基于神经网络和遗传算法的薄壳塑件注塑工艺优化
|
摘要
散热器外壳是电子产品散热器的主要零件之一,由于壁薄,在注塑成型中经常出现壁厚不均、翘曲变形和熔接痕等缺陷。针对该问题,以熔体温度、模具温度、冷却时间、注射压力、注射时间、保压压力和保压时间7个工艺参数为输入量,注塑件的翘曲量作为输出量,建立RBF神经网络模型;利用均匀试验所得的数据作为样本对神经网络进行训练和测试,得到注塑工艺参数与塑件翘曲变形量之间的非线性映射关系。结合遗传算法对工艺参数进行优化,获得最佳的工艺参数为:熔体温度234. 4℃、模具温度31. 5℃、冷却时间23. 8 s、注射压力128. 3 MPa、注射时间4. 7 s、保压压力93. 0 MPa、保压时间14. 1 s,获得预测的最小翘曲变形值为0. 331 875 mm,并使用优化后的工艺参数进行试验。试验结果表明,优化后产品的最大翘曲变形量降低至0. 318 9 mm,与优化前均匀试验所得的0. 378 1 mm相比,得到了明显的改善,降低了15. 7%。
Radiator shell was one of the main parts of the radiator of electronic products. Due to the thin wall,defects such as uneven wall thickness,warpage and weld lines often occurred during injection molding. In order to solve the problem,the RBF neural network which took melt temperature,mold temperature,cooling time,injection pressure,injection time,holding pressure,and pressure holding time as the input parameters and the warpage of the injection molded part as output parameters was established. The neural network was trained and tested to obtain the nonlinear mapping relationship between the injection molding process parameters and the warpage of the plastic parts by using the data obtained from the uniform test as samples. Combining genetic algorithm to optimize the process parameters,the best process parameters were: melt temperature234. 4 ℃,mold temperature 31. 5 ℃,cooling time 23. 8 s,injection pressure 128. 3 MPa,injection time 4. 7 s,holding pressure 93. 0 MPa,pressure holding Time 14. 1 s. The predicted minimum warpage value was 0. 331 875 mm. The optimized process parameters were used for testing. The test results showed that the maximum warpage value of the optimized product was reduced to 0. 318 9 mm,which was significantly improved compared with 0. 378 1 mm obtained by the uniform test before optimization,which was reduced by 15. 7%.
Radiator shell was one of the main parts of the radiator of electronic products. Due to the thin wall,defects such as uneven wall thickness,warpage and weld lines often occurred during injection molding. In order to solve the problem,the RBF neural network which took melt temperature,mold temperature,cooling time,injection pressure,injection time,holding pressure,and pressure holding time as the input parameters and the warpage of the injection molded part as output parameters was established. The neural network was trained and tested to obtain the nonlinear mapping relationship between the injection molding process parameters and the warpage of the plastic parts by using the data obtained from the uniform test as samples. Combining genetic algorithm to optimize the process parameters,the best process parameters were: melt temperature234. 4 ℃,mold temperature 31. 5 ℃,cooling time 23. 8 s,injection pressure 128. 3 MPa,injection time 4. 7 s,holding pressure 93. 0 MPa,pressure holding Time 14. 1 s. The predicted minimum warpage value was 0. 331 875 mm. The optimized process parameters were used for testing. The test results showed that the maximum warpage value of the optimized product was reduced to 0. 318 9 mm,which was significantly improved compared with 0. 378 1 mm obtained by the uniform test before optimization,which was reduced by 15. 7%.
引文
[1]修辉平,徐敏,刘晓红.基于Moldflow和BP神经网络的薄壳注塑件预测[J].塑料科技,2017,45(5):69-72.
[2]吴光辉.塑料手机外壳注塑成型数值模拟及参数优化[J].塑料,2017,46(4):98-101.
[3]余蔚荔.薄壁注塑制品工艺参数优化研究[J].塑料工业,2016,44(7):64-66.
[4] WANG C,HUANG M,SHEN C,et al. Warpage prediction of the injection-molded strip-like plastic parts[J]. Chinese Journal of Chemical Engineering,2016,24(5):665-670.
[5] KITAYAMA S,YAMAZAKI Y,TAKANO M,et al. Numerical and experimental investigation of process parameters optimization in plastic injection molding using multi-criteria decision making[J]. Simulation Modelling Practice and Theory,2018,85:95-105.
[6] HEIDARI B S,DAVACHI S M,MOGHADDAM A H,et al.Optimization simulated injection molding process for ultrahigh molecular weight polyethylene nanocomposite hip liner using response surface methodology and simulation of mechanical behavior[J].Journal of the Mechanical Behavior of Biomedical Materials,2018,81:95-105.
[7] LUO L,YAO Y,GAO F,et al. Mixed-effects Gaussian process modeling approach with applicationin injection molding processes[J]. Journal of Process Control,2018,62:37-43.
