Abstract: Taking housing of a high-voltage connector socket as research object, this paper focuses on mold design and injection molding process optimization. In terms of mold design, point gate scheme was changed to a horn-shaped gate, simplifying mold structure from a three-plate mold to a two-plate mold. Horn-shaped gate was designed as an insert structure, facilitating EDM processing, timely cleaning in case of gate breakage. To improve cooling efficiency, a surrounding water channel design was adopted for cavity, and a well-type water channel design was adopted for core. Due to thinner wall of lower half of product, a combination of round ejector pins and flat ejector pins was used to solve problem of difficult product ejection. Simultaneously, gap between ejector pins and ejector pin holes allows for air escape, resolving issue of air trapping in narrow areas of core. Moldex3D was used to simulate and analyze gating and cooling system designs, verifying their rationality, obtaining preliminary injection molding process parameters. For process optimization, multi-objective injection molding process optimization was performed using standard deviation of volume shrinkage and maximum warpage as responses. An orthogonal experimental design combined with response surface methodology was employed. Optimal multi-objective process parameters were found at a melt temperature of 228 ℃, holding pressure of 237.5 MPa, and holding time of 5.88 s. Further simulation using Moldex3D software yielded a standard deviation of 0.937% for volume shrinkage (a reduction of 45.39%) and a maximum warpage of 0.181 mm (a reduction of 21.30%), indicating feasibility of optimization scheme.
High-voltage connector socket housings (hereinafter referred to as socket housings) are main components used for internal circuit connections in automobiles. Socket housings are mostly manufactured using injection molding. Product requires high dimensional accuracy, with warpage less than 0.2 mm, no obvious dents or flash on the surface, uniform overall shrinkage, and no significant deformation.
Regarding mold design, this paper intends to use Moldex 3D to analyze gating system, cooling system, other structures to verify mold structure scheme and obtain molding process parameters, providing initial data for subsequent optimization experiments. Many researchers have previously used mold flow analysis software to analyze mold structures, providing references for mold design, making structural designs more reasonable, improving production efficiency and product quality. For example, Cai Kaiwu et al. used Moldflow or Moldex3D for mold flow analysis based on structural characteristics of product, meeting requirements for injection molding and surface accuracy of plastic parts. Zhou Shurong used CAE analysis to obtain dimensions of horn gate and point gate based on three-plate mold for precision molding of back shell of mobile phone battery, and completed mold design. In terms of optimizing injection molding process parameters, orthogonal test + response surface test method was used to optimize warpage and shrinkage rate in multiple objectives to improve quality of product. Related studies include Li Li et al. using Moldflow mold flow analysis, taking warpage deformation of product as response target, to obtain ideal molding process parameters and complete mold design. Qiu Tong et al. conducted simulation research on lens injection compression molding process based on Moldflow, taking volume shrinkage rate and warpage deformation as optimization targets, combined orthogonal test method and comprehensive balance method to obtain optimal process parameter combination, and compared it with response surface method process optimization scheme to determine best test method.

Socket shell product is a two-layer structure, as shown in Figure 1. External dimensions are 45 mm * 50 mm * 35 mm. Structure is an asymmetrical upper and lower layer structure, with the thickest part of wall being 5 mm and the thinnest part being 1 mm. Draft angle is set to 1.2°.

 

Fig. 1 Product drawing

Mold cavity layout is a four-cavity mold. To simplify mold structure, a horn-shaped submarine gate is selected, with single gate injection. Runner and gate layout are shown in Figure 2. Horn-shaped submarine gate is used. In addition to having advantages of good pressure holding effect, good melt flow, and good surface quality of point gates, it can also simplify mold structure to a two-plate mold. However, since horn-shaped gate is difficult to process, it is designed as an insert structure, which is convenient for EDM processing, and easy to clean if gate breaks inside mold. A pull rod is set to facilitate ejection of gate after mold opening. Specific structure is shown in Figure 3.

 

Fig. 2 Gating system

 

Fig. 3 Horn gate
(a) Open status (b) Assembly status

To achieve better cooling, based on structure of cavity and core, a surrounding water channel design is adopted for fixed mold, and a water well-type cooling water channel is adopted for moving mold, as shown in Fig. 4.

 

Fig. 4 Cooling water circuit layout

Due to anti-misalignment snap-fit design on the side of product, mold structure needs to be designed with a side core pulling structure, as shown in Fig. 5. Because core pulling distance is small, an external moving mold core pulling mechanism of "slanted guide pillar + slider" is adopted. Slider is driven by two Φ8 mm slanted guide pillars with an inclination angle of 20° and a pulling distance of 7 mm. To avoid interference between wedge block and slider during mold closing, wedge angle is 22°.

