APPLICATION OF A ROBUST MULTI-DATASET FREQUENCY DOMAIN DECOMPOSITION TO ESTIMATE THE GLOBAL MODAL PARAMETERS OF A HIGH-RISE BUILDING

In multi-dataset identification with the classic Frequency Domain Decomposition (FDD) technique, the global mode shapes are obtained by identifying the mode shape parts of different patches separately. Then, the estimated mode shape parts are merged to form the global mode shapes by using the reference mode shape components. However, the presence of closely spaced or poorly excited modes can affect differently the estimation of the mode shape parts; hence, the resulting global mode shapes might not present well-defined modal configurations. This paper describes the application of a correlation spectrum based FDD technique to estimate the modal parameters of a high-rise building by identifying the global mode shape vectors in one step. This is expected to provide mode shape estimates with increased accuracy. To illustrate the performance of the correlation spectrum based FDD technique, the modal properties estimated with this technique is compared to those estimated with the classic FDD one.