Professional Documents
Culture Documents
Optimization of Fiber Orientation Distribution in Fiber Reinforced Composite Injection Molding by Taguchi Back Propagation Neural Network and Genetic Algorithm Particle Swarm Optimization
Optimization of Fiber Orientation Distribution in Fiber Reinforced Composite Injection Molding by Taguchi Back Propagation Neural Network and Genetic Algorithm Particle Swarm Optimization
Abstract
Fiber orientation induced by injection molding of short-fiber-reinforced composites causes anisotropy in material proper-
ties and produces warping. Fiber-orientation distribution is very important to research for mold design and quality to
produce sound molded parts. In this study, three kinds of methods are used to solve the optimization problem. Fiber-
orientation distribution is described by fiber-orientation tensor variation. The objective function is a minimum problem
of the fiber-orientation tensor variation. Parameters such as fiber content, fiber aspect ratio, melting temperature, injec-
tion pressure, holding pressure, and filling time are considered as design variables. Based on orthogonal experiment
design, Moldflow software is used in the fiber-reinforced acrylonitrile butadiene styrene composite injection molding.
The effects of process parameters for the plastic part are studied using the signal-to-noise ratio. The most important
design parameter influencing fiber-orientation tensor variation is determined by finite element analysis results based on
the analysis of variance. The optimization model is established on the basis of the back propagation neural network. The
Taguchi, the particle swarm optimization, and genetic algorithm–particle swarm optimization hybrid algorithm are used
to find the minimum fiber-orientation tensor variation value. Results show that the quality index of the fiber-orientation
tensor variation in the part is improved.
Keywords
Optimization, injection molding, Taguchi method, back propagation neural network, genetic algorithm–particle swarm
optimization hybrid algorithm
Creative Commons CC-BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License
(http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/
open-access-at-sage).
2 Advances in Mechanical Engineering
electrical conductivity, optical properties, and so on. In conditions in injection molding for short-fiber-
the injection molding filling, the fiber is suspended in reinforced polycarbonate composites. The effects of
the melt because of the deformation and effect of flow four main parameters—the filling time, melt tempera-
field, and the orientation will orient along a certain ture, mold temperature, and injection speed—on the
direction and change with different positions and dif- fiber orientation were investigated and discussed. It is
ferent time. Ultimately, the fiber orientation of the found that the dominant parameter is the filling time.
products will present in a complex condition, which In addition, many researchers are committed to
leads to the anisotropy of mechanical properties.1 building the relationship model between the objective
Therefore, the research of fiber orientation has impor- function and the influence factors by artificial neural
tant significance. network, Kriging model, and response surface metho-
In recent years, extensive research on the fiber orien- dology. Kurtaran et al.10 established a predictive model
tation has been conducted using the numerical simula- in injection molding using the artificial neural network.
tion. Hung and Shen2 predicted the fiber orientation It is validated that neural network model is valid for
based on Jeffery’s model during injection molding using predictive capability. Zhang et al.11 employed the
the finite element method. Aurich and Mennig3 pre- response surface methodology to obtain the relation-
sented a thorough experimental study of flax fiber ship model between the objective function and indepen-
orientation. Comparison of the measured and calcu- dent variables. Compared with the reference model, the
lated modulus in tension shows good agreement. Chang performance of the optimized model was significantly
et al.4 presented an effective methodology for the opti- improved. Ozcelik and Erzurumlu12 optimized process
mization injection molding process parameters of short- conditions to minimize the warpage of the part in injec-
glass-fiber-reinforced polycarbonate composites. A tion molding using analysis of variance (ANOVA), arti-
computer-aided engineering (CAE) flow simulation ficial neural network, and GA. Chow et al.13 researched
software was used to simulate the injection molding the global optimization of absorption chiller system.
process and to predict the fiber orientation. On the The neural network was used to model the system char-
basis of the CAE analysis, the optimum process para- acteristics. The results confirmed that this methodology
meters are obtained for the thickest shear layer. could be effective.
