FATIGUE ANALYSIS OF BUTT WELDS. COMPARISON AND VALIDATION OF THE MAIN METHODOLOGIES

Abstract

ABSTRACT: This work presents the main methodologies for the high cycle fatigue design of welded structural components, distinguishing between the use of experimental and design S-N curves, and that of nominal, hot spot and notch stresses. Considering only these distinctions, and not considering other features such as the existence of diverse design curves in the different standards or analysis procedures, six methodologies arise. It is essential, therefore, when designing a given component, to know both the accuracy and the safety associated to each of these methodologies With this aim, and focusing the fatigue analysis on butt welds, a double experimental program is presented that allows, firstly, the experimental S-N curves of the butt joint to be obtained, and secondly, the different methodologies used for different fatigue loading sequences to be validated. The results show how the use of experimental S-N curves provides less conservative results than the use of design curves when the number of loading cycles is very high, whereas the opposite is true when dealing with low cycle fatigue. In the same way, it is observed how the use of nominal stresses, which are more easily obtained than both hot spot and notch stresses, generally provide more satisfactory results, contrarily to what might be expected from the literature. With all this, the use of nominal stresses is recommended with design or experimental curves, depending on the dominant relation between the minimum and the maximum stresses.