 

Figure 5 Side core pulling structure
Lower layer of product has a wall thickness of only 1 mm and a height of 23 mm, which easily causes demolding difficulties. Therefore, four Φ3 mm round ejector pins are set at parting surface, and 1 mm * 2 mm flat ejector pins are set at thin-walled part of bottom of product. Two ejector pins are arranged on each of the other three sides, except for slider side, as shown in Figure 6. Gap between flat ejector pins and ejector pin holes can also serve as an air venting structure.

 

Figure 6 Ejection system

Mold base uses LKM CI type standard mold base CI-2535-A60-B70-C80. Molding parts adopt an insert structure, and ejection method is ejector pin ejection. The overall mold design is shown in Figure 7.

 

Fig. 7 Mold assembly drawing
1—Spring; 2—Core; 3—Slide; 4—Cooling water circuit of the core; 5—B plate; 6—Angle pin; 7—Cooling water circuit of the cavity; 8—A plate; 9—Core insert; 10—Cavity; 11—Cavity insert A; 12—Cavity insert B; 13—Gate; 14—Bottom clamp plate; 15—Support plate; 16—Ejector plate; 17—Flat ejector pin; 18—Ejector pin; 19—Sprue puller; 20—Sprue brush; 21—Locating ring

Product material is an acrylic rubber body and acrylonitrile, styrene graft copolymer (ASA) + polycarbonate (PC) from Chi Mei Corporation of Taiwan, with grade WONDERLOY PC-6220. This material has good fluidity, high surface smoothness after molding, and small deformation. Recommended initial injection process parameters are shown in Table 1.

Table 1 Injection process parameters
Design schemes of plastic parts, gating system and cooling system were imported into Moldex 3D software in STP format and three-dimensional mesh was divided as shown in Figure 8. As shown in Figure 8a, filling time was 1.8 s, filling time of four cavities at equidistant positions from gate was 1.494, 1.478, 1.484 and 1.480 s, respectively. Times were similar, indicating that size and layout of runner and gate were reasonable. In order to ensure cooling effect of mold, temperature difference between inlet and outlet of coolant in circuit should not exceed 5 ℃ under normal circumstances, and temperature difference should not exceed 2 ℃ for plastic parts with high precision requirements. In this analysis, maximum temperature difference between inlet and outlet of coolant in fixed mold water circuit was 0.727 ℃, maximum temperature difference between inlet and outlet of coolant in moving mold water circuit was 0.282 ℃, which met mold design requirements. Analysis results show that standard deviation of volume shrinkage is 1.716% and maximum warpage is 0.2605 mm.

 

Fig. 8 Simulation analysis of gating system and cooling system

This experiment takes standard deviation of volume shrinkage and maximum warpage as objectives. The smaller standard deviation of volume shrinkage, the closer volume shrinkage is to average, and the smaller maximum warpage, the more accurate dimensions.

Based on analysis results of initial injection molding process parameters, and in conjunction with previous literature research, influencing factors of this experiment are determined to be 6: mold temperature (A), melt temperature (B), injection pressure (C), filling time (D), holding pressure (E), and holding time (F). Taking into account economic and time costs, orthogonal experiment is used for factor screening. Based on analysis results from design phase, values of each factor in experiment are shown in Table 2. A 6-factor, 5-level design was used, with standard deviation of volume shrinkage Y1 and warpage Y2 as response values, each with a weight of 50%.

 

Table 2 Six-factor, five-level values table
Orthogonal experiments were conducted on data in Table 2 using Moldex3D software, with 25 trials performed. Indicators examined were standard deviation of volume shrinkage Y1 (%) and maximum warpage Y2 (mm). Experimental results are shown in Table 3. Range values of Y1 and Y2 obtained through analysis are shown in Tables 4 and 5, respectively. Significance ranking of factors affecting volume shrinkage Y1 is B > E > F > D > A > C, and significance ranking of factors affecting warpage Y2 is E > B > F > A > D > C.

 

Table 3 Orthogonal test results

 

Table 4 Shrinkage rate (standard deviation) range table

 

Table 5 Table of maximum warpage range values
Influence trends of each factor on Y1 and Y2 are shown in Figure 9. Using comprehensive balance method, influence trends of B and D on Y1 and Y2 are same, with B1D4 being optimal level. Influence trends of A, C, E, F on Y1 and Y2 are opposite, and their median values are taken, i.e., A3C3E3F3. Therefore, optimal combination for orthogonal experiment is A3B1C3D4E3F3. Inputting these parameters into MOLDEX 3D for simulation yielded Y1 of 1.104% and Y2 of 0.240 mm. Compared to initial analysis results, standard deviation of volume shrinkage rate decreased by 35.66%, and maximum warpage decreased by 7.87%.