As previously mentioned, an extensive research on In recent research, the global optimization approach
the fiber orientation has been conducted using the has become a new trend in optimizing the process para-
numerical simulation method. Advanced methods are meters. GAs and particle swarm optimization (PSO)
highly demanded to model and optimize the injection have been widely used to solve optimization problems.
molding process with the purpose of manufacturing To minimize the warpage on the thin shell plastic
high-quality plastic parts. Many researchers have been part,14 an effective GA was used to obtain the optimum
devoted to the optimization research by their experi- process parameters. Chen et al.15 proposed the hybrid
ence to apply a trial-and-error method or the Taguchi GA-PSO to solve the optimization problem in plastic
approach. Although the methods are time-consuming injection. The final optimal combination of process
and costly, it is important to evaluate design para- parameters was obtained by the hybrid GA-PSO. The
meters. Kuram et al.5 researched that the glass-fiber- confirmation results show that the proposed method
reinforced nylon 6 was reprocessed in the five not only enhances the stability but also reduces the
processing cycles using Taguchi method. The result costs of and time spent in the injection molding.
showed that the method can improve effectively on As Taguchi method is a robust parameter design tech-
mechanical properties. Huang et al.6 presented a hybrid nology, the neural network has become a powerful and
swarm intelligence with Taguchi-genetic algorithm practical method for modeling highly complex nonlinear
(GA) to optimize the inverse kinematics. Wang et al.7 systems. GA and PSO have the broaden application in
investigated the injection molding process parameters various optimization research fields. Therefore, these
for manufacturing a brake booster valve body. The methods have been successfully applied in system model-
optimal parameters are determined by the Taguchi ing and process optimization. This study uses the statisti-
method integrating with the CAE to improve the com- cal method for the first time. In order to determine the
pressive property of the valve body. KC et al.8 best levels for design parameters, the Taguchi design of
researched the optimization of process parameters of experiment (DOE) technique is used to obtain the opti-
sisal-glass fiber hybrid biocomposite in injection mold- mal injection molding conditions. The ANOVA is used
ing by Taguchi method. The results show the optimiza- to find out the main control factors of the injection mold-
tion parameters meet the dimensional requirements of ing process. The neural network method is used to estab-
automotive parts. Chen et al.9 proposed an effective lish the relationship model. Finally, PSO and GA-PSO
methodology integrating the finite element and are, respectively, used in the parameter optimization of
Taguchi methods to determine the optimal design fiber-reinforced composite injection molding.
Li et al. 3
Level Factor
Fiber Fiber Melting Injection Holding Filling time (s)
content (%) aspect ratio temperature (°C) pressure (MPa) pressure (MPa)
A B C D E F
1 10 10 220 60 40 1
2 20 30 240 70 50 2
3 30 50 260 80 60 3
Level Factors
A B C D E F
products’ quality index. As the objective of the content of 30%, a fiber aspect ratio of 50, a melting
research is to minimize the fiber-orientation tensor temperature of 260°C, an injection pressure of 80 MPa,
variation, the S/N ratio related to the smaller-the- a holding pressure of 60 MPa, and a filling time of 1 s
better quality characteristic is selected when calculat- for the fiber-orientation tensor variation.
ing the S/N ratio. The response table of the S/N ratio
and diagram are given in Table 4 and Figure 2,
respectively. The table is used to determine the opti- ANOVA
mal control factor settings in minimizing the fiber- The purpose of ANOVA is to determine the statistically
orientation tensor variation. Table 4 presents the S/N significant parameters influencing the fiber-orientation
response values at three levels and the effect of each tensor variation in the designed experimental study.
