Pressure Weld strength Ex. No Weld time (Sec) Amplitude (µm) (bar) Experiments (N mm-2) 1 3.5 5 30 3.548 2 3.5 4 30 1.451 3 4 4.5 40 3.365 4 4 4.5 50 3.846 5 4.5 4.5 40 4.165 6 4.5 5 50 3.482 7 4 4.5 40 3.154 8 3.5 4.5 40 3.486 9 4.5 5 30 4.764 10 4 4.5 40 3.128 11 4.5 4 50 4.954 12 4 4.5 40 3.357 13 4 4.5 40 3.159 14 3.5 5 50 3.246 15 4 4 40 3.846 16 4 4.5 30 3.316 17 3.5 4 50 3.186 18 4 5 40 3.435 19 4 4.5 40 3.765 20 4.5 4 30 3.589

46 Chinnadurai T., Arungalai Vendan S., Mohan Raj N., Prakash N.

Table 5. Juxtaposition of the optimum experimental and model values of PC-ABS specimens weld strength.

Pressure Experimental RSM Weld Time Amplitude Weld strength Weld strength (sec) (µm) (bar) (N mm-2) (N mm-2) 4 4 40 3.846 3.846

The juxtaposition of the predicted and the experimen- weld time and the amplitude increase when they move tal values referring to the weld strength shows that the from a low level to a high level. Each level of the factors deviation is insignificant. The optimum weld strength affects the response differently. Each factor at its high obtained from the experiments carried out and the RSM level results in higher mean response compared to that at model are shown in Table 5. A plot of the means of the the low level. Alternatively, the factor pressure appears response variable for each level of a factor provides to have a greater effect on the responses, compared with elucidating the main effects of interdependent process the slope. The main effect plots facilitate the visualiza- parameters. The main effect is calculated by subtract- tion of the factors that affect the response. ing the overall mean for the factor considered from the The normal probability plots of the residuals and mean for each level. Fig. 5 shows the locations of the the plots of the residuals versus the predicted response main effects referring to the input parameters and the in respect to the experimental weld strength is shown in weld strength achieved. It is seen that the pressure, the Fig. 6. Residuals generally fall on a straight line imply-

Fig. 5. Main effect plot referring to the experimental weld strength.

Fig. 6. A probability plot referring to the experimental weld strength. 47 Journal of Chemical Technology and Metallurgy, 52, 1, 2017 ing that the errors are normally distributed. This implies that the model proposed is adequate and excludes any possibility of independence violation or a constant vari- ance assumption. The surface plot facilitates attaining the shape of the response surface, the corresponding hills, valleys, and ridge lines. That is why the function f (p, a) from Eq. 2 is plotted versus the levels of p and a as shown as Fig. 7. In this graph, each value of p and a generate a y-value. This three-dimensional graph illustrates the response surface obtained. The contour plots (Fig. 8) are derived using Eq.1. The fitted coefficient is calculated Fig. 8. Contour plot of weld strength vs. pressure and amplitude. by maintaining one of the variables constant.This is visualized by the “hold values” in Fig. 8. The contour plot shows a significant variation in the weld strength as a function of the amplitude and the pressure. This is seen in case of weld strength increase by moving along the pressure axis in respect to a given amplitude value. The curve representing the weld strength at different amplitude and pressure values are self explanatory. In addition, the slope of weld strength curve with respect to the pressure values indicates a minimal change in the amplitude. The dependence of the response on the pressure and the amplitude can be described by the following Fig. 9. Surface plot of the weld strength vs. the pressure equation: and the weld time. = + y f ( p,a) e (2) where, Y is the response, p denotes the pressure, a de- notes the amplitude, while e stands for the error term The function f (p, t) versus the levels of p and t can be described in accordance with: = + y f ( p,t) e (3)

Fig. 10. Contour plot of the weld strength vs. the pressure and the weld time.

