Optimal Wavelet Parameters for System Identification of Civil Engineering Structures

Abstract

Damage-induced changes in structure dynamic properties are commonly tracked with time-frequency representations (TFRs). One of the most widely accepted tools for determining a TFR is the continuous wavelet transform (CWT). The success of CWT analysis is highly dependent on selecting the most appropriate values for the parameters that define the mother wavelet. This article presents a detailed analytical and numerical study to select optimal wavelet parameters using the complex Morlet wavelet (CMOR) and the Gabor wavelet. The results obtained suggest that it is possible to define optimal parameter values based on identification target, instantaneous frequency, or average damping ratio. This reduces the computational cost of a reliable CWT analysis when compared with currently employed iterative methodologies based on minimal Shannon entropy criteria.