Article ID Journal Published Year Pages File Type
783808 International Journal of Mechanical Sciences 2013 9 Pages PDF
Abstract

The finite element simulation is an effective way for the springback study of thin-walled rectangular H96 tube in the rotary-draw bending process. To improve the accuracy and the computational efficiency of the finite element model, the key modeling techniques, such as the introduction of loading boundary conditions for the clamp die and the pressure die, the reasonable modification of the flexible cores, etc., are resolved. Based on the resolved key problems, two finite element models of the whole-tube and the half-tube are established respectively for the springback perdition of rectangular H96 tube. Comparisons of springback perdition results between the experimental values and simulative ones based on two models show that, the half-tube model has greater advantage on the computational efficiency, while its accuracy on the springback angle and springback radius prediction is still in the acceptable range, so the half-tube model is recommended. A further improvement on the accuracy and computational efficiency of the half-tube model is studied. It is obtained that: (1) the improvement for the springback angle prediction can be 9.82% when the Bauschinger effect is considered. (2) The influence of material constitutive model on the springback prediction precision is greater than that caused by simplifying FE model or using different mass scaling factors. (3) In the precondition of meeting the accuracy requirement, a high computational efficiency can be achieved when the mass scaling factor 4800 and the element size 1×1 mm2 are used.

► Modeling techniques about accuracy/efficiency of springback prediction are resolved. ► Springback models of whole-tube and half-tube are established, verified and compared. ► Half-tube model has better computational efficiency but lower accuracy. ► Consideration of Bauschinger effect improves accuracy of half-tube model by 9.82%. ► Half-tube model has good accuracy/efficiency with scaling factor 4800 and mesh 1×1 mm2.

Related Topics
Physical Sciences and Engineering Engineering Mechanical Engineering
Authors
, , , ,