基于神经网络的PP/CaCO_3复合材料的力学性能预测

Prediction Model of Mechanical Properties for PP/CaCO_3 Composite Material Based on Artificial Neural Network

研究利用BP网络建立了对PP/CaCO3复合材料性能进行预测的模型,并在此基础上建立了PP/CaCO3复合材料拉伸强度、冲击强度和复合态结构之间的关系。结果表明,该模型有着较好可信度,与实验现象也有着较好的符合;利用该模型证明了分散相粒子粒间距、EPDM层厚度是影响其材料冲击强度,并决定其脆-韧转变的最重要因素,并证明了EPDM层厚度的临界值大约在0.05μm左右。

A model was established for predicting the mechanical properties of PP/CaCO3 composite materials using neural network. The relationship between tensile and impact strengths and the structure of PP/CaCO3 composites was built up. The calculated results match the experimental data well,and thus the model is reliable. The model reveals that the distance between the dispersed particles and the thickness of the EPDM layer are the most important factors affecting the brittle-tough transition,and the critical thickness of EPDM layer is about 0.05 μm.