Fig。 1。  The  configurations  of  the specimens。

10754 C。-P。    Chen    et    al。  /    Expert    Systems    with    Applications    36    (2013)      10752–10759

Table 2

Simulation factors and factor  levels。

Levels of simulation factors Experimental control factors

A (°C)

Melt temperature B (°C)

Mold temperature C (%)

Inject speed D (MPa)

Packing pressure

1 310 115 65 40

2 320 125 75 45

3 330 135 85 50

2007; SadAbadi & Ghasemi, 2007)。 For a preliminary study, Moldflow® software is utilized to  identify  some  crucial  settings for the injection molding process。 Table 2 lists the process factors and factor levels of the injection molding conditions based on first result of the Moldflow® simulation。 The melting temperature (i。e。, A(°C)), the mold temperature (i。e。, B(°C)), the injection speed (i。e。, C(%)), and the parking pressure (i。e。, D(MPa)) are the factors for the injection molding process, respectively。 Table 3 lists eighteen runs based on an orthogonal array L18(21 × 37), consisting of four experimental factors with three levels for each   factor。

3。4。 Steps for process parameter   optimizations

The following steps are followed for process   optimization: Step 1。   Plan on simulation and experiments based on an appropri-

ate orthogonal array table。

Step 2。 Use Mold-Flow Plastics Insight Package proceeds to conduct simulations based on the orthogonal array table; utilize the ANOVA approach to identify significant factors and then apply regression analysis to model the relation- ship between the injection molding process parameters and targeted output of warpage distortions。 Validate the adequacy of regression equations through residual analysis。

Step 3。 Search for an optimal solution based on a desirability func- tion defined in the Eq。 (1)。 Fig。 2 shows the simulation flow chart。

Step 4。 The experiment process is applied injection machine man- ufacture parts that warpage distortions were coordinate measuring machine (CMM)  measured。

Step 5。 Verify the simulation and experiments results by addi- tional verification runs。

Fig。 2。 The simulation flow chart。

4。 Warpage measurement

The amounts of warpage at the middle locations of the edge of both the x-direction and the y-direction specified by the customers are measured。 3D-CNC CMM (series 500, Mitutoyo Co。, Japan)  with

Table 3

Results of the simulation and experimental runs (in   mm)。

Run no。 A B C D dxsim dysim dxexp