| Abstract: | Urban housing has been provided economically in many parts of the world by large panel precast building systems; in seismic regions, however, special attention must be given to the connections between panels in order to achieve the necessary earthquake resistance. In this paper a comprehensive seismic test programme of concrete panel assemblages done as part of a U.S.-Yugoslavia cooperative research plan in earthquake engineering is described. The test structures were three-storey assemblages of 1/3 scale concrete panels supplied by a Yugoslavian construction company; the models were about 12 ft high and 6 ft long, with the panels interconnected by various types of joint systems along their tops, bottoms and sides. Cyclic pseudostatic testing of the models was done at the IZIIS Laboratories in Skopje, Yugoslavia. A second set of models was assembled at the University of California, Berkeley from identical panels that had been shipped from the Belgrade factory; these models were tested on the 20 ft square EERC shaking table. The purpose of this paper is to compare the earthquake response represented by the two different types of test procedures. A major conclusion was that the response mechanisms in both cases were similar, with the principal deformation being opening of the panel joint at the base of the model due to overturning moments. As the rocking response continued, tension and compression failure occurred alternately at the two ends of the joint. The principal difference between the results was that pseudostatic testing imposed an increasing sequence of symmetric deformation that was independent of the damage being developed in the model, while the shaking table response was greatly influenced by the damage as it occurred—leading to significant unsymmetry. Moreover, the damage occurred in a concentrated region with the shaking table test. This difference emphasizes the fact that pseudostatic test results must be interpreted with caution because they do not fully simulate an actual earthquake response. |
