Recommendations for Experimental Studies on the Seismic Behavior of Steel Components and Materials

The study discussed in this report is concerned with experimental procedures for a seismic performance assessment of steel components and materials. The main objective of this study is to identify parameters and testing programs that will pertain an evaluation of deterioration and closeness to failure of a component which is part of a structure that may be subjected to one or several severe earthquakes of random character.

In order to achieve this objective, the following aspects are conafdared in this study and are discussed in this report: (1) An identification of the purpose of coaponenc experimentation from the viewpoint of performance assessment; (2) A review of testing standards and of experimental procedures employed by the research community; (3) A review of low-cycle fatigue properties of structural steel, recommendations for testing procedures for materials, and a study of the cyclic stress-strain properties of A36 structural steel; (4) A review of low-cycle fatigue damage models and an assessment of their applicability to the problem of performance evaluation of structural components; (5) An experimental study of component performance, considering the deterioration and failure modes of local buckling in beam flanges and of crack propagation at weldments; (6) An analytical study on those seismic response parameters that are needed for damage evaluation and performance assessment of structural components, and for a development of representative cyclic loading histories.

The conclusion drawn from this study is that simple cumulative damage models can be utilized to assess deterioration and failure in structural components. Thus, experimentation should be directed towards a determination of the structural performance parameters needed for cumulative damage modeling. The basic tests for this purpose are constant amplitude tests in which the deformation parameter used in the damage model is kept constant throughout each test. several identical specimens need to be tested because at least two performance parameters must be determined and because these parameters may exhibit considerable scatter. The results of a single test with a preselected loading history cannot be used for a general performance assessment. If a single test is used to check component performance, the applied loading history should be statistically representative of the deformation demands imposed by earthquakes. Recommendations on the selection of such histories and on other aspects of component testing are provided in the last chapter of this report.