Reliability assessment of civil structures under seismic loads requires probabilistic evaluation considering the uncertainty of input ground motion and material properties due to deterioration. However, Monte Carlo calculation for the structural reliability analysis is computationally expensive. This study develops the deep kernel learning surrogate model that can not only reduce the computational cost but also provide explainability for the prediction results. The model extracts the features of seismic loads by the convolutional neural network (CNN) and considers the uncertainty of seismic loads and material properties by the Gaussian process regression with the automatic relevance determination (ARD) kernel. By the incorporating gradient-weighted class activation mapping (Grad-CAM) in the CNN, the parts of seismic load response spectra, where contribute to the constructed surrogate model, can be visualized. The model can also provide which input uncertain parameters of structural properties has relatively influence on the output response by the estimated ARD kernel weights. The developed surrogate model is verified by applying it to the seismic performance analysis of a concrete bridge pier with a seismic rubber bearing under various earthquake loads with different intensity and response spectra. The results show that the developed surrogate model can predict accurate distributions of maximum displacements and can provide reasonable contributions of uncertain inputs to enhance the explainability.
This study developed a framework for real-time dynamic analysis of structural members using physics-informed neural networks (PINN). The interest in the use of augmented reality (AR) and virtual reality (VR) technologies to visualize the results of simulations is now increasing, and many researches are taking efforts to make these simulations more interactive and real-time. However, the application of structural dynamic simulations is limited due to its high computational cost. In this study, the Physics-informed neural networks (PINN) was used to conduct the real-time vibration analysis of a cantilever beam as a basic investigation. Prior to the real-time simulation, a PINN model for solving the cantilever beam undamped free vibration problem was constructed. Sequential trainings and predictions for the real-time simulation were then implemented at fine increment time steps by PINN. The distributions of displacement and bending moment, which were the outputs of the PINN simulations were visualized in AR on the real beam with converting the outputs to color contour for intuitive understanding. The RS framework based on PINN simulation and AR was then recognized to lead to the RS with data assimilation for real-time evaluation of structural condition using measurement data.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.