A new dynamic identification technique: Application to the evaluation of the equivalent strut for infilled frames

Abstract

A new time domain identification technique for systems under Gaussian white noise input is presented, requiring for its application the measurement of the system response but no information about input intensity. The technique proposed is based on the statistic moment equations derived by using a special class of mathematical models named "potential models". These models allow one to determine fundamental properties of the response statistics, making it possible to identify stiffness and dissipation features of a structural system, and also to determine the excitation input. The technique proposed is here applied to the identification of the strut equivalent to the infill of a single story-single bay frame subjected to lateral loads, showing a reduced effort compared to any procedure based on static experimental tests and a higher reliability of the results compared to identification procedures for the strut based only on mathematical assumptions. © 2003 Elsevier Science Ltd. All rights reserved.