[8] OLIAEI E,HEIDARI B S,DAVACHI S M,et al. Warpage and shrinkage optimization of injection-molded plastic spoon parts for biodegradable polymers using Taguchi, ANOVA and artificial neural network methods[J]. Journal of Materials Science&Technology,2016,32(8):710-720.
[9] KITAYAMA S,TAMADA K,TAKANO M,et al. Numerical optimization of process parameters in plastic injection molding for minimizing weldlines and clamping force using conformal cooling channel[J].Journal of Manufacturing Processes,2018,32:782-790.
[10] GINGHTONG T,NAKPATHOMKUN N,PECHYEN C. Effect of injection parameters on mechanical and physical properties of super ultra-thin wall propylene packaging by Taguchi method[J]. Results in Physics,2018,9:987-995.
[11]顾嘉琦,胡宇宣,郭永环,等.基于均匀试验设计的温控器上盖注塑模具工艺参数优化[J].塑料科技,2017,45(8):67-70.
[12]梁德坚,邓其贵,万选明.基于GRNN神经网络的汽车塑件工艺优化设计[J].塑料,2018,47(3):98-102.
[13]孔轶艳,黄力,陆大同.基于CAE和RBF神经网络的注塑工艺优化分析[J].塑料,2017,46(3):121-125.
[14] LAMBA R,MANIKANDAN S,KAUSHIK S C,et al. Thermodynamic modelling and performance optimization of trapezoidal thermoelectric cooler using genetic algorithm[J]. Thermal Science and Engineering Progress,2018,6:236-250.
[15]范克健,刘苏俊,马晓东,等.基于神经网络的车灯副反射器注塑工艺参数优化[J].塑料科技,2017,45(4):89-92.
[16]夏江梅,张黎,文琍.基于代理模型和遗传算法的顺序注塑成型工艺优化[J].塑料,2016,45(4):116-118.
[2]吴光辉.塑料手机外壳注塑成型数值模拟及参数优化[J].塑料,2017,46(4):98-101.
[3]余蔚荔.薄壁注塑制品工艺参数优化研究[J].塑料工业,2016,44(7):64-66.
[4] WANG C,HUANG M,SHEN C,et al. Warpage prediction of the injection-molded strip-like plastic parts[J]. Chinese Journal of Chemical Engineering,2016,24(5):665-670.
[5] KITAYAMA S,YAMAZAKI Y,TAKANO M,et al. Numerical and experimental investigation of process parameters optimization in plastic injection molding using multi-criteria decision making[J]. Simulation Modelling Practice and Theory,2018,85:95-105.
[6] HEIDARI B S,DAVACHI S M,MOGHADDAM A H,et al.Optimization simulated injection molding process for ultrahigh molecular weight polyethylene nanocomposite hip liner using response surface methodology and simulation of mechanical behavior[J].Journal of the Mechanical Behavior of Biomedical Materials,2018,81:95-105.
[7] LUO L,YAO Y,GAO F,et al. Mixed-effects Gaussian process modeling approach with applicationin injection molding processes[J]. Journal of Process Control,2018,62:37-43.
[8] OLIAEI E,HEIDARI B S,DAVACHI S M,et al. Warpage and shrinkage optimization of injection-molded plastic spoon parts for biodegradable polymers using Taguchi, ANOVA and artificial neural network methods[J]. Journal of Materials Science&Technology,2016,32(8):710-720.
[9] KITAYAMA S,TAMADA K,TAKANO M,et al. Numerical optimization of process parameters in plastic injection molding for minimizing weldlines and clamping force using conformal cooling channel[J].Journal of Manufacturing Processes,2018,32:782-790.
[10] GINGHTONG T,NAKPATHOMKUN N,PECHYEN C. Effect of injection parameters on mechanical and physical properties of super ultra-thin wall propylene packaging by Taguchi method[J]. Results in Physics,2018,9:987-995.
[11]顾嘉琦,胡宇宣,郭永环,等.基于均匀试验设计的温控器上盖注塑模具工艺参数优化[J].塑料科技,2017,45(8):67-70.
[12]梁德坚,邓其贵,万选明.基于GRNN神经网络的汽车塑件工艺优化设计[J].塑料,2018,47(3):98-102.
[13]孔轶艳,黄力,陆大同.基于CAE和RBF神经网络的注塑工艺优化分析[J].塑料,2017,46(3):121-125.
[14] LAMBA R,MANIKANDAN S,KAUSHIK S C,et al. Thermodynamic modelling and performance optimization of trapezoidal thermoelectric cooler using genetic algorithm[J]. Thermal Science and Engineering Progress,2018,6:236-250.
[15]范克健,刘苏俊,马晓东,等.基于神经网络的车灯副反射器注塑工艺参数优化[J].塑料科技,2017,45(4):89-92.
[16]夏江梅,张黎,文琍.基于代理模型和遗传算法的顺序注塑成型工艺优化[J].塑料,2016,45(4):116-118.