Damage diagnosis of framework structure based on wavelet packet analysis and neural network

Abstract

In this paper, an approach based on "energy-damage" theory for structural damage diagnosis is presented by use of wavelet packet analysis and improved BP neural network. The damage characteristics of the time domain response signals are more obvious after being transformed by the wavelet. Using the node energy in different frequency bands as the sample of tbe ANN can quite well reflect the damage features. As a numerical example, the benchmark structure given by the ASCE is used for describing the process of the damage detection presented here.