Gowda, S., Deb, A., Kurnool, G., and Chou, C., "Prediction of the Behaviors of Adhesively Bonded Steel Hat Section Components under Axial Impact Loading," SAE Technical Paper 2017-01-1461, 2017, doi:10.4271/2017-01-1461.
Adhesively bonded steel hat section components have been experimentally studied in the past as a potential alternative to traditional hat section components with spot-welded flanges. One of the concerns with such components has been their performance under axial impact loading as adhesive is far more brittle as compared to a spot weld. However, recent drop-weight impact tests have shown that the energy absorption capabilities of adhesively bonded steel hat sections are competitive with respect to geometrically similar spot-welded specimens. Although flange separation may take place in the case of a specimen employing a rubber toughened epoxy adhesive, the failure would have taken place post progressive buckling and absorption of impact energy. The better-than-expected performance of an adhesively bonded component subjected to axial impact load is likely due to the evenly spreading adhesive over the entire flange areas of the hat section as compared to discrete spot welds in a conventional component. In the current study, numerical prediction of the behaviors of axially impacted adhesively bonded double-hat section components has been carried out using explicit finite element modeling and analysis. The formed plates in a double-hat section member are represented with shell elements while the adhesive in between either of the twin flanges is modeled with solid elements. A systematic approach is followed according to which finite element modeling parameters, especially those related to constitutive modeling, are verified by predicting the behaviors of single lap shear and T-peel joints under tensile loading conditions. On obtaining satisfactory correlation for the coupon-level tests mentioned, the modeling procedure is extended to the axial impact simulation of double-hat section components in a drop-weight test set-up, and the capability of predicting load-displacement responses and key crash parameters such as peak load, mean load and energy absorbed is demonstrated.