Comparison of modal parameter identification algorithms based on shaking table model test data

Abstract

Modal parameter identification is the critical component for structural damage detection and structural health monitoring. Although a lot of identification algorithms have been developed, there are still some problems in terms of their accuracy and effectiveness. For example, for rational fractional polynomial method and orthogonal polynomial method, which are based on fitting measured frequency response functions, the former is disturbed by ill-conditioned solution and misfit while the latter is obsessed by precision loss with transformation of base function. On the other hand, for algorithms in time domain, such as eigensystem realization algorithm and stochastic subspace identification, construction of Henkel matrix and determination of system order are the main issues with which no general methods to deal up to now. In order to compare the similarities and discrepancies of different modal identification algorithms, the data taken from a shaking table test on a 12-storey reinforced concrete frame model are processed using selected rational fraction polynomial method in frequency domain and eigensystem realization algorithm in time domain. The comparison of the identification results is further discussed.