企画セッション: リスクと不確実性の定量化

[2A13-21-02] Real-Time Reconstruction Simulation Based on Autonomous Basis Function Selection

*Yu Otake^1,2, taiga Shoji², Yosuke Higo³, Ikumasa Yoshida⁴ (1. Tohoku University, 2. Tohoku University, 3. Kyoto University, 4. Tokyo City University)

Keywords:reconstruction, liquefaction

Providing infrastructure facilities with resilient performance against accidental loads, such as natural disasters, is increasingly recognized. The authors are developing a monitoring design method to improve the resilience performance of infrastructure facilities and a way to predict the overall facility situation from monitoring records immediately. Specifically, we propose a method for immediately predicting (reconstructing) the dynamic behavior of an entire facility using the spatial mode functions (basis functions) of the facility of interest obtained from numerical analysis results. The proposed method is characterized by introducing an adaptive model that immediately determines the number of basis functions to be used for reconstruction. The effectiveness of the proposed method was verified by evaluating the seismic behavior of an embankment on a liquefiable sand layer.