【摘要】 利用SHPB技术和自编的BP神经网络程序 ,以尼龙为代表性研究对象 ,对高聚物在高应变率下的本构模型进行了辨识 .研究表明 :在应变小于 7%范围内 ,以应变与应变率作为输入 ,以应力作为输出 ,BP神经网络能满意地辨识高聚物的动态本构模型 ;而当应变大于7%时 ,除应变和应变率外可取时间作为表征损伤演化的反函数共同作为输入 ,以应力作为输出 ,能获得满意的效果 .更多还原

【Abstract】 Using SHPB technique and a BP neural network program, the constitutive model of polymers at high strain rates is identified. The results show that the constitutive model of polymers can be determined by BP neural network by taking strain and strain rate as input and stress as output, in case of strain less than 7%. if strain is larger than 7%, an additional input parameter-time should be added, which represents the internal damage evolution accompanied with the deformation process of material.更多还原