Table 5 Values of S/N ratio at different levels

shown in Fig。 12 where point 2 located in the thick part and the thickness

S/N ratio Melt temp。 Mold temp。

Packing

press。

Packing

time

was 0。70 mm and point 9 located in the thin part and thickness was

0。37 mm。 At the same time in the thick part around point 2, temperature changed slowly since the lower temperature gradient, so the stress increased slowly。 Meanwhile, the stress of polymer part also rose up due to the influence of packing pressure, which kept the melt filling up。 As packing advanced, the stress in thin part reached the peak rapidly and the stress in thick part kept rising all the time but the growth rate became slow as the temperature gradient  decreased。

Before ejection from the mold, almost the complete product was subjected to stresses。 After releasing from the mold, there were no constrains from the mold, giving the product freedom to distort。 Thus, the stress changed and deformation obtained after ejection like shown in Fig。 13 in which residual stresses of 9 measurement points changed as the temperature of the product reduced from mold temperature to room temperature。 The deformation distribution at the room temperature was shown as Fig。 10(a)。 And simulation results were compared with the 3D Scanning measurement results as shown in Fig。 10(b)。

5。 Optimal Design

5。1 Taguchi method

The purpose of this research  was to minimize the distortion。     As

     Level 1         24。42265        27。91429        25。21323        24。57216   

     Level 2         24。43891        24。67459        24。43213        24。63586   

     Level 3         25。22749        21。50017        24。44369        24。88103   

Delta 0。804833 6。414114 0。769537 0。308867

distortion after ejection。 The process parameters that we considered were melt temperature, mold temperature, packing pressure and packing time。 These parameters were varied。21 The range of the melt temperature was selected as 320~300oC, the mold temperature was selected as 120~ 80oC, the packing pressure was from 180 MPa to 140 Mpa and the packing time was from 2 seconds to 1 seconds。 The ranges of these parameters were given in Table 2。 The ranges were selected based on existing and viable process conditions。 Therefore, a three level orthogonal array L9(34) was employed。 This array specified 9 experimental runs and has 4 columns。 Simulation data were used to analyze the mean response which consisted of the average value and the signal-to-noise (S/N) ratio。 The signal-to-noise (S/N) ratio could minimized the variation in the quality characteristics due to uncontrollable  parameters。22

The purpose of the design is to minimize the distortion of the produce, so deformations of total 9 points listed in Fig。 10(a) were measured as result data。 The distortion was “the smaller the better” type, so the S/N ratio was given as:23

n    2

mentioned before, some of these parameters affecting the quality were controllable while others were noise factors。 The most common method is the trial and error method, which is the most easily applicable in the