Extraction of Impulse Response Data via Wavelet Transform for Structural System Identification

Abstract

Abstract This paper presents a wavelet transform-based method of extracting the impulse response characteristics from the measured disturbances and response histories of linear structural dynamic systems. The proposed method appears to have alleviated some of the most pronounced deleterious aspects of both the time-domain methods that suffer from the matrix ill-conditioning of the input signals and FFT-based methods that must cope with erraneous auto and cross-correlation functions, unless the input signals are rich enough in frequency content The method is found to be effective in capturing very low frequency response components and also far insensitive to output noises than existing methods. The present method has been applied to a variety of problems, which show significant improvements over existing impulse response function extraction methods, especially for limited harmonic excitations and input/output data contaminated with noises.