where, Y is the response, p denote pressure, t denote weld time, e denote error term. The surface plot of the weld strength vs. the pressure and the weld time presented in Fig. 9 indicates the presence of hold values. It is clearly evident that the pressure and the weld time affect significantly the weld strength. The fit Fig. 7. Surface plot of weld strength vs. pressure and amplitude. coefficient is calculated by maintaining the other variable 48 Chinnadurai T., Arungalai Vendan S., Mohan Raj N., Prakash N.

constant. This is shown as “hold values” in Fig. 10. The weld strength increase is observed along the pressure axis as well as the weld time axis. The ANOVA table expresses the same observation as well. The plot pre- sents the weld strength in different pressures and weld time ranges. Besides, the slope of curve shows that the pressure values change considerably as a result of small changes of the weld time. The function f (t, a) can be plotted versus the levels of t and a in correspondence with Eq. 4: y = f (t,a) + e (4) Fig. 12. A contour plot of the weld strength vs. the weld The plots obtained are presented in Fig. 11. In this time and the amplitude. graph, each value of t and a generates a y-value. Surface plots “hold value” is shown in Fig. 11. Weld strength is not affected largely as compared to the previous plots. Weld strength will have minimal changes. The fitted coefficient by maintaining one of the variables constant is shown as “hold values” in Fig. 12. The contour plot indicates that there is the significant difference in weld strength as a result of amplitude and weld time. ANOVA table also validate the observation that the amplitude values are not significant and weld time play a significant role. Analysis shows that pressure contributes more to the weld strength. The curve representing weld strength at different ampli- tude and different weld time ranges are shown.

Response surface analysis The developed second order RSM-based models for Fig. 13. A dependence of the weld strength on. the pressure, weld strength are used to analyze the factor interaction the weld time and the amplitude. effects on pressure, weld time and amplitude by plotting 3D-response surface plots. Since the main objective of of the weld strength for specified input parameters, the the present investigation is to get the optimum values influence of a combination of the pressure along with the remaining process parameters, namely, the weld time and the amplitude is predicted. Fig. 13 shows that the weld strength appears to be high for amplitude values of 30, 40 and 50. However, it is not the amplitude, but the corresponding pressure and weld time that lead to weld strength increase. In fact the amplitude has a negligible contribution to the weld strength.

SEM analysis

The study of the deformed polymer morphology is essential in understanding the form and structure of Fig. 11. A surface plot of the weld strength vs. the weld the welded polymer interface. The deformed ridges time and the amplitude. are prominent features of the welded samples. Their 49 Journal of Chemical Technology and Metallurgy, 52, 1, 2017

A) B) a

C) D)

E)

Fig. 14. SEM images of the welding region of samples A, B, C, D and E. number and the depth are related to the welding time, to the peaks. The higher peaks refer to good melting the amplitude and the weld pressure. At short welding conditions in respect to the energy directors, which are times, the depth of interpenetration across the interface well dispersed in both PC and ABS phases. The continu- is considerably small, while the strength and the energy ous phase is the component of the higher content, while at the welded interface are low. Fig.14 shows SEM mi- the dispersed particles are the minor component. The crographs of PC-ABS samples of different blend input ABS-rich blends have a conventional blend morphology combinations. The white particles in Fig.14 correspond with PC domains dispersed in ABS.