factor on the S/N ratio. Figure 2 illustrates the effects The ANOVA results for fiber-orientation tensor varia-
of control factors by showing S/N values at three tion are given in Table 5. As shown in Table 5, the
level settings for each factor. F-values of factors D and F are all greater than
In Table 4, the best set of combination parameters F0.05, 2, 23 = 3.42. Therefore, the injection pressure and
can be determined by selecting the level with the highest filling time are important parameters. In particular, the
value for each factor. The analysis of the results shows filling time is the most influential factor for fiber-
that the optimal process parameter combinations for orientation tensor variation in injection molding, fol-
the fiber-orientation tensor variation are A3, B3, C3, lowed by the injection pressure and melting tempera-
D3, E3, and F1. These variations correspond to a fiber ture. Chen et al.15 studied optimization of the injection
Li et al. 7
molding process for short-fiber-reinforced composites with the simulation results. The developed BPNN
and found that the filling time is the most influential model gives a satisfactory approximation and can be
factor. This result is consistent with the results of this used as an efficient predictive tool for fiber-orientation
study. Moreover, the fiber content and holding pressure tensor variation.
are not effective factors of minimum fiber-orientation
tensor variation.
Optimization through PSO and GA-PSO
Based on the established relationship model using
BPNN model for fiber-orientation distribution BPNN, the PSO and GA-PSO are, respectively, used to
variation figure out the minimum fiber-orientation tensor varia-
Based on the above-mentioned results, the fiber content tion. The optimum value of fiber aspect ratio (B), melt-
and the holding pressure do not significantly affect the ing temperature (C), injection pressure (D), and filling
fiber-orientation tensor variation. Therefore, factors, time (F) toward minimum fiber-orientation tensor var-
such as fiber aspect ratio, melting temperature, injec- iation can be efficiently obtained by formulating it as
tion pressure, and filling time, are selected as the design an optimization problem, which can be stated as
parameters in establishment of BPNN model for fiber follows
orientation tensor variation because they potentially
affect the fiber-orientation tensor variation. The objec- Find : B, C, D, and F ð9aÞ
tive is to optimize the process parameters such that the Minimize: fiber-orientation tensor variation ðB, C, D, and FÞ
fiber-orientation tensor variation is minimized. The ð9bÞ
neuron number of the input layer of BPNN is deter-
mined by the number of design variables selected, and Subjected to constraints:
the neuron number of the output layer is determined by ð9cÞ
fiber-orientation tensor variation \0:52
the number of the objective indexes. Thus, the number
of input parameters is 4 and the output parameters is 1, In order to figure out the optimal solution, PSO and
that is, the node number of the input layer is 4, and the GA-PSO are used to solve the minimum problem. In
node number of the output layer is 1. The node number this study, a population size of 20, the dimension of the
of the hidden layer is determined with train trials, and particle is 2, and generation number of 100 are
the final value obtained is 10. Therefore, the artificial employed. The evolution of the generations is shown in
neural network configuration is 4-10-1. In this study, a Figure 5.
four-layer BPNN model with two hidden layers is used. Based on PSO algorithm, the figure shows that after
The configuration of BPNN is shown in Figure 3. 50 iterations, the optimal fitness value is 0.4725.
To obtain the training sample library containing However, based on GA-PSO algorithm, the figure
enough samples, the way of circulation exchange is shows that after 10 iterations, the optimal fitness value
used to extend the 27 data sets of orthogonal design in is 0.4307. The corresponding optimized parameter
Table 3 to 81 training sample data sets. combinations are, respectively, given in Table 6.
Among the 81 samples in the previous section, 61
samples are used to train the BPNN. Then, the remain-
ing 20 samples are used to test the BPNN performance. Verification of the optimization result
In Figure 4, the difference of the BPNN forecast and The confirmation experiment is very important in para-
the Moldflow simulation value is shown. The predictive meter design. The numerical experiments have been
value of the established BPNN is in good agreement performed based on the optimized process parameters
8 Advances in Mechanical Engineering
Figure 4. Comparison of fiber-orientation tensor variation with (a) FE analyses and (b) neural network for training data set.
by GA and GA-PSO method. The calculated fiber- and the minimum value is 0.5289 and 0.5051. The mini-
orientation tensor distribution is shown in Figure 6(c) mum value of the fiber-orientation tensor variation is
and (d). The difference between the maximum value 0.5289 and 0.5051. Therefore, the GA-PSO enhanced
Li et al. 9
Figure 5. Fitness curve with (a) PSO algorithm and (b) GA-PSO algorithm.
