4. Characterization tests

Tests are carried out as far as possible on the construction detail that is most representative geometrically of the actual assembly and corresponds to the manufacturing quality. However, certain defects such as porosity are not always easy to obtain deliberately and within the accepted limits. In this case, it is preferable to carry out the tests on sound test pieces but with a representative geometry (connection, excess thickness), the result then being revised using a correction coefficient to correspond to the most unfavorable accepted case.

Since the fatigue strength of welded joints is a highly variable phenomenon, it is necessary to perform a minimum number of tests in order to derive a realistic average and minimum value, taking this variability into account. At least three to four specimens per stress level must be tested in order to understand the variation....