An Application of Generative Adversarial Networks in Structural Health Monitoring

Abstract

In the current work, the use of generative adversarial networks (GANs) in a simulated structural health monitoring (SHM) application is studied. A specific type of GAN is considered, aiming at a disentangled representation of underlying features and clusters of data through some latent variables. This idea could prove useful in SHM, since explanation of how damage mechanisms or environmental conditions affect a structure may be exploited in order to monitor structures more effectively. In a simulated mass-spring example, different damage cases are introduced by reducing the stiffness of specific springs and different damage levels by applying different extents of stiffness reduction. The GAN implementation proves able to capture different damage cases through its categorical latent variables, as well as the damage extent within its continuous latent variables. The results demonstrate that the latent variables are indeed capturing the effect of damage in the structure and can be exploited for the purpose of condition assessment.

Machine learning and structural health monitoring overview with emerging technology and high-dimensional data source highlights

“…With training, the generator improves and produces images that are more real. The idea of using GAN for SHM systems was studied by Tsialiamanis et al 239 It was demonstrated that with prior knowledge, GAN can reflect damage characteristics via categorical and continuous variables despite the presence of EOFs. Therefore, GAN can be promising in training large datasets.…”

A literature review: Generative adversarial networks for civil structural health monitoring

“…To identify and clarify the position of GANs in the civil SHM field, in terms of the type of GAN applications studied by the researchers, an illustrative figure is made (Figure 9) which shows the classification of the applications of GANs in civil SHM and the corresponding studies with the GAN models used in each study. As the main concept of GAN is to learn data domain and data generation, some studies solely studied data generation [ (Kanghyeok and do Hyoung, 2019;Xiong and Chen, 2019;Zhang and Wang, 2019;Tsialiamanis et al, 2020;Xu et al, 2021;Yu et al, 2021;Tsialiamanis et al, 2022a;Heesch et al, 2021;Colombera et al, 2021;Luleci et al, 2022b;Luleci et al, 2023)] (a total of 11 studies) by using original GAN or other GAN variants. Thus, the data generation category is separated from other categories.…”

“…The authors in one study (Tsialiamanis et al, 2020) investigated an application of InfoGAN in SHM where they demonstrated that InfoGAN can capture the damage acceleration responses in a simulated mass-spring application. They induced different extents of stiffness reductions as damage to the system.…”