単層張弦ドームを対象としたデータ駆動型アプローチによる変位予測モデルの構築

抄録

This paper proposes a data-driven modeling method to predict the displacement of beam string structures through the steel erection process. A data assimilation framework is employed to create an as-built simulation model that represents the uncertainties in the real-world structure. Weld shrinkage of the beam and stiffness of the temporary support system are major sources of the model uncertainty, and these parameters are calibrated. The observation data acquired during the erection work are assimilated to update the displacement prediction model. The obtained model can predict the future displacement behavior with less error than the model without data assimilation.