基于贝叶斯后验概率的聚氨酯泡沫材料静力本构模型选择  

SELECTION OF STATIC CONSTITUTIVE MODELS FOR POLYURETHANE BASED ON BAYESIAN POSTERIOR PROBABILITY

作　　者：张元章[1] 牛伟[1] 刘信恩[1] 

机构地区：[1]中国工程物理研究院总体工程研究所,绵阳621900

出　　处：《固体力学学报》2013年第S1期242-246,共5页Chinese Journal of Solid Mechanics

基　　金：国家自然科学基金委员会-中国工程物理研究院联合基金(NSAF)项目(10876100);中国工程物理研究院发展基金项目(2012B0302042)资助

摘　　要：以聚氨酯试样标准单轴拉压实验数据为校准数据,校准线弹性和五参数Mooney-Rivilin、二阶OgdenHyperFoam本构模型的参数.以聚氨酯试样标准三点弯曲实验数据为确认数据,对校准得到的Mooney-Rivilin、Ogden-HyperFoam、线弹性模型进行确认,归一化后得到三种模型的贝叶斯后验概率依次为:0.23、0.75、0.02.在不考虑参数相关性时得到的随机有限元模拟结果方差很大,尚不能供工程实践参考,在考虑参数相关性之后,模拟计算结果的方差明显减小.Linear elastic and Mooney-Rivilin and Ogden-HyperFoam constitutive models were calibrated for polyurethane elastomer with the experimental data on standard tensile and compress experiments.The three models were validated with the experimental data on standard three point bending experiments.The Bayesian posterior probabilities for the Mooney-Rivilin,Ogden-HyperFoam and linear elastic constitutive models are 0.23,0.75,0.02,respectively.The variances of simulation results without parameter correlation considered is too significant for engineering practice,and with correlation considered,the variances are considerably smaller.

关 键 词：贝叶斯后验概率 聚氨酯 本构模型 

分 类 号：TQ328.3[化学工程—合成树脂塑料工业]