Combining direct and indirect sparse data for learning generalizable turbulence models

Learning turbulence models from observation data is of significant interest in discovering a unified model for broad range practical flow applications. Either the direct Reynolds stress or indirect velocity has been used to improve predictive capacity models. In this work, we propose combining and sparse train neural network-based The backpropagation technique augmentation approach are with different ensemble-based framework. These two types can explore synergy constrain training spaces, which enables learning generalizable very data. present method tested secondary flows square duct separated over periodic hills. Both cases demonstrate that observations able generalizability learned similar configurations, compared using only serve as tool due its non-intrusive derivative-free nature.