Wave based damage detection in solid structures using spatially asymmetric encoder–decoder network

Abstract

The identification of structural damages takes a more and more important role within the modern economy, where often the monitoring of an infrastructure is the last approach to keep it under public use. Conventional monitoring methods require specialized engineers and are mainly time-consuming. This research paper considers the ability of neural networks to recognize the initial or alteration of structural properties based on the training processes. The presented model, a spatially asymmetric encoder–decoder network, is based on 1D-Convolutional Neural Networks (CNN) for wave field pattern recognition, or more specifically the wave field change recognition. The proposed model is used to identify the change within propagating wave fields after a crack initiation within the structure. The paper describes the implemented method and the required training procedure to get a successful crack detection accuracy, where the training data are based on the dynamic lattice model. Although the training of the model is still time-consuming, the proposed new method has an enormous potential to become a new crack detection or structural health monitoring approach within the conventional monitoring methods.

Study of wave propagation in discontinuous and heterogeneous media with the dynamic lattice method

“…However, the simulation of large domains with high frequencies requires a substantial number of elements, not only to model the particle size effect but also to reach greater wavelength to elements length ratio and avoid numerical difficulties. One way to overcome this obstacle is to consider the artificial neural networks (ANN) models to predict the location of the discontinuities in the homogeneous and heterogeneous bodies 40 . The recorded displacement wave fields at each reference point in dynamicLEM are considered as training data for developing an ANN method to not only decrease the computational costs but also increase the accuracy of the crack predictions.…”

“…The simulation of crack propagation under dynamical forces by an embedded strong discontinuity approach is also implemented 39 . The simulation results of the dynamicLEM are used to train the artificial neural networks (ANN) model to detect the location of the discontinuities 40 , thus reducing the computational costs.…”

Overall structural seismic damage rapid assessment method based on period and displacement response characteristics

“…That is, assuming that the lateral displacement of the building structure is in a single deformation form under external loads such as earthquake and wind load, the structure has only one degree of freedom in the sense of structural dynamics. For the generalized single degree of freedom system, the generalized mass M z and the generalized stiffness K z associated with the single degree of freedom should be determined first in the process of calculating the natural vibration period 14 .…”

Delamination prediction in composite panels using unsupervised-feature learning methods with wavelet-enhanced guided wave representations

“…This study is performed on a smaller scale, but the current study is much more extensive and detailed. More recent research by Wuttke et al [18] have also used wavelet-transformation for propagating wavefields in a plate-like structure. A numerical dataset is collected from a dynamic lattice model, and an asymmetric encoderdecoder network is used to study crack initiation.…”