Effect of mesh size on calculation of strain non-uniformity index in drawn sheet metal parts

The spatial variation of the true major, minor and thickness strains is the combined outcome of a number of material, tool, geometric and deformation variables (many of which cannot be measured) and their complex interactions over various levels. When one calculates the strain nonuniformity index from an FEM simulation, the mesh size can be expected to influence the value of the strain nonuniformity index leading to a wrong prediction of failure. This is similar to the effect of grid circle size and circle spacing in experimentally measured strain distribution. The present paper attempts to bring out a minimum mesh size for the critical strain nonuniformity index to remain unaffected by the meshing effects. This will enable the user to set an adequate level of mesh refinement during adaptive meshing without an excessive increase in the computational cost. Comparatively, the 'constraint factor', which denotes the ratio of the difference between the minor strains at the peak and the pole to that between the major strains at these two locations, is likely to be much less affected by the mesh size. strain nonuniformity index reaching a critical value is accompanied by the constraint factor moving towards zero. (C) 2014 Published by Elsevier Ltd.