基于CAE和T-S神经网络的储物箱注塑工艺优化

Optimization Design of Injection Molding Process Based on CAE and T-S Neural Network

针对塑件翘曲变形过大而导致塑件注塑失效的问题,通过运用CAE分析得出了影响翘曲变形过大的主要因素为收缩不均;采用正交试验方法获得了初步优化后参数,为Tθ(230℃)Ts(65℃)PI(70 MPa)ti(3.5 s)Ph1(60 MPa)th1(10 s)Ph1(75 MPa)th1(12 s)tc(6 s),对应的翘曲值为5.53 mm。在此基础上,再次运用GSO算法对改进的T-S模糊神经网络进行预测,得到了进一步优化的翘曲值,为3.49 mm,对应优化后的工艺参数为Tθ(230℃)Ts(68℃)PI(70 MPa)ti(4 s)Ph1(65 MPa)th1(8 s)Ph1(75 MPa)th1(14 s)tc(4 s),将优化后的工艺参数应用于实际注塑后,塑件的实效问题得到了有效解决,具有较强的实践参考价值。

In view of the problem of injection failure caused by large warping deformation of plastic parts,CAE analysis was used to find out that the main influencing factor of large warping deformation was uneven shrinkage.Used the orthogonal experiment method for preliminary improve the optimization of the parameters for Tθ(230℃)Ts(65℃)PI(70 MPa)ti(3.5 s)Ph1(60 MPa)th1(10 s)Ph1(75 MPa)th1(12 s)tc(6 s),the corresponding buckling value was 5.53 mm,base on the optimization of the paraneter,again used the GSO algorithm improved T-S fuzzy neural network to forecast,obtained the further optimization of warp value 3.49 mm,corresponding to the optimized process parameters for Tθ(230℃)Ts(68℃)PI(70 MPa)ti(4 s)Ph1(65 MPa)th1(8 s)Ph1(75 MPa)th1(14 s)tc(4 s),was applied to the actual injection,plastic parts of practical problem was resolved,it had the strong practical reference value.