Development of Stress-Based Forming Limit Curves for Predicting Crack Occurred during Deformation of UHSS DP980's Parts

Unless the strain-based forming limit diagram (e-FLD) is a useful tool in failure prediction for sheet metal, the complex shape can lead to a non-linear strain history, which e-FLD cannot adequately describe. This issue can be improved by using the stress-based forming limit diagram (s-FLD). However, determining the s-FLD through direct theoretical calculation is complex for beginners. This research uses a forming simulation software named PAM-STAMP to transform e-FLD to s-FLD by simulating the Nakajima stretch test. Because a yield criterion and hardening model are required in s-FLD determining procedures, this research also studies the accuracy of Hill48, Barlat89, and Yld2000 (in the absence of balanced biaxial condition test results) when integrated with the Swift hardening model and the Yoshida-Uemori (Y-U) model. The thickness and springback of an automotive part named Panel-RF FRT HRD were used to compare the accuracy of the forming simulation. It was found that the s-FLD predicted a crack on the workpiece better than the e-FLD, no matter whether there were differences in combination in yield criteria or hardening models. In the thickness and springback prediction, the r-based Hill48 couple Y-U model showed the best result, followed by Barlat89 and Yld2000, and was less accurate when Hill48 was coupled with the Swift hardening model.