No abstract

Herein, we investigate the structure-property relationships of soft magnetorheological elastomers (MREs) filled with remanently magnetizable particles. The study is motivated from experimental results which indicate a large difference between the magnetization loops of soft MREs filled with NdFeB particles and the loops of such particles embedded in a comparatively stiff matrix, e.g. an epoxy resin. We present a microscale model for MREs based on a general continuum formulation of the magnetomechanical boundary value problem which is valid for finite strains. In particular, we develop an energetically consistent constitutive model for the hysteretic magnetization behavior of the magnetically hard particles. The microstructure is discretized and the problem is solved numerically in terms of a coupled nonlinear finite element approach. Since the local magnetic and mechanical fields are resolved explicitly inside the heterogeneous microstructure of the MRE, our model also accounts for interactions of particles close to each other. In order to connect the microscopic fields to effective macroscopic quantities of the MRE, a suitable computational homogenization scheme is used. Based on this modeling approach, it is demonstrated that the observable macroscopic behavior of the considered MREs results from the rotation of the embedded particles. Furthermore, the performed numerical simulations indicate that the reversion of the sample’s magnetization occurs due to a combination of particle rotations and internal domain conversion processes. All of our simulation results obtained for such materials are in a good qualitative agreement with the experiments.

No abstract

Herein, an artificial neural network (ANN)-based approach for the efficient automated modeling and simulation of isotropic hyperelastic solids is presented. Starting from a large data set comprising deformations and corresponding stresses, a simple, physically based reduction of the problem’s dimensionality is performed in a data processing step. More specifically, three deformation type invariants serve as the input instead of the deformation tensor itself. In the same way, three corresponding stress coefficients replace the stress tensor in the output layer. These initially unknown values are calculated from a linear least square optimization problem for each data tuple. Using the reduced data set, an ANN-based constitutive model is trained by using standard machine learning methods. Furthermore, in order to ensure thermodynamic consistency, the previously trained network is modified by constructing a pseudo-potential within an integration step and a subsequent derivation which leads to a further ANN-based model. In the second part of this work, the proposed method is exemplarily used for the description of a highly nonlinear Ogden type material. Thereby, the necessary data set is collected from virtual experiments of discs with holes in pure plane stress modes, where influences of different loading types and specimen geometries on the resulting data sets are investigated. Afterwards, the collected data are used for the ANN training within the reduced data space, whereby an excellent approximation quality could be achieved with only one hidden layer comprising a low number of neurons. Finally, the application of the trained constitutive ANN for the simulation of two three-dimensional samples is shown. Thereby, a rather high accuracy could be achieved, although the occurring stresses are fully three-dimensional whereas the training data are taken from pure two-dimensional plane stress states.

No abstract