Damage Detection of Orthogonal Pyramid Space Truss Based on Wavelet Packet Energy-Neural Network

An approach of wavelet packet energy combined with neural network is proposed to research damage detection of space truss structures in this paper. Energy distribution will change when damage emerged. The dynamic parameters difference between damaged and intact structure can be decomposed into different frequency bands and reconstructed by wavelet packet, then the corresponding wavelet packet energy are obtained. The damage locations are determined by wavelet packet energy. At the same time, the damage degree can be received by RBF neural network, considered wavelet packet energy as samples. An orthogonal pyramid space truss with assumed damages was detected by this method. The results show that damage location and damage degree can be determined accurately by this method and its valuable to the damage identification of space truss structure.