An improved springback model considering the transverse stress in microforming

Abstract

To satisfy the higher accuracy requirement in the microforming of thin-walled structures, the springback behavior under miniaturizing dimensions need to be accurately predicted and controlled. However, as the feature size approaches the thickness in microforming, many assumptions in conventional springback model for macroforming are not valid making these models can hardly provide effective springback analysis in microforming. In the present work, the applicability of a springback analytical framework of conventional stretch-bending was first explored via bending tests of 0.1 mm-thick SS316L sheets under different holding conditions with a punch radius from 3 to 5 mm which is much greater than the thickness. The analytical springback angle results agree with the experimental ones with the maximum error of 4.2% for different punch radii, grain sizes and stretching forces conditions, indicating the correctness of the model. The model was further verified in the microforming of parallel channels with a fillet radius of 0.15 mm. Finite element simulations were conducted to obtain the stretching loads for each channel. However, a significant underestimation of the channel height after springback for over 15% was found especially for the middle channel. The transverse stress is revealed to be no longer neglectable due to the strong stretching load and the small die radius during the microforming of parallel channels, which substantially contributes to yielding, hardening and thickness reduction in the stretch-bending area. In addition, the uneven stretching also affects the springback angle and channel height, which is caused by the nonuniform plastic deformation of the adjacent channels. By incorporating the effects of the transverse stress and the non-uniform stretching loads, an improved springback model was developed for microforming of parallel channels in this work. The shift distance of the neutral plane is better estimated by the improved model than by the conventional one. The predictive heights by the improved model are also revealed to agree with the experimental ones during the microforming of parallel channels with a maximum error of less than 5%.