Generalized Parameter Identification for Finite Viscoelasticity
Generalized Parameter Identification for Finite Viscoelasticity
复制标题
有限粘弹性的广义参数辨识
DOI：
10.1016/j.cma.2007.03.010
复制
发表时间：
2007
影响因子：
7.2
通讯作者：
P. Steinmann
中科院分区：
文献类型：
--
作者：
Bernd Kleuter;A. Menzel;P. Steinmann
在工程应用中，弹性体和其他橡胶类材料经常同时承受短期和长期载荷。在建立粘弹性介质在不同时间尺度上的通用模拟工具时，只有考虑到适当数量的实验数据，才能确定合适的材料模型及其相应的材料参数。在这项工作中，我们提出了一种大应变粘弹性材料参数识别的算法，其中考虑了多个实验数据。基于这种方法，可以缩短长时间试验的加载时间间隔，并且参数识别过程现在参考短时间和长时间载荷下的试验数据，而不会因为不同的时间尺度而分离参数。所采用的粘弹性材料定律是基于非线性演化定律，并且在远离热力学平衡的情况下是有效的。在给定(测量)的力荷载作用下，通过比较实验和有限元模拟得到的非均匀位移场，通过最小化最小二乘泛函来实现识别。在此优化过程中，所有材料参数通过基于梯度的方法同时识别，并对该方法进行了半解析灵敏度分析。一个有代表性的数值例子是基于非泡沫型聚氨酯的短时间和长时间测试的测量数据。作为优势，所开发的识别方案仅呈现一组单一的材料参数。
Elastomeric and other rubber-like materials are often simultaneously exposed to short- and long-time loads within engineering applications. When aiming at establishing a general simulation tool for viscoelastic media over these different time scales, a suitable material model and its corresponding material parameters can only be determined if an appropriate number of experimental data is taken into account. In this work we present an algorithm for the identification of material parameters for large strain viscoelasticity in which data of multiple experiments are considered. Based on this method the experimental loading intervals for long-time experiments can be shortened in time and the parameter identification procedure is now referred to experimental data of tests under short- and long-time loads without separating the parameters due to these different time scales. The employed viscoelastic material law is based on a nonlinear evolution law and valid far from thermodynamic equilibrium. The identification is carried out by minimizing a least squares functional comparing inhomogeneous displacement fields from experiments and FEM simulations at given (measured) force loads. Within this optimization procedure all material parameters are identified simultaneously by means of a gradient based method for which a semi-analytical sensitivity analysis is calculated. A representative numerical example is referred to measured data based on short-time and long-time tests of a non-cellular polyurethane. As an advantage, the developed identification scheme renders solely one single set of material parameters.