 

Fig. 9 Main effect diagram

Based on experimental results of orthogonal experiment, warpage still does not meet requirement of less than 0.2 mm, so response surface experiment is used for further optimization. Top three factors in terms of significance are selected as response surface experiment factors: melt temperature (B), holding pressure (E), and holding time (F), indicators are standard deviation of volume shrinkage rate Y1 and maximum warpage Y2.
Response surface experiment design is an experimental condition optimization method, which is suitable for solving nonlinear multivariate problems. This method uses multivariate quadratic regression equation as a tool to express each influencing factor and response target through a function, and makes reasonable values for each factor so that response target value reaches optimal. For second-order polynomial response surface, expression between process parameters and corresponding response values at different levels is shown in equation (1).

 

In equation (1): y' is target value of response; is regression coefficient of response surface; x is design variable; i, j are variable subscripts; n is number of design variables; ε is statistical error.
Common experimental design methods for response surfaces include Box-Behnken (BBD) method and Central Composite (CCD) method. Based on results of orthogonal experiments, three factors were selected as experimental factors for response surface: melt temperature (B), holding pressure (E), and holding time (F). Indicators were standard deviation of volume shrinkage Y1 and maximum warpage Y2. Since BBD method does not have axial points and requires fewer experiments when number of factor levels is same, BBD method was chosen to optimize injection molding process parameters. Values of response surface design are shown in Table 5, and experimental results of response surface are shown in Table 6.

 

Table 6 Response surface experimental design and results

 

Table 5 Response surface design value table
Variance analysis was performed on results, and residual probability plots were observed. As shown in Figures 10a and 10b, most points are close to a straight line, indicating that residuals conform to a normal distribution and model is valid. R-sq and R-sq (adjusted) of standard deviation of volume shrinkage rate model Y1 are 99.93% and 98.73%, respectively, R-sq and R-sq (adjusted) of maximum warpage amount Y2 are 99.99% and 98.95%, respectively, indicating that model is not overfitted. Standard deviation of volume shrinkage rate and maximum warpage amount were optimized by response optimizer. Optimization objective is to minimize value. Optimization prediction results are shown in Figure 11.

 

Figure 10 Residual normal probability plot

 

Fig. 11 Optimization prediction diagram
Optimization results for Y1 and Y2 show a good agreement of 0.8349, indicating good numerical matching. Optimization results are as follows: at a melt temperature of 228 ℃, a holding pressure of 238.005 1 MPa, and a holding time of 5.88 s, standard deviation of volume shrinkage rate is 0.934 8%, and maximum warpage is 0.181 5 mm.

Numerical values of optimization scheme were input into Moldex3D for simulation. Standard deviation of volume shrinkage rate was 0.937%, and maximum warpage was 0.181 mm. These results are close to predicted optimization results, indicating that optimization model is feasible.

(1) Mold flow analysis of product was performed using Moldex3D software. A horn-shaped submarine gate was adopted for gate. Cooling scheme is a surrounding water channel with a diameter of 8 mm for fixed mold and a well-type water channel with a diameter of 16 mm for moving mold. Mold structure was designed based on flow channel and cooling scheme, and parameter ranges of each factor in subsequent optimization experiments were determined according to mold flow analysis results.
(2) Multi-objective optimization experiment adopted a combination of orthogonal experiment and response surface methodology. Using standard deviation of volume shrinkage and maximum warpage as responses, a 6-factor, 5-level orthogonal experiment was used for factor selection, identifying three main factors: plastic melt temperature, holding pressure, and holding time. Response surface methodology used BBD method. Fifteen sets of experimental warpage deformation values were obtained through Moldex3D analysis. Optimized predicted values using response optimizer were: at a plastic melt temperature of 228 ℃, a holding pressure of 237.5 MPa, and a holding time of 5.88 s, standard deviation of volume shrinkage was 0.9348%, and maximum warpage was 0.1815 mm.
(3) Predicted process parameters were input into Moldex3D for simulation analysis. Under this process condition, standard deviation of volume shrinkage was 0.937% and maximum warpage was 0.181 mm, which was close to optimized prediction results. Compared with initial simulation values, standard deviation of volume shrinkage was reduced by 45.39% and maximum warpage was reduced by 30.52%. Compared with orthogonal test results, standard deviation of volume shrinkage was reduced by 15.33% and maximum warpage was reduced by 24.5%, indicating that optimized model is feasible.