機械学習とデータ同化を組み合わせた振動現象のパラメータ推定

Key Points

• The aim is to develop an efficient method for estimating parameters related to the vibration behavior of structures.
• Combination of data assimilation techniques and machine learning methods
• Used ensemble data assimilation with variational autoencoders
• Conducted numerical experiments to validate the proposed method
• Proven effectiveness compared to traditional ensemble data assimilation methods
• Improved convergence speed in parameter estimation
• Achieved high accuracy while reducing computational costs