生物软组织黏弹性参数反演方法研究  

Inverse Estimation for Viscoelastic Parameters of Biological Soft Tissue

作　　者：程兴凯[1] 陈科[1] 林江莉[1] 

机构地区：[1]四川大学材料科学与工程学院,成都610065

出　　处：《实验科学与技术》2012年第5期4-6,88,共4页Experiment Science and Technology

基　　金：国家自然科学基金项目资助(30970781/H1805);教育部博士点基金项目资助(20100181110002)

摘　　要：针对黏弹性参数反演中的模型选择问题,提出应用黏弹性材料的应力松弛曲线作为待反演的参数,不涉及任何黏弹性本构模型,采用有限元仿真与人工神经网络相结合的方法:通过有限元仿真得到不同应力松弛曲线下的剪切波传播速度,然后将仿真得到的数据作为训练样本输入到人工神经网络中进行训练,建立起生物组织黏弹性参数和剪切波传播速度的神经网络输出模型,有效解决了模型选择问题。最后通过粒子群优化算法反演得到组织的黏弹性参数。As for the mechanical model choice problem of inverse estimation for viscoelastic parameters,in our research,the stress relaxation curve are regarded as inverse estimation parameters which aren＇ t related to any viscoelastic constitutive model,then the combination of finite element method and artificial neural network are applied to estimate viscoelastic parameters inversely.Multiple sets of shear wave speeds of relaxation curve parameters are generated from FEM simulation experiments.Then the data obtained from FEM simulation experiments are input into the artificial neural network for training to establish the neural network model of viscoelastic parameters and shear wave speeds,which has solved the mechanical model choice problem.Finally,particle swarm optimization algorithm is applied to estimate viscoelastic parameters of biological tissues.

关 键 词：黏弹性 有限元法 反演 人工神经网络 

分 类 号：O242.21[理学—计算数学]