Investigation of time efficiency in wavelet-based Markov parameter extraction methods

摘要

This paper investigates the time efficiency of using a wavelet transform- based method to extract the impulse response characteristics of a structural dynamic system. Traditional time domain procedures utilize the measured disturbances and response histories of a system to develop a set of auto and cross correlation functions. Through deconvolution of these functions, or matrix inversion, the Markov parameters of the system may be found. By transforming these functions into a wavelet basis, the size of the problem to be solved can be reduced as well as the computation time decreased. Fourier transforms are also used in this capacity as they may increase the time efficiency even more, but at the cost of accuracy. This paper will therefore compare the time requirements associated with a time, wavelet, and Fourier- based method of Markov parameter extraction, as well as their relative accuracy in modeling the system.