Monte Carlo simulation to predict fatigue SN curve in Aluminum Friction Stir Welds
Antonio Miranda1; Scott Walbridge2
1 UnB; 2 University of Waterloo
doi:10.20906/CPS/CILAMCE2015-0729
Resumo
Recent papers on the fatigue design of aluminum welds have highlighted a lack of guidance in existing codes and standards for the fatigue assessment of aluminum alloy structures containing joints made by friction stir welding (FSW). Structural design code writers in North America and Europe are starting to address this need. In addition to provisions on fatigue performance, codes and standards are also beginning to appear that provide guidance on welding procedures and quality control. However, these provisions seem to be prescriptive from the point of view of fatigue, and the influence of defects resulting from the welding process on the fatigue performance have not been extensively studied. Against this background, the goals of the current paper are: 1) to perform a statistical analysis of available fatigue life data for FSW joints and compare the results to design stress-life (S-N) curves in various codes and standards using Monte Carlo approach; 2) to present a review of the literature on FSW joints with a focus on the material and geometric parameters required to perform a fracture mechanics analysis of these joints; and 3) to show how probabilistic fracture mechanics, given these parameters, can aid in establishing and refining the quality control guidelines for cyclically loaded FSW joints. Based on the presented analysis, parameters are identified for which additional data collection would be beneficial for improving our understanding of the fatigue behavior of FSW joints and how this behavior is affected by the weld quality.
Palavras-chave: probabilistic analysis; aluminum friction stir weld; fatigue parameter data