Virtual Field Characterization for Ratcheting Effect Under Cyclic Loading

Abstract

Ratcheting is an important mechanical behavior of metals and alloys, which is caused by the repeated accumulations of tensile and compressive strain in circle load. However, the current characterization methods of ratcheting effect are mostly based on standardized testing and uniform data, and more comprehensive field measurement data cannot be used. This paper focuses on how to make full use of field measurement data to characterize ratcheting effect and identify the corresponding kinematic constitutive. A nonlinear virtual field method that can invert the parameters of Chaboche constitutive from strain field data is proposed. And a return mapping strategy driven by iteration of internal variables is used to reconstruct the stress field, which ensures the convergence speed and global convergence in the black box search of the nonlinear virtual field method. By using the finite-element model to generate the strain field data, the numerical experiment shows that the ratchet path identified by the nonlinear virtual field algorithm is basically consistent with the prior ratchet path generated by the finite-element simulation. The adaptability of the algorithm to data density and noise amplitude was also verified: under lower data noise interference, more strain field training data makes the inversion results more accurate; but in the case of high sound amplitude, it is necessary to reduce the data size to obtain accurate fitting results of the ratchet path. By training the measured field data from 3D-digital image correlation, it is shown that the algorithm can also run effectively under the complex working conditions of non-uniform deformation.