動的濡れ―スケールと非定常性の効果

動的な濡れ現象はインクジェットプリントなどの液滴の固体面への衝突やコーティングなど工学的にさまざまな分野で見られる．しかしながら，依然として理論と実験値は乖離することが多く，理論では考慮されていない各種パラメータの影響の評価が急務となっている．本講演では特に動的濡れに与える非定常運動の効果について現状と課題を述べる．

Development of a Multiscale Reduced-Order Modelling Framework for CFD-DEM Simulations

This talk presents a multiscale reduced-order modeling framework for multiphase granular flow systems, aimed at accelerating simulations of industrial gas-solid flows. The framework combines a nonintrusive POD-based reduced-order model (ROM), with a signed distance function-based graph neural network (SGN), referred to as ROM-SGN. While CFD-DEM is an established method for simulating gas-solid flows, it is often computationally heavy for industrial-scale applications. The proposed ROM-SGN framework achieves several orders of magnitude speedup while maintaining acceptable accuracy in capturing gas-solid dynamics. This approach offers a promising and efficient pathway for constructing digital twins of multiphase processes.