高应变率下高聚物本构模型的BP神经网络辨识  

Constitutive Model Identification of Polymers at High Strain Rates Based on BP Neural Network

作　　者：徐明乔[1] 施绍裘[1] 王礼立[1] 

机构地区：[1]宁波大学工学院,浙江宁波315010

出　　处：《宁波大学学报（理工版）》2001年第4期38-43,共6页Journal of Ningbo University：Natural Science and Engineering Edition

基　　金：国家自然科学基金 (10 0 310 10 )资助项目

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 .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.

关 键 词：BP神经网络 高聚物 本构模型 高应变率 损伤 辨识 材料性能 

分 类 号：TB324[一般工业技术—材料科学与工程] TP183[自动化与计算机技术—控制理论与控制工程]