50 Chinnadurai T., Arungalai Vendan S., Mohan Raj N., Prakash N.

CONCLUSIONS stress constraint on the deformation and fracture of polymers, Acta Mechanica Sinica, 18, 5, 2002, 480-493. This study reports the development of a second order 9. M.N. Machmud, M. Omiya, H. Inoue, K. Kishimoto, mathematical model using RSM aiming to predict the Rubber particle size effect on impact characteristics maximum weld strength of PC-ABS specimens with em- of PC/ABS (50/50) blends. In Fracture of Nano and bedded energy directors. Ultrasonic welding is applied. Engineering Materials and Structures, 2006, 505-506, Experiments are carried out to identify the parameters Springer Netherlands. affecting the weld strength and to optimize the range 10. Z.N. Yin, L.F. Fan, T.J. Wang, Experimental inves- of all parameters studied. RSM is further employed to tigation of the viscoelastic deformation of PC, ABS investigate the parametric interdependency. The results and PC/ABS alloys, Materials Letters, 62, 17, 2008, reveal that the weld strength depends primarily on the 2750-2753. pressure and the weld time. The good coherence between 11. Z.N. Yin, T.J. Wang, Deformation of PC/ABS alloys at the model and the experimental data indicates valida- elevated temperatures and high strain rates, Materials tion, which may be extrapolated to studies of joining Science and Engineering: A, 494, 1, 2008, 304-313. processes with the participation of other materials. 12. P.A. O’connell, M.J. Bonner, R.A. Duckett, I.M. Ward, The relationship between slow crack propa- REFERENCES gation and tensile creep behavior in , Polymer, 36,12, 1995, 2355-2362. 1. A. Norouzi, M. Hamedi, V.R. Adineh, Strength 13. Troughton, J. Michael, Handbook of plastics joining: modeling and optimizing ultrasonic welded parts of a practical guide, William Andrew, 2008. ABS-PMMA using artificial intelligence methods, 14. Montgomery, C. Douglas, Design and Analysis of The International Journal of Advanced Manufacturing Experiments: Response surface method and designs, Technology, 61,1-4, 2012, 135-147. New Jersey: John Wiley and Sons, Inc., 2005. 2. A. Grewell David, A. Benatar, J.B. Park. Plastics and 15. M.H. Raymond, A.I. Khuri, W.H. Carter, Response composites welding handbook, v. 10, 2003. surface methodology, 1966–l988, Technometrics 31, 3. Z. Zhang, X. Wang, Y. Luo, Z. Zhang, L. Wang, 2, 1989, 137-157. Study on heating process of ultrasonic welding for 16. V.N. Gaitonde, S.R. Karnik, B.T. Achyutha, B. thermoplastics, Journal of Thermoplastic Composite Siddeswarappa. Genetic algorithm-based burr size Materials, 23, 5, 2010, 647-664. minimization in drilling of AISI 316L stainless steel, 4. S.J. Liu, I.T. Chang, S.W. Hung, Factors affecting the Journal of materials processing technology, 197, 1, joint strength of ultrasonically welded 2008, 225-236. composites, Polymer Composites, 22, 1, 2001,132-141. 17. N. Bradley, The Response Surface Methodology, 5. I.F. Villegas, Strength development versus process Dissertation, Indiana University South Bend, 2007. data in ultrasonic welding of thermoplastic compos- 18. A. Kumar, P. Shailesh, S. Sundarrajan, Optimization of ites with flat energy directors and its application to the magnetic arc oscillation process parameters on mechani- definition of optimum processing parameters, Com- cal properties of AA 5456 Aluminum alloy weldments, posites Part A: Applied Science and Manufactur- Journal of Materials and Design, 29, 2008, 1904-1913. ing, 65, 2014, 27-37. 19. Q.Z. Fang, T.J. Wang, H.G. Beom, H.P. Zhao, 6. A. Levy, S. Le Corre, A. Poitou, Ultrasonic welding Rate-dependent large deformation behavior of PC/ of thermoplastic composites: a numerical analysis of ABS, Polymer, 50, 1, 2009, 296-304. the mesoscopic scale relating processing parameters, 20. G. Padmanaban, V. Balasubramanian, Optimization of flow of polymer and quality of adhesion. Int. J. Mater. laser beam welding process parameters to attain maximum Form., 7, 1, 2014, 39-51. tensile strength in AZ31B magnesium alloy, Journal of 7. M. Notomi, K. Kishimoto, T.J. Wang, T. Shibuya. Ten- Optics and Laser Technology, 42, 2010, 1253-1260. sile and fracture behaviors of PC/ABS polymer alloy, 22. W.G. Cochran, G.M. Cox, Lattice designs, Ex- In Key Engineering Materials, 183, 2000, 779-784. perimental designs, John Wiley and Sons, New 8. T. Wang, K. Kishimoto, M. Notomi, Effect of triaxial York, 1992, 396-438. 51