S /N = –10log 1---∑yi

ni=1

where yi  is the response value for measurement point。

(12)

foundries environment。 However, it demands extensive experiment work and causes excessive losses in time and money。 Taguchi method is an ideal tool for continuous and rapid quality improvements, which uses a few of experimental designs, provides a clear understanding of the variation。19,20

This paper analyzed various significant process parameters of the process of injection molding with metal-inset in order to minimize the

The result values of measurement point and the average value and the S/N ratio for each case were shown in Table 3。 The average value of distortion for each parameter at level 1~3 and the difference between the lowest value and the highest value were shown in Table 4 and plotted in Fig。 14。 The S/N ratios of various parameters at the different levels were given in Table 5 and plotted in Fig。 15。 It was clear from Fig。 14 that distortion was minimum value at the third level of parameters: melt

Fig。 14 Average values of distortion for each parameter at level from 1 to 3

Fig。 15 S/N ratios of various parameters at the different levels

temperature, mold temperature and packing time and at the first level of packing pressure。 It was also clear that from Fig。 15 that distortion was minimum value at the third level of mold temperature and packing time and at the first level of melt temperature and packing pressure。 From Figs。 14 and 15, we concluded that mold temperature was the most significant parameter and melt temperature and packing pressure also had some influence on distortion。

5。2 Distortion  minimization

The influence parameters affecting the final distortion have been found out。 Minimizing the distortion should be taken in the next step。 Response Surface Method (RSM) helped us to open a door to get optimal value,24,25 which explored the relationships between explanatory variables and response variables。 The surface plot of a signal response for a desired pair of independent factors was shown in Fig。 16, and the response function was obtained as follows:

Fig。 16 Surface plot of a pair of independent factor: (a) response between mold temperature and melt temperature; (b) response between mold temperature and packing pressure。

f = 24。696 – 0。716x1 + 3。138x2 + 0。817x3

– 0。209x4 – 0。865x1x2 + 0。138x1x3 + 0。628x2x3

(13)

where, f is the response variable, x1 stands for initial temperature of polymer before injecting, x2 stands for mold temperature, x3 stands for packing pressure, and x4  stands for packing time。

From Fig。 16, we knew that small mold temperature, melt temperature and big packing pressure could get small response value that    meant

Fig。 17 Distortion minimization results: (a) Distortion pattern of optimization; (b) Comparison graph of distortion

getting small distortion。 To solve optimization problems, Conjugate Gradient Method (CGM) was used。26,27 The range of each variable were shown in Table 2。 The optimal results were a melt temperature of 309。569oC and a mold temperature of 80。098oC, a packing pressure of 158。818 MPa and a packing time of 1。467 s that were quite similar to the results obtained by Taguchi method。 In Fig。 17, the distortion of the product was declined obviously。

6。 Conclusions

The capability of controlling distortion of injection molding with metal-insert is a highly challenging in manufacturing that only can be solved if the mechanisms of distortion development and its effects on the process are well understood。 In our research, the following conclusions could be obtained: