Search Results

For this latest SAE Fatigue Design and Evaluation project, fatigue tests were run by loading, in bending, both welded and machined T-Joint specimens that have the same geometry. The test rig setup consisted of a horizontally mounted actuator, with pinned joints at both ends, where the load is applied to the top of the vertical leg of the “upside down T” of a T-Joint specimen, while the horizontal legs of the “upside down T” were clamped to the bedplate. Specimens were tested until failure or until the specimen was unable to carry the commanded load. They were cycled under constant amplitude (at several load levels and R ratios), block cycle, and variable amplitude loadings. Welded and machined T-Joint specimens of the same geometry were included in the test plan such that fatigue life predictions could be compared to test lives for each case. Those comparisons would demonstrate the methodology’s relative predictive ability to manage welds, residual stress, etc...

With the increased use of tubular steel products, especially for automotive hydroforming applications, there is increased interest in understanding the mechanical properties measured by tensile tests from specimens of different orientations in the tube. In this study, two orientations of tensile specimens were evaluated -- axial specimens with and without flattening and flattened circumferential specimens. Three steels were evaluated -- two thicknesses of aluminum killed drawing quality (AKDQ) steel and one thickness of high strength low alloy (HSLA) steel. Mechanical property data were obtained from the flat stock, conventional production tubes and quasi tubes. Quasi tubes were produced from the flat stock on a 3-roll bender, but the quasi tube was not welded or sized.

Electron beam (EB) welding process is characterized by an extremely high power density that is capable of producing weld seams which are considerably deeper than width. Unlike other welding process, heat of EB welding is provided by the kinetic energy of electrons. This paper presents a computational model for the numerical prediction of microstructure and residual stress resulting from EB welding process. Energy input is modeled as a step function within the fusion zone. The predicted values from finite element simulation of the EB welding process agree well with the experimentally measured values. The present model is used to study an axial weld failure problem.

The Society of Automotive Engineers Fatigue Design & Evaluation Committee [SAE FD&E] is actively working on a total life project for weldments, in which the welding residual stress is a key contributor to an accurate assessment of fatigue life. Physics-based welding process simulation and various types of residual stress measurements were pursued to provide a representation of the residual stress field at the failure location in the fatigue samples. A well-controlled and documented robotic welding process was used for all sample fabrications to provide accurate inputs for the welding simulations. One destructive (contour method) residual stress measurement and several non-destructive residual stress measurements-surface X-ray diffraction (XRD), energy dispersive X-ray diffraction (EDXRD), and neutron diffraction (ND)-were performed on the same or similarly welded samples.

The presence of residual stresses affects the fatigue response of welded components. In the present study of thick welded cantilever specimens, residual stresses were measured in two A36 steel samples, one in the as-welded condition, and one subjected to a short history of bending loads where substantial local plasticity is expected at the fatigue hot-spot weld toe. Extensive X-Ray Diffraction (XRD) measurements describe the residual stress state in a large region above the weld toe both in an untested as-welded sample and in a sample subjected to a short load history that generated an estimated 0.01 strain amplitude at the stress concentration zone at the weld toe. The results show that such a test will significantly alter the welding-induced residual stresses. Fatigue life prediction methods need to be aware that such alterations are possible and incorporate the effects of such cyclic stress relaxation in life computations.

A new total fatigue life methodology was utilized to make fatigue life predictions, where total fatigue life is defined as crack initiation and subsequent crack propagation to a crack of known size or the component’s inability to carry load. Fatigue life predictions of an A36 steel T-joint geometry were calculated using the same total fatigue life methodology for both welded and machined test specimens that have the same geometry. The only significant difference between the two analyses was the inclusion of the measured weld residual stresses in the welded specimen life predictions. Constant amplitude tests at several load levels and R ratios were analyzed along with block cycle and variable amplitude loading tests. The accuracy of the life predictions relative to experimental test lives was excellent, with most within a factor of +/- two.