Figure 6. Fiber-orientation tensor distribution: (a) recommended parameter, (b) Taguchi method, (c) PSO optimization, and (d)
GA-PSO optimization.
time of 1 s provide the minimum fiber-orientation ten- 2. Hung CF and Shen YK. Numerical simulation of fiber
sor variation. Furthermore, this study indicates that orientation in injection mold filling. Int Commun Heat
the optimization methodology can also improve the Mass 1995; 22: 791–802.
other parts manufactured in fiber-reinforced composite 3. Aurich T and Mennig G. Flow-induced fiber orientation
injection molding. in injection molded fit fiber reinforced polypropylene.
Polym Composite 2001; 22: 680–689.
4. Chang SH, Hwang JR and Doong JL. Manufacturing
Declaration of conflicting interests process optimization of short glass fiber reinforced poly-
The author(s) declared no potential conflicts of interest with carbonate composites in injection molding. J Reinf Plast
respect to the research, authorship, and/or publication of this Comp 2000; 19: 301–321.
article. 5. Kuram E, Tasci E, Altan AI, et al. Investigating the
effects of recycling number and injection parameters on
the mechanical properties of glass-fibre reinforced nylon 6
Funding using Taguchi method. Mater Design 2013; 49: 139–150.
The author(s) disclosed receipt of the following financial sup- 6. Huang HC, Xu SD and Wu CH. A hybrid swarm intelli-
port for the research, authorship, and/or publication of this gence of artificial immune system tuned with Taguchi-
article: This work was supported by the Fundamental genetic algorithm and its field-programmable gate array
Research Funds for the Central Universities (WUT: realization to optimal inverse kinematics for an articu-
2017III020) and Science and Technology Research Project of lated industrial robotic manipulator. Adv Mech Eng
Jiangxi Provincial Department of Education (No. 2016; 8: 1–10.
GJJ160711). 7. Wang YQ, Kim JG and Song JI. Optimization of plastic
injection molding process parameters for manufacturing
a brake booster valve body. Mater Design 2014; 56:
References 313–317.
1. Yang YK. Optimization of injection molding process for 8. KC B, Faruk O, Agnelli JAM, et al. Sisal-glass fiber
mechanical properties of short glass fiber and polytetra- hybrid biocomposite: optimization of injection molding
fluoroethylene reinforced polycarbonate composites: a parameters using Taguchi method for reducing shrink-
case study. J Reinf Plast Comp 2006; 25: 1279–1290. age. Compos Part A 2016; 83: 152–159.
Li et al. 11
9. Chen CS, Chen TJ, Chen SC, et al. Optimization of the ANOVA, neural network model and genetic algorithm. J
injection molding process for short-fiber-reinforced com- Mater Process Tech 2006; 171: 437–445.
posites. Mech Compos Mater 2011; 47: 359–368. 13. Chow TT, Zhang GQ, Lin Z, et al. Global optimization
10. Kurtaran H, Ozcelik B and Erzurumlu T. Warpage opti- of absorption chiller system by genetic algorithm and
mization of a bus ceiling lamp base using neural network neural network. Energ Buildings 2002; 34: 103–109.
model and genetic algorithm. J Mater Process Tech 2005; 14. Kurtaran H and Erzurumlu T. Efficient warpage optimi-
169: 314–319. zation of thin shell plastic parts using response surface
11. Zhang X, Yu S, Gong Y, et al. Optimization design for methodology and genetic algorithm. Int J Adv Manuf
turbodrill blades based on response surface method. Adv Tech 2006; 27: 468–472.
Mech Eng 2016; 8: 1–12. 15. Chen WC, Nguyen MH, Chiu WH, et al. Optimization
12. Ozcelik B and Erzurumlu T. Comparison of the warpage of the plastic injection molding process using the Taguchi
optimization in the plastic injection molding using method, RSM, and hybrid GA-PSO. Int J Adv Manuf
Tech 2016; 83: 1–14.