Cornell, C. Allin
Mexico : UNAM, Instituto de Ingenieria, paper presented at Simposio Luis Esteva, UNAM, 12 de septiembre de 2005., 2005-09-12, PDF
The apparently simple act of selecting accelerograms for use in conducting nonlinear dynamic analysis has been the subject recent study and conclusions are emerging. This paper attempts to elucidate those conclusions by starting from very ideal and simplified cases to deduce observations that are supported by more through numerical studies of nonlinear MDOF structural models. A direct and an indirect approach are discussed. To insure both lack of bias and adequate confidence limits in the estimation of response and response likelihoods, the selection of the spectral shape and number of the records is found to be dependent on the site, the structure and ground motion level. With proper choice of the spectral shape scaling of the records should not cause significant bias.
See also:
Simposio Luis Esteva - tendencias y retos de la ingenieria sismica